14
יין
1
هم
',
t
4
*
す
​་་.
༑་་་ ་ ་ ་
་་་གེའནའ
P326
800
2657
C. 2.
2017
}
་་
1
པ་་་བ
·
+
14
S
*
n
.}
ra
EX
1
บ
14
F

་
:
ARTES
1837.
LIBRARY
VERITAS
SCIENTIA
OF THE
UNIVERSITY OF MICHIGAN
TUEBOR
SI QUÆRIS PENINSULAM·AMŒENAM
CIRCUMSPICE
H
800
P32%
±592
S
Copy 2

PRINCIPLES
OF THE
HISTORY OF LANGUAGE
1
A HISTORY OF THE GERMAN LANGUAGE
By H. A. STRONG, M.A., LL.D., Professor of Latin in the
Liverpool University College, and
KUNO MEYER, Ph. D., Lecturer on Teutonic Languages,
Liverpool University College.
8vo. Cloth, 6s:
'This book supplies a long-felt want.... The authors have spared no pains to
include in their volume the results of the most recent researches of German
philologists.'-Literary World.
LONDON ;
SWAN SONNENSCHEIN, LOWREY, & CO.

PRINCIPLES
7.11. Scatt
OF THE

HISTORY OF LANGUAGE
BY
HERMANN PAUL
PROFESSOR OF THE GERMAN LANGUAGE AND LITERATURE IN THE
UNIVERSITY OF FREIBURG
TRANSLATED FROM THE
OF THE
EDITION
SECOND EDITION
ORIGINAL
BY
H. A. STRONG, M.A., LL.D.
PROFESSOR OF LATIN IN UNIVERSITY COLLEGE, LIVERPOOL
。
NEW YORK
MACMILLAN & CO.
1889
010 313 131
PREFACE TO THE AMERICAN EDITION.
ONE by one the new sciences outgrow the methods of
the placer. Their material becomes ever more recondite,
and the necessity for both specialization and co-operation
becomes ever greater. The ambition of research no longer
satisfies itself with brilliant combinations of surface-facts,
and startling surprises for the mob. The amateur and the
adventurer are gradually shouldered out. They can have
no sympathy with the minute and laborious processes
which have, with the advance of the science, become neces-
sary, and, even if they had, their tools and equipment.
would be inadequate to the work.
The rapid increase of the materials for a science of
language within the last few decades has acted on the one
hand to repress amateurism, and on the other to check
arbitrariness of method on the part of professional lin-
guists.
The first combinations, out of which were developed
the earliest results of the science, were made without
the guidance of principles, without any determination of
method, and without any well-defined knowledge of the
scope of the science, of its relation to kindred sciences,
iv
PREFACE TO THE AMERICAN EDITION.
or of the real nature of its own material. The infant
science simply groped its way into its first vague orien-
tations.
It was a sort of general intuition only, that dictated the
selection of the more probable out of the mass of the less
probable combinations, and it was the sum of these accepted
combinations that gradually availed to establish in the
mind of the individual investigator that ever finer stand-
ard of tests which made up his unconscious principles of
method.
The growth of the science has been distinctly character-
ized by continual approximation toward each other of the
methods and standards of the various leading investiga-
tors. A review of the opinions of different scholars in the
period prior to the appearance of Schleicher's Compendium
(1861) shows an almost total lack of common understand-
ing among investigators. The personal ipse dixit of the
great masters and the individual it seemeth good constituted
the sole standards of test for this age of particularism.
Opinions came somewhat less in conflict in those days, it
is true, for the vastness of the unoccupied territory gave
elbow-room enough. It was undoubtedly owing to this
circumstance that Professor Curtius in his later days came
to regard the period prior to the introduction into this
world of the Bpis of neo-grammarianism as idyllic for its
peace and unity, like the days before Babel and the dis-
persion.
During the last two decades the chief activity in lin-
PREFACE TO THE AMERICAN EDITION.
V
guistic studies has centered about Berlin, Göttingen, and
Leipzig, and the keepership of the oracles of method has
transferred itself largely into the hands of the three so-
called "schools,”—of Berlin, of Göttingen, and of Leipzig.
The sect of the neo-grammarians is essentially the Leipzig
school. An unprejudiced survey of the work of the differ-
ent savants of whatever adherence will show, despite all
the acrimony of recent discussions, a rapid approximation
during the last fifteen years toward an essential unity of
method and a common standard of judging results. It is
the Leipzig school that has made the loudest calls for state-
ment of method, and it is perhaps largely due to natural
dislike of a scheme of method drawn up and promulgated
by a rival school, that the two other schools have offered
such sturdy opposition. In terms of actual scientific work
these schools in reality no longer exist. In their inclina-
tion to make stricter application of the laws of sound-and-a
wider use of the principles of analogy in the practical pro-
cedures of investigation, the Berlin and Göttingen scholars
are essentially at one with the neo-grammarians. The real
progress of the science has paid but little heed to the bar-
riers of the sects.
Up to the beginning of the present decade no formal
attempt had been made to define and characterize the
science, to state its principles, and formulate its methods.
A variety of treatises having for their purpose the clarifi-
cation of common thinking concerning the nature and life
of language, and the dislodgment of popular misconcep-

4
vi
1
PREFACE TO THE AMERICAN EDITION.
1
tions of the subject, had appeared during the two pre-
ceding decades in England and America. Of all these.
incomparably the ablest and most widely influential were
those from the pen of our distinguished countryman,
Professor William D. Whitney. Especially in his treatise
on the Life and Growth of Language he calls repeated
attention to the necessity of some general formulation of
the principles of the science in its relation to its peculiar
material.
It is a common article of Anglo-Saxon creed, that the
average German savant, immured among the mountains of
his accumulated data, fails in his sense of perspective; and
the German, it is interesting to note, repays the compliment
by attributing to his Anglo-Saxon neighbor an undue
readiness to estimate the forest before knowing much about
the trees therein.
Be that as it may, the Principien der Sprachgeschichte
by Professor Paul represents a thorough-going attempt of
one of the greatest living German scholars to formulate.
into the principles of a science the determined facts of
language-history. It is written in the interest of perspect-
ive by a master of the most recent detail-study. As the
first work of its kind, its appearance may safely be said
to determinę a distinct period in the development of the
science; as the most profound discussion of the material
involved, it becomes the fundamental handbook of the
science, presenting a definite standard of doctrine toward
which the specialist and every man who would have any
1
PREFACE TO THE AMERICAN EDITION.
vii
word in the linguistic discussions of to-day, is forced to take
attitude, and a wide range of accurate and discriminating
statement in which the general student will find the safest
Introduction to the subject.
CORNELL UNIVERSITY, ITHACA.
BENJAMIN IDE WHEELER.
CONTENTS.
INTRODUCTION,
!
Necessity for a general theoretical science or system of principles to
accompany the History of Language, not less than other branches
of the Science of History - Closer definition of its task - The
doctrine of principles at the same time the basis for the doctrine of
method--Transference of the method of observation usual in natural
science to the science of culture-The science of language admits
of more perfect methodical treatment than other historical sciences
-Co-operation of psychical and physical factors in all development
of culture--The science of culture always a social science-Criticism
of the 'popular psychology' of Lazarus and Steinthal-Reciprocal
influence of mind upon mind only indirectly possible by physical
aid-Transformation of indirect associations into direct-Peculiari-
ties of the science of language as contrasted with other sciences-
Scientific treatment of language only possible by means of histori-
cal observation.
!
PAGE
xxi
CHAPTER I.
GENERAL OBSERVATIONS ON THE NATURE OF THE
DEVELOPMENT OF LANGUAGE.
Subject of the science of language-Organisms of idea groups the basis
of all linguistic action-Such organisms the true media of historical
development-Requisites for the description of a condition of lan-
guage-Change of usage is the result of the ordinary exercise of
speech-Language develops by gradations-Classification of the
changes-Grammar and Logic,
I
X
CONTENTS.
CHAPTER II.
THE DIFFERENTIATION OF LANGUAGE.
Analogies from organic nature-Statement of the problem-Linguistic
change and differentiation-Varying states of intercourse-Indivi-
dual spontaneity and social influence-Mutual independence of
each differentiation--The image of a genealogical tree inadequate—
Dialectical divisions graduated-The essential mark of a dialect its
sounds-Technical and poetic language,
PAGE
21
}
CHAPTER III.
ON SOUND-CHANGE.
The active factors in the production of sound-Slight consciousness of
the element of a word-The word a continuous series of infinitely
numerous sounds-Limits to power of distinguishing sounds-
Deviations from the prescriptions of the motory sensation inevitable
-Modification or displacement of the motory sensation-Causes
of the divergence—' Convenience' a secondary cause, the motory
sensation the primary cause-Control through the sound-picture-
Relation of the individual speaker to his linguistic milieu-Sound-
changes which do not depend on a displacement of the motory
sensation-Uniformity of the laws of sound-change,
CHAPTER IV.
CHANGE IN WORD-SIGNIFICATION.
Change of meaning consists in a widening or narrowing of its scope-
Distinction between 'usual' and 'occasional' meanings-Distinc-
tion between 'abstract' and 'concrete' meanings--Plurality of
meaning-Means by which abstract words may receive in their
occasional uses a concrete meaning-Means by which meaning is
specialised-The occasional meaning does not necessarily include
་
1
і
1
1
CONTENTS.
xi
all the elements of the 'usual' meaning-Transference to things
spatially, temporally, or causally connected with the original denota-
tation-Change of usage effected by 'occasional' modification—
Classification of changes of meaning-Specialisation-Increase of
connotation-Limitation to a part of the original connotation-
Transference to things spatially, temporally, or causally associated
-Combination of the above classes-Change of meaning in groups
of words-Scope of meaning conditioned by the state of culture of
the individual speaker-Scope of meaning conditioned by the state
of culture of the whole society,
•
PAGE
65
CHAPTER V.
ON ANALOGY.
Material- and formal-groups-Proportion-groups-Material-Formal in-
fluence of proportional groups upon speech (analogical creation)
-Analogical creation in syntax-Analogy in inflexion and word-
formation Divergence of analogical formation from usage-
Analogy in sound-substitution,
92
>
CHAPTER VI.
THE FUNDAMENTAL FACTS OF SYNTAX.
A sentence is the linguistic expression of the combination of several
ideas— Means of denoting the combination-Means of distinguishing
subject and predicate--Emphasis, position, etc.-Concrete and
abstract sentences-Sentences of apparently one element only-
Impersonal verbs-Negative sentences-Sentences of assertion
and demand--Interrogative sentences--Extension of the sentence--
Double subject-Double object-Use of a sentence as subject or
predicate Union of dependence and independence-Indirect
speech-Sentence in apposition to a noun-A noun in apposition to
a sentence-Parataxis-Graduated approximation to hypotaxis—
Transition from demand or interrogation to hypotaxis,
III
xii
CONTENTS.
CHAPTER VII.
CHANGE OF MEANING IN THE DEPARTMENT OF SYNTAX.
Comparison of syntactical with verbal change of meaning-The genitive
-The accusative-The prepositions-Apposition and partitive geni-
tive-Subject of a verb-Substantive and adjectival predicate-
Participial constructions--Conjunctions,
PAGE
147
CHAPTER VIII.
ON CONTAMINATION.
Definition-Contamination (1) Phonetic-(2) Syntactical contamination
-Anomalies (1) Momentary-(2) Usual-Pleonasms arising from
contamination,
•
160
CHAPTER IX.
ORIGINAL CREATION.
The conditions of new creation still present-Creation has never wholly
ceased-Application to original creation of the results won in other
departments of the history of language-Interjections-Nursery
language-The first creations belonged to no grammatical category,
denoted complete intuitions, and were uttered with no thought of
communication-Incapacity of primitive man to produce speech at
will-Reproduction necessary to the conception of language,
174
CHAPTER X.
ON ISOLATION AND THE REACTION AGAINST IT.
Possibility of a regularly observed system of grouping in each linguistic
period-System conditioned solely by correspondence in form and
meaning-Causes of isolation-Destruction of groups-(1) Etymo-
logico-phonetic groups-(2) Syntactical groups-(3) Formal and
material groups (a) by change of meaning, (b) by change of sound-
CONTENTS.
xiii
Reaction by aid of unification-Unification of differences which
result from difference of position in the sentence-Unification of
phonetically differentiated forms which belong to the same stem—
The qualities of unification owing to favouring and hostile conditions
—(2) Varying cohesion of the etymological groups-Influence of the
formal grouping-Conversion of a merely formal difference into a
difference of meaning- Conversion of elements of the stem into
inflexions-All phonetic changes involuntary,
CHAPTER XI.
FORMATION OF NEW GROUPS.
Obliteration of distinctions by sound-change-Union of unrelated words
in material groups (simplest form of popular etymology)-Phonetic
transformation (complex kind of popular etymology)—Union in
formal groups (a) where function is the same-Union in formal
groups (6) where the function is different,
•
PAGE
190
229
•
251
CHAPTER XII.
ON THE INFLUENCE OF FUNCTIONAL CHANGE ON
FORMATION BY ANALOGY.
Transference into another group changes the direction of analogical for-
mation-Consequences of the mutation of an appellative into a
proper name, of a case into adverb, of the fusion of a syntactical
combination into a verbal unity-Crystallisation-Operation of
change of meaning on the construction-Construction differently
understood owing to the influence of a synonym,
CHAPTER XIII.
DISPLACEMENTS IN THE GROUPING OF WORDS
ETYMOLOGICALLY CONNECTED.
The grouping of etymologically connected words diverges in the course
of time from that required by their formation; hence springs analo-
gical creation on a new basis-Examples-Fusion of two suffixes-
Displacement of relations in compounds,
•
261
xiv
CONTENTS.
CHAPTER XIV.
ON THE DIFFERENTIATION OF MEANING.
Causes of superfluity in language-Tendency to eliminate it-Elimina-
tion and utilisation-Phonetic differentiation for the purpose of
differentiating meaning, only apparent-Works on doublets-Cases
of apparent differentiation-Examples of real differentiation-Cog-
nate processes as a result of partial identity of meaning-Syntactical
differentiation, .
PAGE
272
CHAPTER XV.
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
The original harmony between psychological and grammatical cate-
gories is in course of time disturbed; tendency to remove the dis-
crepancy-Gender-Number-Tense—Voice,
CHAPTER XVI.
288
•
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION.
The divergence of psychological and grammatical distribution-Duality
and multiplicity of elements-The psychological predicate--Psycho-
logical subject and copula-Elements which are regularly psycho-
logical subject or predicate-The discrepancy avoided by peri-
phrasis-The discrepancy eliminated-Psychological treatment of
adverbial determinants-Rarity of discrepancy in languages of slight
formal development-Change of functions between the determinant
and the determinate-Separation of elements grammatically con-
nected-Genitive and adjective-Genitive and substantive-Verb
and adverb-Infinitive and an element dependent on it-Origin of
connecting words-Indirect reference becomes direct-- An element
equally related to two other elements is attracted to the first-Com-
pound sentences-Transition from dependence to independence-
Inversion of the relation of principal and subordinate sentence-
Impossibility of drawing a hard and fast line between principal and
subordinate sentences,
•
1
WI
CONTENTS.
XV
CHAPTER XVII.
ON CONCORD.
Concord started from cases in which one word came to agree with
another without regard for the latter, and was then analogically
extended to other cases-Cases in which the secondary growth of
concord is historically traceable—Variation of concord between
two parts of a sentence-First rudiments from which concord
proceeded,
PAGE
339
CHAPTER XVIII.
ECONOMY OF EXPRESSION.
Relative economy of expression depends on the need-Ellipse either to
be assumed in a minimum of cases, or else to be recognised as part
of the essence of expression-Words or phrases supplied-Absence
of links-A psychological subject or predicate may be drawn from
the situation—Isolated sentences in form dependent,
351
CHAPTER XIX.
RISE OF WORD-FORMATION AND INFLEXION.
Original new creation in word-formation secondary in its nature-
Growth of composition-Transition from syntactical juxtaposition
to true composition of a verb with an adverb-Of a verb with an
objective-accusative-With a prepositional determinant-Verbal
complexes which have the character of compounds-Co-ordination
between the elements of a compound and an independent word-
Sound-change with isolating effect-Limits within which the com-
pound appears as such-Origin of derivative and inflexional suffixes
-Formation of suffixes constant and unceasing-Rise of new suf-
fixes-Rise of inflexional suffixes-Criticism of the analysis of Indo-
European 'grund-formen,'
367
xvi
CONTENTS.
CHAPTER XX.
THE DIVISION OF THE PARTS OF SPEECH.
The division of the parts of speech due to consideration not of any logi-
cal principle, but of three points-Meaning, function, and inflexion-
Criticism of the usual division—Finite verb-Connecting-words, con-
junctions and prepositions-Intermediate stages between classes
of words-Substantive and adjective-Adverbs-Adjectives used as
a predicatival attribute-Origin of prepositions and conjunctions-
The preposition-Conjunctions-Difference between preposition and
conjunction in the simple sentence,
PAGE
403
CHAPTER XXI.
LANGUAGE AND WRITING.
Advantages and shortcomings of writing as compared with spoken.
language-Alphabets in use not as useful as they might be made—
Dialectical differences concealed by writing-Incapacity of writing
to communicate alterations of pronunciation-Effect of fixity of
orthography-Natural tendency in orthography towards greater
fixity-Analogies between the development of writing and that of
language-Abolition of variations in sound-signs of identical value
—Effect of etymology, .
433
CHAPTER XXII.
ON MIXTURE IN LANGUAGE.
Mixture in the wider and the narrower sense-Mixture of distinct
languages, dialects, and stages of language-Mixture originates in
individuals-Bilingualism-The chief modes of influence-A. Adop-
tion of foreign material-B. Usage of native material after a foreign
model-Causes of the adoption of words-Stages in the process of
adoption-Treatment of foreign phonetic material-Assimilation of
words already adopted-Competition among languages in the con-
tribution of loan-words-Pleonastic combination of a native with a
CONTENTS.
xvii
foreign suffix-Borrowing of derivative suffixes-Influence on the
inner linguistic form of language—Mixture of dialects-Borrowing
from an older period,
PAGE
456
CHAPTER XXIII.
THE COMMON-LANGUAGE.
Intercourse between areas lying separate from each other-For a common-
language to be created the necessity for it must be felt-A common-
language must be based upon an existing dialect-Circumstances
which must necessarily precede the formation of a common-language
-Development necessary even in an artificial language,
INDEX,
a
475
503
LIST OF ABBREVIATIONS.
A.S. =
Anglo-Saxon.
Andr. Volkset. = Andresen, über deutsche volksetymologie. 4th edition.
Heilbronn, 1883.
Andr. Spr. = Andresen, Sprachgebrauch und sprachrichtigkeit im
deutschen. 3d edition. Heilbronn, 1883.
=
Delbrück SF.
Diez. =
Draeg. or Draeger
Delbrück, Syntaktische forschungen.
Diez, Grammatik der romanischen sprachen. 4th
edition.
Draeger, Historische syntax der lateinischen sprache.
2d edition.
DWb. = Deutsches Wörterbuch von Jac. und Wilh. Grimm.
GOE.
Goethe.
I.E. Indo-European (the Indo-Germanic of the author).
LE. =
Lu.
M.H.G.
Madvig, Kl. schr. =
Mätzner engl.
Michaelis
=
Lessing.
Luther.
Middle High German.
Madvig, Kleine schriften.
Mätzner, Englische grammatik. 2d edition.
Mätzner franz. = Mätzner, Syntax der neufranzösischen sprache.
Caroline Michaelis, Romanische wortschöpfung.
Morphologische untersuchungen auf dem gebiete der
indogermanischen sprachen von Osthoff und
Brugmann.
Morph. Unt. =
N.H.G.
New High German.
O.F. Old French.
=
O.H.G.
Old High German.
O.N.
Old Norse.
SCHI.
SH.
Schiller.
Shakespeare.
Steinthal, Haupttyp. = Steinthal, Characteristik der haupttypen des mensch-
(or Typen.)
lichen sprachbaus.
Wegener = Wegener, Untersuchungen über die grundfragen
Ziemer =
des sprachlebens. Halle, 1885.
Ziemer, Junggrammatische streifzüge im gebiete
der syntax. Colberg, 1882.*
Ziemer, Comp. = Ziemer, Vergleichende syntax der indogermanischen
Zschr. f. Völkerps.
comparation. Berlin, 1884.
Zeitschrift für Völkerpsychologie, herausg. von
Lazarus und Steinthal.
ERRATA.
P. 382, line 2, delete there is a tendency.
P. 414, line 15, for eykoteîs read kykoteîs.
L
INTRODUCTION.
a general
principles
with the
language.
ANGUAGE, like every other production of human culture, Necessity of
falls under the cognisance of history; but the history of doctrine of
language like every other branch of the science of history has, side by side
running parallel with it, a science, which occupies itself with the history of
general conditions of the existence of the object historically developing,
and investigates the nature and operations of the elements which
throughout all change remain constant. This science lacks as yet
a title generally applicable and admittedly suitable. The name
'philosophy of language' implies something rather different.
/
And in any case there seem to be good grounds for preferring
to avoid this expression. Our unphilosophical age readily sus-
pects under such a title metaphysical speculations of which the
historical investigation of language needs to take no count, The
truth is that the science of which we are thinking is philosophy
in the same way as physics or physiology is philosophy, neither
more nor less. Least of all are we justified in opposing the
historical portion, as empirical, to this general portion of the
science of language. The one portion is just as empirical as the
other.
function
It is very seldom that a knowledge of the laws of any Its special
single simple experimental science suffices to enable us to under- defined.
stand the process of historical development: it is rather in the
essence of all historical movement, especially where this is con-
nected with any department of human culture, that very numerous
xxii
INTRODUCTION.
forces come simultaneously into play-forces which act and re-
act on each other; and it is the task of very different sciences
to investigate their essence. It is thus natural that any such
general science, standing as an exact correlative to any given
historical science, is unable to present such an isolated entirety
as the so-called exact sciences, like mathematics or psychology.
It is truer to assert that it forms a conglomeration of different
pure exact sciences, or, as a rule, of different segments of such
sciences. It may be that we shall be disposed to suspect any
such compound result which always bears the stamp of chance
upon it, and hesitate to apply to it the name of science at all./
But whatever we may think of it, the study of history undeniably
demands, as an indispensable auxiliary, that such various elements
should be dealt with as a whole; if it can dispense with inde-
pendent investigation, it cannot dispense with the appropriation
of results attained by others. / But it would be a great mistake
to suppose that the mere setting together of fragments of different
sciences would result in the kind of science which we have in
view. On the contrary, tasks are imposed upon it with which
the exact sciences which it employs as auxiliaries do not concern
themselves. The exact sciences no doubt compare the single
processes, disregarding, however, their temporal relation to each
other, and merely caring to discover where they agree and where
they differ; and by their aid to find what is unchanging and
constant amid every change. The conception of development
is absolutely strange to them—nay, it seems irreconcilable with
their principles; and they thus stand in sharp antithesis to the
historical sciences. To bridge over this antithesis some method
of treatment seems required which would deserve the name of
historical philosophy more truly than what we commonly denomi-
nate as such. We will, however, abide by custom, and prefer
C
INTRODUCTION.
xxiii
to avoid the word 'philosophy,' employing the phrase 'science
of principles' instead. This science has to solve the difficult
problem: How, under the assumption of constant forces and
relations, is a historical development still possible, or a progress
from the simplest and most primitive to the most complicated
groups? Its method of procedure differs in still another im-
portant way from that of the exact sciences, as I have already
hinted. The latter make it their object to follow nature, and
to endeavour perpetually to isolate the operation of each single
force from the general drift of forces, so as to take cognisance
of each in its simple nature, and then by co-ordinating similar
with similar to build up a system. The historical doctrine of
principles, on the other hand, has to investigate exactly those
points in which the single forces mutually interfere, and to inquire
how even the forces most differing in nature, about whose reci-
procal relations the exact sciences hardly concern themselves at
all, are able to steer to a common goal by means of perpetual
reciprocal operation. It is evident that in order to understand
the interpenetration of the manifold factors we ought to have as
clear views as possible about the single forces which are in opera-
tion, and of the nature of their operations. Analysis must precede
synthesis. For as long as we reckon with unsolved complications
we cannot be held to have arrived as yet at a scientific way of
treating our material. It is thus clear that the science of prin-
ciples, in the sense in which we employ the phrase, rests on the
base of the experimental exact sciences, under which head I
also reckon of course psychology; it contains, however, an im-
portant addition which entitles us to assign it an independent
position by the side of the former. ___
This large science falls naturally into as many branches as
there are branches of special history, the word 'history' being
xxiv
INTRODUCTION.
The doctrine
of principles
the doctrine
of method.
used in its widest sense, and not confined to the development of
the human race. It is presumable a priori that certain general
conditions of fundamental importance will be found to constitute
the necessary basis for every kind of historical development; it
is, however, more certain still that the development of each object
must be in a special manner conditioned by its particular nature.
Whoever undertakes to lay down the principles of any historical
course of study must never lose sight of the other, especially the
most closely allied branches of the science of history, so as thus
to grasp the most general leading features, and not lose sight
of them again. On the other hand, however, he must beware of
losing himself in mere generalities, and thereby failing to notice
the special application to the special case. He must beware, too,
of transferring metaphorically the results attained in different
departments, a process whereby the real circumstances which
form the strict object of investigation are merely hidden.
It is not until such sciences of principles are founded that the
the basis of special investigation of history finds its true value. Not till then
does historical research rise above the mere process of stringing
together apparently accidental dates, and in the general applica-
bility of its results approach the exact sciences, which would fain
dispute the validity of its title. Now if the science of principles
appears to be the highest goal to which all the endeavours of
special sciences are directed, we must remember, on the other
hand, that the former is the indispensable guide of the latter, with-
out whose aid it cannot advance a step with certainty beyond what
is simply laid down, and which never appears in any other way
than, on the one hand, as fragmentary, on the other hand, in con-
fused confiscations which have previously to be solved. The
effectual scrutiny of the conditions of historical growth, taken in
conjunction with general logic, gives at the same time the basis for
INTRODUCTION.
XXV
the doctrine of method which has to be followed in the verification of
each single fact.
to the science
or civilisa-
treatment
the natural
It cannot be maintained that, up to the present time, a spirit of Application
equal earnestness and equal thoroughness has been displayed with of culture,
respect to questions of principle in all the departments of historical tion, of the
research. This spirit has been displayed in a far greater degree common in
with respect to the historical branches of natural science than to sciences.
that of the history of culture. One main reason for this difference
is that the difficulties that present themselves in the latter case are
much more serious than in the former. It has, as a rule, to deal
with far more complicated factors, and the confused thread of
these as long as it remains unravelled renders an exact knowledge
of the causal connexion an impossibility. We have further to
remember that its most important basis, viz., experimental psycho-
logy, is a science of very recent date, and has only lately begun to
be brought into any kind of relationship with history. On the
other hand, however, in the same proportion as the difficulty has
shown itself a greater one, the necessity has become less, or at any
rate less sensible. The history of the human race has always
regarded data as to facts proceeding in each case from cotem-
porary witnesses (though it may be arrived at through several
intermediate links) as its proper source, and has only regarded as
of secondary importance memorials, the products of human
culture, which have approximately maintained the form given to
them by the latter. In fact we even hear the expressions a 'his-
torical' and a 'prehistorical' date employed, and the limit is fixed
by the commencement of historical tradition. For the former,
therefore, the picture of a historical development is already given,
distorted as this picture may be; and it is easily intelligible that
science may deem itself fully occupied with a critical process of
correcting this picture, and even proceed deliberately to reject all
‘his-
?
xxvi
INTRODUCTION.
speculation which ventures beyond it. The question is very dif
ferent with the prehistoric period of human culture, and even with
the history of the development of nature, organic and inorganic,
which reaches back to epochs infinitely more remote. In this
case, too, hardly the least historical element is given as such.
All attempts at a historical apprehension-with the exception of
the scanty traditions handed down by older times depend
upon inferences alone. And there is absolutely no result to be
attained without confronting the questions of principle, and fixing
definitely the general conditions of historical growth. These
questions of principle have therefore always stood in the very
centre of research; they have ever been the rallying point in the
conflict of opinions. At present the domain of organic nature is
the scene of the fiercest struggle; and it must be acknowledged
that it is here that the thoughts most fruitful for the understanding
of all spiritual development, not excluding that of the human race,
have first attained to a certain lucidity.
The tendency of science is at present apparently to extend this
speculative method of regarding subjects to the history of culture
as well, and we are persuaded that this tendency will assert itself
more and more in spite of all opposition, active and passive, that
may be brought to bear against it. We have already fully con-
ceded that such a method is not so indispensable a need for the
science of culture as it is for natural science, and that for the
former we have no right to expect any such far-reaching results as
for the latter. But this does not exempt us from the duty of
testing exactly what results we can attain; and even the possi-
bility of a negative result of this testing process does not prevent
the exact definition of the limits of our knowledge attainable from
possessing what may turn out under certain circumstances of
considerable value. Besides, we have as yet no cause to despair
INTRODUCTION.
xxvii
of actually attaining positive results, at least for certain depart-
ments of knowledge. In any case we should be wholly unjustified
in thinking lightly of the methodological gain accruing from a dis-
tinct statement of the questions of principle. We merely deceive
ourselves, if we seek to prove the most simple fact in history
without some accretion of speculation. Indeed we do speculate,
even unconsciously, and we have to set it down to a fortunate
instinct if we hit the right mark. We may, very probably, main-
tain that hitherto even the methods of historical research in vogue
have been found out rather by instinct than by any many-sided
reflection penetrating the inmost essence of things.
natural result of this is that a quantity of personal fancies obtrude
themselves, giving rise to an endless strife of opinions and schools.
There is only one way out of the difficulty: we must earnestly
apply ourselves to carry these methods back to the first funda-
mental principles, and cast aside all which cannot be deduced
from these. Now these principles, as far as they are not actually
in the nature of logic, result precisely from the examination of the
essence of historical development.
And the
capable
historical
perfect
There is no branch of culture in which the conditions of its The science
of language
development permit of being apprehended with such exactness as the most
that of the science of language, and it follows that there is no among the
science of culture whose method can be brought to such a degree sciences of
of perfection as that of the science of language. No other has as method.
yet been able to transcend so far the limits of tradition; no other
has proportionately proceeded at once speculatively and con-
structively. It is mainly owing to this peculiarity that it appears
to be closely related to the historical natural sciences; and this
has led to the misdirected attempt at excluding them from the
circle of the sciences of culture. In spite of this position which
the science of language has occupied from its very foundation, we
xxviii
INTRODUCTION.
Combination
of psychical
seem far removed from the time when we can say that its method
is worked out to the degree of perfection of which it is capable.
Even now a tendency has set in which makes for a thorough
subversion of that method. In the course of the arguments which
have been brought forward on both sides, it has clearly come out
how great is the confusion in the minds of many investigators of
language, even as to the elements of which their science is com-
posed. It is precisely these arguments which have immediately
prompted this treatise. Its object is to contribute its best to
infuse lucidity into these views, and to aim at attaining an under-
standing at least among those who bring to their task an un-
prejudiced sense for truth. It is to this end of main importance to
set forth the conditions of the life of language, scrutinising them
from as many sides as possible, and thus to draw the fundamental
lines for a general theory of the development of language.
We divide the historical sciences, taken in their widest sense,
and physical into the two main groups of historical natural sciences and sciences
all culture. of culture. The characteristic mark of culture must be defined as
elements in
the presence of the activity of psychical factors. This seems to
me to be the single possible delimitation of its area as against the
objects of natural science pure and simple. In the same way we
must be prepared, doubtless, to recognise a certain culture in the
animal world, and we must reckon the history of the development
of the art-impulses and of social organisation among animals as
belonging to the sciences of culture. Such a course can only be
beneficial to the right appreciation of these facts.
The psychical element is the most essential factor in all movements
of culture: everything turns on it: and it follows that psychology is
the most important foundation of the whole science of culture taken
in its highest sense. It does not, however, follow that the psychical
INTRODUCTION.
xxix
element is the only factor: no culture is possible on a purely psychical
basis: and hence it seems, to say the least of it, very inaccurate to
define the sciences of culture as mental sciences. The truth is that
there is only one pure mental science, that is, Psychology regarded
as an exact science. As soon as we enter the area of historical
development we have to deal with physical side by side with
psychical forces. The human mind must always work in harmony.
with the human body and with its environing nature in order to
bring forth any product of culture; and the secret of its growth,
the way in which it comes to its completion, depends upon
physical no less than on psychical conditions; and both these sets
of conditions must necessarily be known in order to gain a perfect
appreciation of historical growth. A necessity is thus imposed of
mastering not merely psychology, but also the laws according to
which the physical factors of culture move. The natural sciences
besides, and mathematics, are a necessary foundation for the
sciences of culture. If we are not commonly conscious of this, the
reason has to be sought in the fact that we generally speaking
content ourselves with an unscientific observation of daily life, and
hence it fares ill with what we commonly understand under the
name of history. The psychical side has fared in much the same
way, and notably up to the most recent times. But it is incon-
ceivable that, without the aid of a number of experiences made as
to the physical possibility or impossibility of a process occurring,
any one should be in a position to understand any event of history
or to practise any kind of historical criticism. It seems therefore
to follow that the main task for the doctrine of the principles of the
Science of Culture is to expound the general conditions under which
the psychical and physical factors, following their own special laws,
succeed in co-operating for a common purpose.
The task of the doctrine of principles presents itself in a some-
XXX-
INTRODUCTION.
V
of culture
is always
a social
science.
The science what different light from the following point of view. The science
of culture is always a social science. Not till society is formed is
culture possible; society gives the first impulse to make man a
historical being. It is no doubt true that a human mind in a
state of entire isolation has an historical development of its own,
and this too with respect to the relation to its body and its
environment; but even the most gifted human mind could only
succeed in arriving at a very primitive degree of development;
and this would be cut short by his death. It is not until what an
individual has gained has been transferred to other individuals,
and till several individuals co-operate to the same end, that a
growth beyond these narrow limits is possible. Not merely the
art of agriculture, but every kind of culture depends upon the
principle of the division of labour and upon co-operation. The
most special task imposed upon the doctrine of principles regarded
as a social science—the task whereby it maintains its independ-
ence as against the exact sciences which lie at its base—seems
to be that of showing how the single individual is related to the
community; receiving and giving; defined by the community and
defining it in turn; and how the younger generation enters on
the heritage of the elder.
For the rest, the history of the development of organic nature
approaches very closely the history of culture. Every higher
organisation comes to perfection by the agency of a quantity of
cells which co-operate according to the principle of division of
labour, and are differentiated in their configuration according to
this principle. But this principle is active even within the single
cell, the most elementary organic formation; and it is owing to this
that maintenance of the form is found possible amid the change
of material. Each organism breaks up sooner or later; but each
may leave behind it remnants detached from its own essence, in
INTRODUCTION.
xxxi
which the formative principle, to which it owed its own formation,
actively operates, and to which every step forward which it has suc-
ceeded in making in its own formation contributes; always assum-
ing that no disturbing influences coming from without interfere.
(
Lazarus and
Völker-
It might seem as if the doctrine of the principles of Social Criticism of
Science as set forth by us were identical with what Lazarus and Steinthal's
Steinthal denominate Völkerpsychologie,' which they seek to psychologie.
represent in their journal. But the two are far from coinciding.
From our remarks already made it is quite evident that our know-
ledge has to interest itself to a great extent in what does not
belong to psychology at all. We are able to bring the effects
which the individual experiences from society, and which he on
his side, in conjunction with the rest, makes felt, into four main
categories. In the first place, psychical groups, or collections of
ideas are called into being in his mind to which he would never
have attained at all, or at all events much less readily if others had
not prepared the way for him. In the second place, he learns to
carry out certain appropriate movements with the different parts
of his body, which eventually serve to set in motion foreign bodies
or tools: it holds true in the case of these also that without the
example given him by others he would have been much slower in
learning them, and perhaps never have mastered them at all. We
thus find ourselves upon physiological but at the same time upon
psychological ground. Movement in itself is physiological; but
the attainment of the necessary power to control it at will, which
is the point here concerned, involves the co-operation of psychical
factors. In the third place, natural objects worked by the aid of
the human body, or merely transferred from the spot where they
were produced with the view of serving some practical purpose
and thus becoming tools or capital, are transferred from one
individual to another, from an elder to a younger generation; and
xxxii
INTRODUCTION.
there is a common participation of different individuals which takes
place in the process of working or shifting these objects. In the
fourth place, individuals exert a moral force on each other which
may undoubtedly operate as much to the disadvantage as to the
advantage of progress, but which is inseparable from the nature
of culture.
Of these four categories the first is at all events the only
one in which 'völkerpsychologie,' in the sense in which it is
used by Lazarus and Steinthal, is interested. It is therefore
conceivable that the part of our doctrine of principles, which
has reference to this first category, might be supposed to fairly
tally with this. But setting aside the fact that this category can-
not be regarded merely as isolated from the rest, it remains true
that what I have in my mind is very different from what Lazarus
and Steinthal, in the introduction to their periodical (vol. i. p. 1,
73), lay down as the object of 'völkerpsychologie.'
Much as I appreciate the important services rendered by both
these writers to psychology, and especially to the psychological
method of treating history, still it appears to me that the definitions
proposed in this introduction are untenable, in part misleading, and
tending to conceal the real circumstances. The fundamental
thought pervading the whole work is that popular psychology bears
exactly the same relation to the single people, and to humanity
as a whole, as that which we denominate by the simple name of
psychology bears to the individual. Now precisely this thought
seems to me to be based on a series of logical confusions; and I
am compelled to see the reason of these confusions in the fact
that the fundamental difference between an exact and a historical
science has not been held fast, but the two are continually
invading each other's domain.
1
1 It is no doubt true that this difference is hinted at on p. 25 sqq., where a distinction
INTRODUCTION.
xxxiii
The conception of popular psychology itself fluctuates between
two essentially different ways of viewing the subject. On the
one hand, it is regarded as the doctrine of the general conditions
of mental life in society; on the other hand, it is regarded as
the characteristic of the mental peculiarity of different nations,
and as an inquiry into the reasons from which this peculiarity
took its origin. On p. 25 sqq. these two views of science are
presented as two portions of universal science, of which the first
forms the synthetic foundation of the second. Now according
to neither of these views does popular psychology stand in the
assumed relation to individual psychology.
If we turn, in the first place, to the second view, there can
is drawn between the 'synthetic or rational' and the 'descriptive' studies of natural
science, and a corresponding division of popular psychology is attempted. But complete
confusion reigns, e.g. on p. 15 sqq. From the fact that there exist only two forms of all
being and growth, viz., nature and spirit, the authors conclude that there can be found
only two classes of real sciences-one having for its object nature, and the other spirit.
In this, therefore, no account is taken of the circumstance that it is possible for sciences
to exist whose task it is to examine the reciprocal effects of nature and spirit. It seems
even more questionable when they proceed. . . 'Accordingly we find opposed to each
other natural history on the one hand, and the history of humanity on the other.' In
this case, in the first place, 'history' must be taken in a sense very different from that
which we commonly connect with the word, viz., as the science of events and processes.
But how does the 'humanity' come to stand all at once in the place of 'spirit'? The
contents of the two words are far from tallying. Further, it is laid down as a distinction.
between nature and spirit that nature moves in a perpetual round of her processes
according to fixed laws, in which the different courses remain isolated and independent,
in which it was always the past that was repeated, and nothing new could arise; while
spirit, it is said, lives in a series of interconnected creations, and displays progress. This
distinction, thus abstractly stated, is unquestionably inadequate. Nature as well as
spirit-organic nature at any rate-moves in a series of interconnected creations; and in
nature there is such a thing as progress. On the other side it is maintained by the
authors that spirit moves in one regulated course, in a perpetual repetition of the same
fundamental processes. Now here we have a confusion of two contrasts which must be
kept completely apart: that between nature and spirit on the one hand, and between
procedure according to fixed law and historical development, on the other. We can
only ascribe it to this confusion that the authors have been able at all to call in question
whether psychology is to be accounted a branch of the science of nature or of spirit, and
that they finally decide on giving it a place between the two. This confusion is, to be
sure, the traditional one; but it is time that we broke away from it after the instances of
progress made by psychology on the one hand, and the science of organic nature on the
other.
C
xxxiv
INTRODUCTION.
surely be nothing but the characteristics of different individuals to
answer to the characteristics of different nations. This, however,
is not called psychology at all. Psychology has never anything
to do with the concrete shaping of any individual human spirit,
but with the general conditions of spiritual processes. What
then justifies us in employing the name of this knowledge for
the description of the concrete form assumed by the mental
peculiarity of a people? What the authors are thinking of is
nothing but a portion, and no doubt the most important portion,
but one that admits of no isolation, of what has been elsewhere
called the history of culture or 'philology,' placed however on a
psychological basis, a demand which must necessarily be made at
the present day for all researches based upon the history of culture.
It is, however, no exact science like psychology, and no doctrine
of principles, or, to employ the expression of the authors, no
synthetical basis of the history of culture.
The incorrect parallels thus drawn have led to further con-
sequences not easily admitted. 'Popular psychology,' say the
authors, ‘has to deal with the mind of the entire community, which
is again different from all the different minds which belong to it,
and which sways them all' (p. 5). Further, we read (p. 11): 'The
facts which fall under the view of popular psychology exist partly
in the popular mind, conceived as a unity, and in the relations
between its several elements (as, for instance, religion and art, the
state and morality, language and intelligence, etc.), and partly in
the relations between the single minds which make up the people.
Here, therefore, the same fundamental processes appear as in
individual psychology, only in a more complicated and extensive
way.' Surely this is to conceal the true nature of the processes
involved by assuming a series of abstractions. All psychical
processes come to their fulfilment in individual minds, and no-
INTRODUCTION.
XXXV
where else. Neither the popular mind, nor elements of it, such
as art, religion, etc., have any concrete existence, and therefore
nothing can come to pass in them and between them. Away,
then, with these abstractions! For 'away with all abstractions!
must be our last word if we wish to attempt in any place to define
the factors of that which actually happens.1
I will not very
seriously reproach the authors with a fault often met with at
every step of science, and which even the most careful and pro-
found thinkers do not always avoid. Many an inquirer who feels
himself at the high level of the nineteenth century smiles with
lofty contempt at the strife of the nominalists and realists of the
middle ages, and fails to understand how people can have been
led to explain the abstractions of the human understanding as
actually existing things. But the race of unconscious realists
is far from extinct, even among the students of nature.
among our students of civilisation they flourish and thrive most
markedly, and especially among the very class which piques itself
on talking in Darwinian metaphors. But quite apart from this
abuse, the epoch of scholasticism-indeed the epoch of mythology
-is far from lying as far behind us as seems to be thought; our
sense is too much entangled in the toils of both, because they
control our language, which cannot break loose from them. The
man who fails to employ the necessary effort of thought to liberate
himself from the despotism of the mere word will never rise to
an unprejudiced contemplation of things as they are. Psychology
became a science at the moment when she ceased to acknowledge
And
1 Misteli, Zeitschrift für Völkerpsychologie, xiii. 385, has, curiously enough, so mis-
apprehended me as to suppose that I would hear of no abstractions at all being made.
Of course I mean merely that no abstractions must be allowed to interpose an obstruction
between the eye of the observer and the actual things, so as to prevent him from grasping
the connexion of cause and effect among the latter. The instruction which he imparts to
me on the value of abstracting is therefore just as superfluous as his critical remark there-
upon, to the effect that I actually make more extensive abstractions than others.
xxxvi
INTRODUCTION.
1
The natural
influence of
only possible
mediation.
the abstractions of mental powers as something really existing.
Thus, perhaps, it will come into our power to gain important
conquests in many domains, solely by setting resolutely aside
abstractions bearing the stamp of realities, which obtrude them-
selves between the eye of the observer and the concrete pheno-
mena, and disturb his vision.
I beg that these remarks may not be regarded as a mere
digression. They refer to what we ourselves have in the following
pages to observe with regard to the development of language, and
what, on the other hand, the exposition of Lazarus and Steinthal
does not permit us to recognise as likely to be of any service. We
pass naturally from this point to the criticism of the first apprecia-
tion of the conception of folk-psychology.
As we, of course, cannot here reckon with a general mind, and
minds with elements of this general mind, it follows that in popular
by physical psychology we cannot possibly have any concern except with
relations existing between single minds. But the assertion that in
this the same fundamental processes display themselves as in
individual psychology is, for the reciprocal operation of these
single minds, only permissible if understood in a certain sense, and
this would require a closer explanation than we can give. At any
1 In spite of this express request, L. Tobler, in the Lit. Blätt. für germ. und rom.
phil. 1881, sp. 122, remarks upon my introduction: 'All these introductory definitions
of conceptions are rather proper to a philosophical review, and exercise no influence upon
the further course of the exposition.' And Misteli, u.s. p. 400, supports him, and thinks
he might have added 'fortunately.' I must confess that it is disappointing to me that
two savants who, in spite of their special knowledge, profess to take interest in general
questions, have so completely failed to recognise the real point on which my whole work
turns. The very essence of it, in my view, is to trace the development of language from
the reciprocal effects which individuals produce on each other. Thus a criticism of the
views of Lazarus and Steinthal, whose fault consists precisely in their inattention to this
reciprocal influence, is closely bound up with the general tendency of my book. Misteli
especially is of the view that my general theoretical positions needed no attention on the
part of the philologist, and that the latter would find sufficient material in the traditional
categories of grammar. He thus gives his approbation to the old dualism of philosophy
and science, which it is our duty at the present day most strenuously to oppose.
INTRODUCTION.
xxxvii
rate it is not the case that the ideas pass beyond the limits of the
single mind, and operate upon the ideas of other minds, as they
operate upon each other within the limits of a single mind. And
just as little can we speak of the entire groups of ideas in
individual minds acting upon each other, as single ideas do in the
mind of an individual. It is rather an axiom of fundamental
significance which we must never lose from sight, that all purely
psychical reciprocal operation comes to its fulfilment in the individual
mind alone. All intercourse of mind with mind is merely indirect,
and is rendered possible merely by physics showing the way. If
we therefore apprehend psychology in Herbart's sense as the
science of the relation borne by ideas to each other, there can
exist only an individual psychology, to which no 'popular
psychology,' or whatever else it may be called, can properly be
opposed.
We find, however, in the exposition of individual psychology, a
second special part added to this general one, which treats of the
history of the development of the more complicated groups of
ideas which, in the process of experience, we find in an essentially
similar way in ourselves, and in the individuals whom we have to
contemplate. No objection can be taken to this, so long as we
only remain conscious of the fundamental contrast which prevails
between the two parts. The second is no longer exact science,
but history. It is easy to see that these more complicated groups.
owe the possibility of their formation to the fact that an individual
lives in society with a number of other individuals. And in order
to penetrate more deeply into the secret of their origin, we must
endeavour to picture clearly to ourselves the different stages which
they have successively passed through in individuals that have
lived before. It is from this point that Lazarus and Steinthal
have evidently arrived at the conception of popular psychology.
xxxviii
INTRODUCTION.
But just as we have no right to give the name 'psychology'
to a historical representation which exhibits the actual course of
this development, so we have none to apply that term to the
science of principles, which sets forth the general conditions
under which such a development is possible. What is psychical
in this development comes to its fulfilment within the individual
mind according to the general laws of individual psychology.
But everything whereby the operation of one individual upon
another is rendered possible, belongs to another domain than that
of psychology.
In speaking of the different stages in the development of the
psychical groups, I have employed the ordinary metaphorical
method of expression. In accordance with our previous explana-
tions, we cannot suppose that an image-formation, as it has taken
shape in the individual mind, can actually be the real foundation
whence the formations in other minds take their rise. The fact is
rather that each mind must start from the beginning.
We can
place nothing already formed in it, but everything must be
created anew in each mind; the primitive ideas by means of
physiological excitations, the groups of intuitions by relations
which the primitive ideas have assumed towards one another
within the mind itself. In order to evoke in one mind a train of
ideas corresponding to one which has taken its rise in another
mind, the mind can do nothing but create by the action of the
motor nerves a physical product, which on its side calls forth the
corresponding ideas, correspondingly associated in the mind of the
other individual by exciting his sensitive nerves. The most im-
portant of the physical products which serve this purpose are
precisely the sounds of language. Besides these, there are the
tones of other kinds, facial expressions, gestures, pictures, etc.
The means by which these physical products are qualified to
INTRODUCTION.
xxxix
!
serve as a medium for transmitting ideas to another individual is
either an inner and direct reference to the ideas in question (as for
instance, a cry of pain, a gesture of passion), or a connexion depend-*
ing on an association of ideas, in which process the idea standing in
direct relation to the physical instrument forms the connecting
link between this and the idea imparted; this is the case with
language.
Now, the matter of ideas, which is the point in question, can
never be created in the mind by this kind of communication. It
must, on the contrary, be already existent in the mind, through
the agency of physiological excitations. The effect of communica-
tion can be no other than this, that certain masses of ideas reposing
in the mind are awakened thereby, and raised to the threshold of
consciousness, by which process under certain circumstances new
connexions between them are created or old ones cemented.
The matter of ideas then as such is incommunicable. All that
we imagine that we know about the ideas of another individual
depends exclusively upon conclusions drawn from our own. We
assume in this statement that the mind of the other individual
stands in the same relationship to the exterior world as our own,
that the same physical impressions produce on it similar ideas to
those in our own, and that these ideas connect themselves in a
similar way. A certain degree of correspondence in mental and
bodily organisation, in surrounding nature, and in the circum-
stances of life, is accordingly the condition precedent for the
possibility of an understanding between different individuals.
The greater the correspondence, the easier the understanding.
Conversely, every divergence in this relation makes not merely
possible but necessary, either a failure to understand, or an im-
perfect understanding.
The power of understanding is carried furthest by those
xl
INTRODUCTION.
Transforma-
tion of
direct associ-
physical means which stand in direct relationship to the ideas
communicated; for it proceeds often from the general points of
correspondence in human nature. On the other hand, where the
relation is an indirect one, it is presupposed that in the different
minds the same association is united, and this presupposes a
correspondence in experience. Accordingly we must presuppose
as an axiom, that all communication among mankind began with
the former kind, and that it only passed over to the second when
this was attained. At the same time, it must be insisted on
that the resources of the former kind are definitely limited, while
in the case of the second, a wide scope for action presents itself,
seeing that when association is voluntary countless combinations
are possible.
If we now ask on what the fact depends that the individual, in
indirect into spite of the fact that it is compelled to create for itself its own
ation. range of ideas, still receives by means of society a defined direc-
tion for its mental development, and a far wider extension than it
could attain to in isolation, we must fix as the essential point on
the conversion of indirect into direct associations. This conversion
fulfils itself within the individual mind; but the result attained is
transferred to other minds, of course owing to physical mechanism
in the way described. The gain then consists in this, that in these
other minds the groups of ideas do not require to make the same.
détour in order to come together, as in the case of the first mind.
This is, therefore, especially a gain in cases where the connexions,
which serve as the immediate links, are of minor importance com-
pared with the connexion finally resulting. It is owing to this
economy of labour and time to which one individual has assisted
another, that the latter is in his turn in a position to employ the
result of this economy to set up a further connexion, for which the
first individual had no time at his disposal.
INTRODUCTION.
xli
With the transference of a connexion converted from an
indirect one into a direct, it does not follow that the movement of
ideas which has originally conduced to the origin of this con-
nexion is transferred as well. If, for instance, the Pythagorean
maxim is transmitted to any one, he does not learn from this
transmission the way in which it was originally discovered. He may
then simply come to a standstill at the direct connexion imparted
to him, or he may, by means of his own creative combination,
arrive at the maxim by the agency of other mathematical maxims
already known to him, in which process he has a much easier task
than the original discoverer. If, however, as is here the case, various
means of arriving at the result are possible, it does not necessarily
follow that he must light upon the same as the original discoverer.
It appears, therefore, that in the course of this important pro-
cess, seeing that the starting and final points of a series of ideas
are transmitted in direct connexion, the connecting links which
originally aided in setting up this connexion must, to a large
extent, be entirely lost for the following generation. This is in
many cases a healthy riddance of useless ballast, whereby the
space necessary for a higher development is secured. But the
difficulty of arriving at a knowledge of the genesis of the ideas is,
of course, very much increased.
After these remarks, which hold good for the entire develop-
ment of culture (the special application of which to the history of
language has to engage our attention at a later time), we will now
endeavour to bring out the most important peculiarities whereby
the science of language differs from the other sciences of culture.
As we take close cognisance of the factors with which it has to
reckon we shall at once be in a position to justify our assertion
that the science of language, of all historical sciences, is able to
give us the surest and most exact results.
xlii
INTRODUCTION.
Peculiar
character-
science of
language.
Every science depending on experience attains greater accuracy
istics of the in proportion as it succeeds in observing, in the phenomena with
which it has to deal, the operation of the single factors when isolated.
It is precisely in this that the difference between the scientific and
the popular method of looking at things lies. Isolation is, of
course, more difficult to effect, the more intricate the complications
are in which the phenomena, as such, are presented. In this
respect we are singularly favoured in respect of language. This
certainly does not hold true if we compare, in our view, the whole
material contents deposited in language. For we, no doubt, find
there that all which has in any way touched the human mind, the
organisation of the body, its environing nature, the entire range of
culture, all the experiences and circumstances of life, have left
behind effects on language, and that language, therefore, when
looked at from this point of view, is dependent upon the most
manifold factors, indeed on every kind of factor imaginable. But
the peculiar task of the science of language is not to observe these
material contents. It can only offer its contribution to this end in
conjunction with all the other sciences of culture. Its task, strictly
speaking, is to observe the relations which this subject-matter of
idea assumes towards definite groups of sound. Thus we find that
of the four categories of social influence given above (p. xxxi.),
only the two first come into consideration. We need also, as of
main importance, only two exact sciences as the basis of the
science of language, viz., psychology and physiology, and of the
latter certain portions only. What we commonly understand by
the physiology of sound or phonetics certainly does not comprise
all the physiological processes which belong to linguistic affinity,
not, for instance, the excitement of the motor nerves whereby the
organs of language are set in motion. Further, acoustics con-
sidered as a branch of physics, as of physiology, will come under
INTRODUCTION.
xliii
its purview.
Acoustic processes, however, are not directly
influenced by physical processes: they are so influenced only
indirectly by the processes connected with the physiology of
sound. They are defined by these to this extent: that after the
impulse once given, their course, generally speaking, receives no
further impulses to deviation, at all events none such as are of
importance for the essence of language. Under these circum-
stances any deeper research into these processes is, at any rate,
not so indispensable as the knowledge of the movements of the
organs of language. I do not mean to assert by this that there
may not be many conclusions to be drawn from acoustics as well.
The relative simplicity of linguistic processes comes out plainly
if we compare with them, say, those of economic science. In this
case we have to deal with the reciprocal operation of the entirety
of physical and psychical factors with which mankind enters into
any relation whatever. The most earnest endeavours will never
succeed in expounding with absolute accuracy the part played by
each single one of these factors in the process.
Another point of great importance is the following. Every
linguistic creation is always the work of one single individual only.
Several no doubt may create similar products, but neither the act
of creation nor the product is affected by that. It never happens.
that several individuals create anything by working together with
united forces and divided functions. This is quite different from
what we see in economic or political sciences. In the develop-
ment of economics and politics it grows ever more difficult to
observe closely the different relations, the more that the union of
forces and the distribution of duties proceeds; and thus the very
simplest relations in these departments are at once less evident
than the linguistic ones. No doubt, so far as a linguistic creation
is transferred to another individual and transformed by him, and
xliv
INTRODUCTION.
'
as this process goes on repeating itself ever anew, a division of
labour and a union of labour is apparent here; and indeed with-
out such, as we have seen, no culture is conceivable. And where
in our transmission a number of intermediate stages are wanting,
there the student of language is also compelled to resolve perplexed
complications arising from the co-operation no less than from the
successive labours of different individuals.
It is further of great importance, from this point of view also,
to remark that linguistic formations are created without pre-
conceived intention, at all events without any intention of
establishing anything lasting, and without any consciousness on
the part of an individual of his creative activity. In this respect,
the formation of language is distinguished especially from all
artistic production. The involuntary character which we here
attribute to linguistic creations is certainly not so generally
recognised, and still needs proving in detail. We must, however,
while considering this characteristic, draw a distinction between
the natural development of language and the artificial which
comes to be formed by means of a conscious directing interference.
Such voluntary efforts are almost exclusively connected with the
foundation of a general language on an area split up into dialects.
In what follows, we must in the first instance put such efforts
entirely out of consideration, in order to acquaint ourselves with
the simple course of natural development, and then and not till.
then we must deal with their operative power under a special
head. This course of procedure is not merely legitimate but
necessary. We should otherwise be acting like a zoologist ör
botanist who, in his endeavours to explain the origin of the animal
and vegetable world of the present day, should in every case
proceed on the assumption of artificial breeding and improvement.
This comparison is indeed in the highest degree appropriate.
INTRODUCTION.
xlv
Just as the cattle-breeder or the gardener can never create any-
thing absolutely at his own will from nothing; as these are con-
fined in all their experiments to an alteration of natural growth
possible only within certain definite limits; just so does an artificial
language rise only on the foundation of a natural one. Just as no
artificial grafting or breeding can neutralise the operation of the
factors which determine the natural development, so no voluntary
regulation can produce similar effects in the department of lan-
guage. These factors, interfere as we may, work constantly and
consistently, and everything which is formed artificially and
adopted into language gives way to the play of their forces.
It would now be natural to show how far the involuntary nature
of linguistic processes facilitates our examination of their nature.
The first conclusion again to be drawn is, that these must be
relatively simple. In the case of each change, a short step only
can be taken. How, indeed, could it be otherwise, seeing that
each change is secured without any previous calculation, and
commonly speaking without any suspicion on the part of the
speaker that he is producing something not produced before? No
doubt we are next called upon to trace the evidence whereby these
processes make themselves felt as closely as possible, step by step.
It follows, however, also from the simplicity of linguistic processes
that there is no room for individual peculiarities to make themselves
prominent therein. The most simple psychical processes are similar
in the case of all individuals, their peculiarities depend only upon
varieties in the combinations of these simple processes. The great
resemblance of all linguistic processes in the most different individuals
is the most essential foundation for an exact scientific knowledge of
these processes.
It is thus then that the process of learning language falls into
an early period of development, in which, generally speaking, in
xlvi
INTRODUCTION.
Scientific
treatment
of language
only possible
all the psychical processes there is very little voluntary effort and
consciousness, and very little individuality displays itself. And
the case is much the same with the period in the development of
the human race which originally created language.
Were not language so completely reared on the basis of the
common properties of human nature, it would not be the fitting
instrument for general communication that it is. Conversely, the
fact that it is so entails the necessary consequence that it rejects
everything of a purely individual character which seeks in any
way to force itself upon it, and that it accepts and retains nothing.
but what is sanctioned by the agreement of a number of individuals
in connexion with each other.
Our position that the involuntary nature of the processes
favours the attainment of an exact scientific knowledge is easily
confirmed from the history of the other branches of culture. The
development of social relations, of law, of religion, of poetry, and
generally speaking of the other arts, exhibits the greater uni-
formity, and makes a stronger impression of natural necessity, in
proportion as the stage of development concerned is the more
primitive. While in these departments voluntary design and
individualism have become increasingly conspicuous, language in
this respect has remained far nearer to the original state of things.
And this circumstance is one more proof that language is the
original basis of all higher mental development in the individual
as in the entire race.
I have briefly to justify my choice of the title 'Principles of
the history of Language.' It has been objected that there is
by a histori- another view of language possible besides the historical.¹ I must
contradict this. What is explained as an unhistorical and still
scientific observation of language is at bottom nothing but one
cal method.
1 Cf. Misteli, z.s., p. 382 sqq.
INTRODUCTION.
xlvii
incompletely historical, through defects partly of the observer,
partly of the material to be observed. As soon as ever we pass
beyond the mere proof of single facts and attempt to grasp the
connexion as a whole, and to comprehend the phenomena, we
come upon historical ground too, though it may be we are not
aware of the fact. No doubt a scientific treatment of language
is possible, not only in cases where different stages of develop-
ment of the same language are before us, but also where the
materials at our disposal occur side by side. The most favourable
case for our purpose is, when several related languages or dialects.
are known to us. In that case it is the task of science not merely
to determine what reciprocally corresponds in the different
languages or dialects, but as far as possible to reconstruct the
fundamental forms and meanings which have not come down to
us from those which have. But in this process comparative
immediately passes into historical observation. But also in cases
where only one definite stage of development of a single dialect is
before us, scientific observation is a possibility up to a certain.
point. How? it may be asked. If we compare, for example, the
different significations of a word with each other, we attempt to
establish which of these is the fundamental one, or to what funda-
mental signification, now obsolete, they point. If, however, we define
a fundamental signification from which the others are derived, we lay
down a historical fact. Or we compare the related forms together,
and deduce them from a common fundamental form. In this case
again, we lay down a historical fact. Indeed, we cannot even
assert that related forms are derived from a common basis with-
out becoming historical. Or we lay down a phonetic sound-shift-
ing (lautwechsel) between related forms and words. If we would
explain these we are necessarily drawn to the conclusion that such
change is the after effect of a phonetic change (lautwandel), in
xlviii
INTRODUCTION.
other words, of a historical process. If we attempt to characterise
the so-called inner form of language in the sense in which it is
employed by Humboldt and Steinthal, we can only do this by
going back to the origin of the forms of expression employed, and
to their fundamental meaning. And so I cannot conceive how any
one can reflect with any advantage on a language without tracing
to some extent the way in which it became historical. The only
element which might still claim exemption from historical obser-
vation would be general reflexions upon the individual employ-
ment of language, and about the relation of the single to the
general usage of language. However, the following pages will
prove that precisely these reflexions are to be closely connected.
with the observation of historical development.
CHAPTER I.
IT
GENERAL OBSERVATIONS ON THE NATURE OF THE
DEVELOPMENT OF LANGUAGE.
the Science
Tis of main importance for the historian that he should clearly Subject of
and precisely appreciate the extent and nature of the subject of Language.
he has undertaken to investigate. This remark seems a truism,
and yet the Science of Language is only now beginning to realise
its truth, and to remedy the neglect of many past years.
2. Historical Grammar took its rise from the older Descriptive
Grammar, and retains even now much from its predecessor. It has
maintained, in the arrangement of its grouping at least, absolutely
the old form. It has merely laid down a series of descriptive
grammars parallel to each other. In fact comparison, and not
explanation of development, is regarded as in the first in-
stance the proper characteristic of the new Science. Indeed
Comparative Grammar, concerning itself as it does with the re-
ciprocal relations of cognate families of speech whose common
source is lost for us, is actually opposed formally to Historical
Grammar, which traces from the starting-point given by tradition
all further development. And even now many philologists and
scholars hardly think that both of these are merely a single science,
with the same task and the same method, only that the relation
between the given material and the combining intelligence takes
different forms. But the same kind of comparison has been applied
A
2
PRINCIPLES OF THE HISTORY OF LANGUAGE.
in the department of Historical Grammar in the narrower sense;
descriptive grammars of different periods have been tacked to-
gether. Practical utility has dictated this method of procedure,
which is necessary for systematic exposition; and this utility will
always, to a certain point, assert its claims. But it cannot be
denied that the entire conception of the development of language
is directly influenced by this method of exposition.
3. Descriptive Grammar has to register the grammatical forms.
and grammatical conditions in use at a given date within a certain
community speaking a common language; to take note, in fact,
of all that can be used by any individual without his being mis-
understood and without his utterances striking him as foreign. Its
contents consist not in facts, but merely in abstractions from
observed facts. If we make such abstractions at different times
within the same linguistic community, we shall find the results
different. It is through comparison that we obtain the certainty
that revolutions in the language have occurred; we discover too,
perhaps, a certain regularity in the reciprocal relations of such
revolutions; but this method sheds absolutely no light on the true
nature of these. The connexion of cause and effect is hidden as
long as we calculate by means of these abstractions only, as if
one had actually taken its rise from the other. For there is no
such thing as a connexion of cause between abstractions; cause
and effect exist only between real objects and facts. As long as
we are content with descriptive grammar in the case of abstrac-
tions, we are far indeed removed from a scientific apprehension of
the life of language.
4. The true object of philological study is rather the entire
sum of the products of the linguistic activity of the entire sum of
individuals in their reciprocal relations. All the groups of sound
ever spoken, heard, or represented, with the associated ideas, whose
1
3
GENERAL OBSERVATIONS.
symbols they were; all the numerous relations entered into by
the elements of speech in the minds of individuals—all these
belong to the history of language, and must, properly speaking, all
be thoroughly apprehended to render a full apprehension of its
development a possibility. It need hardly be said that to solve
such a problem is an impossibility. It is good to state the ideal
aim of a science in all its bareness of outline. By so doing we
become aware of the gulf between our powers and our possibilities.
We learn that we must in many questions content ourselves with
an avowal of ignorance; and that super-acuteness, which imagines
that it can explain the most complicated historical developments
by a few ingenious aperçus, is humbled. But it is for us an inevit-
able necessity to get a general idea of the play of the forces at
work in this huge complex-forces which we must always keep
before our eyes, if we would endeavour to arrange correctly the
few scanty fragments which we can really claim to possess out
of it.
Idea-Groups
all Linguistic
Action.
5. A part—and a part only-of these operating forces is Organisms
visible to us. Speaking and hearing are not the only processes the basis of
bound up with the history of language. Nor, again, does this
history stop at the ideas awakened by language and the language-
pictures which pass through consciousness in unspoken thought.
No; probably the greatest progress made by modern psychology
consists in the recognition of the fact that numerous psychical
processes fulfil themselves unconsciously, and that everything which
has ever been in consciousness remains as an efficacious factor in
unconsciousness. The recognition of this fact is of the greatest
importance for the Science of Language as well, and has been
used by Steinthal on a large scale for the purpose of that science.
All the utterances of linguistic activity flow from this dark chamber
of the unconscious in the mind. All the linguistic material at
4
4
PRINCIPLES OF THE HISTORY OF LANGUAGE.
L
the disposal of any individual is there to be found; in fact, we
might say, even more than he can dispose of under ordinary
circumstances. It is to be found as a highly-complicated psychical
formation, consisting of many groups of ideas, confused and inter-
penetrating with each other. It is no business of ours here to
examine the general laws under which these groups are found.
We may refer our readers for this to Steinthal's Introduction to
Psychology and the Science of Language. We are merely concerned
to set forth their contents and their activity.
6. These groups are the product of all that has entered into
our consciousness whether through listening to the utterances of
another, through our own speaking, or through thought clothed in
the forms of language. Through these groups what has once been
in consciousness can again, under favourable circumstances, be
recalled to consciousness, and also what has been once under-
stood or uttered can again be either understood or uttered. We
must, according to the general rule already laid down, accept as
truth that no idea which has once been introduced by linguistic
activity into consciousness disappears and leaves no traces, though
these traces may often be so weak, that certain special circum-
stances, which maybe never occur at all, are needed to impart to
this idea the conscious faculty of acceptance. Ideas are introduced
in groups into consciousness, and hence as groups remain in
unconsciousness. Ideas awakened by sequences of sound associate
themselves into a series; and ideas called up by movements of
the organs of language associate themselves into a sequence.
Series of sounds associate themselves with series of movements
of the organs of speech. The ideas for which they serve as
symbols associate themselves with both alike; not merely the
ideas of meanings of words, but likewise those of syntactical
relations.
}
GENERAL OBSERVATIONS.
5
7. And not merely do single words, but larger sequences of
sound-nay, entire sentences, associate themselves immediately
with the constituent parts of the thought which they clothe.
These groups, furnished at least originally by the exterior world,
now proceed to organise themselves in the mind of each individual
into far fuller and more complicated combinations, which come to
their fulfilment for the most part unconsciously, and then proceed to
operate unconsciously-which in by far the greater number of cases
never arrive at clear consciousness, and which still effectually operate.
Thus it is that the different uses, in which we have come to be
acquainted with a word or a phrase, associate themselves with each
other. Thus, too, the different cases of the same noun, the different
tenses, moods, and persons of the same verbs, the different deriva-
tives from the same root, associate themselves, thanks to the relation-
ship between their sounds and their meaning; further, all words
of similar functions-e.g. all substantives, all adjectives, all verbs;
further, derivatives formed from different roots with the same suf-
fixes; further, forms of different words with similar functions-
e.g. all plurals, all genitives, all passives, all perfects, all conjunc-
tives, all first persons; further, words similarly inflected-e.g. in
NHG. all weak words as contrasted with all strong words, all
masculines which form their plural by means of umlaut as con-
trasted with those that form it otherwise, also words only partially
inflected, may group themselves in contrast to such as depart more
violently from the regular use; and further, clauses alike in form
or in function similarly associate themselves. And thus there are
besides these a quantity of associations connected by manifold
links-associations of greater or less importance for the life of
language. These associations may one and all rise and operate
without consciousness, and they must by no means be confounded
with grammatical categories, which are the result of conscious
6
PRINCIPLES OF THE HISTORY OF LANGUAGE.
abstraction, though they not unfrequently cover the same
ground.
8. It is not less significant than natural that this organism of
groups of ideas is in a state of perpetual change in each individual.
In the first place, each impetus which receives no accession of
strength by repetition of its impact, or by a fresh introduction
into consciousness, loses force, and this perpetually. In the
next place, every activity of language, hearing, or thought adds
something new. Even in the process of the exact repetition of an
earlier activity certain impulses at least-of the organism already
existing are strengthened. And however rich be the record of
an individual's past activity, still the occasion for something new
to arise is perpetually occurring, at least in the form of new varia-
tions of old elements (irrespective of the fact that something hither-
to unusual may make its appearance in language). In the third
place, the relations of association within the organism are constantly
displaced, by the weakening as by the strengthening of the old
elements, and finally by the addition of new ones. Thus, if the
organism of the adult, when contrasted with the stage of develop-
ment of earliest boyhood, exhibits a comparative degree of stability,
it still remains exposed to many oscillations.
9. Another point equally natural and equally important is the
following: The organism of the group of ideas which depend
on language takes a peculiar development in the case of each
individual, and thus in each takes a peculiar form; should such
organism be compounded of different elements drawn from similar
sources, still these elements will be introduced to the mind in
different sequence, differently grouped, with different intensity, in
some cases more, in others less frequently; and the reciprocal
relation of their force, and thus the manner of their grouping, will
result differently, and this even though we take no account of the
GENERAL OBSERVATIONS.
7
difference in the capacities common to all, and those peculiar to
the individual.
10. The mere consideration of the boundless variability, and of
the peculiar conformation of each single organism, is sufficient
to make us realise the necessity of a boundless variability of
language as a whole, and of a growth of dialectic varieties not
less vast.
Idea-Group
Media of
Historical
ment.
II. The psychical organisms here described are the true media organisms
of historical development. What has been actually spoken has no the true
development. It is misleading to say that one word has arisen Develop
from another word spoken at some previous time. The word-
as a product of our physical organs-disappears, and leaves no
trace when once the organs it has set in motion have returned
to their state of repose. And in the same way the physical
impression on the hearer passes away. If I repeat the same
movement of the organs of speech which I have once made, a
third, a fourth, or a fifth time, there is no physical connexion of
cause between these four similar movements; but they are con-
nected by the psychical organism, and by this alone. In this alone
remains the trace of the past, whereby further procedure of the
kind can be caused in this alone lie the conditions of historical
development.
12. The physical element of language has exclusively the
function of communicating the effects of the single psychical
organisms to each other; but it is for this purpose indispensable,
because, as has already been insisted on in the Introduction, there
is no such thing as a direct influence of mind upon mind. In
itself but a transitory phenomenon, it still, by its co-operation
with the psychical organisms, enables these to leave effects even
after it is past and gone. As its effect ceases with the death of
the individual, the development of a language would naturally be
8
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Requisites
for the
of States of
Language.
confined to the duration of a single generation, were it not that
new individuals gradually supervene in whom new organisms of
language are produced under the influence of those already in
existence. The vehicles of the historical development of a language
always disappear, and are replaced by new ones after a compara-
tively brief space of time. This is a simple truth, but not the less
important, and not the less often overlooked.
13. Let us now consider-the nature of the object being what
Description it is-the task of the historian. He cannot avoid describing states
of language, seeing that he is concerned with large complications
of elements which lie simultaneously near each other. If, how-
ever, this description is ever to become a really useful basis for
historical contemplation, it must attach itself to the real objects—
i.e. the psychical organisms just described. It must give as true a
picture of these as possible; it must not merely give an exhaustive
list of the elements of which they are composed, but must also realise
their relation to each other, their relative strength, the numerous
connexions they have formed with each other, the degree of the
closeness and durability of such connexions; it must, to express
ourselves in more popular fashion, show us how the instinct of
language is setting. To describe the condition of a language
adequately, it would be, strictly speaking, necessary to observe
with full accuracy every individual belonging to one community of
speech, to note the character of such groups of his ideas as
depend upon language, and to compare with each other the results
gained in each individual case. As a matter of fact, we have to
content ourselves with something far less perfect than this-some-
thing falling short of our ideal more or less-always, however
considerably.
14. We are often confined to the observation of some few indi-
viduals—sometimes even to that of a single one; and we are able
1
GENERAL OBSERVATIONS.
9
only partially to become acquainted with the organism of lan-
guage of these few or of this single one. By comparing indi-
vidual organisms of language, we obtain a certain average, by
which the strictly normal part of language—namely, its usage-is
defined. This average is, of course, more infallible the more indi-
viduals studied, and the more fully each can be observed. The
less perfect the observation, the more doubts remain as to what
is to be set down to individual peculiarity and what is common
to all or most. In any case, usage-to the exposition of which
the efforts of the grammarian are almost exclusively directed
-governs. the language of the individual to a certain degree
only; but side by side with this there is always much which
is not so defined-nay, which actually is in direct opposition to
usage.
15. Even in the most favourable cases, the observation of an
organism of language is beset with the greatest difficulties. In no
case can it be observed directly. It is an unconscious something
reposing in the mind. It is cognisable by its effects only-the
single acts of linguistic activity. A picture of the groups of ideas
lying in unconsciousness can only be obtained by the aid of
many inferences.
16. Of the physical phenomena of linguistic activity, the acoustic
are the most readily accessible to observation. Still, no doubt the
results of our aural apprehension are for the most part hard to
gauge and to define, and it is even harder to give an idea of them
except again by means of direct communication to the ear. The
movements of the organs of speech are less immediately accessible
to observation, but are capable of a more exact definition and
description. No proof is needed at the present day that no exact
representation of the sounds of a language exists except that which
teaches us what movements of the organs are necessary to produce
ΙΟ
PRINCIPLES OF THE HISTORY OF LANGUAGE.
them. The ideal of such a manner of representation can only
approximately be obtained when we are able to make observa-
tions on living individuals. When we unfortunately cannot do this,
we must always keep this ideal before our eyes, and endeavour
to realise it as nearly as we can by restoring the living sound
to the best of our powers from the substitute of writing. This
endeavour, however, can only result successfully for one who is
in some degree versed in phonology, and has already made
observations on living languages which he can transfer to dead
languages; and also, further, has conceived a correct idea of the
relation between writing and language. Thus here already we
have a wide field open for power of combination; here already
familiarity with the conditions of the life of the object is an indis-
pensable demand.
17. The psychical side of linguistic activity is, like everything
psychical, to be directly learnt by self-observation only. All
observation of other individuals supplies us in the first instance
with physical facts only. We can only succeed by the aid of
analogical inferences in referring these to psychical facts on the
basis of observations made on our own mind. Thus, self-observa-
tion ever applied anew, and scrupulous analysis of our individual
instinct of language, are indispensable for the training of the
philologist. Then analogical inferences are naturally most easy
in the case of objects which approach most nearly to the actual
ego. Thus the true nature of linguistic activity is more readily
apprehended in the case of the mother-tongue than in any
other. Further, we are naturally in a better position when we
can employ observations on the living individual than when
we are referred to the casual remains of the past. Only in the
case of the living individual can we obtain results free from
every suspicion of adulteration; only in this case can we at
GENERAL OBSERVATIONS.
II
pleasure verify our observations and make methodical experi-
ments.
18. Thus, to describe a linguistic condition, which may give
a thoroughly trustworthy basis for historical investigation,¹ is no
easy task, and may, under certain circumstances, be a very difficult
one. To fulfil this task adequately, we must have clear views
as to the conditions of the life of language, and this is the more
necessary in proportion as the material at our disposal is less full
and less trustworthy, and as the resemblance there is between the
language to be described and the mother-tongue of the describer
is less. We cannot, therefore, be surprised if the ordinary
grammars fall far short of our claims. Our traditional grammatical
categories are most unsatisfactory as a means for enabling us to
realise the way in which the elements of language are grouped.
Our grammatical system is far from being finely enough differ-
entiated to meet the requirements of the psychological groups.
We shall again have frequent occasion to show in detail the
inadequacy of this. It tempts us, besides, to transfer, when the
circumstances are not applicable, what is abstracted from one
language to another. Even if we confine ourselves to the circle
of the Indo-European languages, the employment of the same
grammatical watchword causes numerous discrepancies. The
picture of a particular condition of language is often blurred
when the beholder happens to be acquainted with a language
nearly related to the object of his consideration, or with an older
or more recent stage of its development. The greatest care is, in
this case, necessary to prevent the intrusion of any foreign material.
Historical Philology has fallen markedly short in this very par-
1 What we have claimed as indispensable for Scientific Grammar is no less indispens-
able for Practical Grammar, subject always to the necessary restrictions imposed by the
capacity of the learner; for the aim of practical grammar is nothing but to familiarise
the learner with the instinct of a foreign language.
12
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Change of
Usage
the ordinary
Exercist of
Speech.
ticular; it has transferred wholesale its philological abstractions,
drawn from researches into an older stage of language, into a more
recent one. Thus, for instance, the signification of a word is made
to depend upon its etymology even though all consciousness of
this etymology has long since vanished, and an independent
development of the signification has set in. Thus, in the doctrine
of inflexions, the grammatical divisions applicable to the most
primitive period have been maintained through all later times ;
a proceeding thanks to which the after effects of the original
conditions are doubtless brought out, but not the new psychical
organisation of the groups.
19. When descriptions of different epochs in the life of a
language are drawn up conformably to our requirements, a con-
dition is fulfilled which renders it possible, from the comparison of
the various descriptions, to form an idea of the processes which
have taken place in the past. This will naturally enough be the
more successful the nearer the circumstances which enter into
comparison. But even the slightest variation in usage is com-
monly found to be the result of the co-operation of a series of
single processes which, in great part, or indeed entirely, escape our
observation.
20. Suppose that we now endeavour to answer the question,
results from What is the real cause of the change of usage in language?
Changes produced by the conscious intention of single individuals
are not absolutely excluded [cf. the history of the words gas,
etc.]. Grammarians have endeavoured to reduce written languages
to regularity. The terminology of sciences, arts, and trades is
settled and enriched by teachers, investigators, and discoverers.
Under a despotism, it may happen that the caprice of the monarch
has in a single point prevailed. But it has not in most cases
aimed at an absolutely new creation, but only at the settlement
GENERAL OBSERVATIONS.
13
of some point on which usage had not yet decided; and the
significance of such capricious decisions is as nothing compared
with the slow, involuntary and unconscious changes to which the
usage of language is perpetually exposed. The real reason for
the variability of usage is to be sought only in the normal linguistic
activity. From this all voluntary influence on usage is excluded.
No other purpose operates in this, save that which is directed to the
immediate need of the moment-the intention of rendering one's
wishes and thoughts intelligible to others. For the rest, purpose
plays in the development of language no other part than that
assigned to it by Darwin in the development of organic nature,—
the greater or lesser fitness of the forms which arise is decisive
for its survival or disappearance.
21. If usage is displaced by linguistic activity without any
one's voluntary interference, this, of course, depends on the fact
that usage does not perfectly dominate linguistic activity, but
always leaves a certain measure of individual freedom. The
part played by this individual freedom reacts on the psychical
organism of the speaker, though it acts as well on the organism
of the hearer. The result of the sum of a series of such displace-
ments in the single organisms, assuming that these displace-
ments tend in one direction, is a displacement of usage.
new usage forms from what was in its origin individual only,
and this, in certain cases, succeeds in expelling the other. Side
by side with this we find a quantity of similar displacements
in the single organisms which have no such thorough-going
success, simply because they do not mutually support each
other.
A
22. The result of this is that all the doctrine of the principles of
the history of language centres round the question, What is the
relation between linguistic usage and individual linguistic activity?
14
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Language
develops by
Gradations.
How is the latter defined by the former, and how does this con-
versely react on it?¹
ی او
23. It must, then, be our object to bring under the head of com-
mon categories the different changes of usage which occur in the
development of language, and to examine each single category
according to its growth and the different stages of its development.
To arrive at this goal we must confine ourselves to the cases in
which these stages of development display themselves to us with
the greatest possible fulness and precision. It thus happens that,
generally speaking, modern periods of language afford us the
most useful materials. But even the smallest variation in usage
is already a complicated process, which we do not comprehend
without taking count of individual modifications of usage. Where
ordinary grammar makes it its business to divide or to draw lines
of demarcation, we must exert ourselves to detect every possible
intermediate step and connecting process.
24. A graduated development is possible in every department
of the life of language. This gentle graduation manifests itself on
one side in the modifications undergone by individual languages;
on the other, in the relations borne by individual languages to each
other. The object of this work is to manifest this in detail. It is
well to point out at once in this place, that the single individual
may possibly have, with reference to the linguistic material of the
linguistic community of which he is a member, a relation partly
1 It is hence clear that philology and linguistic science must not define their
several territories in such a way that the one might properly concern itself with merely
the finished results of the other. The only tenable distinction between linguistic science
and philological handling of language would be this, that the former deals with the
general and permanent facts of speech, the latter with their individual application. Yet
the work of an author cannot be properly estimated without a just view of the relation of
his productions to the whole complex of his linguistic perceptions, and of the relation of
this complex to general usage. And conversely, the modification of usage cannot be
understood without a study of individual speech. For the rest I refer to Brugmann, Zum
heutigen Stand der Sprachwissenschaft, p. 1 sq.
GENERAL OBSERVATIONS.
15
active, partly only passive-i.e. he does not himself employ all that
he hears and understands. Besides, we must notice that, of the
linguistic material employed by many individuals conjointly, one
prefers one part, another prefers another. It is on this that the
variation mainly depends, even between the individual languages
which stand nearest to each other, and the possibility of a gradual
displacement of usage.
A
25. The changes in language fulfil themselves in the individual,
partly through his spontaneous activity, by means of speaking
and thinking in the forms of language, and partly through
the influence which each individual receives from others.
change in linguistic usage can hardly be brought about without
the co-operation of both. The individual always remains exposed
to lasting influences from others, even when he has thoroughly
absorbed what is the common usage of language. But the main
period for the exercise of such influence is the time of the first
acquisition-of the learning of language. This is in principle not to
be separated from any other kind of influence; and it, generally
speaking, follows in the same way; and it cannot be said that
there is in the life of any individual a particular point after which
we could reasonably maintain that the power of acquiring language
is excluded. But the difference in different stages is, after all,
very great. It is quite clear that the processes in the course of
learning language are of the highest importance for the explanation
of the variations in the usages of language-that they afford the
weightiest reason for these variations. If we compare two epochs
divided by a long space of time from each other, and say that the
language has changed in such and such points, we are not describing
the true state of matters: the case stands rather thus-the language
has undergone a new creation; and the new creation is not quite
identical with the former, which has now perished and disappeared.
16
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Classification
of the
Changes.
26. In the classification of the changes in the usage of language we
may proceed from various points of view. There is one important
difference of general application which should be noted. The
processes may be either positive or negative-i.e. they consist either
in the creation of what is new, or in the disappearance of what is
old; or, in the third place, they consist in a replacement-i.e. the
disappearance of the old and the appearance of the new are
due to the same act. The last is exclusively the case in sound-
change. The semblance of this displacement shows itself in other
departments as well. The semblance is occasioned by the fact
that the intermediate stages are disregarded, an appreciation of
which would show that in truth we have to deal with a succession
of positive and negative processes. The negative processes always
rest on the fact that there is some creation not new in the
language of the younger generation which was in existence in the
language of previous ones; thus, to speak accurately, it is not
with negative processes, but with the non-occurrence of processes
that we have to deal. But this non-occurrence must be con-
ditioned by the fact that what later disappears altogether has
already become a rarity in the case of the older generation :
a generation, which has a merely passive relation to it, inter-
venes between one with an active relation and one with none
at all.
27. On the other hand, the changes in usage might be classified
with respect to whether sounds or the signification were affected.
Thus we meet at once with processes affecting the sounds without
the signification coming into consideration at all; and we meet
with processes affecting the signification without the sounds being
similarly affected-i.e. we thus obtain the two categories of sound-
change, and of change in signification. Every change in meaning
presupposes that the group of ideas which has reference to the
GENERAL OBSERVATIONS.
17
distribution of sounds is felt as the same; and in the same way
every change of sound presupposes that the signification has
remained unchanged. This does not, of course, exclude the pos-
sibility of change occurring in the sound as well as in the meaning.
But the two processes do not stand in any relation of cause and
effect with each other; the one is in no way caused by the other,
nor are both the result of the same cause. In the case of other
changes the distribution of sounds and signification together
comes into consideration. Under this head we may at once class
the very primitive combination of sound and meaning, which
we may designate as 'original creation.' It is, of course, with this
that the development of language began, and no other processes
are possible excepting on the basis of what this original creation
has produced. Under the same head, however, will come different
processes which have this in common, that the sound-elements of
language already existing enter into new combinations on the basis
of the signification they have received. The most powerful factor
in this process is analogy, which plays indeed a certain part in the
department of pure sound, but finds its main scope where significa-
tion co-operates with it as well.
28. If our method of observation be correctly carried out, it will
follow that its general results will be applicable to all languages
and to all stages of their development; and, indeed, generally
speaking, to the first beginnings of language. The question as to
the origin of language can only be answered on the basis of the
doctrine of principles. There are no other means of answering the
question. We are unable to give a historical picture of the origins
of language on the basis of tradition. The only question that we
can answer is, How was the origin of language possible? This
question can only be satisfactorily answered if we succeed in
deducing the origin of language exclusively from the activity of
B
18
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Grammar
those factors which we still see in activity in the further development
of language. Besides, no tenable contrast can be drawn between
original creation of language and its mere further development.
As soon as the first start has been made, we have language and its
further development. There exist merely graduated differences
between the first origins of language and later epochs.
29. There is one point to which attention must be drawn. In
their opposition to the former customary method of dealing with
language, in which all grammatical relations were simply derived
from logical relations, some have gone so far as to wish to exclude
those logical relations which find no expression in grammatical
form. This is not to be approved of. For, necessary as it is to
mark the difference between logical and grammatical categories,
it is equally necessary on the other side to define the relations of
either to the other. Grammar and logic do not coincide, because
the formation and application of language does not proceed on
the basis of strict logical inquiry, but by the natural and untrained
movement of the groups of ideas, which either follows or not the
agency of laws more or less logical according to the natural or
acquired capacity of each individual. But the linguistic form of
the expression does not always tally with the true movement of
the groups of ideas with their now greater, now less, logical con-
sistency. Besides this, psychological and grammatical categories
do not cover the same ground. It follows from this that the
philologist must keep the two apart; but it does not follow that
in analysing human utterances he is at liberty to disregard psychical
processes which fulfil themselves in speaking and hearing, without,
however, manifesting themselves in linguistic expression.
30. It is not until he has made a full consideration of what does
not indeed actually lie in the elements of which individual utterances
are composed, but which the speaker still has before his mind,
GENERAL OBSERVATIONS.
19
and which is understood by the hearer, that the philologist can
arrive at an acquaintance with the origin and transformations of
linguistic forms of expression. Whoever considers grammatical
forms as merely isolated, without marking their relation to indi-
vidual mental activity, can never arrive at an understanding of the
development of language.
CHAPTER II.
THE DIFFERENTIATION OF LANGUAGE.
Analogies
from Organic
Nature.
IT
T is established past all doubt by Comparative Philology
that several different languages have often developed out
of a single essentially uniform language, and that these again
have not remained uniform, but have split up into a series of
dialects. It would be natural to expect that the observation of
this process would, more than any other possible circumstance,
force on our notice the analogies to be drawn from organic nature.
It is surprising that Darwinian philologists have not specially
thrown themselves into this view. The parallel, when not pressed
beyond fixed limits, is indeed justifiable, and instructive as well.
If we would press this parallel a little, we must compare the
language of the individual-in other words, the entire materials of
language of which he disposes—to the individual animal or plant ;
and the dialects, languages, families of language, etc., with the
species, genera, classes, etc., of the animal and vegetable world.
32. There is another important point in which it is of moment to
recognise the absolute justice of the parallel. The great revolution
through which zoology has passed in modern times depends to
great extent upon the recognition of the fact that nothing ha
real existence except single individuals that the families, ganera,
and classes are nothing but comprehensions and divisions formed
by human understanding, which may result according to individual

THE DIFFERENTIATION OF LANGUAGE.
21
.
caprice; that divisions between species and those between indi-
viduals differ not in essence, but only in degree, In judging of
dialectic differences we must put ourselves on a similar basis: we
have, strictly speaking, to differentiate as many languages as there
are individuals. If we comprehend the languages of a definite
number of individuals in one group, and exclude those of other
individuals as contrasted with this group, we abstract in this
process on each occasion something from certain differences while
we set value upon others. Thus some room is always left for the
arbitrary arrangement. We have no right to start by assuming
that individual languages, speaking generally, must admit of being
brought under a classified system. We must be prepared to find
that into however many groups we may divide languages, there
will always remain a number of individuals which we shall be
uncertain to ascribe to one or to the other of two nearly related
groups. And the same dilemma becomes really formidable when
we try to combine the smaller groups into larger, and to mark
these sharply off from each other. A sharp and well-marked
division is not possible until the community of intercourse has
been broken off for several generations in succession.
33. If, then, we speak of the severing of an originally uniform
language into different dialects, the phrases ill express the true
nature of the process. The truth is that at any given moment
within any given community there are as many dialects spoken.
as there are individuals to speak them; and, what is more, dialects
each having its own historical development, and each being in a
state of perpetual change. Severance into dialects means really
nothing more than the growth of individual variations beyond
a particular limit.
34. Another point in which we may establish a parallel is this
-the development of an individual animal depends on two factors.
22
PRINCIPLES OF THE HISTORY OF LANGUAGE.
On the one side it is conditioned by the nature of its parents,
whereby it receives a definite impulse imparted by inheritance; on
the other side we have to set the chance effects of climate, food,
way of life, etc., to which each individual is exposed in its own
existence. One of these entails an essential resemblance to the
parents; the other the possibility of a certain departure from this
within definite limits. And thus the language of each individual
shapes itself on one side according to the influences of the
languages of the several companions with whom he holds com-
merce, which, from our point of view, we may regard as the
progenitors of his own; on the other side, according to peculi-
arities wholly independent of these, and the special motives
supplied by his spiritual and bodily nature. Here, too, we find
agreement in this respect that the first-named factor is always by
far the more powerful of the two. A violent change in the type is
not manifest till after a lapse of time, and is only produced by the
fact that each modification of the nature of the individual, which
varies from the original drift imparted to it, aids in imparting the
drift to the following generation. It is just so in the history of
language. We may further make this statement about language
as about every animal organism: the lower the grade of develop-
ment, the stronger will be the second factor in comparison with
the first.
35. On the other hand, we must not overlook the great differ-
ences which exist between linguistic and organic generation. In
the latter the direct effect of the parent ceases at a definite point;
after this we have merely the tendency imparted, which operates
afterwards. In the generation of the language of an individual,
the surrounding languages maintain their share till his end, though
it may be true that their influences are most powerful in the earliest
childhood of the language in question, and grow weaker and weaker
THE DIFFERENTIATION OF LANGUAGE.
23
•
the more this grows and gains strength. The generation of an
animal organism is due to an individual or to a pair. In the
generation of the language of an individual, the languages of a
great quantity of other individuals are concerned; in fact, of all
with whom he has in the course of his life come into linguistic
contact, though, of course, in different degrees. And, what makes
the matter more complicated still in this process of generation,
it may happen that the different individual languages may be
active and passive at once in their relations to each other; the
parents may be the children of their own children. Finally, we
must remark that, even if we speak of the language of an in-
dividual, we have to deal not with a concrete being, but with an
abstraction, except when we understand by the expression the
entirety of the groups of ideas combined in the mind, with their
manifold and complicated relations, and depending upon linguistic.
activity.
36. It is by intercourse, and nothing else, that the language of
the individual is generated. The family origin claims our notice.
only as far as it influences the physical and mental capacity of the
single individual. This, as has been remarked, is no doubt a
factor in the form' taken by language, but a very secondary one
as compared with the influences of direct communication.
the Problem.
37. If we start from the undeniable truth that each individual Statement of
has his or her own language, and that each such language has its
own history, the problem is not so much how from a language
essentially uniform different dialects arise-the rise of variations
seems a mere matter of course. The problem which challenges
solution is this: How comes it that while the language of each
individual has its own special history, this degree of agreement—a
certain greater or less-maintains itself within this miscellaneously
constituted group of individuals?
24
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Linguistic
Change and
38. The increase of dialectic varieties depends, of course, on
Differentia the change in linguistic usage. The more violent the change, the
tion.
Varying
States of
Intercourse.
But
greater the opportunity afforded for the growth of variations.
the degree of this growth is not conditioned merely by the violence
of the change, for no change necessarily entails a lasting differen-
tiation; and the circumstances which operate towards the main-
tenance of the agreement, or towards the speedy restoration thereof,
may exist in very different degrees.
39. The life of a language cannot be conceived of apart from
continuous differentiation. If it were conceivable that in one
linguistic centre every individual language were absolutely iden-
tical, in the very next moment the impulse would be given for the
formation of varieties among them. The spontaneous develop-
ment of each individual must strike out a particular path, according
to peculiarities in the natural bent and experiences of the being
who serves as its vehicle, and the circumstances of his life. The
influence exercised by or upon the single individual never extends
to more than a fraction of the whole, and within this fraction we
find important graduated variations. Correspondingly, no doubt,
there is a continuous tendency to cancel the differentiations which
have appeared-a tendency operating thus: departures from a
former usage are either rejected or are transferred to individuals
who have never developed them spontaneously.
This process
of cancelling, however, is never quite perfect. It only becomes
approximately so within a circle in which a continuous inter-
change of communication takes place. The less active the inter-
course the more differences arise, and are maintained. The possi-
bility of differentiation goes still further if all direct communication
has ceased, and there is only an indirect connexion by means of
connecting links.
40. If the activity of intercourse were perfectly uniform in
THE DIFFERENTIATION OF LANGUAGE.
25
degree at all points in one linguistic centre, we should have nothing
but individual languages, of which those which stood in close con-
nexion with each other would differ in each case from each other
but slightly, while between the furthest points of contrast strongly
marked differences might have arisen. It would then be impossible
to comprise a number of individual languages in a single group
which we might set over as a whole, perfect in itself, against any
other such comprehension. Each individual language would
necessarily be apprehended as a connecting link between several
others. But such a state of matters has nowhere existed, and
never does exist in fact. It would be conceivable only if no
natural boundaries existed, no political and religious unions-if,
let us say, the whole people lived in a plain without any large river,
in isolated houses, at approximately identical distances from each
other, without any central place of assembly. Even under these
circumstances the process of grouping into family languages would
take place. But in reality we find either a collective life as in
towns and villages, or perhaps, in the case of nomadic tribes, in
hordes; or, where the system of isolated homesteads prevails, we
find at least smaller or greater political and religious unions with
places of assembly. In mountainous countries single valleys are
more or less completely shut off from each other. Islands are
divided by the sea. Even where no such barriers exist, unculti-
vated stretches of land, woods, moors, swamps, etc., lie between the
single settlements. Individual languages, therefore, are driven to
form groups according to the natural environing circumstances
which determine the relations between them, as well as according
to their political and religious circumstances. These groups are,
comparatively speaking, uniform, and are sharply marked off from
others outside of them. Thus, such groups are in the first instance
formed by the smallest unions-by single townships. Where the
26
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Individual
and Social
Influence.
inhabitants of the place live together, each single individual will
stand nearer to the other than to those who belong to another
place. It is thus possible for a real border to form here-a
boundary-line not obscured by any connecting grades. Under no
other circumstances can variations, at once plainly marked and
enduring, take their rise, such as cannot maintain themselves,
at least permanently, between inhabitants of the same place.
As long, however, as neighbouring places keep up a vigorous
communication with each other, it may also be that no plainly
marked and enduring difference arises between them; in any
case the differences will remain comparatively insignificant. If
we, however, endeavour to group round each local dialect such
local dialects as are in lively intercourse with it, we shall find as
a result a quantity of groups which reciprocally intersect each
other. It is possible that the process of grouping may have a
slightly different result for each single place. Places may be
added or subtracted, and the communication may be to some
extent modified with respect to those which remain.
41. Every change in linguistic usage is a product, on the one
Spontaneity side, of the spontaneous impulses of single individuals, and on
the other of the conditions of intercourse above described. In cases
where a spontaneous tendency is diffused evenly over an entire
linguistic area, affecting the majority of the speakers, it will be found
to affect the generality in a comparatively short time. It may be,
however, that in different divisions this impulse is distributed with
different strength. Under such circumstances, in districts lying far
from each other, and connected by no communication, the cancelling,
as far as it is necessary, must conduce to different results. Between
these the strife will continue to rage, and will not be easily settled,
because one side has a more vigorous influence upon one part, the
other side on the other. This 'no man's land' forms a boundary
THE DIFFERENTIATION OF LANGUAGE.
27
wall through which the influences cannot pass from one side to the
other, or only when so weakened as to be practically ineffective.
42. A border-land like this must be forthcoming, if the continuity
of the communication through the whole linguistic area were uniform
-if no
bars were placed to free communication by distance
in space, by natural obstacles, or by political border-lines. As
the reciprocal influence of areas divided by such barriers is reduced
to a small proportion, distinct limits may also form for dialectic
peculiarities. An entire cessation of intercourse is not necessary
for the purpose.
It need necessarily be only so weak that it
remains ineffectual without a certain degree of spontaneous re-
ciprocity. Thus it is, too, that a dialectic frontier which has
long existed may be gradually again removed if the spontaneous
reciprocity, which was at first wanting, manifests itself later, or if,
again, similar circumstances proceed from different sides.
Indepen-
each Differ-
43. Each linguistic change, and, further, the origin of every Mutual
linguistic peculiarity, has its own special history. The limit which dence of
marks the prevalence of one is not decisive for the limit of the entiation.
other. Were the relative activity of intercourse alone decisive, no
doubt the borders of the different dialectic peculiarities would
necessarily coincide absolutely. But the spontaneous tendencies
to change may distribute themselves in a fashion essentially dif-
ferent, and the result of the reciprocal power of influence must
accommodate itself to this. If, for instance, a linguistic area
divides itself in respect of a dialectical difference into the groups
a and b, it may happen, and, in fact, will often do so, that a
division made in respect of another peculiarity coincides with it;
but it may also happen that a part of a unites with b, or vice
v rsa; or even a part of a and b may conceivably stand in contrast
to another part of a and b.
44. If, therefore, we draw border-lines for all the dialectic
28
PRINCIPLES OF THE HISTORY OF LANGUAGE.
of a Genea-
logical Tree
inadequate.
The Image varieties which may occur in one continuous linguistic area, we get
a very complicated system of manifold lines crossing each other.
A neat partition into main groups, which we might divide again
into so
many sub-groups, is not possible. The simile of a
genealogical tree, commonly adopted to enable learners to appre-
hend the facts, is always inexact. We can only employ this by
voluntarily seizing on certain differences as essential, and by
overlooking others. If the prominent characteristics are actually
seized upon, we cannot perhaps deny that a genealogical tree may
aid us to realise the actual circumstances, only we must beware
of encouraging the delusion that an exact representation of the
facts is thereby given.
45. The difficulty of representing the facts by means of a genea-
logical tree is enhanced if we endeavour, by its means, to give the
chronology of each development, as is natural for a genealogy
to do.
46. Since communication and the power of reciprocally influ-
encing each other, as between neighbouring districts, has not been
abolished by the appearance of certain differences, should changes
set in at a later period, the development may still be a common
one. Thus changes may make their way in an entire linguistic
area after it has been already strongly differentiated, or they may
set in in several parts already characterised by peculiarities. Thus,
for instance, the lengthening of the short root vowels (cf. MHG.
lesen, geben, rěden, etc.) is, in the case of the Low and Middle
German dialects, carried out in an essentially uniform manner,
whilst many older changes have attained a much less extent.
We must keep this always before us in our consideration of the
older linguistic periods, which we are enabled to endeavour to
picture to ourselves by inferences alone. We are too much
accustomed to consider all such changes of an original condition of
THE DIFFERENTIATION OF LANGUAGE.
29
language as pervade an entire area as older than such as are con-
fined to single portions of this area: and from this point of view
we assume something like a common European period, a Slavo-
Germanic, a Slavo-Lettic, an original Germanic, an East and West
German original language. Certainly it is not to be disputed that,
commonly speaking, the wider extent of a linguistic peculiarity
justifies the assumption of its higher age, but nothing more than a
mere assumption. Further, besides the cases in which we can
positively prove it, there will be various others in which the more
widely diffused change is more recent than that which is confined
to a more limited area.
47. It likewise does not always follow that the peculiarities which
seem most striking are necessarily the oldest. The fashion now in
vogue, of dividing German into the three main heads of Upper,
Middle, and Low German, depends upon the conditions of sound-
change. This change probably did not set in before the seventh
century A.D., and continued to operate till the ninth, and indeed in
certain points even longer. But actually before this date striking
differences were in existence which are thrown into obscurity by
the present classification. For instance, under the head of Low
German, three not unessentially different groups are comprised-the
Frisian, the Saxon, and a part of the Franconian; the Franconian
is divided between Low and Middle German.
48. Further, we cannot lay it down as a principle of universal
application, that the groups which have earliest begun to differ-
entiate themselves as against each other must therefore be most
strongly differentiated, or, conversely, that in the case of the most
strongly differentiated groups the differentiation must have begun
earliest. The degree in which communication is maintained may
change. The geographical position of the groups in reference to
each other may alter. Even apart from this, spontaneous agree-
30
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Dialectical
Divisions
ment may cause new changes to make their way past older
limits, while they themselves perhaps find a limit where before
none existed. Or it may be that a special district, whose develop-
ment has been essentially identical with that of an adjoining one,
while differing from all the others, is affected by specially violent
changes, while the district which hitherto has followed the same
track remains with the rest in the earlier stage.
49. Since the levelling effect of communication forbids the
graduated. operation of too violent changes between districts closely situated
which enjoy a regular communication, each small group shows
an intermediate stage between the groups which bound it on
different sides. It is a false idea, which is still very prevalent,
that intermediate stages owe their origin to a secondary contact
between two dialects hitherto separated. We cannot, of course,
maintain that such has never been the case. A transitional stage
may be formed by a group, either by the fact that it represents the
actual connecting link between two distinct formations already
existing in neighbouring groups, or that it represents them side
by side, or by the fact that it has certain dialectic peculiarities in
common with one group and another set with the other. When
the relations of dialects stand thus with respect to each other,
the mutual understanding between neighbouring districts need
never be stayed, because the divergences are too trivial, and
because people accustom themselves to these too easily; and
there may thus nevertheless exist differences between those more
remote which make mutual understanding impossible.
50. This relation can be observed in the most widely differing
languages. It is specially obvious in German. A Swiss finds it
impossible to understand a Holsteiner-nay, even a Hessian or
a Bavarian, and yet he is indirectly connected with these by
unchecked streams of communication. The gradual shading off
THE DIFFERENTIATION OF LANGUAGE.
31
of the German dialects on a large scale may be seen to great
advantage in the so-called High German process of sound-
shifting.¹ The same process of gradual shading off can be easily
observed by a casual examination of Firmenich, Germaniens völker-
stimmen. A clearer idea still of the wonderful multiplicity of
this shading-off process, may be obtained by the Atlas of Lan-
guages prepared by G. Wenker. The circumstances are much
the same not merely within the single Romance languages, but
actually within the entire Romance linguistic territory. The
territories of the separate nations are to be defined with any
accuracy by the written languages, and by these alone, and not by
the dialects. Thus, for instance, the North Italian dialects have
some very important peculiarities in common with the French ;
and are nearer the neighbouring dialects of France than to the
Italian written language or the dialect of Tuscany. The Gascon
dialect again in many respects is the transition stage from the
Provençal (South French) to the Spanish; the Sardinian shows the
transition step from the Italian to the Spanish, etc.
51. This sketch of the development of languages presupposes
that the individuals are settled. Each change of an individual-still
more so, of groups of individuals-produces modifications which we
have to treat in Chapter XXII. as mixtures. The existence of a
written language operates equally as a moderating factor; of which
we shall have to speak in Chapter XXIII.
Languages.
52. The case may also, of course, present itself that the com- Separation of
munication between the several parts of a linguistic community is
completely broken by strongly marked natural or political limits,
by the emigration of one part, by the interposition of a foreign
people, etc. From this moment the language of each single part
develops itself likewise independently, and violent contrasts arise
1 Cf. Braune, Beiträge, i. p. 1 sqq., and Nörrenberg, ib. ix. p. 371 sqq.
32
PRINCIPLES OF THE HISTORY OF LANGUAGE.
without transition stages to connect them. It thus happens that
several independent languages arise out of one, and this process
may repeat itself several times.
53. It is scarcely conceivable that no noteworthy changes should
have occurred through the whole linguistic territory down to the
moment when such a division of a single speech into several has
taken place. It is inconceivable that a language should exist
ranging over a fairly wide district, and having a fairly long
development behind it, without some dialectic variations.
We
must therefore regard, as a rule, the independent languages which
have developed out of a common original language as continua-
tions of the dialects of the original language; and we may assume
that a part of the differences prevailing between them dates
back to the period of their continuous connexion.
The same
would hold true of this part as holds true, generally speaking,
with respect to dialectic differences in an unbroken linguistic
area. It might therefore happen, if we designate the dialects
which have developed into independent languages by the letters
of the alphabet, that a might have something in common with
b in contrast to c and d, something else in common with e in con-
trast to b and d, and again something else with d in contrast to
b and c, etc.; and these agreements might rest upon a real causal
connexion. From this point of view the relations of the Indo-
European families of language, for instance, ought to be regarded.
But in single cases it is hard to decide whether reciprocal influ-
ence has really and truly contributed to the agreement in the
development. The impossibility of a coincidence even in the
case of quite independent development can hardly ever be proved.
54. The separation, further, need not necessarily tally with
ancient dialectic limits, and particularly not when it is caused by
migration. It is possible for a portion of a group which agrees
THE DIFFERENTIATION OF LANGUAGE.
33
in the most essential points to separate, while the other portion
remains in connexion with the other groups which are further
removed from it. It is possible again for portions of different
groups to detach themselves at once. For instance, Anglo-Saxon
is originally closely connected with Frisian,-indeed, it seems.
improbable that it ever existed on the Continent as a separate
dialect; but it seems to have owed its origin to the time when
Frisian hordes broke from their homes, and united with them-
selves certain elements of other Teutonic stocks. But Anglo-
Saxon then received its own special development, while the Frisian
has remained connected with other German dialects. There is a
sharply-defined limit between English and German, while no such
limit exists between Frisian and Low Saxon.
tial Mark of
Sounds.
55. The character of the sounds remains, and always must The Essen-
remain, the characteristic factor in the dialectic distribution of a Dialect its
a district linguistically united. The reason of this is that in their
formation everything depends upon the direct influence gained by
immediate personal intercourse. Vocabulary and signification,
the formal and syntactical parts of language, may be transferred
without difficulty. Whatever has here arisen that is new may
wander far and wide without any essential alteration, supposing
that it strikes a responsive chord. But sound, as we shall see in
the following chapter, is never exactly handed on as it is received.
Where a wide gulf exists, the influencing power in regard to
changes of sound entirely ceases. Here then far more violent
differences develop themselves than in the case of the vocabulary,
inflexions and syntax, and these differences pass more uniformly
through long spaces of time than in the case of sounds. On the
other hand, if a real severance of language has set in, the differences
between the different languages may make themselves felt as
characteristically in other departments as in that of sound.
C
34
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Technical
and Poetic
56. The least characteristic feature in language is the voca-
Language bulary and its application. In this case, more than in any other,
we find elements transferred from one dialect into another, just
as we do from one language into another. In this case there
are more individual variations than in any other. Nay, differences
may actually occur here which have nothing to do with dialectic
differences, and which even cross with these and interpenetrate
them. In each higher stage of culture, technical expressions arise
for the different trades, arts, and sciences, which, mainly or indeed
exclusively employed by one particular professional class, are
hardly understood, or not understood at all, by others. In the
case of the formation of such technical languages, similar relations
are observed as in the case of the rise of dialects. Under this
head will come also the difference between the language of poetry
and that of prose, which likewise extends to what is formal and
syntactical in language. Peculiar circumstances in the case of
ancient Greece actually led to a purposely artificial employment
of differences of sounds. There may, however, be such a thing
as a poetical language (and this is the most common case), which
in the most different dialectical combinations of sound contrasts
uniformly with the language of prose.
57. All natural development of language conduces to a continual
unbounded growth of dialectic differences. The reasons which
impel to this are given with the common conditions of the life of
language, and are completely inseparable from these. The idea
is unfortunately still met with in philological works which
enjoy a high reputation, that the earlier centrifugal movement, by
the operation of which dialects were said to have arisen, was
counteracted in a higher stage of culture in a time of more lively
intercourse, by a backward centripetal impulse. This idea rests
upon imperfect observation. The formation of a common language,
THE DIFFERENTIATION OF LANGUAGE.
35
which writers have in view when using this language, does not
come about by a gradual assimilation of the dialects to each other.
The common language does not proceed from the single dialects
by the same process as that by which a later form of the dialect
arose from a previous one. It is rather a foreign idiom to which
the dialect is sacrificed. On this head more will be said in Chapter
XXIII.
CHAPTER III.
ON SOUND-CHANGE.
The Active
Factors in
the produc-
tion of
Sound.
IN
N order to understand the phenomenon which we usually
designate as sound-change, we must get a clear idea of the
physical and psychical processes which operate in the production
of groups of sound. If we disregard—as in this case we ought, and
indeed must do the function which these subserve, the following
points challenge our consideration :-In the first place, the move-
ments of the organs of language, as originated by the excite-
ment of the motor nerves, and the muscular activity thereby
awakened; secondly, the series of sensations by which these
movements are necessarily accompanied the 'motory sensation'
(bewegungs-gefühl, as Lotze and, following him, Steinthal have
named it); thirdly, the sensations of tone produced in the hearers,
among whom, under normal circumstances, the speaker himself
must be reckoned. These sensations are, of course, not merely
physiological processes, but psychological as well. Even after
the physical excitement has passed away, these sensations leave
a lasting psychical effect, viz., in the shape of memory-pictures,
which are of the greatest importance for sound-change. For these
are the only means of connecting the single physiological pro-
cesses, and these set up a connexion of cause and effect between
the earlier and later production of the same combination of sounds.
The memory-picture left behind by the sensation of the move-
i
ON SOUND-CHANGE.
37'
ments carried out before is that which renders possible the repro-
duction of similar movements. Motory and sound sensations
need necessarily stand in no intimate connexion with each other.
But both enter into an external association, since the speaker hears
himself speaking at the same time. The mere act of listening to
others/gives no motory sensation, and thus gives no capacity of
reproducing the combination of sounds once heard. For this
reason, an effort and a certain amount of practice is necessary, in
order to enable us to reproduce in speaking any sound which we
have not been hitherto accustomed to utter.
4
[
sciousness of
of a Word.
59. The question naturally presents itself- What is the Slight Con-
analysis of the motory and sound sensations respectively, and to the Element
what grade are the special factors in their analysis consciously
perceived? Probably nothing has hindered a correct appreciation
}
of the nature of sound-change/so much as the fact that the extent
and the distinctness of consciousness in this subject has been over-
rated. It is a great mistake to suppose that for the apprehension
of the right sound of any word in its peculiarity—in fact, for the
possibility of an excitement of the ideas bound up with it-the
single sounds composing the word/need come into consciousness
at all. Indeed, it is not always indispensable to the apprehension
of an entire sentence that even the single words composing it
should come into consciousness according to their sound and their
signification. The self-deception under which grammarians labour/
depends on their having regarded the word not as a portion of the
living language-audible for a moment, and then passing away—
but as something independent to be analysed at leisure, with a
view to its leisurely dismemberment. A further source of deception
lies in the habit of starting not from the spoken, but from the
written word. In writing, no doubt, the word seems separated into
its elements, and it may appear requisite that every one who writes
38
PRINCIPLES OF THE HISTORY OF LANGUAGE.
course.
should presuppose this dismemberment. But in real truth the
matter is somewhat different. No doubt when writing was dis-
covered, and each time that it was applied afresh to a language not
hitherto expressed by its aid, such dismemberment must necessarily
have been presupposed. Further, it must continuously happen that
each time that handwriting is learned anew, an exercise in the
spelling of spoken words must go hand in hand with it. But after
a certain facility has been attained, the process in writing is not
exactly in the first place the dismemberment of each word into its
single component sounds, and then for each single sound the setting
down of its proper letter. The speed with which the process is
carried out excludes the possibility of the single factors coming
clearly into consciousness, and demonstrates at the same time that
it is unnecessary that they should do so for a regular and normal
But a really abbreviated process comes in, whereby writing
is to some extent emancipated from language-a process which we
shall on a later occasion have to consider rather more closely. And
if we observe a little more accurately the facts connected with this
dismembering faculty of the man who can write, it will clearly
force itself on our notice how little consciousness intrudes into the
elements of word-sound. We can daily make experience of the
fact that the manifold discrepancies between writing and pro-
nunciation pass to a great extent unheeded by the members of
any given linguistic community, and strike the foreigner first,
though he can give himself no satisfactory explanation of what
these discrepancies repose on. Thus every German who has not
enjoyed a training in the physiology of sound is convinced that he
writes as he speaks. Suppose, however, that he really is justified
in entertaining this conviction as against the Englishman and the
Frenchman, still, to omit niceties, there are plenty of striking
instances in which the pronunciation differs greatly from the
ON SOUND-CHANGE.
39
writing. It occurs to few that the final consonant in tag, feld, and
lieb is a different sound from that caught in tages, feldes, liebes, or
that the n in the German word anger represents a sound essentially
different from that of the n heard in land. In the common pro-
nunciation of ungnade we have a guttural nasal, in that of unbillich
a labial nasal; but no one thinks of this. It excites actual sur-
prise to assert that in the German word lange, g is not heard; that
in the second syllable of legen, reden, ritter, schütteln, e is not heard ;
that the final consonant of leben in the ordinary pronunciation
contains no n, but is anm with no e preceding it. Indeed, it is safe
to assert that most people will dispute these facts, even after their
attention has been drawn to them. This holds true in many cases,
even of good scholars. We see from this how entirely the analysis`
of the word is learnt with the writing, and how small is the con-
sciousness of the actual elements of the spoken word.
continuous
Series of
infinitely
numerous
Sounds.
60. A real analysis of the word into its proper elements is not The Word a
merely extremely difficult, but is actually impossible. A word is
not a united compound of a definite number of independent sounds,
of which each can be expressed by an alphabetical sign; but it is
essentially a continuous series of infinitely numerous sounds, and
alphabetical symbols do no more than bring out certain character-
istic points of this series in an imperfect way. The remainder,
which remains undenoted, no doubt necessarily reveals itself from
the definition of these points, but reveals itself only up to a cer-
tain point. The continuity of sound is seen with the greatest clear-
ness in the case of the so-called diphthongs, which exhibit such a
series of very numerous elements (cf. Sievers' Phonetik, § 19, i. a).
Sievers was the first to expressly bring out the significance of the
transitional sounds. But it follows from this continuity of the
word that an idea of the individual parts cannot be a self-yielded
result, but must be the fruit of scientific reflection, however primi-
40
PRINCIPLES OF THE HISTORY OF LANGUAGE.
tive this may be, and it is the practical need of writing to express
sounds which has conduced to this.
61. What is true of the sound-picture is also naturally true of the
motory sensation. Indeed, we must go further in this point. No one
can maintain that the individual ever has any idea of the different
movements made by his organs in the act of speaking. It is plain
that these can only be ascertained after the most careful scientific
observation, and that scientific men are not agreed upon many
points in connexion with this question. Even the most superficial
and roughest views as to these movements are not possible without
a voluntary habit of careful and protracted observation. They
are superfluous, and not needed to produce sounds and sound-
groups which we are trained to produce. The process seems to
be the following. Each movement excites in a definite manner
certain sensitive nerves, and thus evokes a feeling which associates
itself with the direction of the movement of their centre by means
of the motor nerves. If this association is sufficiently established,
and if the memory-picture left by the feeling is sufficiently strong
-a condition which, as a rule, is not reached without practice.
(ie. without frequent repetition of the same movement, varied, it
may be, with many vain attempts)-then the memory-picture of
the sensation may have power to reproduce the movement asso-
ciated therewith as its reflection; and if the sensation called up
thereby corresponds with the memory-picture, then we may
also rest assured that we have carried out the same movement as
formerly.
62. But we might concede that the degree of consciousness which
the single factors of the sound-picture and of the motory sensation
attain by dint of mastery of writing, and, further, by reflection, was
even greater than it is;—we might concede that an absolutely clear
consciousness of these elements was absolutely necessary for a
ON SOUND-CHANGE.
4I
mastery of the mother-tongue as of any foreign language (and
certainly a higher degree of clear consciousness, is necessary than
in the application of what has been learnt by practice); but from
this it would not follow that the same degree of clearness must be
attained in common daily discourse. It rather lies in the nature
of the psychical organisation that all ideas which originally oper-
ated merely by consciousness' receive by practice the capacity of
operating automatically; and that this automatic operation is the
first and indispensable condition of the speedy course of ideas
demanded in every position of daily life and in language as well. ·
Even the professed student of the physiology of sounds will speak
much and hear much without a single sound revealing itself dis-
tinctly to his consciousness.
63. For the proper judgment, then, of the natural life of lan- ·
guage, regulated by no species of pedantry, we must cling to the
fundamental maxim that sounds are produced and taken cognisance
of without any clear consciousness. This statement contradicts all
such explanatory theories as presuppose in the minds of indi-
viduals) an idea of the sound-system of language; under which
head come several hypotheses as to the German sound-shifting
process [as, for instance, that of Max Müller].
Speech.
64. On the other hand, however, the unconsciousness of the Control of
elements/does not exclude an exact control over them.
We may
utter or hear a group of sounds to which we are accustomed
without ever thinking that it is in fact precisely this group, made
up in such and such a way; but as soon as in a single element a
departure from the usual is observed-which departure needs to
be but very slight it is noticed, unless indeed any extraordinary
obstacles supervene to prevent it; and each departure from the
accustomed unconscious course of ideas naturally forces itself
upon our consciousness. Of course it does not follow that, with
42
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Limits to
Power of
the consciousness of the departure, the consciousness of its nature
and cause is also given.
< 65. The possibility of control extends as far as the power of
distinguish perceiving differences. This, however, is not limitless, while the
ing Sounds.
possibility of gradual transitions in the movements of the organs
of speech, and of course also in the sounds produced thereby, is
certainly limitless. Thus, between a and i, as well as between a
and u, there lies an unbounded number of possible transitions of
vocal sound. In the same way the places of the articulation of all
the lingual-gum sounds lend themselves to representation by the
picture of a continuous line in which each point may be the one
preferred. Between them and the labial sounds certainly such an
imperceptible transition is impossible; still the denti-labials stand
⚫ in close relationship to the denti-linguals (th, f).
the transition from final sound to fricative, and
gradually brought about; for complete closure
possible narrowing process go closely together. All differences of
quantity, of pitch, of energy in articulation, as in expiration, are
conceivable in all possible transitions. And so with much besides.
It is this circumstance specially which renders sound-change
intelligible.
In the same way
vice versa, may be
and the greatest
66. Now, if we reflect that it does not depend merely upon the
differences in those sounds into which we commonly, (though
inaccurately,)divide the word, but also upon the differences in the
transition sounds, in the accent, the time, etc.; and further, if we
reflect that unequal portions may always be brought into combina-
tion with a series of equal portions,-it must then be clear that
a manifold variety is possible in the groups of sound, and this
even in the case where the actual difference is comparatively
small. For this reason it is possible that strikingly different
groups may be still conceived of as essentially identical, thanks to
ON SOUND-CHANGE.
43
their, on the whole, superior resemblances; and this is what renders
an understanding possible betwen speakers of different dialects, so
long as the differences do not pass beyond a certain grade. But
for this very reason a number of variations may set in whose
difference it is hard to note, or indeed impossible, until attention
has been specially drawn to them.
from the
of the Motory
inevitable.
67. Early childhood is for every one an experimental stage, in Deviations
which the individual gradually learns by manifold efforts to Prescriptions
imitate what has been spoken before him by those who surround Sensation
him. When the greatest amount of success has crowned his
efforts/a period of comparative rest ensues. The former important
vacillations cease, and there exists from this time forward a great
uniformity in the pronunciation and freedom from disturbing
causes, unless indeed the evident influences of foreign dialects, or
of a written language, come between. This uniformity, however,
can never become absolute. Less important vacillations in the
pronunciation of the same word in the same place in the sentence
are inevitable. For, speaking generally, in the case of every
movement of the body, however much such movement may be
the result of training, however fully the impulse may have been
developed, there still remains a certain amount of uncertainty; it
still remains left to chance (in a certain degree, however small),
whether the pronunciation be uttered with absolute exactness, or
whether a slight deviation from the correct path towards one side.
or the other manifests itself. Even the most practised marksman
misses his mark sometimes, and would miss it in most cases/if it
were a mere point with no extension, and if his weapon had only
a single point which could touch the goal. Any one's hand-
writing may be as defined and characteristic as you please, and
his general peculiarities may be at once recognisable, still he will
not reproduce, each time he writes, the same letters and the same
44
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Modifica-
tion, or Dis-
the Motory
Sensation.
combinations of letters in absolutely the same way. It must be
the same with the movements whereby sounds are produced.
This variability of pronunciation, which remains unnoticed because
of the narrow limits in which it moves, gives the key to our com-
prehension of the otherwise incomprehensible fact that a change.
of usage in the sounds of a language sets in and comes to its
fulfilment without the least suspicion on the part of those in
whom this change is being carried out.
68. If the motory sensation were always to remain unchanged as
placement, of a memory-picture, the insignificant deviations would always centre.
round the same point with the same maximum of distance. In fact,
however, this sensation is the product of all the earlier impressions
received in the course of carrying out the movement in question,
and, according to a common law, the impressions, not merely
those which are absolutely identical, but also those that are
imperceptibly different from each other, are fused into one. Cor-
respondingly to their difference, the motory sensation must be
somewhat modified with each new impression, to however
insignificant an extent. It is, in this process, of importance that
the later impressions always have a stronger after-influence than
the earlier. It is thus impossible to co-ordinate the sensation with
the average of all the impressions received during the whole.
course of life; rather, the numerically-speaking inferior may, by the
fact of their freshness, outbalance the weight of the more frequent.
It must, however, be observed that supposing the distance of the
possible divergence to remain the same, a displacement of the
limits of this divergence is brought about with each alteration
of the sensation.
69. Let us now take a line in which every point is exactly fixed
as the proper normal path of movement to which the motory
sensation leads; then, of course, the distance from that point
ON SOUND-CHANGE.
45
which is possible as maximum when the movement is really
carried out without conflicting with the sensation, is commonly
as great upon one side as upon the opposite. But it does not
follow from this that the deviations which really set in must be
uniformly divided on either side according to number and
extent.
70. These deviations, which are not defined by the motory sensa-
tion, have, as is natural, their independent causes-causes, moreover,
wholly unconnected with the motory sensation. If such causes act
at the same moment, with exactly the same force, in opposite direc-
tions, then their operations cancel each other, and the movement
is carried out with absolute exactness. This case will occur very
seldom indeed. In by far the most numerous cases the balance
will incline to one side or the other. It is, however, possible for
the relation of the forces to undergo manifold changes according
to circumstances. If this change is as favourable for one side
as for the other; if a deviation towards one side 'always alternates.
with a corresponding deviation towards the other side, in this case
the very smallest displacements of the motory sensation will be
immediately arrested. Matters are, however, very different when
the causes which impel to one side have the preponderance over
those which have an immediately opposite tendency, whether this
be in each particular case or only in the generality. The original
deviation may have been ever so insignificant, the motory sensation
having suffered thereby the slightest possible displacement, still for
the next time a somewhat greater displacement from the original
is rendered possible, and with this coincidently a displacement of
the sensation// There thus gradually arises, by adding together all
the displacements) (which we can hardly imagine small enough),
a notable difference—whether it be that the movement progresses
steadily in a special direction, or that the advance is regularly
46
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Causes of the
Divergence,
1
interrupted by relapses, if only the latter are less frequent and
smaller than the first.
71. The reason why the inclination to deviation is greater on
one side than on the other must be probably sought in the fact
that the deviation towards the side to which it tends/is in some
respect more convenient. The examination of the nature of this
greater or less degree of convenience is a purely physiological task.
It must not, however, be supposed that it is not at the same time
conditioned by psychology. Accent and time, which are of such
decisive significance in the process, and also the energy dis-
played in muscular activity, are essentially dependent on psychical
conditions, but their operation upon sound relations/is neverthe-
less physiological. In the process of progressive assimilation/ it
can be nothing but the idea of the sound yet to be uttered which
operates upon the preceding one; but this is psychical relation of
a very simple kind manifesting itself uniformly throughout, while
all special definition of the process of assimilation must be based
upon an examination of the physical generation of the sounds in
question.
72. For the task which we have set ourselves,/it is sufficient to
point to certain general points of view. There are a great number
of cases in which we may say quite simply, This sound-group is
more convenient than that. Thus the Italian words otto and cattivo
are without any doubt easier to pronounce than the Latin octo, and
the NHG. empfangen than a form *entfangen, unaffected by assimi-
lation, would be. Assimilation, either partial or entire, is a
phenomenon occurring in all languages. When, on the other
hand, the single sounds come into question, hardly any general
principles can be laid down as to the greater or less facility of
pronouncing one rather than another, and all theories on this point
based on abstractions from narrow grounds show themselves worth-
ON SOUND-CHANGE.
47
less when confronted with a fuller experience. And, further, no
perfectly general definitions can be given for the combination of
several sounds. Facility depends to a great extent) upon condi-
tions of quantity and upon the accentuation, expiratory as well
as musical. One sound-group is convenient in the long syllable
another in the short, one in the syllable which bears the stress
another in that which has no stress; while the circumflex makes
other demands than the grave or the acute. But, further, the
measure of convenience adapts itself to a quantity of circumstances
which may be different for each individual, but which may attach
themselves to larger groups as well in the same or in a similar
way without being shared by the others. One point requires
specially emphasising in this case. A certain harmony of the
sound-system is found existing in all languages. We see from
this that the direction in which a sound deviates must be partially
conditioned by the direction taken by the other sounds. Much
depends, as Sievers has shown, in this case on the so-called neutral
position of the organs. Each variation in this entails, of course,
also a variation in relation to the convenience of single sounds.
A gradual displacement of this neutral position will have to be
judged precisely after the analogy of what we have said above
about the similar displacement of the motory sensation.
ence' a
Cause the
Sensation
Cause.
73. It is of great importance never to lose sight of the fact 'Conveni-
that the consideration of convenience in each production of sound Secondary
affords in every case only a very subordinate and secondary cause; Motory
the motory sensation always remains the really decisive motive the Primary
power. One of the commonest errors is the supposition that a
change which has arisen in a long period by numerous small dis-
placements is to be referred to a single act resulting from a desire
for convenience. This error partly results from the method in
which rules for sound 'are apprehended in practical grammar, and
48
PRINCIPLES OF THE HISTORY OF LANGUAGE.
even in grammars which claim to teach on scientific principles.
For instance, it is commonly said that if a sonant consonant
appears as a final, it takes the form in this language of the corre-
sponding surd (cf. MHG. mîde—meit, rîbe—reip), just as if we had
to do with a change occurring each time occasioned by the fact
of the surd being better adapted to the close of the word. The
truth is that it is in this case the motory sensation developed
by tradition which produces the surd, while the gradual reduction
of the voice-tone to absolute annihilation, and strengthening of the
pressure of expiration connected therewith, belong to a period
perhaps long past and gone. It is equally mistaken to refer the
appearance of a sound-change in each case to some particular
manifestation of laziness, weariness, or neglect, and to ascribe its
non-appearance in other cases to some special care and observa-
tion. It may well be that the motory sensation is not in every case
developed to the same degree of certainty. But there is no such
thing as a conscious effort made to prevent a sound-change. For
those who are affected by the change have no suspicion that there
is anything to guard against, and they pass their lives always in
the simple belief that they speak to-day as they spoke years ago,
and that they will continue to the end to speak in the same
way. Were any one able to compare the movements which his
organs made in the utterance of a word many years before/with
? those which he makes at present, he would most likely find a
striking difference. But to make any such real comparison would
? be an impossibility. The only possible test is in each case the
motory sensation; and this is correspondingly modified-in fact,
exists no more in the mind as it existed at the previous time.
Control
through the
'Sound-
Picture.'
74. There is, however, a controlling source which opposes a
powerful barrier to the development of the single individual just
described—that is, the sound-picture. Motory sensation is formed
ON SOUND-CHANGE.
49
from the movements of the speaker only; the sound-picture, on the
other hand, takes shape not merely from our own utterances, but
also from all that we hear from those with whom we enter into
communication. Now, if it were the case that a notable displace-
ment of the motory sensation were to occur, accompanied by no
corresponding displacement of the sound-picture, a discrepancy
would be felt between the sound produced by the first and the
sound-picture obtained by the previous sensations. Such a dis-
crepancy is avoided by the motory sensation correcting itself after
the sound-picture. This happens in the same way as the motory
sensation directs itself at first in childhood according to the sound-
picture. It belongs to the very essence of language as a means of
communication, that the single individual should always find him-
self in agreement with the companions with whom he communi-
cates. Of course no such thing as a conscious effort at this result
exists, but the demand for such agreement remains, as something
self-intelligible, unconscious. This demand cannot either be com-
plied with with absolute exactness. If the motory sensation of the
individual cannot fully master his movements, and is actually
exposed to slight deviations, the free room for the movement
which finds play within a group of individuals must of course be
still greater, for it will certainly never be in the power of the
motory sensation of each individual to satisfy completely the
sound-picture which floats before him. And there is this further
consideration, that this sound-picture as well must take a some-
what different shape in the case of each individual, thanks to the
differences which exist in sound-sensations, and is likewise subject
to perpetual vacillations. But these vacillations, within a group
connected by active communication, cannot pass beyond rather
narrow limits. They are in this case unnoticed, or, even should
they be noticed on nearer observation, they still hardly admit
D
50
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Relation
of the
individual
speaker
to his
linguistic
milieu.
(
of definition, or indeed of expression, even by the aid of the
most perfect alphabet. This is not merely a matter of à priori
suspicion, but lends itself to objective observation in the case of
living dialects-of course not in the case of those which show a
graduated influence of the written language. If deviations more
or less violent in the case of an individual are found-for instance,
as the result of an organic fault-this makes little difference in the
result of the whole.
75. As long, then, as the single individual with his tendency to
deviation stands alone opposed to his companions in intercourse
he can only yield to this tendency in a very limited measure,
seeing that its operations are always counteracted by counter-
operations, which regulate the result. A displacement of greater
extent can only appear if it prevails throughout the entirety of
the individuals in a group which is to some extent) secluded from
all external influences, at least in comparison to the activity of
the communication prevailing within its circuit. The possibility of
such a process needs no demonstration in cases where the deviation
suits the convenience of all, or almost all, the organs of speech
better than the strict conservancy of the direction of the motory
sensation. It must be specially noticed in this connexion that the
already existing correspondence in accent, time, etc., /gives an
impulse towards the same path. The same holds good of corre-
spondence in the neutral position. But this is not nearly sufficient
to explain the whole proceeding. We see, of course, that manifold
developments proceed from the same starting-point, and this
without (necessarily in every case) being conditioned by changes
in accent or other circumstances of any kind which claim as their
exciting cause psychology; and we must ever put the question
anew-How comes it that precisely the individuals composing
this group/undergo in common the influence of such and such
.
·
ON SOUND-CHANGE.
51
change? Similarity of climate, of soil, of life has been cited to
explain the difficulty. We have, however, to state with reference
to this) that up to the present date not even the first steps have
been taken for methodically collecting materials relating to those
which might tend to prove dependence of the development of
language on such influences. What is advanced in favour of this
theory in individual cases may be easily reduced to a reductio ad
absurdum. It hardly admits of doubt that peculiarities in the
organs of speech/are transmitted by inheritance, and hence a degree
of relationship, closer or more remote, is to be reckoned among the
other factors which condition a greater or less correspondence in
the construction of the organs. But this is not the only cause on
which the latter depends. And just as little does the development
of language depend solely on the construction of the organs. In
addition, however, dialectic separation and dialectic reconciliation
seem in
very many instances to run counter to the corporeal relation-
ship. It will then be labour spent in vain to endeavour to explain
the fact of the agreement of all the individuals in a single group as
a spontaneous result, and therein to overlook the other factor,
which is operative side by side with this spontaneity, viz., the force
of community of intercourse.
76. If we start from the assumption that each individual has
his special bent and his special development, the possibility of very
numerous variations is certainly admitted. But if we take each
factor which comes under our consideration as isolated, then the
number of the possible variations is indeed very limited. If we
observe the changes of each single sound taken singly, and if we
again differentiate in this process the displacement of the locality
of articulation, transition of closure to narrowing, and vice versa,
strengthening or weakening of the pressure of expiration, etc.,
we shall often be in the position of obtaining two possible cases,
52
PRINCIPLES OF THE HISTORY OF LANGUAGE.
1
and only two, of deviation. Thus, for instance, the a sound may
gradually change into that of all possible vowels; but the direction
in which it moves can still(in the first instance) be only that towards
2, or that towards u. Now it can certainly easily occur that the
two or three possible directions may, (in a large linguistic area,
all things considered, be fairly balanced. But it is very unlikely
that this should occur in all the different points at every time.
The case that, in an area held together by an extremely active
intercourse, one tendency should gain the upper hand may easily
occur, solely by the caprice of chance-i.e. even if the agreement
of the majority is not conditioned by a more close inner connexion
as against the individuals who stand outside the group, and if
the causes which impel to this definite direction are different, as
they may be, in the case of different individuals. The fact of
the prevalence of a tendency in such a narrow circle suffices to
prevail against the opposing barriers. The active cause is, that the
displacement of the motory sensation to which the majority leans
entails a displacement of the sound-picture according in the corre-
sponding direction./ The individual is, in fact, not dependent on
the entirety of the members of the whole linguistic community
with respect to the arrangement of his ideas of sound, but only, as
an invariable rule, on those with whom he enters into intercourse,
and again not on these uniformly, but in very different degrees,
according to the frequency of the communication, and according
as each individual brings his activity to bear in the process. It
does not matter from how many persons he hears such and such
peculiarity of language; the whole consideration is how often
he hears them. We must, (while on this topic,) observe that the
speakers who deviate from the commonly adopted standard /may
again differ among themselves, and that their several influences
reciprocally may thus neutralise each other. If, however, a definite
ON SOUND-CHANGE.
53
displacement of the motory sensation has set in owing to the
removal of the retarding influence of communication, /we then
find that in the course of this tendency/a further slight deviation is
rendered possible. Meantime, however, the minority as well is
swept into the current by the movement. Precisely the same
causes which prevent the minority from departing too far from
the common usage in their progressive movement, forbid also
that it should lag much behind the advance of the majority. For
the superior frequency of any pronunciation is the only measure
for its correctness and fitness to serve as a standard. Thus the
movement progresses in this way-there is always' a part some-
what in front of the average and another part somewhat behind it;
but the whole advances with so little difference between its parts.
that a striking contrast never occurs between individuals who
stand in equally close communication with each other.
77. In this way it will always be found that the displacements
which occur within the same generation/are slight and scanty.
More notable displacements do not occur until an older generation
has been thrust aside by a new one springing up. In the first
place, if a displacement has already penetrated to the majority,
while a minority still opposes it, it will be found that the coming
generation will naturally adapt itself to the majority, especially
when the majority has the more convenient pronunciation. Even
if the minority in these cases should cling to the old custom, it
must yet die out. It may, however, be the case that the impulse
of the younger generation may set in a special direction differing
from the elder one. The same motives which in the case of the
elder generation impel to a particular kind of deviation, from the
impulse already formed,/must in the case of the younger generation)
operate at the very outset upon the shape to be taken by their
language. It may therefore be properly said that the main occasion
1
1
L
Sound-
changes
54
PRINCIPLES OF THE HISTORY OF LANGUAGE.
of sound-change consists in the transmission of sounds to new
individuals. For this process, then, the expression 'change,' if we
would remain true to facts, is quite inapplicable; we have rather
to deal with a new creation deviating from the old form.
78. In the process of mastering language the sounds alone/are
transmitted, and not the motory sensations as well. The agree-
ment of the sounds/which are self-generated with those heard from
others/gives the individual the assurance requisite that he is
speaking correctly. That the motory sensation has taken an ap-
proximately identical form/can only be assumed on the supposi-
tion that approximately similar sounds can only be produced by
approximately identical movements of the organs of language. If
it is possible to produce an approximately identical sound by
means of different movements, it must also be possible that the
motory sensation of any learner of language may take a different
shape from that of the persons from whom he learns it. (For a
few particular cases) such deviation of the form taken by motory
sensation/must be conceded as possible, Thus, for instance, the
dorsal t and s sounds are not very different from the alveolar in
sound, although their articulation is essentially different. Lingual
and uvular are still fairly easy to distinguish, and it seems that
in different German dialects /the one or the other prevails all
through; but the transition of the one into the other can hardly
be explained in any other way than by the fact that deviating
utterances were not corrected because the sound-deviations were
not sufficiently marked.
79. There are other sound-changes which do not depend upon
which do not the displacement or deviating form taken by the motory sensation,
displacement which, therefore, have to be separated from sound-change in the
depend on a
of the Motory
Sensation.
}
narrower sense hitherto described. These changes, (however,) have
this much in common with that, that they proceed to their fulfil-
་
1
ON SOUND-CHANGE.
55
1
ment without any regard to the function of the word. The effect?
of these processes is not the change of the elements of which
the sentence is composed, by substitution, but merely an inter-
change of these elements in certain definite cases.¹
80. The first of the changes which fall under this head is
that of metathesis. Of this there are two main divisions. The
first of these is when two sounds immediately following are trans-
posed, as in the case of AS. fix = OHG. fisc; first = frist; irnan =
rinnan. The second is when two sounds not immediately follow-
ing change their places; cf. the case of OHG. erila by the side
of elirà = NHG. erle and eller; AS. weleras, the lips, as against the
Gothic wairilos; OHG. ezzih, which must have had the sound
of *etik before the sound-shifting process set in Lat. acetum ;
Ital. dialectically grolioso = glorioso; crompare = comprare; MHG.
kokodrille Lat. crocodilus.
81. Under this head, too, must be ranged assimilations
between two sounds not related, as Lat. quinque from * pinque ;
original German finfi (five) = * finhwi, etc.__
82. We more commonly find dissimilations between two
similar sounds not in contact; cf. OHG. turtiltuba, from the Lat.
turtur; marmul, from Lat. marmor; MHG. martel) with marter,
from martyrium; prîol with prior: and conversely, MHG. pheller
with phellel, from Lat. palliolum; OHG. fluobra, consolation,'
as against OS. frôfra and AS. frôfor; MHG. kaladrius with kara-
drius; Middle Lat. pelegrinus from peregrinus.2
(
83. Further, the falling out of a single sound may be regarded
as assimilation, if this be caused by the fact of the same sound.
occurring in its neighbourhood: cf. Greek Spúpaктos 'wooden
δρύφακτος
¹ Cf. Brugmann, Zum heutigen Stand der Sprachwissenschaft, p. 50.
2 Cf. Bechtel, Ueber gegenseitige Assimilation und Dissimilation der beiden Zitter-
laute. Göttingen, 1876. I cannot, however, accept all Bechtel's examples as properly
referred to this head.
$
56
PRINCIPLES OF THE HISTORY OF LANGUAGE.
barrier, derived from φράσσω, and ἔκπαγλος from πλήσσω. It is
just so with the disappearance of an entire syllable occurring near
a similar one closing with the same consonant: cf. μédiμvov,
a double of ἡμιμέδιμνον, ἀμφορεύς οἱ ἀμφιφορεύς, κελαινεφής
instead of * keλawvoveons; Lat. semestris for * semimestris.
84. These processes seem to admit of no other explanation than
that they are based upon repeated mispronunciations affecting
spontaneously a considerable part of the linguistic community. It
is a well-known circumstance that in the act of speaking the order
of the words, syllables, or single sounds becomes displaced, owing
to one element forcing itself prematurely into consciousness. It is
equally well known that of two similar elements one is easily
dropped. It is further notoriously difficult to pronounce (with
accuracy) a succession of similar and yet different sounds uttered
quickly. It is on this that the joke depends about If Peter Piper
picked a peck of pickled pepper, etc.; Round about the rugged rock
the ragged rascals ran, etc. It will also be impossible to deny that
conditions favourable to certain mispronunciations exist, and that
hence these mispronunciations occur in the mouths of quite
different persons, and this repeatedly. The mispronunciations
may then pass by inheritance as a normal form to the younger
generation. These processes are most easily understood when
they affect foreign words which contain series of sounds repugnant
to the genius of the language which adopts them. In these cases
inexact perception and defective recollection will likewise occur.
The phenomena often therefore resemble those which will be dis-
cussed, under the head of Sound-substitution, in Chapter XXII.
below. It suggests itself also often for our consideration if popular
etymology is not also at work in producing them. Everything
is not yet clear in these processes.
85. We have now to answer the important question, which
ON SOUND-CHANGE.
57
of the Laws
change.
has been in recent times the subject of so much dispute: Can Uniformity
we assert uniformity of sound-laws? In the first place, we must of Sound-
fully understand what we mean, generally speaking, by a sound-
law. The word 'law' is itself used in very different senses, and
this very fact induces errors in its application. The idea of sound-
law is not to be understood in the sense in which we speak of
'laws' in Physics or Chemistry, nor in the sense of which we were
thinking when we contrasted exact sciences with historical sciences.
Sound-law does not pretend to state what must always under
certain general conditions regularly recur, but merely expresses
the reign of uniformity within a group of definite historical
phenomena.
86. In the statement of sound-laws the rule has been to start
with a comparison. The circumstances of one dialect have been
compared with those of another; those of an older stage of
development with those of a more recent. Abstractions have
been made by comparing the relations of one dialect with those.
of another, those of an older stage of development with those of
a later one. Sound-laws have also been formed by abstraction
by comparing the different relations within the same dialect
existing at the same time. The rules commonly adopted even
into practical grammars are of the latter kind. Thus, to cite
a sentence taken word for word from Krüger's Greek Gram-
mar-A T-sound, followed by another, passes regularly into σ.
Examples : άνυσθῆναι from ἀνύτω, ἐρεισθῆναι from ἐρείδω, πεισ-
θῆναι from πείθω. It has been before remarked that we must
not allow ourselves to be led away by such rules to adopt the
view that the sound transitions in question are each time.
effected anew by the creation of the one form out of the other.
The forms in question which stand in such a relation to each
other are either both taken up by the memory, or one is formed
}
58
PRINCIPLES OF THE HISTORY OF LANGUAGE.
from the other by analogy (cf. Chapter v.). This relation will be
designated not as sound-change (laut-wandel), but as sound-substitu-
tion (laut-wechsel). Sound-substitution is not identical with sound-
change, but is merely an after-effect of it. Accordingly we ought
never to apply the term 'sound-law' to sound-substitution, but only
to sound-change. A sound-law can no doubt reflect itself by means
of the effects left behind in the relations of a language as they
exist side by side; but as sound-law/it never applies to this, but
invariably to a historical development carried out in a definite
period.
87. If we, therefore, speak of the uniform efficacy of sound-
(laws, this can only mean that in the case of sound-laws occur-
ring within the same dialect, all the separate cases, in which the
same sound-conditions occur,/are treated uniformly. It must
either happen, therefore, that where the same sound existed pre-
viously, the same sound always remains in the later stages of
development as well; or, where a separation into different sounds
has occurred, there must be a special reason to be assigned;
and, further, a reason of a kind affecting sound alone-such as the
effect of neighbouring sounds, accent, place of syllable, etc.
for the fact that in one case one sound has arisen and in an-
other/a different one. No doubt we must take (into account)(in
this) all the different factors in the production of sound.
Especially we must regard the word not as isolated, but in the
light of its place within the sentence taken as a whole. Not till
then is it possible to recognise the real uniformity of sound-
changes.
ነ
88. (It is not difficult, after the enunciation of these prin-
ciples, to show the necessity of this uniformity as far as it turns
on actual sound-change depending on a gradual displacement of
the motory sensation; to speak more accurately, we should (no
ON SOUND-CHANGE.
59
F
doubt) rather say, not uniformity,' but the occurrence of all devia-
tions from it/within limits too fine for our detection.
89. It must be plain to every one who acknowledges (in all
phenomena) the operation of general laws, that the process of de-
velopment works uniformly to its fulfilment. A motory sensation
does not form itself specially for every word, but in every case
where the same elements recur in language/their production is
guided by the same sensation. Should, then, the motory sensation.
suffer displacement by reason of the pronunciation of an element
in any word, then this displacement is also a precedent for the
same element in another word. Thus the pronunciation of this
element in the different words/vacillates just as much as does
that of the same word within the same narrow limits. Vacillations
of pronunciation caused by quicker or slower, louder or gentler,
more careful or more negligent utterance, will always affect the
same element in the same manner, no matter in what word it may
occur, and these must always move in corresponding distances
from the normal.
90. As far as the development of a single individual goes,/it
is always an objection which is urged against the uniform work-
ing of sound-laws. It is maintained that our etymological con-
sciousness—our regard for related forms, stands in the way of
the operation of a sound-law. Whoever maintains this/must, in-
the first place, clearly understand that it involves no denial of
the continuous activity of the factor which impels to sound-change
-only the supposition of a factor of an entirely different nature
which operates against this. It is decidedly not a matter of indif-
ference whether we assume that a factor is at one time operative,
and at another time inoperative; or whether we maintain that it is
(under all circumstances) operative, and that its operating power is
counteracted by that of another factor. But how are we to con-
安
​60
PRINCIPLES OF THE HISTORY OF LANGUAGE.
}
(
ceive of the chronological relation in the operation of these factors?
Do they both operate simultaneously, so that it comes to no change,
or does the one operate after the other, so that the operation of
the latter is always again cancelled? The first could only be
conceivable on the assumption that the speaker knew something
of the threatened change, and sought to guard himself against it
beforehand. I think that I have made it sufficiently clear that there
can be no question of this at all. If we, however, allow that the
effect of the factor of sound makes its influence first felt, but is
then counteracted by the other factor, which we shall have in a
following chapter to characterise more closely, the uniformity of
the sound-laws is hereby admitted. We can advantageously dis-
pute this point at most: Whether it is the rule that the analogy
makes itself felt immediately after the appearance of a very slight
difference between the etymologically connected forms, or (if it
does not show itself operative until the gap has declared itself
more unmistakeably. In principle, this is no difference. That
the latter is in any case very frequent, we can see by experience,
on which we shall speak at greater length later. But it lies in
the nature of the case that differences which are not yet felt
as such/cannot diminish the feeling for etymology, and are not
diminished by this.
91. Equally inadmissible is the supposition that considerations
as to the clearness and intelligibility of a form stand in the way of
a transition of sound. We come sometimes upon relations which
seem to prove the affirmative. Thus, for instance, in NGH. the
intervening e of the weak perfects and participles after t and dis
maintained as in redete, rettete, while it is in other cases rejected.
But if we revert to the sixteenth century we find that double forms
are the rule in every verb;/on one hand we find zeigete by the
side of zeigte; on the other, redte by that of redete. Sound-change
ON SOUND-CHANGE.
61
1
has thus made its appearance with no consideration of utility;
and the greater utility of the one set of forms has decided merely
their continued existence.
92. Thus the question can only remain whether the communi-
cation of the different individuals among each other/can occasion
breaches of uniformity. This would only be conceivable under the
supposition that the single individual were to be exposed simul-
taneously to the influence of several groups of persons who had
plainly parted from each other by a different system of sound-
development,/and that he learnt some words from one, and others
from the other group. This, however, presupposes a wholly
exceptional state of things. Commonly speaking, there exist no
such differences within the linguistic community within which the
individual grows up, and with which he stands in far more intimate
connexion than with more distant associates. Where it does not
happen that,(in consequence of special historical causes,) larger
groups are detached from their original dwelling-place, and thrown
together with others where the population is, (at most,) modified by
slight accessions or departures, but remains constant as far as its
main body is concerned, no differences can be developed, which
are apprehended as such. Even if A pronounces a somewhat
different sound from B in the corresponding place,/still the
apprehension of the one sound as well as that of the other/fades
into the sound-picture which the hearer already carries in his
mind; and thus it follows again that only the same motory
sensation can correspond with them. It is absolutely impossible
that, owing to two differences so slight,) two different motory
sensations should form in the same individual. As a general rule/
it would be impossible, even supposing that the extremes which
occur within a small linguistic domain/were the only existing ones.
But even if the hearer were in a position to apprehend the differ-
4
62
PRINCIPLES OF THE HISTORY OF LANGUAGE.
"
ence between these two, still the series of fine transition-steps,
which he always hears without any break, would render it im-
possible to keep a frontier line unbroken. And so, even assuming
'that he hears one word more frequently and earlier from people
who tend to this extreme, and the other more frequently and
earlier from people who tend to the other extreme,/still this can
never produce the result that, in his efforts to repeat the word,
the production of a sound in one word should be guided by a
different motory sensation from the production of a sound in the
other word, if the same individual would set an identical sound. in
both places.
93. Thus within the same dialect/no irregularity develops,
excepting either as the result of a mixture of dialects, or, as we
shall have to detail with more accuracy, as the result of the
borrowing of a word from a foreign dialect. We shall have later
to examine to what extent and under what conditions such
borrowing appears.
Of course in our statement of sound-laws
we do not have to reckon with any such apparent irregularities.
94. The attempts which have been made to explain sound-
change as dependent on individual caprice or on an inaccurate ear
are hardly worth mentioning. A single inaccuracy of ear cannot
possibly have any lasting results for the history of language. If I
do not accurately catch a word from any one who speaks the same
dialect as myself, or another with which I am well acquainted,
but if I guess his meaning from the context of his discourse,
then I supply the word in question according to the memory-
picture which I have in my mind. If the connexion is not
sufficient to explain clearly the meaning, it may be that I shall
supply a wrong meaning, or I may supply nothing at all, and
satisfy myself with understanding nothing, or I may ask again.
But how I should come to think that I have heard a word of
ON SOUND-CHANGE.
63
different sound, and still to set this word in the place of the one I
understand/is to me incomprehensible. Certainly it may more
easily happen to a child who has never yet heard a particular
word/to apprehend that word imperfectly, and to reproduce it
again imperfectly. But the child will also frequently reproduce
imperfectly that which it has apprehended more correctly because,
in its case, impulse is not yet adequately formed. Its apprehen-
sion, as its reproduction, will correct itself, if it always hears the
word anew; if this be not so, it will forget it. Wrong apprehen-
sion of sound is only regularly seen when people who belong to
different dialectic areas or different languages converse with each
other, and the shape in which foreign words are adopted /is no
doubt much influenced by this circumstance, but certainly more
by the want of motory sensations for the sounds which are wanting
to their own dialect.
95. There remain now certainly some kinds of sound-changes
in which uniformity of action/cannot theoretically be proved
necessary. These, however, make up a relatively small part of the
entire sound-changes, and they admit of sharp delimitation. Thus
on the one hand, we have to reckon (under this head) the cases in
which a sound is imitated by means of a deviating articulation.
On the other side/we must include the metatheses, assimilations,
and dissimilations referred to on p. 55, sq. For the rest, even in this)
case, we find(as a matter of fact, to some extent) that regularity is
the rule, especially in the case of the metathesis of sounds immedi-
ately following each other, and further, e.g., in the case of the
dissimilation of the aspirates in Greek, as in kéɣνка, πéþevya, and
elsewhere.
96. The question, as to how far sound-laws are to be regarded
as admitting of no exceptions,/cannot be absolutely decided by the
materials of language before us, because there are changes in
64
PRINCIPLES OF THE HISTORY OF LANGUAGE.
language which, (although in their nature absolutely different from
sound-change,)still produce corresponding results with this. Thus
our question is intimately connected with the second question:
How far does the operation of these other changes extend, and how
are they to be divided from sound-change? Of this we shall have
to treat below.
CHAPTER IV.
CHANGE IN WORD-SIGNIFICATION.
Sou
Meaning
widening or
of its scope.
OUND-CHANGE is effected by repeated substitution of a Change of
sound or sounds almost imperceptibly differing from the consists in a
original. In this process the disappearance of the old is simul- narrowing
taneous with the appearance of the new. In the case however of
change in signification, the maintenance of the old is not excluded
by the appearance of the new. It consists invariably in a widening
or narrowing of the extent of the signification, corresponding to
which we find an impoverishment or an enrichment of the con-
tents. No signification absolutely different from the original can
be formed but from the successive processes of widening and
harrowing.
of' Usual
casional'
98. Change in signification, however, resembles sound-change in Distinction
this, that it is effected by a departure in individual usage from the and 'Oc-
common usage, which departure passes only by gradual stages into Meanings.
common usage. The possibility, we may even say the necessity, of
change in signification, springs from the circumstance that the sig-
nification which attaches to a word each time that it is employed,
is not necessarily coincident with that which by usage attaches to it
considered in itself. As it seems desirable to adopt distinct names
for this discrepancy, we shall employ the expressions 'usual'
and 'occasional' signification; possibly 'general' and 'individual'
night serve as well. We understand then by 'usual signification'
E
66
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Distinction
of 'Abstract'
and 'Con-
crete'
Meanings.
the entire contents of the conception bound up in any given word
as it presents itself to the member of any body of individuals
speaking one common language: by the term 'occasional signifi-
cation' we understand the contents of the conception which the
speaker, in uttering the word, connects therewith, and which he
expects the listener to connect with it likewise.
99. The occasional signification is very commonly richer in
contents than the usual one, and narrower in extent. In the
first place the word in its 'occasional' sense may denote some-
thing concrete, while in its 'usual' sense it denotes only some-
thing abstract, i.e. some general conception under which different
concrete ones may be ranged. By a 'concrete' conception I shall
be understood to mean something pre-supposed as really existing,
subject to definite limits of time and space: by an 'abstract' one
I understand a general conception, the contents of a mere idea and
nothing more, freed from all trammels of time and place. This
distinction has nothing to do with the ordinary division of
substantives into abstract and concrete. The substantive appella
tions which we call 'concrete' denote a conception as general
as the so-called abstract nouns; and, conversely, the latter are
occasionally used as concretes, as they are used to express a single
quality or activity defined by limits of space and time.
100. By far the greater number of words are capable of bearing
in occasional use abstract and concrete significations indifferently.
There are indeed some which are by their very essence intended to
denote something concrete, but which do not refer to something
immediately and definitely concrete, this reference being given them
by individual application. Such words as these are the personal,
possessive, and demonstrative pronouns, the demonstrative adverbs,
and such words as now, to-day, yesterday.
101. Words like I, here, dieser, serve no other purpose but to
CHANGE IN WORD-SIGNIFICATION.
67
define some one's position in the concrete world;¹ but in them-
selves they contain nothing definite, and it requires the aid of
individualising elements to give them such definiteness. Then
take the case of proper names. These doubtless denote a single
object or person; but seeing that the same name may attach to
different persons and different localities, the result is that there
remains a variation between occasional and usual signification.
Finally, a small number of words have to be considered express-
ing an object regarded as existing once, and once only; such
are God, devil, world, earth, sun. These are at the same time
general and proper nouns, but only when understood in a certain
way and regarded from a definite, not from a general, point of
view. Conversely there are some words which from their very
nature must be ranged with the general and not with the con-
crete: such are the adverbs and pronouns-ever, any; the German
je, irgend; MHG. ieman, dehein; Lat. quisquam, ullus, unquam,
uspiam; but the general character even of such words as these
suffers certain limitations in the occasional usage-cf.: If he ever did
it at all; If he ever really should do it; Have you been at any trouble?
102. A further and very important difference between usual Plurality of
and occasional signification is the following. The 'usual' significa-
tion of a word may be various; its 'occasional' meaning is always
single except in cases where ambiguity is intentionally attached to
the word, either to deceive, or to point a witticism. Attempts have
been made, as by Steinthal, Zschr. f. völkerpsych. i. 426—though
unsuccessful we believe-to prove that there are no such things
as words with several significations. Under this head must be
ranged first those words which correspond in sound, but differ in
meaning, such correspondence being purely accidental; such is
1 Even our demonstrative pronouns-and the word the-may be employed as
denoting abstract conceptions: e.g., The whale is a mammal. In German er can be used
in the same way, e.g., Er bringt lebendige junge zur welt.
Meaning.
68
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the German acht = diligentia-proscriptio-octo. These cases are
naturally enough excluded by Steinthal, who maintains that here
the words in question are not the same, but different words. In
sound, however, they are identical, and whoever hears such a
combination of sounds spoken without any connection is power-
less to recognise which of the significations inherent in the
word is in the mind of the speaker. We thus have, if we keep
to the actual state of things, and add nothing which does not
strictly belong to it, a word which usually receives several signi-
fications. But in the many cases in which we have identity
not merely of sounds, but of etymology as well, we are driven
to recognise actual plurality of signification. Take, for instance,
the word fuchs in modern High German. It means, in the first
and most common case, ‘a fox'; then 'a horse of foxy colour;'´a
red-haired person;' 'a sly fellow ;' 'a gold coin;' 'a freshman at
college.' Bock, a 'he-goat;' bock, 'the outside seat of a coach,' and
'a mistake;' futter, 'food for cattle' and 'lining;' mal, 'a 'spot,' a
'token,' a point of time;' messe, 'a mass' and 'a fair ;' ort, 'a place'
and a shoemaker's tool;' rappe, 'a black horse' and 'a coin;
stein, ‘a stone' and 'a weight,' and 'a disease;' geschick, 'fate’
and 'dexterity;' geschickt, 'sent' and 'dexterous;' steuern, 'to
pilot a vessel,'' to pay taxes,' and 'to hinder ;' MHG. beizen 'to bait'
and 'to chase with a falcon ;' erbeizen, 'to dismount from horseback ;'
weide, 'pasture,' 'hunting ground,' 'fishery,' 'time' (as in ander-
weide, another time'); Latin, examen, ' swarm' and 'examination.'
Steinthal recognises the primary signification in such cases as the
only one to those which have in the course of history sprung
from this he denies the title to independence. But his view is
correct only as a statement of what exists at the time when
the derived signification first springs from the primary meaning.
This state of things is however merely transitory. In most
CHANGE IN WORD-SIGNIFICATION.
69
of the cases cited above it is impossible to recognise, without
some historical knowledge, the original connexion between the
various significations, and these bear the same relation to each
other, as if the identity in sound were purely accidental. This is
especially true when the primary meaning has disappeared. But
in many cases also where the relationship of the derived to the
primary signification is still recognisable, we cannot but acknow-
ledge the independence of the former, especially in all cases where
it has become the 'usual' one. There is one sure criterion of these
cases, viz., whether a word 'occasionally' used in the derivative
sense in question, can be understood without the aid of the
primary meaning, i.e. without the necessity of the primary mean-
ing forcing itself on the consciousness of the speaker or hearer.
There are further two negative tests whereby we may judge that
a word has not a simple but a complex signification; the first
of these is that no uniform definition can be framed wherein the
whole extent of the meaning, neither more nor less, can be in-
cluded; and the second is that the word cannot, if employed
'occasionally,' be used in the whole extent of its signification. The
reader may apply these tests to the examples cited above.
103. Even in cases where the usual signification may be
regarded as simple, the individual meaning may vary from this,
and still not become concrete, proceeding as it may on one of
the different bye-meanings included in the general conception.
Thus the simple word nadel can in single cases be understood as
stecknadel, nähnadel, stopfnadel, stricknadel, häkelnadel, etc.
104. All understanding between individuals reposes on the cor-
respondence in their psychical attitude.¹ For the common under-
standing of the usual meaning no more correspondence is needed
1 The following illustrations proceed much on the same lines as the views put forward
by Wegener in his book, Aus dem leben der sprache; and to a certain extent resemble
those of Bréal, Les Idées latentes du Langage (Paris, 1868).
70
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Means by
which Abs-
than such as naturally exists between all the members of a given
number of individuals who speak the same language, assuming
always that they have thoroughly mastered their language. When,
however, in the 'occasional' use the signification attaching
to a word is specialised and still claims to be understood, such
claim can only be based upon a closer accord between the
speakers. The same words may either be perfectly intelligible or
unintelligible, or, again, may be liable to misconception according
to the frame of mind of the person accosted and the chance
environment of other circumstances, according as certain aids.
to understanding are present or absent.
need to be of linguistic nature at all.
to form a clear notion of these in detail.
These aids do not
We must endeavour
105. Words which properly have an abstract signification
tract Words may be brought into relation with something concrete by connect-
ing them with such words as those described (p. 66), whose func-
may receive,
in their
Occasional'
Uses, a
Concrete
Meaning.
tion it is to express the concrete, more especially those with the
article where it is developed. But the use of the article has
generally developed in such a way that it is not confined to the
function of individualising, but is also attached to the noun where
it expresses the general conception. Languages which have
developed no article employ abstract words (without any special
mark of denotation) for the concrete.
106. Whether the reference to the concrete is expressly
denoted or no, in any case other means must be adopted for its
closer definition. The first of these is the perception common to
the speaker and hearer. The latter recognises the fact that the
former in speaking of tree or tower means a definite single tree
or tower, that is, assuming that they both have before their eyes
the object in question. The perception may be supported and
defined by signs made by the eyes or hands, or by other

CHANGE IN WORD-SIGNIFICATION.
.71
gestures. Such signs as these may likewise serve to point to
objects not directly cognisable by the senses, so long
direction in which they lie is known.
as the
107. A second method whereby the word is made to refer to
something definite and concrete is found in the past utterances
of the speaker, or, it may be, in his special explanation. If it
is understood that a word once bears a concrete sense, it is
possible that this same sense may continue to attach to the word
throughout the further course of the conversation. In this case
the remembrance of the previous utterance takes the place of
immediate perception. This reference to the past can again be
supported by the demonstrative pronouns and adverbs. Originally
employed to express a certain perception, the transference of
their function to denote previous utterance serves admirably to
render the individualising of the signification, intended by the
speaker, clear and intelligible to the hearer.
108. In the third place, we have to take account of the
special force which may attach itself to the representation of
anything concrete, even when this force dispenses with the aid
of perception or mention of a word previously used in a par-
ticular sense. This force may make`itself jointly and commonly
felt in the mind of the interlocutors. Such agreement or cor-
respondence is dependent for its existence on the fact of common
residence, common age, common position, business, and manifold
common experiences of the speakers. An instance of this is
seen in the use of the rhetorical usage commonly known as кaт'
ἐξοχήν. Countrymen talk of going to town without more
definite specification, meaning thereby the town nearest to
where they happen to live. Words like the town-hall, the market,
are used by the inhabitants of a particular town for the town-
hall or market of that particular town; the kitchen, the dining-
72.
PRINCIPLES OF THE HISTORY OF LANGUAGE.
room, in the mouths of the members of a given family, refer
solely to the kitchen or dining-room of their particular house.
Thus again, when we speak of Sunday we mean the nearest
Sunday to the day on which we happen to be speaking, and it
only needs to be indicated, in order to fix the Sunday with perfect
accuracy, whether the day referred to is past or future. Words
expressing the relationship of one person to another are naturally
referred to persons who stand in such relationship to the hearer
and to the speaker alike: and, further, the singular, in such
cases, is perfectly clear, as long as there is only one person who
could be properly so described. Thus for the purposes of inter-
course between the sisters and brothers of a family, the concrete
reference of the words father and mother is sufficiently intel-
ligible, as is the employment of such words as the Queen or the
President to the British or the Americans respectively. Nay, even
in cases where the relation exists on one side only-on that of the
speaker, or on that of the hearer-it is still possible, under the
influence of auxiliary circumstances, that the reference should be
unmistakeable. Such a question as How is the wife? may be as
intelligible as How is your wife? If a concrete object have once
at an earlier period in any way acquired a signification appre-
hended alike by the speaker and the hearer, it can be called into
consciousness by the word that denotes it. This is especially
true if it is still fresh in the memory, or when a situation recurs
similar to that in which the object presented itself, as challenging
special observation. Suppose, for instance, that two friends are out
walking together, and they meet a lady previously unknown to
either, about whom they have exchanged remarks. They take the
same walk again, and one asks, Shall we meet the lady again?
In this case the reference in the mind of the other is clear and
unmistakeable.
1
CHANGE IN WORD-SIGNIFICATION.
73
109. In the fourth place, a closer definition may be summoned
to the aid of the speaker. But it is commonly found that such
a definition per se produces no concrete sense, but only when used
concurrently with the other factors already described. By means
of such definition the word to which the definition is attached
must either have received a reference to a group of concrete objects,
out of which by definition a further selection is made; or a con-
crete reference must thereby be given to the defining word. Both
processes may in fact happen at once. Thus the epithet old, if
attached to duke, gives per se no concrete sense. But if the facts
of the situation point the reference to some definite ducal family,
the person referred to is distinctly defined. The word castle gains
no concrete meaning by the addition of the word royal or king's
unless the known facts have previously given the word king a
concrete reference. But the phrase the king's castle comes to
mean one object and no more when it can be assumed that a single
castle of the royal personage in question exists, or if there are any
facts in the situation which tend to single out an individual; for
instance, if any one be referred to a particular place in which he
must necessarily suppose the castle in question to lie.
10. Finally, concrete sense spreads from one word to others
placed in relation to it. In sentences like John never drew bridle;
I never laid hand upon him; I took him by the arm; You hit me
on the shoulder,-the words bridle and hand gain a concrete
meaning owing to the subject, and arm and shoulder owing to
the object.
III. In the same way as general names receive a definite
concrete reference, proper names which belong to different indi-
viduals come to denote but one. The simple appellation Charles
is sufficient to identify an individual, supposing that he is in our
presence, or that we have recently mentioned him. He is
74
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Means by
which
sufficiently identified then, even apart from this, within the circle
of his family, or of a small body of acquaintances who have no
other Charles within their body. Under other circumstances we
define him more closely-e.g.: Charles the Sixth of France-
Charles the First of England. In the same way the same name
is given to many places; but one name is sufficient to define it
for the whole neighbourhood, and even for larger circles when
the place meant is by far the most important of all the places
bearing the same name (such as Melbourne, London, Strass-
burg): otherwise a nearer definition has to be used, as Stony
Stratford, Newton-le-willows.
112. The same factors which serve to impart to a word a
Meaning is concrete reference serve equally to specialise its signification.
specialised.
On
hearing a word, the natural tendency is to think of the most ordi-
nary of its various significations, or of its primary one. These
tendencies sometimes coincide. Where, however, it happens that
several meanings, each tolerably common, stand side by side, it
will be found that the primary meaning will present itself to con-
sciousness before those derived from it; and this will often be
found true where the derived meaning is in the more common
use. This is in accordance with an ordinary psychological law.
But the matter is at once different if certain groups of ideas are
awakened in the mind of the hearer before the word is uttered,
or will be awakened simultaneously with its utterance, ideas which
are more closely connected with a derived or comparatively un-
usual meaning. It makes a great difference whether I hear the
word sheet in a haberdasher's shop, or on a yacht, or at a
publisher's; as it does if I hear the words to bind at a book-
seller's, or in a harvest-field. When different men of different
trades and professions-as joiners, gamekeepers, medical men-
meet and converse, they are given to apprehend each word as
CHANGE IN WORD-SIGNIFICATION.
75
it comes most nearly home to their profession. The connexion
in which a word occurs is of great importance. By means of
such connexion the different possibilities of the apprehension of
a word are minimised. Take such instances as a good point, a
point of view, a point of honour; the bar of an hotel, the bar of
justice; the foot of a mountain, the foot of the table; the tongue
of a woman, the tongue of a balance; a well-attended ball, a
football club; a bay and a grey; the cock is turned on; it costs a
crown; the train starts, a train of thought; a clear voice, a clear
day; clean linen, a clean heart; John is a donkey; the money
goes, the milk goes; to stand still, to stand upon ceremony, to stand
at ease.
'Occasional
does not
include all
of the
Meaning.
113. In the cases we have mentioned the variation of the occa- The
sional meaning from the usual one consisted in the fact that the Meaning
former contained all the elements of the latter, and at the same necessarily
time something beyond. There are, however, cases where the the elements
occasional meaning does not contain all the elements of the usual Usual'
meaning, while it still may contain something which does not,
strictly speaking, belong to the latter. The fundamental condition
for the possibility of such partial utilisation of the usual significa-
tion of a word is given most ordinarily by the fact that the latter
is in most cases made up of different elements which may be
separated. Any idea of any substance must necessarily com-
prehend the idea of several characteristics. But there are also
many ideas of qualities and activities which we can designate by
a single word, which are likewise compound. For instance, the
names of colours denote (of course speaking from a psychological
point of view) perfectly simple qualities; blue, red, yellow, white,
black. And even these may well be employed to denote qualities
which, according to their proper meaning, are not fully adequate.
For each colour may, of course, be mixed with each other colour

76
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Transference
to things
Spatially,
Temporally,
at will; and thus there arises an infinite number of transition stages
which cannot possibly each receive a definite name. And the result
is that we are content to leave unimportant admixtures without any
name, so that the limit within which a name expressive of colour
is applicable remains uncertain and shifting. But a much wider
room for inadequate application is given us by words whose signi-
fication consists in a complex assembly of ideas.
114. Such, among others, are all so-called metaphorical ex-
pressions. We are accustomed to say that, for a comparison to
be made, there must necessarily be, besides the two objects
compared, a tertium comparationis, or third object with which
they are compared. But this tertium is nothing new, nothing
added to what we have already, but it is that part of the contents
of the two combinations (groups) of ideas compared with each
other which they have in common. If we say of a man he is
like a fox, we do not identify the two qualities as in the case of
a mathematical equation; nothing more is meant than this, that
one of the characteristic qualities of which the conception 'fox'
is made up is likewise implied in the idea that we have of this
particular man: i.e. as a rule, of his craftiness. And so we are
able to say more accurately by expressing the tertium as well-
He is as sly as a fox. On the other hand, we can say even more
simply he is foxy, in which case the adjective again does not
indicate the full extent of all the qualities of a fox, but only a
selection of these; and lastly, simplest of all, we can say he is
a fox.
115. There is another possibility which enables a word to pass
beyond the limits of its own proper signification; ie. into a
or Causally usage again of course at the outset purely occasional. It consists

associated
with the
original
denotation.
in this, that something which is according to ordinary experience
connected with the usual contents by relations of space, or time, or
CHANGE IN WORD-SIGNIFICATION.
77
cause, is understood together with the word, or, it may be, under-
stood alone, when the word is uttered. Under this head falls the
figure of speech known in Latin treatises on style as pars pro toto
[as where trabs or, again, carina stands for a ship'].
:
116. On each occasion when the word oversteps the limits of
its usual signification, a directing impulse must lend its aid—that
is, if the reference is to be rightly understood. Such an impulse.
is in this case more necessary than where nothing further is
required than to ascertain which of several significations already
recognised as usual is intended (cf. above, pp. 74-5). We never
feel ourselves impelled to understand a word in a sense which
does not include all the elements of its usual meaning, as long
as we are not reminded by something that such a sense is im-
possible. It becomes then a matter of necessity if we would
arrive at the right sense, that this reminder should set us on
some positive right track. In such proverbs as Speech is silvern,
but silence is golden, we should not think of the predicates as
used metaphorically, if they could be understood as connected
with their subjects in their proper sense. When Shakespeare
talks about the majesty of buried Denmark,' we guess from the
combination of majesty, buried, and Denmark, the right sense which
we are to attach to each.
of Usage
117. These departures of the occasional meanings from the Change.
usual meaning are starting-points of true change of signification. effected by
? Occasional
As soon as these departures repeat themselves with a certain regu- Modification.
larity, what was individual and momentary becomes gradually
generic and usual. The border line between what constitutes the
occasional and usual signification of a word is shifting. In each
individual case the beginning of the transition from an occasional
to a usual meaning is made as soon as, on the employment or
apprehension of the former, the recollection of an earlier employ-
78
PRINCIPLES OF THE HISTORY OF LANGUAGE.
ment or apprehension comes into play: the full transition may
be deemed accomplished as soon as such recollection only is'
effective, and when employment and apprehension alike follow, with-
out any reference to the usual signification of the word. Between
these two we may have a manifold gradation. Within the circle.
of friends who hold inter-communication, the different individuals
who compose it may find themselves at different stages in this
transition process. It is quite impossible for the process to fulfil
its course, in the case of any individual, while his interlocutors
remain totally unaffected by its influence. For it belongs to the
very essence of the process that it should pass to its fulfilment by
repeated and regular employment of a signification which in its
commencement was but occasional, and this must pre-suppose an
apprehension on at least one side of the circle of interlocutors;
and this apprehension is, for these again at least, a commence-
ment of the process. The process will, however, in the case of
one individual, not easily pass to its fulfilment, unless the influence
which he exercises on his associates is reciprocated by these.
Such reciprocal action will of course most readily arise where there
is not only external influence at work, but where a spontaneous im-
pulse exists to employ the word in the same occasional sense as it
naturally follows from the common mutual relations of the indivi-
duals. One of the most powerful efficients in the change from an
occasional meaning into a usual one is the first tradition to a
new generation. The process of learning the meaning of words is
not commonly assisted by any definition by whose aid the usual
signification is specified in its contents and full extent; such defin--
ing process is in fact an impossibility until a fair knowledge of
language is acquired, and must be considered even then excep-
tional. A child learns only occasional applications of a word, and,
what is more, learns in the first place nothing but its relations

CHANGE IN WORD-SIGNIFICATION.
79
to a concrete object. In spite of this the child, when once
it has acquired the application of the word, begins at once to
generalise it. And this is quite natural. The reference to the
single concrete object cannot in any case be permanent. For
the mental impression left by such object contains in itself
nothing whereby, on a new cognition of the object in question,
its actual identity or non-identity with the object once taken
cognisance of, can be recognised. The right appreciation of such
object depends on a succession of conclusions drawn from a suc-
cession of instances, and cannot be gained without this process
-nay, in many cases cannot be gained at all. For the simple
and unreasoning mind of childhood it is sufficient if the con-
tents of the idea presented to it agree, to warrant an identifying
process on his part, whether real identity exists or not. And
further, such agreement needs to be partial only-indeed under
some circumstances merely trivial to cause this identification;
that is, as long as the mental impression is still vague and
confused.
Thus it is that from the very commencement of
the process of acquiring language, the custom grows of defining
by the same word, not merely a single object but several-not
merely objects which actually resemble each other, but such
as bear even a remote resemblance. And this custom maintains
its ground, even when differences, disregarded previously, are
noticed, since it is strengthened and forwarded by the example
of the adults. But this course of procedure implies that no
clear conception can exist of the contents and extent of the usual
meaning. The child makes numerous mistakes from the fact that
it connects with the word a conception sometimes too wide, some-
times too narrow, and attaches to it an application correspond-
ingly too narrow or too wide. He will more commonly err on
the side of width than of narrowness of application, and the more
80
PRINCIPLES OF THE HISTORY OF LANGUAGE.
so according as his stock of words is more limited. A child will
include a sofa under the name of a chair, an umbrella under that
of a stick, a cap under that of a hat;—and this repeatedly!
Another impulse to inexact appreciation of meaning proceeds
from the fact that the objects indicated are frequently parts of
a larger whole, or are indissolubly connected with other objects
in the speaker's mind. In this case the child cannot but be un-
certain how the part which the word is intended to define, when
eliminated from the whole idea, should be limited. It will draw
these limits now wider, now narrower, than use demands; not
unfrequently omitting what of right belongs to the meaning,
and including what does not so belong. Besides, the process of
learning new words and new methods of applying old ones is far
from confined to early childhood. Expressions of rare occurrence,
or which denote complex ideas, or which pre-suppose a high or a
special standard of culture, have to be mastered by the adult no
less than by the child; and supposing that he learns them merely
by their occasional application, he is exposed to the same errors as
the child. All these inaccuracies in the apprehension of the usual
meaning are, taken singly, of no account, and are, as a rule, cured
by time.
Yet it cannot but be that in particular instances the
union of a large number of individuals in the same misapprehen-
sion, leaves behind it lasting traces. We shall thus have to recog-
nise a kind of change of meaning depending on the fact that for
the meaning which among the elder generation is usual, a meaning
only partially corresponding with it is substituted. But we shall
have to confine the domain of this change to the rarer sorts of
conceptions which more easily defy exact determination, as, in the
case of others, a gradual process of correction according to the
actual use cannot fail to ensue.
118. In most cases the impulse to change of meaning proceeds

CHANGE IN WORD-SIGNIFICATION.
81
from the elder generation which holds already complete control of
the usage. But the younger generation has likewise a special share
in forwarding the process, from the fact that the different applica-
tions of a word-group themselves from the very outset somewhat
differently from the way in which they do this in the case of
the older generation. Each method of application, apprehended
as it is for each immediate case, may be mastered for itself with-
out regard to other possible ones; and hence each may receive a
greater degree of independence than it enjoyed in the minds of the
older generation. The derived is often learnt before the primary
meaning: and this fact contributes not a little to fixing its inde-
pendence. A child may for instance often hear a horse spoken
of as a bay, or a stupid fellow as an ass. In such cases the
primary meaning is from the very outset dispensed with as an
aid to right comprehension. Until an individual has fully mastered
the usage of a word, he is not often qualified to distinguish whether
a particular manner of application which comes before him is
usual or merely occasional, and will thus be inclined to adopt the
occasional meaning,—if circumstances have strongly impressed that
meaning on his mind-not less readily than the usual one.
of Changes
1. Speciali-
Increase of
119. Seeing that the change in the usual signification takes Classification
its rise from modifications in the occasional application, we find of Meaning:
in both cases the same kinds of change. The first and main sation.
kind, accordingly, is specialisation of the meaning by narrowing Connotation.
the comprehension of the word, and the enriching of its contents.
The German word schirm is an instructive instance of the differ-
ence between occasional and usual specialisation. The word can
be used of any object employed as a 'screen.' In the occasional
usage it may signify a 'fire-screen,' or 'lamp-screen;' a 'screen'
for the eyes, an 'umbrella,' a 'parasol,' etc. But while it is
only by some definite situation that we are led to think of the
1
F
82
PRINCIPLES OF THE HISTORY OF LANGUAGE.
(
word as 'fire-screen' or 'lamp-screen,' it immediately occurs to
us to conceive of it as 'umbrella' (regen-schirm), or 'parasol'
(sonnen-schirm), and we then think less of the ordinary function of
'screening' than of an object of special shape and construction.
Thus we must perforce acknowledge that this meaning has dis-
severed itself from the more general meanings, and stands as a
special and independent one; and it matters not whether it can
logically be ranged under them or not. For this logical subordina-
tion is only possible by discarding factors not less essential to the
meaning than what we immediately have in view. Other examples
are the use of frumentum for corn' in Latin, of fruit for the pro-
duce of certain trees as compared with 'the fruits of the earth;'
corn in English is restricted to wheat, while in German korn denotes
any species of grain; dach was used in MHG. for any kind of
covering; it is now restricted to the covering of a house [fowl in
English means specially a 'barn-door fowl']. And a special
usage of this kind is seen in the names of materials used for pro
ducts of materials-as glass, horn, gold, silver, copper, paper, as
when we talk of paper money, etc. It is the business of the
lexicographer to distinguish, in the enumeration of the special
usages of a word, between such as have become usual and such as
are purely occasional: a distinction which is commonly neglected.
120. Proper names take their rise from the change of the
occasional concrete meaning of certain words into usual meanings.
All names of persons and places took their origin from names
of species; and the usage kar' oxýv forms the starting-point for
this process. We are able to watch this process distinctly in
numerous instances of place-names. Such commonly recurring
appellations are especially instructive, as the following:-Field,
Hill, Bridge, Townsend, Hedges, Church, Stone, Meadows, Newton,
Villeneuve, Newcastle, Neuchatel, Neuburg, Milltown, etc. Such

CHANGE IN WORD-SIGNIFICATION.
83
appellations served, in the first instance, merely to indicate to
dwellers in the vicinity the person or town to be indicated; and
they were sufficient to distinguish the persons or places intended
from others in the immediate neighbourhood. They passed into un-
mistakeable proper names from the moment when they were taken
over and apprehended by neighbours at a greater distance in this
concrete sense; or when they were more sharply divided from the
names of species originally identical with them, by agencies which
isolated them still further. No doubt there are besides these a large
class of names of places, which from the very outset resemble the
nature of strict proper names, in the fact that they are derived from
the names of persons, or are at least influenced by personal names.
121. There is also one kind of specialising process which
begins to operate immediately that the word is used at all. This
is especially found in the case of words which can be derived at
will from other words in common use according to the laws of
any language, but which do not come into actual use until a
special need calls them into play. Such words as these are often
only to be referred from the first stage in their career to the root-
word, with a more special meaning than the derivative per se
expresses. Thus the substantive formations in er, MHG. ære,
coming from other substantives in English and German, denote
properly a person who stands in some relation to the idea of the
root-word, of whatever kind this relation may be; but in the case
of single words thus terminated the most varied instances of
specialisation are seen. In the case of the MHG. æhtære, from
ante (NHG. acht, 'persecution'), the word denotes alike the 'pur-
suer' and the 'pursued;' but in the individual application the
two can never be understood together. In the case of scholar
(Latin scholaris) the idea of 'schoolmaster' might well have been
included, but there is no trace of the word ever having been

84
*PRINCIPLES OF THE HISTORY OF LANGUAGE.
2. Limita-
tion to a
part of the
Original
employed in that sense, or in any other sense than that which it
bears to-day. [Similarly a poulterer is never used otherwise than
in the sense of a vendor of poultry; a fisher always denotes
one who tries to catch fish; a burgher, the dweller in a burgh (or
borough); a falconer, one who keeps falcons for the chase; while a
pensioner denotes one who receives a pension.] It is the same
with verbs like the German bechern, buttern, haaren, hausen, herzen,
köpfen, mauern, etc. In the case of many words we are not in a
position to decide whether an application in a more general sense
has preceded or not. This original specialisation must, of course,
be itself in the first instance only an occasional one, seeing that
the word, as such, only points to the general idea evolved by the
combination of the root-word with the derivative suffix, and it is
the common situation of the speaker and hearer which adds, for
the first time, anything to the range of the meaning. It is, in
this case also, only gradually that the usage can be formed, and
according to the general fundamental necessary conditions of
language.
122. In every case where the need of denoting a conception
hitherto undenoted makes itself felt, one of the most obvious
Connotation. expedients is to choose a word easily formed, expressive of an
important part of the contents of a conception; in fact, some
prominent characteristic. Etymology teaches us that many sub-
stantives have thus proceeded from the appellations of more
simple qualities. But the conclusion is certainly not warranted
that all substantives took their origin in this way-for example,
that all are necessarily derived from verbs.
123. The second main kind of change in signification—a
change contrasting with the former is its reduction to one part of
the original contents; though commonly such reduction is accom-
panied by amplification on another side. It is hardly possible
CHANGE IN WORD-SIGNIFICATION.
85
to reduce the great mass of phenomena occurring under this
head into definite groups. A few of marked peculiarity may be
however mentioned. In many cases the appearance presented to
the eye gives the motive to the appellation, e.g., in the case of the
eye of a potato; the head or heart of a cabbage, the arm of a river,
the cup of a flower. A statue or a picture is named directly after
what it represents, as an Apollo, a Laocoon, the Adoration of the
Magi. We call a part of one object after the part of another object
which by its position corresponds with it; for instance, we talk of
the neck or belly of a bottle, the shoulder of a mountain, the foot
of a mountain, the tail of a kite; we call a measure by the name
of some object which has a size, or length, or breadth in some
measure corresponding, as a cubit, an ell, a foot. The correspon-
dence of the function fixes the name in the German word feder
for 'steel pen.' The analogy between place and time renders it
possible that we should transfer expressions formed to express
ideas of conceptions of time to those of place; e.g., long and short;
before, after; behind, and numerous other adverbs and prepositions.
The analogy between different modes of sense-perception renders
possible the transference of the impression made on one sense to
that made on another: take such instances as sweet, beautiful,
clear (originally applicable to hearing alone), and the Latin clarus
(originally applicable to the sense of sight alone). The words
used to denote sensual and corporeal ideas are transferred to
spiritual and intellectual [as in such cases as apprehension, re-
flection-nay, the word spirit itself, and the Latin anima]. Take
also such words as to feel, to see, sweet, bitter, lovely, taste, dirty,
great, small, lofty, low, warm, fire, to sting, to thrill, etc. Words
again which, strictly speaking, denote one species only, are con-
verted into symbols denoting a wider extension, as cat, crab, apple,
rose. By confining ourselves to one prominent characteristic, we

86
PRINCIPLES OF THE HISTORY OF LANGUAGE.
3. Trans-
ference to
may make proper names pass into class names, as when we talk
of a man as a Cicero, an Agamemnon, a Cato; and we have such
further developments as a cannibal, a vandal; Tom, Dick, and
Harry; John Doe and Richard Roe. Such adjectives as romantic,
Gothic, etc., may also serve as illustrations of the development.
124. The third main head of change in meaning is the trans-
ference of the idea to what is connected with the fundamental
Temporally, conception by local, temporal, or causal relations.
things
Spatially,
or Causally
associated.
125. The simplest subdivision of this is where a part is sub-
stituted for the whole. The part is in such cases always a
prominent characteristic, and is only as such capable of denoting
the whole. Take such instances as bow, blade, flèche, and the
MHG. rant (NHG. rand), 'rim,' used in epic poetry for shield.
Persons and animals are often named after characteristic parts
of the body and mind: as grey-beard, curly-head, thick-head;
red-breast, fire-tail; a good soul, a bright spirit; in French,
blanc-bec, grosse-tête, rouge-gorge, rouge-queue, pied-plat, gorge-
blanche, mille-pieds; esprit fort, bel esprit. Similar to this
usage is the application of names for objects commonly found
in connexion with others instead of those with which they are
connected. Such are names taken from garments, as blue-
stocking, green domino, a red-coat, a blue-jacket.
other appellations which are transferred from one object to that
included or contained within it, such as 'the town' in such cases
as the talk of the town; the Cabinet, the Church, the Court, etc.
Conversely we find a transference of the idea from the object to
its surroundings, as in the Round Table, the Porch, the Moun-
tain. Very commonly it happens that the name of a quality is
transferred to the person or thing possessing the quality, as in
age, youth; plenty, plain, desert, bitters: in German a quantity
of words in -schaft, as mannschaft, verwandtschaft [a termination
There are
CHANGE IN WORD-SIGNIFICATION.
87
answering to ship in English; and the usage may be paralleled
by such expressions as his worship in speaking of a magistrate].
There are others in -head, originally denoting condition, as God-
head; but there are many more instances in German of words
ending in -heit or -keit, which are similarly employed, such as
christenheit, mehrheit, neuigkeit; and the use of titles generally,
such as, your highness, his majesty, his excellence, his holiness, etc.
As the examples show, collective names take their rise in this
way as well as names for single persons and things; but the
words in question do not invariably form names for substantives.
The same truth holds good about the so-called nouns of action
as about the names of qualities. By nouns of action we mean
appellations of activity and conditions which are derived from
verbs—e.g.: overflow, train, income, government, providence, gilding,
warning; in German such words as rat, fluss, vortrab, zukunft.
In these cases the name of the action has been transferred to
its subject; but it might equally well be transferred to its
object, if object be taken in the widest sense. Thus it may be
transferred to the inner object, the consequence of which is the
denotation of the result―e.g.: rift, spring, growth, a rise (out of
a plain), assembly, union, education; or to the exterior object
which in any way is affected by the activity, thus, seed, speech,
doings, lamentations, bewailings, resort, excuse, dwelling; add the
German einfahrt, ausflucht: by this method names also are created
for the place where anything happens, for the means whereby
anything is brought to pass, etc. Here, too, we may classify the
practice of denoting writings by the name of their author, as
Have you read Shakespeare ?-or works of art by the name of
their painter or sculptor, as a Raphael, a Michael Angelo; further,
that of calling a person by some nickname derived from some
favourite word of his own, as Heinrich Jasomirgott [and Cedo

88
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Combination
of the above
Classes.
Change of
Meaning in
Groups of
Words.
alteram in Tacitus]; the nursery names of animals, as a bow-wow,
a gee-gee, etc.; besides these we may add the names of such
plants as noli me tangere, forget-me-not, etc.
126. Of course the different kinds of change in meaning may
follow each other, and so combine. Thus the word abendmal
has in one sense gained in its comprehension, seeing that it is
confined to the sacrament of the Lord's Supper, and the solemn
ceremony derived from it: on the other hand, it has lost some of
its meaning, seeing that it is used of a solemn festival which
does not take place in the evening. Rosary is used of a special
kind of necklace composed of beads, and used for a definite
sacred purpose, but has lost in its meaning all connexion with
roses. A horn is a wind instrument which may be made of horn
but the name may equally well apply to an instrument made out
of other materials.
127. It is common enough to find that some idea foreign
to the essence of a word, but only accidentally connected with
it, gradually becomes absorbed into its signification as a mere
accessory; and this is then thought of as the proper meaning,
the primary meaning passing out of memory. Thus names of
relations of place and time gradually pass into causal words: as
consequence, purpose, end (in such phrases as He did it to the end
that, etc.) means, way.
128. Seeing that all language proceeds in sentences-in fact,
that the unit of language, alike in thought and in sound, is
the sentence and not the word-it is natural enough that the
change in meaning should affect not merely the separate words,
but also groups of words and entire sentences.
These are
capable, of course, of assuming a meaning at first occasional,
which presently, however, by repetition becomes usual: a meaning
which is not covered by the words which we receive when we
CHANGE IN WORD-SIGNIFICATION.
89
connect the meanings of the words which compose the whole
group. A very few examples of this process may suffice,
as we shall have to recur to this peculiarity in Chapter XIX.
There are many combinations in German, in which the word
hand plays a great part, in which, however, the signification
hand is hardly thought of, unless special attention is drawn
to it.
Such are auf der hand liegen, an die hand geben, bei
der hand sein, unter händen haben, etc. We might parallel these
expressions in English by such phrases as well in hand, offhand,
hands off, at hand, etc. We cannot say in these cases that special
meanings of the word hand have been developed; they have rather
been obscured by the attention which we have come to pay to the
meaning of the phrase as a whole. English, like German, is full of
such turns of expression. In many of these the sense can only be
derived from the meanings of the several words by the aid of an
historical knowledge of language. Take such phrases as das bad
austragen, einem ein bad zurichten, einem das bad gesegnen (from
Clytemnestra's murder of Agamemnon in the bath), einen bären
anbinden, einen bock schiessen, einen ins bockshorn jagen (from
the tall talk of sportsmen) weder hand noch fuss haben, einem
einen korb geben, maulaffen feil halten, einem etwas auf die nase
binden, einem das pelz waschen, einem ein x für ein u machen, etc.
[In English we may cite such cases as to dine with Duke
Humphrey, to tell a cock and bull story, all his geese are swans,
to stuff one up, to give one the sack, to be half seas over, etc.]
Meaning
by the state
129. The entire store of ideas in the human mind strives to Scope of
attach itself to the vocabulary of language. But the circle of the conditioned
ideas of single individuals in any society differs widely from that of culture
of others in the same society, and what is more, the circle of individual
the ideas of each individual is always liable to expansion or
contraction. Hence it follows that a quantity of individual
of the
speaker.
90
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Scope of
Meaning
conditioned
of culture
of the whole
community.
peculiarities must necessarily be found in the ideas attached to
the vocabulary ;-peculiarities which are not of the least account
in the common definitions of the meaning in the case of the
single words, and groups of words. For instance, it is doubtless
true that the word horse has the same meaning for everybody,
in so far as everybody refers it to the same object; but it is
equally true that each man in his own particular line, a rider, or
a coachman, or a zoologist will connect therewith a richer quan-
tity of conceptions than any one who has nothing to do with
horses. The idea of the position of a father with respect to his
own child is composed of a series of factors, which are not always
present in their entirety when the word father is employed. We
can propose a definition of the word father which may satisfy
the claims of physical and juristic science; but the very point
which, according to this definition, makes up the very essence of
the fathership, is not contained at all in the ideas which a little
child connects with the word. The differences in the ground
covered by feeling and ethical judgment are the most sensible
and remarkable. What different individuals understand by good
and bad, by virtue and vice, is impossible to bring under one
definition, indisputable and undisputed.
130. As the circle of ideas of each individual connects itself
with the store of words at his disposal, it follows that the meaning
by the state of the entire store of words in a language must arrange itself
according to the entirety of ideas found in any people, and must
change as these change. The meaning of the words adapts itself
to the standard of culture attained from time to time by each
nation. This comes to pass, not merely by the fact that new
words are created for new objects and new relations, or that
kindred though different meanings become attached to the old
words (as in the instance of stahl and feder, 'a steel pen,' literally

CHANGE IN WORD-SIGNIFICATION.
91
'a steel feather '), but there are a quantity of unobserved changes.
which are not commonly remarked as such at all, and which are
the immediate result of a change in the whole culture of a nation.
[Take such words as humility, talent, spirituality, and the numerous
other words to which Christianity has given a deeper and more
spiritual significance.] Then again, such a word as ship may
have taken its origin when there were only the most primitive kind
of ships in existence, and the word may have lived on after a
time when ships are constructed of enormous size and complexity.
We do not in such cases assume any change in meaning, but still
it must be conceded that the ideas attaching to the word ship
have changed considerably. And thus it fares with all the products
of culture, whether they be objects cognisable by the senses, or
purely mental or intellectual conceptions.
CHAPTER V.
ON ANALOGY.
Material-
and Formal-
groups.
A
S has been noticed in Chapter I., single words attract each
other in the human mind, and the result is the appearance
of a quantity of larger or smaller groups. The reciprocal attrac-
tion depends always upon a partial correspondence of the sound
or of the meaning, or of the sound and the meaning conjoined.
The separate groups do not always run in separate parallel lines.
There are larger groups which comprise several smaller groups,
and a constant process of reciprocal crossing is taking place among
the groups. We must distinguish two main divisions which we
will denominate as the 'material' and the 'formal.'
132. A material-group is formed by, for instance, the different
cases of a substantive. This group again admits of division into
smaller groups according to two different principles; into cases on
the one hand of the singular, dual, and plural; or into the forms of
nominative of the singular, plural, and dual, and genitives, datives,
etc., of the same numbers; and these two methods of grouping
cross each other. A more complicated system of groups principal
and subordinate is supplied us by the forms of the verb, especially
in Greek. Larger material-groups with looser connexions arise
next by the connexion of all words which correspond in their
meaning. As a rule, partial agreement in signification is accom- ·
panied by partial agreement in the form taken by the sounds; and
ON ANALOGY.
93
this again commonly depends upon the etymological connexion of
the word. There are, however, material-groups based solely on
meaning and not upon sound: such as sein—werden, be—was,
here-there, good-better, am-is, parvus-minor, ópáw-eidov-
ὄψομαι.
133. We understand by formal-groups the sum of all nouns of
action taken together, of all comparatives, of all nominatives, of all
first persons of the verb, etc.
134. There are also in this case larger groups, which again
comprise smaller; for instance, the last-named 1st singular indica-
tive present may be grouped with the 1st singular subjunctive
present. Further, the connexion may be of a closer or looser
character as it may happen. The connexion of a functional cor-
respondence with one of sound is in the case of the formal-groups
not nearly so commonly formed as in the case of the material ones.
Commonly the formal-groups fall into several smaller ones, each
single one of which is held together by correspondence of its
sounds, whilst they differ among themselves; cf., the datives libro,
anno-mensae, rosae-paci, luci, etc. Then, according to the
measure, greater or smaller, of correspondence in sound, there arises
again a subordination of lesser groups to larger; cf.: gab, nahm—
bot, log-briet, riet-held, fell, which correspond with each other, as
contrasted with instances like sagte, liebte, etc. (and said and loved).
135. The material-groups are all the way through crossed with
the formal.
Groups:
Formal.
136. Not merely is it the case that single words collect themselves Proportion-
into groups, but analogous proportions between different words do Material-
the same. The motive to the creation of such proportion-groups,
which form at the same time an equation of proportions, is given
by the interpenetration of,material- and formal-groups just spoken
of. The basis for the comparison is in this case the correspon-
94
PRINCIPLES OF THE HISTORY OF LANGUAGE.
:
dence of the signification of the material element in the one direc-
tion and the formal in the other; so that we may call this division
the material-formal proportion-groups. A correspondence in sound
may possibly occur in both directions; cf. tag: tages: tage=arm:
armes : arme=fisch : fisches : fische; führen : führer : führung=
erziehen: erzieher: erziehung, etc.; cf. lead, leader, leading; ride,
rider, riding;—or with the permutation, which is possible in the
case of all proportions, of the connecting limbs, tag: arm:
fisch=tages armes fisches, etc. The correspondence in sound
may, however, be confined to the material element; cf. gebe: gab=
sage: sagte=kann: konnte; Lat. mensa : mensam : mensae=
mensae=hortus :
hortum : horti=nox : noctem : noctis, etc.; rauben: raub=ernten :
ernte=säen : sat=gewinnen: gewinst; as against gebe: sage : kann
=gab: sagte: konnte, etc. Equations, however, in which the cor-
respondence of sound is limited to the formal element, are of
much less importance, as gut: besser schön schöner, as are those
in which no correspondence of sound occurs at all, as in the case
of bin : was = lebe : lebte ; ὁράω : εἶδον = τύπτω : ἔτυψα ; am : was =
live: lived-go: went; do and did.
:
137. It is possible for proportion-groups to form themselves
inside the forms which belong to a material-group, as soon as an
arrangement of these is possible from different points of view.
Thus, in the case of the noun, the cases of the singular may be set
in proportion with those of the plural, as in hortus: horti: horto=
horti: hortorum: hortis. A verbal system gives still more mani-
fold proportions. We may, for example, propose equations like
amo: amas: amavi: amavisti=amabam: amabas, etc. There exists
here then no difference of the material element in the correspond-
ing members, as in the case of the material-formal proportion-
groups; but in its place a partial difference manifests itself in
the function of the formal element by the side of the partial
ON ANALOGY.
95
correspondence. A correspondence in sound may possibly con-
nect itself with the correspondence in the function; cf. amabam:
amabas=amaveram : amaveras (the tense stem being ama).
138. Another kind of equation of proportions depends on the
change in sound; cf. klanges (phonetically written klaññes): klang
(klank)=singe: sang=hänge: hängte, etc., or spruch: sprüche = tuch :
tücher=buch : büchlein (change between guttural and palatal ch),
[cf. wife: wives=calf : calves, etc.].
139. The members of each proportion here consist of words
connected etymologically, and hence showing connexion in their
material element with respect to their signification and the arrange-
ment of their sounds; showing at the same time, however, a
difference in sound, which manifests itself correspondingly in all
other proportions. The connexion of the formal elements in this
case does not enter into consideration. As long as we only con-
sider cases like klanges: klang=sanges: sang=dranges: drang, we
cannot definitely decide if we have not rather to do with a material-
formal equation of proportions. The change in sounds must—if it
is to be cited here-show itself in cases which as far as the relations
of function are concerned have nothing in common, and must
thereby show themselves as independent of the meaning. We
distinguish this division of proportion-groups as the material-
sound-groups, or etymological-sound-groups.
This
140. A further kind arises from syntactical connexions.
differs from those hitherto described, in the fact that the con-
nexion of the members of which the single proportions are made
up, forces itself on the mind from without, i.e. by hearing or speak-
ing. The connexion of analogous proportions with each other
must likewise be called into existence by attraction in the mind
itself. For instance, sentences like Rose a nurse of ninety years,
associate themselves with others like Out spake the mighty Appius ;
96
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Influence of
Proportional
Speech
Creation).
the predicate in both cases preceding. Again, associations come
about like that of pater mortuus with filia pulchra and caput
magnum (where in each case there is agreement in gender, number,
and case); and by such associations we get the equation rose:
nurse=spake : Appius, and pater : mortuus=filia: pulchra=caput :
magnum. The feeling for a particular function associates itself
with the outer form of syntactic connexion, and this function then,
in common with the exterior form, constitutes the bond which
holds the proportions together. All syntactic functions can be
abstracted from such proportions, and such alone. And hence it
arises that the syntactical proportion-groups are in some degree
the necessary antecedent for the creation of the formal-groups
and of the material-formal relation-groups. For instance, it is
impossible for genitives to group themselves together unless
such connexions as the house of the father, Charles' brother, etc.
do so antecedently.
141. There is hardly a word in any language which can be said
Groups upon to stand completely out of the groups here sketched. Others are
(Analogical always to be found in some respects like them, to which they can
attach themselves. But with respect to the greater or less variety
of the connexions into which a word enters, and with respect to
the intimacy of the connexion, there are notable differences. The
process of grouping comes to its accomplishment so much more
easily, and is so much the more durable, on the one hand, the
greater the correspondence in signification, and in the shape taken
by the sounds ;-on the other, the more intensely the elements
adapted to form groups are pressed on our notice. In the last
respect we have to take into account for the proportion-groups
on the one side, the frequency of the occurrence of single-words;
on the other, the number of possible analogous proportions has to
be taken into account. Where the single elements are not impres-
ON ANALOGY.
97
sive enough, or where their inter-correspondence is too weak, they
either fail to unite at all, or their union is at best a loose one.
There are again in this case numerous intermediary grades possible.
142. Those proportion-groups which have gained a certain
degree of solidity are of supreme importance for all linguistic
activity, and for all development of language. It is unjust to
this important factor in the life of language to neglect to take it
into any account, until it produces an actual change in the use
of language. One of the fundamental errors of the old science of
language was to deal with all human utterances, as long as they
remain constant to the common usage, as with something merely
reproduced by memory, and the result of this has been that we
have not been in a position to form any right conception of the
share taken by proportion-groups in the alteration of language.
True it is that W. v. Humboldt insisted on the fact that speaking
is a perpetual creation. But even down to the present day we
find ourselves confronted by spirited and often unintelligent
opposition as soon as we seek to draw the consequences of this
view.
༄་
143. The fact is that the mere reproduction by memory of what
it has once mastered is only one factor in the words and groups.
of words which we employ in our speech. Another hardly less
important factor is the combinatory activity based upon the exist-
ence of the proportion-groups. The combination consists to some
extent in the solution of an equation between proportions, by the
process of freely creating, for a word already familiar, on the
model of proportions likewise familiar, a second proportional
member. This process we call formation by analogy. It is an
incontrovertible fact that a quantity of word-forms and syntactic
combinations which have never been introduced into the mind
from without, are able not merely to spring into being by the
G
98
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Analogical
Creation in
Syntax.
aid of the proportion-groups, but also from this time forward are
confidently produced without the speaker having any conscious-
ness that he is leaving the safe ground of what he had learnt.
For the nature of this proceeding it is immaterial whether
anything, during the process, comes out, which was usual in
language before, or something which was not there before.
Further, it makes no essential difference whether the new stands
in any kind of contradiction to that which prevailed hitherto it
suffices that the individual who is uttering a word feels no con-
tradiction with what he has already learned. In other cases it is
doubtless true that a reception has taken place from outside; but
the after-effects of this would be too weak for the matter received
to be again recalled into memory, were it not that the proportion-
group into which it has been ranged comes to its aid.
144. It will be readily conceded that a minimum of the sen-
tences which we utter is learnt by heart as such that most of
them, on the contrary, are composed on the spur of the moment.
If
we try to master a foreign language methodically, rules are given us,
according to which we join together the single words into sentences.
But no teacher, who has mastered the rules of his profession, will
fail, as he goes on, to give examples to illustrate the rules, i.e. with
a view to the sentences which are to be independently framed.
Rule and example mutually supplement each other in their activity,
and we can see by this procedure of the teachers that certain
advantages attach to the concrete example which are lacking to
the abstract rule. In the process of naturally mastering one's
mother-tongue no rule, as such, is given, but only a number of
examples. We hear gradually a number of sentences which are
connected together in the same way, and which hence associate
themselves together into one group. The recollection of the special
contents of the single sentences may grow less and less distinct in
ON ANALOGY.
99
the process; the common element is always strengthened anew by
repetition, and it thus comes about that the rule is unconsciously
abstracted from the examples. Precisely because no rule is given
from without, no single example suffices, but only a group of
examples whose special contents appear a matter of indifference.
For the idea of the general applicability of the examples cited, which
gives each individual the feeling that he is authorised to compose
sentences on his own account, becomes developed only by this
process. As soon as a rule learnt by rote has been frequently
enough applied, the time has arrived when it can begin to operate
automatically. It is unnecessary to call either the rules or
a particular example into consciousness, in order to still form
perfectly correct sentences. In fact, as far as the common practice
in practical exercise goes, the learner arrives by a circuitous route
at the very same point attained by the man who has enjoyed no
grammatical training at all.
145. One of the main disadvantages entailed on the individual
who has received his ideas by example only, as opposed to him.
who has received the rule and example at once, consists in the
fact that he is not, like the latter, fully instructed as to the appli-
cability of his examples. For instance, whoever hears the German
preposition in repeatedly employed with an accusative, will be
easily led to apprehend this as the ordinary usage of in; and
whoever hears the same word employed sometimes with an accusa-
tive, and sometimes with a dative, will need at least a certain
time before he has rightly found out the difference, and meanwhile
will use one or the other promiscuously. In a case like this, a
definite rule simplifies matters. Such a combination of two
groups which by usage should be kept apart, becomes more
possible the finer the logical distinction needed for the purpose,
and the greater the room left thereby for subjective apprehension.
100
PRINCIPLES OF THE HISTORY OF LANGUAGE.
146. But a group is in the best position for extending
its precedent over the area of a related group, if it decidedly
surpasses the latter in respect of the frequency of the cases
which occur. And besides, much that
that occurs in the usage
of language stands isolated and alone, and neither fits in with
any consciously abstracted rule, nor with a group that has uncon-
sciously arisen. But all that part of language which lacks the
support of an environing group, or which enjoys it only in a
limited measure, proves, unless impressed by repeated usage
intensely upon the memory, not strong enough to withstand the
power of the larger groups. Thus, to cite one example, in German,
as in other IE. languages, the rule is, that in cases where two
objects depend on one verb, one stands in the accusative, the
other in the dative. But there are accompanying this rule certain
cases, and there were formerly more, in which a double accusa-
tive stands instead. These cases have to be learnt by heart, and
always must have been. The instinct of language becomes.
hesitating in consequence of the contravention of the common
rule, and this may finally lead to the disappearance of the con-
struction as it becomes isolated. We hear at the present day
almost as often the expression 'Er lehrt mir die kunst, as 'Er
lehrt mich die kunst;' and no one says any longer 'Ich verhehle
dich die sache,' after the MHG. model, but only 'Ich verhehle dir?'
Analogy in
Inflexion
譬
​Formation.
147. But the creative activity of the individual is very marked
and Word in the area of word-formation, and still more in that of inflexion.
Very few of the verbal and noun forms that we pronounce are due
to reproduction by a mere effort of memory; there are many which
we have never before spoken nor heard; others so seldom, that
without the aid of the groups with which they have connected
themselves, they could never be recalled into consciousness. In
any case the ordinary rule is that production and reproduction
ON ANALOGY.
ΙΟΥ
operate together, and this in very different relative proportions to
each other.
148. The operations of analogy force themselves on our atten-
tion with special clearness in the case of the grammatical adoption
of the inflexional forms of a foreign tongue. A quantity of para-
digms are committed to memory, and then of the single words
only so many forms are impressed on the mind as are sufficient
to fix the fact that they belong to such or such paradigm. Some-
times a single form is sufficient. The other forms are created at
the moment when they are wanted, according to the paradigms—
i.e. according to analogy. In the beginning the paradigm which
has been committed to memory will be that which each will have
before his eyes.
But after a fairly large number of words has
been formed after this model, and these have left traces in the
mind, the act of creation follows even without the word which has
served as paradigm coming into consciousness.
The forms at
an earlier period shaped from other words now co-operate, and
the result is that only the formal element common to all alike
comes into consciousness, while the different material elements
reciprocally bar each other's way. And thus now the relation-
ship of the speaker to the inflexion-forms at the moment of his
application of them is much the same as that gained in the
natural process of mastering the mother-tongue. This natural
process of learning leads, by a less direct but in the end not less.
certain path, to the same goal. In this process there is no special
adhesion of the formal elements to a special single material one,
and the totality of the possible forms never ranges itself in a
definite sequence into one series. The principle is never laid down.
that such a word must accommodate itself to another. The cir-
cumstance that a quantity of forms of different words have like
relationships to each other is sufficient to produce the feeling that
102
PRINCIPLES OF THE HISTORY OF LANGUAGE.
it
Divergence
we are justified in carrying this likeness still further. As soon as all
the forms of a quantity of words have imprinted themselves on the
mind and closed into groups, the instinct of language regards it as
self-evident that the forms of other groups belong as well to such
groups: e.g., that the other cases of a substantive belong as a
necessary complement to the nominative or genitive of a sub-
stantive. And hence it comes also that we do not apprehend
each case and each verbal form as a special word, but that we
comprise at once the whole series of forms under the ordinary
titular form (nominative, infinitive) of a substantive or a verb.
[Cf. Welsh plant, ‘children.']
149. The circumstances attendant on word-formation resemble
only partially those of inflexion. It is true that numerous methods
of such formation come into being as easily and simply as the
flexional forms, as in the case of the comparative and super-
lative from the positive. In other cases the words transmitted
call forth creations by analogy only in a limited quantity, and
again in the case of others they call forth none at all. This
difference of relation is simply conditioned by the different
adaptability of the material transmitted to group-formation.
150. Seeing that most of the forms common in language
of Analogical permit of being reduced to similarly related groups, it is natural
from Usage. enough that by the aid of proportions, groups should often be
Formation
created which were before common in language. But if this was
to be invariably the case, on the one hand, all forms which could
have been formed according to proportion would have to be already
once formed; and, on the other hand, such a perfect harmony
of the form-system would have to exist as we nowhere find in
fact; or at least, where different methods of formation existed
side by side (different classes of declensions or conjugations, different
methods of forming a noun of the agent from a verb, etc.), corre-
ON ANALOGY.
103
#
sponding forms from different classes would never take analogous
forms; but every single form would indicate, without the least
ambiguity, to which of the existing classes the word in question
belonged. As soon as a form is capable, in respect of its struc-
ture, of belonging to more classes than one, it becomes likewise
possible to construct the other forms attaching to it upon more
than one scheme. Which of the two alternative schemes shall
then prevail depends solely upon their relative strength.
A proportional formation meets with no check in the mind
where no expression for the function for which it is created
is as yet in existence. Neither does it where a divergent
expression is already in use, which has not, however, been
transmitted to the person in question; and this is common
enough in the rarer class of words. If, however, the ordinary form
has once been received into the memory, it is a mere question of
power if, at the moment when a particular function is to be
exercised, a form for this purpose is raised into consciousness by
means of simple reproduction or by the aid of a proportion. It
may in this case happen that the proportion asserts itself in the
first place, but that the connexion formed at an earlier period.
with the remembered image of the usual form is still strong
enough to render perceptible the contrast between the new creation
and the image in the memory. One begins then to reflect that one.
was going to say, or has already said, something wrong. This is,
then, one of the various ways according to which one can commit
an error in speech. We shall also have to admit such error when
the speaker, after he has spoken, does not of himself become aware
of the contrast with the picture on his memory, but immediately
recognises it when his attention is called to it by a gentle hint.
But the power of the memory-picture may be so small that it has
no power to contend against the proportional formation, and the
latter prevails unchallenged.
104
PRINCIPLES OF THE HISTORY OF LANGUAGE.
151. By the operation of the groups each individual has thus
at once the power and the disposition to pass out of the usual in
language, and both in a high degree.
152. We have now to notice that everything which is created
in this manner leaves behind it a lasting effect. Should this not
be from the very outset strong and tenacious enough to render an
immediate reproduction possible, still it renders a future repetition
of the same creative process more easy, and contributes towards
the further pushing back of the possible opposing obstacles. By
repetition like this, there can afterwards be added to the new
formation such power as it wanted in order to be directly repro-
duced.
153. But each such transgression of usage, when confined to the
case of an individual, where it adds something to the usual sense
without contradicting it, gives the impression of a certain auda-
city; where it does contradict it, it appears as an unmistakeable
fault. It is possible for such fault to remain isolated, and not to
pass into any sort of custom; it is likewise possible for it, even
after it has passed into a custom, to be laid aside again by the
force of intercourse, whereby the usual is again appropriated,
whether this be for the first time or anew. Assuming, however,
that it is not laid aside, it commonly perishes with the individual,
and does not easily pass to any other. It is much easier for a
creation which conflicts with no previously existing one so to
pass; in this case it is far easier for a single individual to give the
impulse. On the other hand, it fares just the same with the re-
placing of the hitherto-usual by the new as with the change in
sound and meaning. It is only on an occasion when, within a
limited area of intercourse, the same new creation makes its influ-
ence completely felt, and this spontaneously, in a large number
of individuals, that a change of usage can shape itself.
The
ON ANALOGY.
105
possibility of such a spontaneous correspondence in the case of
several persons depends on the overwhelming agreement in the
organisation of those groups of ideas which influence human speech.
The greater the number of those among whom the new formation
makes its appearance, the easier will be the process of transmis-
sion to others; in fact, the faster will what at first seemed an
error gain in authority.
154. As with respect to the sound-relations, and with respect
to the meaning attributed to the words, we find the most violent
departures in new analogical formations from ordinary usage in
the language of children. The less finished, and the weaker the
impression made by the single words and forms, the less the
restrictions put upon the new creation; the wider play, in fact,
will it be found to have. Thus all children have a tendency to
employ the regular and usual methods of composing words for the
irregular and unusual modes, which do not yet adhere to their
memory. In NHG. [or English], for instance, they are disposed
to treat all verbs as weak. If, with the increasing development
of the individual, the process of new formation grows more and
more rare, this is of course not in consequence of any disappear-
ance of a store of wealth which was originally there, but of the
diminished necessity for such formation; since, to serve the very
purpose for which they were at an earlier period called into being,
forms caught up by memory tend to present themselves more and
more. Commonly it will be found that in this particular area the
aberrations of children's talk leave behind them no particular
consequences for the further development of language in general;
but here and there traces remain behind even of this. Particularly
in cases where the adults are disposed to new word-formations will
a corresponding inclination manifest itself even more strongly in
the children; and these will give themselves free career in gratify-
106
PRINCIPLES OF THE HISTORY OF LANGUAGE.
ing the inclination as soon as ever the necessary restriction due to
the language of the adults is removed.
155. An analogical new formation has no power to drive out of
the field at a single blow a pre-existent form of similar meaning.
It is scarcely conceivable that the picture of the latter should
at one and the same moment become so faint in all the indivi-
duals of a community that analogical formation could proceed on
its course without let or hindrance. On the contrary, it always
happens that certain individuals persistently maintain the old form,
while others avail themselves at once of the new formation. But
as long as ever an unbroken intercourse is maintained between
the two, there must be some kind of compromise as well, and
thus, in a smaller or more numerous group of individuals both
forms must necessarily become current. Not till after a long and
obstinate struggle between both forms can the new formation be
invested with undisputed sway.
156. Seeing that the new analogical formation is the solution of
an equation between proportions, it follows that at least three
terms must be present to enable such an equation to be instituted.
Each must be capable of being compared in some point with the
other-i.e. in this case an agreement must appear, with the one
in the material, with the other in the formal element. Thus in
Latin such an equation may be instituted as this,-animus: animi
=senatus: x, but not animus: animi=mensa : x. And thus it happens
that a word can only undergo any influence depending upon analogy,
when it agrees with these in the formation of one form or of several.
No doubt it is true that influence at times makes itself felt without
such agreement, but this can then hardly be called formation by
analogy in the strict sense of the word. A flexional termination.
may be felt owing to the special frequency of its occurrence as the
regular and usual termination for a flexional form. It then extends
ON ANALOGY.
107
probably to other words even without the support of words simi-
larly formed. Of this nature is, in Attic Greek, the extension of the
genitive termination -ou from the second declension to the mascu-
lines of the first: we find ToλíTOU instead of Tоλíтew, as we should
expect, from the Homeric -ao and Doric -a; the coincidence of
both in gender has in this case sufficed to set the influence at work.
[There are many similar instances in modern Greek.] The genitive
dual of the third declension in Greek has borrowed its termination
from the second ; ποδοῖν is formed after the analogy of ἵπποιν.
In German the genitive ending in s is transferred to the feminine
proper names with the termination a, as Berthas, Claras.
157. Of course new formations come into existence as well on
the basis of the proportion-groups described before, which combine
out of forms of a similar material-group. In MHG. the third per-
sons plural run as follows:-Indic. pres. gebent, conj. geben; indic.
pret. gåben, conj. gæben. In NHG. the form geben has also been
adopted in the indic. pres., after the analogy of the other three
forms; in late MHG., ent has found its way into the other forms
by a reverse process. The second person sing. indic. pret. of the
strong verb, which was formed after a peculiar fashion in MHG. (du
gabe, ware) is altered after the analogy of the other second persons.
Sound-
158. The creative effect of analogy in the region of sound-sub- Analogy in
stitution has hitherto received very little attention. Sound-substi- Substitution.
tution is in the first instance, as we have seen, an effect of sound-
mutation, which sets in when a similar sound or group of sounds
has split into several, in consequence of different conditions of
sound. As long as these conditions are maintained, and, besides
this, no disturbance of the effects of the sound-transformation
comes in through other influences, it is possible that the forms.
which have arisen through the process of sound - mutation may
arrange themselves into proportion-groups (see examples, p. 95).
108
PRINCIPLES OF THE HISTORY OF LANGUAGE.
In this case we may speak of the change in sound as a living one.
If, however, the conditions disappear which caused the different
treatment of the sound, then no more etymological sound-pro-
portions are able to form themselves; in fact, the process of
sound-change is petrified. Thus, for instance, the change between
h and g in ziehen - zug, gedeihen-gediegen, is conditioned no
longer by relations in the speech of the present day-the reason
which originally called into existence this substitution, że. the
changing Indo-European accent, has long been put out of account.
The change between hoher-hoch, sehen-gesicht, geschehen -
geschichte, coincides no doubt with a change of the position within
the syllable; but seeing that in most cases where a perfectly
analogous change of position takes place, no sound-change any
longer occurs (cf. rauher—rauh, sehen—sah and sieht, geschehen,
-geschah, and geschieht), this change is likewise dead. It is
different in MHG., where it is an all-prevailing principle that the
sound of a ch answers to an h in the beginning of a syllable where
it stands after the sonant of the syllable. Thus we find rûher
-rûch, sehen-sach, geschehen-geschach; before s and t in early
MHG. we certainly find an h written as well (as in sihst, siht); while
in later MHG. we find it likewise denoted by ch (sichst, sicht).
159. The material-sound proportion-groups are productive, it will
be observed, in the same way as the material-formal. It is for instance
inconceivable that both the different pronunciations of the German
ch are specially learnt by each individual to meet each special case;
rather do memory-impressions and creation-by-analogy operate
simultaneously; and, indeed, without the co-operation of the latter
the invariability in the change between both could not be acquired
which actually exists.
160. The operation of analogy in the case of sandhi-phenomena
is past all doubt. What would on any other hypothesis be the
ON ANALOGY.
109
}
explanation of the fact that in French the final consonants s, z, t, n
are treated differently according as the word connected with them
begins with a consonant or a vowel? It is no doubt possible that
a quantity of such combinations as nous vendons-nous aimons,
un fils-un ami, have perpetuated themselves by memory from
generation to generation, but certainly this cannot be the case
with nearly all that come into use, and have at an earlier period
done so.
In spite of this, the change is accurately observed even
by those quite unversed in grammar, and in the case of any new
combination at will.
161. The action of etymologico-phonetic proportional groups
commonly produces such forms as would also be produced by
the fundamental sound-mutation. Still it sometimes happens
that new forms are produced which would be impossible ac-
cording to the strict law of sound-change. The reason of this
is either a reversal of the proportions not properly justifiable, or
a confusion of the relations by means of later sound-mutation.
162. The sound-law in many High and Middle German dialects
holds good that n in the auslaut of syllables has disappeared, but has
maintained itself at the end of the word, when the following word
commences with a vowel, and it has transferred itself to this: thus
in Swabian ę ros (ein ross)—e-n ôbęt (ein abend), i duę (=MHG. ich
tuon)—duę-n-i. The speakers are therefore accustomed to find an
ʼn in many cases apparently intruding between a vowel-ending and
a vowel-commencement; and in consequence of this, the n trans-
fers itself to cases where in the older time no n existed. Thus in
Switzerland we find combinations like wo-n-i, 'where I;' sę-n-išš,
'so it is;' wię-n-e, 'how a;' so-n-e, 'so a;' bî-n-em, 'by him;' tsię-n-ęm,
'to him.' The same phenomenon meets us in Swabia, e.g. in the
dialect of the country round Horb: bei-n-ęm, 'by them;' zuę-n-enę,
'to them;' dî mâ-n-i-dich mag ich, lô-n-ęms—lass es ihm, 'leave it
IIO
PRINCIPLES OF THE HISTORY OF LANGUAGE.
to him;' gei-n-çms—gib es ihm, ‘give it him;' while similarly in
Bavarian Swabia, and in a neighbouring portion of the regular
Bavarian area si-n-ist=sie ist, wie-n-i=wie ich, etc. In the Carin-
thian dialect we also find ba-n-enk-bei euch. In the old Provençal
we find the form fon, a variant of fo, formed after the analogy of
bon-bo. Under this head comes also v épeλKVσтIKÓν, so far as
it is not etymologically justified.
163. [Familiar examples of this transference of n are seen in
cases like newt in English for an eft, etc.]
164. The same law which holds good in the Alemannic and
Swabian for n holds good in Bavarian for r. The Bavarian, there-
fore, says der arm, but de jung; ęr is, but ę hât; meï bruędęr odęr i,
but i odę mei bruędę. In consequence of this, we see combina-
tions arise like wię-r-i—wie ich, ge-r-e—gehe er, dâ się-r-i—da sehe
ich, kâę-r-i—kann ich, aę-r-i—abhin=hinab. Correspondingly we
find MHG. jârâ, nûrd (from já), nû+â, which must be explained
after the analogy of the relation da (from an older dâr) to dârane,
wa to wârane, hie to hierane, sâ to sârie.
165. Double formations in the phonetics of sentences are doubt-
less the domain in which this kind of analogical creation most
commonly manifests itself. But it is not confined to these. If in
late MHG. after the final e is dropped from zahe, geschahe, hæhe, etc.,
the forms zach, geschach, hæch take their rise, we must probably not
suppose a transition of the h sound into ch: the forms have rather
attached themselves to the analogy of the change which existed
before between hoch-hohes, geschehen-geschach, etc. It will be
precisely the same in cases like sicht, geschicht (written in olden
times siht, geschikt), from sihet, geschihet.
CHAPTER VI.
1
A
THE FUNDAMENTAL FACTS OF SYNTAX.
is the
Expression
combination
of several
ideas.
LL linguistic activity consists in the formation of sentences. A Sentence
The sentence is the linguistic expression or symbol, denot- Linguistic
ing that the combination of several ideas or groups of ideas has of the
been effected in the mind of the speaker; and is at the same time
the means of reproducing the same combination of the same
ideas in the mind of the hearer. Any narrower definition of the
sentence than this must be rejected as insufficient. Among the
common errors touching the essence of a sentence, we must reckon
the notion that it must contain a finite verb. Combinations such
as omnia præclara rara; summum ius summa iniuria; träume ·
schäume; I a liar?-I thank You?-YOU talk to ME?-are in
every respect sentences as good as The man lives; he is dead.
!
denoting the
tion.
167. We possess the following means of expressing in language Means of
the combination of ideas: (1) The juxtaposition of the words Combina-
which correspond to the ideas by themselves; (2) the order of
these words; (3) the gradation amongst these in respect to
energy of utterance, to the relative strength of emphasis, as, Charles
is not COMING-Charles is NOT coming; (4) the modulation of the
pitch-cf., Charles is coming, as spoken in a mere assertion, and
Charles is coming?-as an interrogative sentence; (5) the time,
which is commonly found to stand in close correspondence with
the energy and the pitch; (6) link-words, such as prepositions,
I 12
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Subject and
Predicate,
Psycho-
logical and
Gramma-
tical.
conjunctions, and auxiliary verbs; (7) the inflexional modifica-
tion of the words, whether (a) the method of the combination be
more closely defined by the inflexional forms themselves (patri
librum dat), or (b) the connexion between the words be denoted
by formal agreement, as in anima candida. It is self-evident that
the two last methods have only been able to shape themselves
gradually by means of a somewhat prolonged historical develop-
ment, while the five first-named stand at the disposal of the
speaker from the very outset. But (2) to (5) inclusive are
not always directly determined by the natural course of the ideas
and feelings they represent, but are capable of a traditional de-
velopment [cf. the method of distinguishing homonyms in Chinese,
by lowering or raising the tone].
168. The way in which the ideas are capable of being connected
with each other is fixed with more or less exactness according to
the frequency and definiteness of the methods employed. It is
with regard to the method of combination, just as with respect to
the single idea. The linguistic expression for this does not need
to be equivalent to the psychical relation as it exists in the mind.
of the speaker, and as it is to be produced in the mind of the
hearer. It may be much less definite.
169. It thus happens that every sentence consists of at least two
elements. These elements are related to each other, not as exact
equivalents, but are differentiated according to their function.
They are termed subject and predicate. These grammatical
categories repose on a psychological, a logical, relation. No doubt
we have to distinguish between the psychological and the gram-
matical subject or predicate, as the case may be-since the two
do not always correspond, as we shall presently see in detail. But
still it remains true that the grammatical relation is built up solely
on the foundation of the psychological.
THE FUNDAMENTAL FACTS OF SYNTAX.
113
170. The psychological subject is the group of ideas which is
first present in the consciousness of the speaker or thinker; with this
a second, the psychological predicate, unites itself. The subject is,
to speak with Steinthal, the apprehending portion; the predicate
the apprehended. Von der Gabelentz¹ is quite right in defining
both elements from the point of view of the hearer. The psycholo-
Gabelentz
gical subject, according to him, is that on which the speaker wishes
to make the hearer think, and to which he would direct his obser-
vation; the psychological predicate that which he wishes him to
think about it. Still, such definition of the predicate might easily
lead to an erroneously narrow conception, such as is commonly
current in our grammars. We must abide by this principle, that
it depends only upon this, that one idea is in consciousness con-
nected with the other idea.
171. We are accustomed nowadays to attach a narrower sense
to the relationship of the subject to the predicate. If the predicate
is a noun, we demand for the normal structure of the sentence
that it should either be identified with the subject, or that it should
denote the wider conception to which the narrower one of the subject
should be subordinated, or that it should announce a quality inher-
ing in the conception of the subject. But in the case of proverbs
and proverbial expressions relations of quite another nature are
expressed by the grammatical form of the juxtaposition of subject
and predicate; take such expressions as one man one vote; much
cry and little wool; first come first served; a word to the wise;
like master like man; better aught than nought; in German, gleiche
brüder gleiche kappen; viel feind viel ehr'; viele köpfe viele sinne; alter
fuchs alte list; klein geld kleine arbeit; neuer arzt neuer kirchhof;
heisse bitte kalter dank; kurz gebet tiefe andacht; roter bart untreue
art; gevatter übern zaun gevatter wider herüber; glück im spiel unglück
1 Zeitschrift für völkerpsychologie, vi. 378.
H
t
[14
PRINCIPLES OF THE HISTORY OF LANGUAGE.
in der liebe; mit gevangen mit gehangen; früh gesattelt spät geritten;
allein getan allein gebüsst: correspondingly in other IE. languages,
as in French we find bon capitaine bon soldat; bonne terre mauvais
chemin; longue langue courte main; brune matinée belle journée ;
froides mains chaudes amours; fèves fleuries temps de folies; soleil
à la vue bataille perdue; point d'argent point de Suisse, etc.
Certainly we are accustomed to apprehend such sentences as
abbreviated hypothetical periods, and accordingly to set a comma
between the two component parts: but the fact that they may be
paraphrased by a hypothetical period (such as where there is much
cry there is little wool) does not concern us. The grammatical form
of such is precisely the same as that of borrow sorrow; most haste
least speed. In the first sentences formed by children, the mere
collocation of words serves for the expression of all possible
relations.
172. Examples taken from experience are given by Steinthal
(Einl. pp. 534-6): cf. papa hat-papa has a hat on; mamma baba=
I will sleep with mamma. Where a foreign language has to
be employed, with which the speaker is only imperfectly ac-
quainted, he turns in his need to the same primitive method of
rendering himself intelligible, and is understood, supported as he
is by the situation. He says, for instance, wine table, and is under-
stood to mean that he wishes the wine to be placed on the table.
The conditions which cause the production of such sentences, and
which enable the hearer to guess the unexpressed relation of the
conceptions, occur, of course, not merely in the commencement of
each man's linguistic activity, or in that of mankind generally speak-
ing, but at all times. If they come to be employed in the stages.
of higher development only in a limited degree, the reason of this
is merely that more perfect means of expression are at hand.
173. Originally there was one method, and one only, of mark-
THE FUNDAMENTAL FACTS OF SYNTAX.
115
tone.
distinguish-
and Predi-
Emphasis,
ing the difference between subject and predicate-i.e. stress of Means of
In the case of the isolated sentence, the psychological ing Subject
predicate is always the more strongly accented, as the more im- cate:
portant portion of the sentence, and as containing the new matter. Position, etc.
This may be regarded as a law pervading all nations and all ages.
The place in the sentence may have afforded a second means of
distinction. Von der Gabelentz, in the treatise mentioned above
(p. 376) lays it down that the order subject-predicate (both regarded
as psychological categories) may be held as admitting no exceptions.¹
This view does not seem to me quite correct. In pronouncing judg-
ment on this question, we must leave out of consideration the
language, and the cases in which a strict rule has been laid down by
tradition for the position of the grammatical subject and predicate.
We may only cite cases in which both may change their place—
cases therefore in which the position is conditioned, not by gram-
matical, but solely by psychological rules. The view held by Von
der Gabelentz, that a preceding grammatical predicate is always the
psychological subject, no doubt meets in many instances the case,
as, for instance, in this passage from Goethe: Weg ist alles, was
du liebtest, Weg, warum du dich betrübtest, Weg dein glück und
deine ruh.' If, however, we say 'A puff of wind caught the hat,
and away it went, we cannot possibly take away as a psycho-
logical subject. In the same way there exists a correspondence
between the psychological and grammatical subject if, in reply to
the remark, 'John seems a sensible man,' the answer should be
given, "Tis an ass that he is;' and so in many cases.
The idea
of the subject, no doubt, always precedes in the consciousness of
the speaker; but as soon as he begins to speak, the more significant
idea of the predicate may press so far into the foreground that it
1 Wegener, on the other hand (p. 31 sqq.), holds the precedence of the predicate to
be the natural order, a view which I cannot adopt. [Cf. Herbert Spencer's Essay on
the Philosophy of Rhetoric for a discussion of the question.]
style
116
PRINCIPLES OF THE HISTORY OF LANGUAGE.
must be uttered in the first place, and the subject not be added
till afterwards. This often occurs when the idea of the subject
was previously present in the conversation, as in the instances just
mentioned. Then the person addressed has, as a general rule,
while he is hearing the predicate, the subject belonging thereto in
his mind, which, indeed, may as well remain unexpressed: cf.,
'What is John?'-A merchant.' But even when the person
addressed is not prepared for the subject, a desire to produce.
superior emphasis may cause the predicate to come to the front.
The speaker, then, in his interest in the presentation of the
main idea, holds the accurate guidance of his interlocutor as
of secondary importance; and it does not occur to him till later
that such guidance is necessary. It is a similar psychological
process when the subject is expressed at first by a pronoun,
whose relation is not self-evident to the person addressed, and
is not expressed more definitely till later, as, She is coming, my life,
my dear (TENNYSON); Ist sie blind, meine liebe (LESSING); sie
hindert nicht allein nicht, diese binde (ib.); was für ein bild
hinterlässt er, dieser schwall von worten (ib.); MHG. wie jâmerlîche
ez stât, daz hêre lant (WALTH. v. d. VOGELWEIDE); si ist iemer unge-
schriben, diu fröude die si hâten (HART. v. AUE); Fr. elle approche,
cette mort inexorable. It is clear from the citations given above
that the sentences with a psychological predicate preceding have
a relationship with the sentences to be considered further, in which
the predicate alone is expressed. They are an anomaly, as against
the custom of placing the subject first, which obtains in less
passionate narration or description; but they are an anomaly
which still cannot be set aside, and an anomaly of not unfrequent
occurrence. Thus the position of the words cannot be regarded as
a means of marking the separation between subject and predicate
given with the commencement of sentence-construction.
THE FUNDAMENTAL FACTS OF SYNTAX.
117
Abstract
174. Just as single words may possess concrete and abstract Concrete and
significations, so may sentences as well. A sentence is concrete Sentences.
when one of the main members of it, the psychological subject
or the psychological predicate, is concrete. Normally the subject
which gives the sentence its concrete nature would be concrete.
Concrete and abstract sentences do not need to be different
according to the form of their expression. We can say generally
with respect to the nature of a horse, The horse is swift, just as
with respect to some particular horse we can say, The horse is
worthless: and the different nature of the sentences can only be
told by the connexion and the situation. In the first sentence
we might employ a plural means of expression, Horses (or, all
horses) are swift. But it cannot be said in that case to remain
properly abstract; for 'all horses' must mean all horses which
exist, and is consequently to be looked on as a concrete
expression. If the subject is concrete, the sentence cannot be
abstract. No doubt the different possibility exists that the
predicate can be thought of as something simply attaching to
the subject, something stationary or repeating itself, or as some-
thing attaching itself only at a particular time. In the first case
there exists in some sense a middle step between an abstract and
a concrete sentence, and it may be allowed in default of a better
to employ for this kind of sentences the expression 'abstract-
concrete.' This difference, however, does not require to be met
by a corresponding difference in form of expression. He speaks
quickly may denote 'He speaks quickly at this moment,' and 'He
commonly speaks quickly; He is lazy, may denote a fault on
a given occasion, or a lasting characteristic.
apparently
175. Our assertion that two members at least go to make up Sentences of
a sentence seems to be contradicted by the fact that we find one element
sentences consisting of only a single word or of a group forming a
only.
118
PRINCIPLES OF THE HISTORY OF LANGUAGE.
<
unit. The contradiction is explained by the fact that in this case
one member of the sentence (as a rule the logical subject) is taken
for granted, and finds no expression in language. It may be
completed from what has been said before. We ought specially
to notice that in the course of conversation it often needs to
be taken from words of the other. The answer commonly con-
sists of a predicate alone; the subject is either contained in the
question, or the whole question is the logical subject. (1) Who
struck you?'-' John.' (2) 'Was it you?'-'Yes' (No, certainly,
surely, of course). Similarly remarks like the following may serve
as predicate to a sentence spoken by another, Avowedly, all right,
very possibly, strange enough, no wonder, nonsense, stuff, balderdash,
etc. In other cases, it is the object of perception common to
speaker and hearer alike-the situation-that forms the logical
subject, to which the attention may be still more pointedly
directed by gestures.
This object of perception may be the
speaker or the person accosted; cf., your servant, most obedient
servant; all right, welcome; so sad! why so sad? Besides these we
have many exclamations of astonishment and alarm and appeals
for aid, like fire !-thieves !-murder 1-help! and challenges,
like Friend or foe? Questions, too, like Straight on or not?
-Right or left? When the Prince in Lessing's Emilia begins
with Klagen, nichts als klagen! Bittschriften, nichts als bittschrif
ten!—these are mere predicates: the subject is formed by the
letters which he takes up in his hand. In the case of such
sentences, which as far as their mere expression goes consist of
a single member, what for the speaker is the psychological pre-
dicate becomes for the hearer the subject instead. For the man
who, on seeing a house on fire, calls out 'Fire!' the situation is
the subject, and the common idea fire is the predicate; on the
other hand, for the man who hears 'Fire!' cried before he him-
THE FUNDAMENTAL FACTS OF SYNTAX.
119
1
self sees a fire, the idea of fire is the subject, and the situation
is the predicate. There may also occur sentences in which, for
both sides alike, what is uttered is the subject, and the situation
is the predicate. Supposing, for instance, that any one sees a
child in danger, he naturally cries out to the person intrusted
with its custody merely-' The child!' Hereby the object alone
is denoted to which attention is intended to be called-i.e. the
logical subject: the predicate, on the other hand, has to be
gathered by the person addressed from what he sees, if he follows
the direction to which his attention is called. Or, supposing
that one of two companions in travel remarks that the other has
forgotten his umbrella, the mere exclamation, Your umbrella!'
is quite sufficient to make the latter complete the predicate. The
vocative, pronounced as such, to summon or warn, or entreat any
one, or to call his attention to his turn of action, is such another
sentence-lacking a verbal predicate, but not a psychological
one. On the other hand, by the side of a verb in the second
person without subject pronoun, the vocative may be appre-
hended as the subject to this. We commonly punctuate' Charles,
come!' and 'Come, Charles!'-but, on the other hand, You come!'
and 'Come you!' and yet the relations are in the two cases the
same.
Verbs.
176. This is the place to determine the position of the so- Impersonal
called impersonal verbs. It is a much-disputed question whether
these are to be regarded as lacking a subject or not. A critical
discussion of the current views upon it is to be found in Miklosich's
treatise, Subjectlose sätze (second edition, Vienna, 1883); a treatise
by Marty in the Vierteljahrsschr. f. wissenschaftliche philos. viii.
56 sq., is chiefly based upon Miklosich's material.
177. In order to answer the question properly, a strict division
must be made between the grammatical form and the logical
1.
I20
PRINCIPLES OF THE HISTORY OF LANGUAGE.
1
relation denoted thereby. If we regard the first merely, it
cannot be doubted that sentences like es rauscht; il gèle; Low
Servian vono se błyska (it lightens') have a subject. But all
efforts have proved fruitless to treat this es, il, vono, as a logical
subject, and to give it a definite interpretation. Again, in sen-
tences like the Latin pluit, Greek veɩ, Sanscrit varšati (‘it rains'),
Lithuanian sninga (it snows'), we may assume that the formal
subject is not wanting; for such subject may be contained in the
verbal termination under which a personal he, she, or it may be
understood.
It might certainly be said for the opposite view
that in the languages in question the third person can stand
also by the side of an expressed subject [as Jupiter pluit, ỏ Zeûs vel,
etc.]. But it is impossible to prove that the impersonal did not
arise before this form of applying it. It seems most natural in
this case to recognise a formal subject. It is with the personal
ending just as with the independent pronoun. The sentence, as
it is brought into the normal form, has received a formal subject
which has nothing to do with the psychological. We must pre-
suppose an older stage, in which the simple verbal stem was set
down; a stage which is actually seen in Hungarian at the present
day, where the third person singular has no suffix (see Miklosich,
p. 15).
And we can form a lively idea of this stage of language
after the analogy of the sentences just discussed, which consist
of a single (not verbal) word. These are really and truly, as far
as the linguistic expression goes, subjectless.
178. We may then lay it down that the psychological subject
is as little expressed in the sentence it is hot as in the sentence fire.
But we must guard against being misled into the view that none
exists. In this case too we have a junction of two ideas. On one
side stands the perception of a concrete phenomenon, on the other
the idea of burning or of fire; an idea already reposing in the
THE FUNDAMENTAL FACTS OF SYNTAX.
121
mind, under which the perception in question may be ranged.
The word fire can be a sentence only as an imperfect expression
for the connexion of these two elements. It is conceivable that in
the case of the verb in the corresponding application, the infinitive
might have become usual instead of the impersonal. And indeed
the infinitive is actually employed where an order or demand is
in question. For instance, as a word of command, aufsitzen stands
on the same plane as marsch! and psychologically speaking it
may be regarded as the imperative to the impersonal, es wird
aufgesessen [cf. the use of the infinitive in ancient Greek, and in
modern French in such instances as voir, pp. 10 and 50].
179. Miklosich and Marty dispute the existence of a psycho-
logical subject for impersonal sentences. They hold these as
true sentences composed of a single member in accordance
with Brentano's psychology, and they see in these a proof
of the theory that the logical judgment does not need to con-
sist of two members. It seems that Marty was brought to this
view partly by the consideration that for the expression of per-
ception in a concrete sentence linguistically consisting of two
members, something more needs to be demanded than the connec-
tion of the two members. If we say, for example, This pear is hard,
we must have previously brought the object of which we wish to
say anything under the common category of pear, and the quality
which we have remarked as attaching to it under the common
category of hard. We must thus, in order to pronounce our judg-
ment, have formed two additional auxiliary judgments. Thus one
might imagine from this point of view that the impersonal sentence
really contains no more than the predicate of a normal sentence;
and as the latter is denoted as consisting of two members, it then
seems but commonly consistent to denote the former as consisting
of one member only. According to this view, however, the fact
122
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Negative
Sentences.
Sentences of
Assertion
mortal.
is overlooked that what in one case was nothing but an auxiliary
judgment, in the other has become an end in itself. We might
with equal right neglect the difference which exists between the
member of a sentence the mortal man, and the sentence man is
But under all circumstances a sentence like Fire, it is
hot, is of two members. It is hard to form an intelligible idea
of judgments consisting of but a single member, and logicians
should not cite language as a proof of the existence of such :
otherwise, they show that their very thoughts are dependent on
linguistic expression-the very thing from which they ought to
seek to emancipate themselves.
180. According to our investigations hitherto pursued, it is
clear that impersonal sentences, consisting according to their
linguistic expression of a single member, are always concrete,
never abstract. For their very object and end is to produce a
concrete intuition with a common conception.
181. If we have defined the sentence as the expression for the
connexion of two ideas, negative sentences seem to contradict this,
since they rather denote a separation. Still, such a separation
finds no expression unless the ideas in question have met in the
consciousness of the speaker. We may define the negative sentence
of affirmation as the expression of the fact that the attempt to
establish a relation between two ideas has failed. The negative
sentence is in any case of later birth than the positive. Probably
negation finds in every case a special expression in language. One
might, however, very well imagine that negative sentences might be
formed in a primitive stage of development of language, in which
the negative sense might be indicated by nothing else than the
stress and the accompanying gestures.
182. What can only be laid down as a possibility in the case of
and Demand. the difference between positive and negative sentences is certainly
THE FUNDAMENTAL FACTS OF SYNTAX.
123
true of the difference between sentences of assertion and sentences
of demand. The term 'sentences of demand' seems appropriate
as the most neutral. Under the term demand we comprise of course
requests, commands and forbiddings, advice and warning, encour-
agement, concession, declining, and deprecating. No examples
are necessary to show that for all this the same expression of
language can be employed, and that the different shades of mean-
ing attaching to the word can be only recognised by the different
tones indicative of feeling. We must however add to these,
sentences implying a wish. A wish may be expressed with the
expectation that its utterance may have an influence upon its
realisation; then it is equivalent to a demand; but it can equally
well be uttered without any such expectation at all.
This is a
difference which is not always remarked by the naïve consciousness
of a child or uncivilised man. The language of poetry, and indeed
that of untutored conversational language, finds it natural to raise
mere wishes to demands, and to express these by the imperative.
Willing and demanding touch one another even more in con-
junctive or optative forms of expression.
183. We are at present accustomed to regard the affirmation as
the only normal sentence. The sentence of demand is however as
old, if not older than this. The earliest sentences which are uttered
by children--the very earliest of course consist of a single word only
-have reference to their needs. They serve to express demands
or affirmations made to indicate a need which requires to be satis-
fied. It may be assumed that the circumstances were similar
at the earliest stage of speech development. Thus no special
linguistic resource was needed originally to characterise the sen-
tence of demand; the simple correlation of subject and predicate
sufficed for this case as well as for the sentence of affirmation; the
stress of the voice was the only means by which the difference
124
PRINCIPLES OF THE HISTORY OF LANGUAGE.
could be told. Even now we employ such sentences of demand in
numbers; sentences in which the demand is not characterised as
such. These are sentences without a verb, such as: Eyes right !—
Attention!-Hats off!-This way!—All aboard!-Foking apart―
An eye for an eye-Peace to his ashes—A health to all good lasses-
Away with him-Here with it! etc. Then there are sentences
composed of a single member, as far as their expression in
language goes, such as: Hush!-Quick!-Slow!-Forwards!—
Up!-To work!-Confusion! etc. Sentences of demand especi-
ally occur in this primitive form, which as a rule does not hold
good for sentences. It is from this negative circumstance that the
consequence springs, that these negative sentences may be imme-
diately recognised by us as demands. Still there are instances
which are capable of bearing either meaning, as 'Fire!' as a cry
of alarm and 'Fire!' as a word of command.
184. Besides this, instead of a definite characteristic form of the
verb, a form essentially indefinite may be employed for demands.
Thus, the past participle, as in German rosen auf den weg gestreut,
alles harms vergessen (HÖLTY), In die welt, in die freiheit gezogen
(SCHILLER). The infinitive is more common still; cf. the German
use of absitzen, schritt fahren [and that of the French voir, where
we say see (see p. 121)]. In Italian the infinitive is usual in
negative prohibitions-as non ti cruciare; and the same usage is
common in Roumanian, Provençal, and Old French. [Cf. Diez,
iii. 212; Gospel of Lyons: Es biens dou monde ne te croire; i.e.
Ne crois pas en les biens du monde.] Jolly (Geschichte des infin., pp.
158-209) would explain these infinitives by the original dative
function of the infinitive, and this explanation must be held to
account for the imperative use of the infinitive in Greek. But the
usage in German and Roumanian is of modern origin, and must not
be connected with conditions which took their origin in IE. times,
THE FUNDAMENTAL FACTS OF SYNTAX.
125
the consciousness of which had long passed away from the instinct
of language. For the epoch in which this usage has grown up the
infinitive is nothing more than the denomination of the verbal
conception per se; and hence these infinitive sentences are to be
placed upon the same plane as sentences like 'March!' It is
noteworthy that the second person singular of the IE. imperative
exhibits the pure tense stem (Aéye).
Sentences.
185. The interrogative sentences are commonly ranged as a Interrogative
third class by the side of the sentences of affirmation and demand.¹
No single principle can, however, be found for such a triple
division of sentences, and these three classes cannot be co-ordi-
nated with each other. We must rather assume a double method
of division into two. The sentences of affirmation, as well as
those of demand, have their pendant in interrogative sentences;
cf. the Latin quid faciam with quid facio? We employ the
term 'deliberative questions' to express this use, but we might
designate them simply as interrogatory-demand sentences.
186. Of the two main divisions of the questioning sentence,
that in which a single member of a sentence is called into question
is certainly of later origin than that in which the whole sentence is
questioned. The former requires a special interrogative pronoun
or adverb which is not required by the second. The interrogative
is in the IE. languages indefinite as well. There seems to be no
criterion by which we may decide which is the original of these
two functions. To suppose the latter to have taken its origin
from the former causes no difficulty. But the reverse process
1 Cf. for what follows, Imme, Die fragesätze nach psychologischen gesichtspunkten
eingteilt und erläutert. Cleve, 1879, p. 81.
2 No one has yet succeeded in discovering a perfectly satisfactory terminology for
these two kinds. Delbrück, SFI., p. 75, calls the former 'verdeutlichungsfragen;' the
second bestätigungsfragen.' Imme, u. s. i. 15, adopts the second term, and replaces
the former by 'bestimmungsfragen.' It appears to me, however, that the expression
'bestätigungsfragen ' is unsuitable, as it includes the expectation of an affirmative answer.
126
PRINCIPLES OF THE HISTORY OF LANGUAGE.
is conceivable, and thus we should have a path from the older
form of interrogative sentence to the later form. To the question
'Ist jemand da?'-we may answer (Ja) der vater, or '(Nein)
niemand. Now if we discard our usual method of framing an
interrogatory, Jemand ist da?'-the point where this touches
'Wer ist da?' is obvious. Questions with an interrogative pro-
noun stand even nearer to questions with an indefinite pronoun
in cases where a negative answer is looked for as the natural one;
cf.: Who will do that?-Will any one do that?-What can I
answer?—Can I answer anything?— Where is such a man to be
found? Is such a man anywhere to be found?
187. The question to which the simple answer yes or no is
expected is in many languages characterised by a special particle,
in the Teutonic and Romance languages by the position of the
words. Thus-ne in Latin [li in Russian] serves to mark an
interrogation. But the interrogative position of words is not
confined ab initio to the interrogative sentence. We find
it, for instance, in OHG., OS., and AS., commonly in the sentence
of affirmation; cf.: verit denne stuatago in lant, holoda inan
truhtîn, etc.
188. Accordingly the interrogation was not to be recognised
by the position alone, and the interrogative stress was the only
decisive characteristic whereby it differed from affirmation. We
have at the present day interrogations in which this stress or tone
is the only criterion, namely, those which contain no verb; cf.:
Any one there?—All ready?—A glass of beer, Sir? (spoken by
a waiter); in French, Votre désir? [Monsieur désire ?]
189. We can thus easily imagine that there may have been
long sentences before any other determining method was found for
them besides the interrogative tone. It thus seems that the
interrogative form of the sentence is possible even in a quite
THE FUNDAMENTAL FACTS OF SYNTAX.
127
primitive stage of language, though it must be regarded as more
recent than the assertive and demanding forms.
190. The interrogation pure and simple occupies in some ways
a half-way position between positive and negative assertion. It is
neutral in the sense that it can make no essential difference
whether it be cast in a positive or negative form, except that the
positive form is naturally taken in preference as the simpler, and
the negative form assumes the function of expressing a modi-
fication of the question pure and simple.
191. There are, in fact, various modifications, whereby the
interrogation can be more or less made to approach the character
of the assertive sentence. It may thus, for instance, become a
dubious assertion, in which one is disposed, at the outset, to a
definite assumption, and expects nothing but a final confirmation
from another. In this case the negative interrogatory form
comes in where a positive answer is expected-Were you not
there? I thought I saw you.
It makes no essential difference as
far as the sense goes, if we employ instead of this the positive
interrogatory form, with a tone of query-You were there?—
You are contented? Thus we may arrive at this intermediate
stage from both sides.
192. The circumstances attending the expression of admira-
tion or surprise are similar to these. Surprise is the subjective
incapacity to take in one collection of ideas by means of another,
in spite of an external instigation to do so, whether proceeding
from one's own observation or from the suggestion of another.
For this again we may employ either the interrogative form, or
the assertive form, with the interrogatory tone-Is Francis dead?
-Francis is really dead?-Are you here again? You are here
again? Sentences without a verb are neutral in this respect :
-You my long lost brother?-What, THAT to ME?-Already
128
PRINCIPLES OF THE HISTORY OF LANGUAGE.
here?—So soon? And infinitival clauses are similarly used, as,
He to turn out such a rascal! Expressions of surprise also occur
in which the psychological subject and predicate are connected by
'and, as in German-So jung und schon so verderbt?-A maid,
and be so martial? (SHAKESPEARE.) The expression of surprise
is weakened into a mere formula wherewith to introduce a con-
versation; cf. Ausgeschlafen?-So vergnügt?-Noch immer bei
der arbeit? etc. [As we might ask in English, A good nap?—
Always in good spirits ?—As merry as ever? etc.]
193. A special case is to be noted in the surprised or indignant
repudiation of an assertion. For this the primitive form of
expression without any finite verb is specially common: I a liar?
Er und bezahlen? Latin-Ego lanista? (CIC.) French-Moi
vous abandonner?
MY pardon?-How? not know the friend that served you?
Besides this the indignant repudiation of advice tendered occurs,
as in such cases as, Ich dich ehren ? (GOETHE.) A sentence such
as this must probably be reckoned amongst the interrogatory
demands.
Italian-Io dir bugie? English-SHE ask
194. Of course the provocation to an interrogation is originally
a need felt by the interrogator. There are, however, questions-of
course of later origin-in which the proposer of the questions has
no doubt about the answer which suits, and merely wishes to give
the person addressed occasion to find out the answer for himself.
Such cases are questions put by teachers to the taught. If a
hint is added as to the nature of the answer expected by the
interrogator, we have the class of questions commonly described
by the vague term rhetorical. The interrogator by the process
compels the person addressed to admit a truth on the strength
of his own conviction gained after consideration; and thereby
the process is brought home to him more emphatically and
THE FUNDAMENTAL FACTS OF SYNTAX.
129
energetically than it would have been if merely communicated
to him.
of the Sen-
195. The relation of subject and predicate in the wide sense Extension
indicated above is the relation from which the other syntactic con- tence.
ditions take their rise, with one sole exception, namely, the copula-
tive connexion of several elements into a single member of a
sentence. This connexion may in the case of developed languages
be denoted by a particle, but the mere collocation in succession
suffices, and thus we cannot be surprised that in a primitive stage
of language it was found possible to dispense with any special
linguistic expression for coupling.
Subject.
196. Every other mode of extending sentences is effected by a Double
recurrence of the relation between subject and predicate. Two
main cases may be distinguished. It may either happen that two
members unite themselves as contemporaneous with a third-¿.e.
that two subjects attach themselves to a single predicate or two
predicates to a single subject, which might be represented by the
formula (a + (b) + c). Or a connexion of subject and predicate.
may present itself as subject or predicate in relation to a further
member, a case which might be represented by the formula
(a+b)+c. This further member may of course be compounded
again. Suppose that in the first case the logical relation of
the two subjects to the common predicate, or that of the two
predicates to the common subject, is exactly alike, then such a
sentence, consisting of three members, may be replaced without
any essential change in the sense, by one of two members, one
of which will be a copulative combination. From this there arise
points of contact between these two kinds of sentence and con-
fusions between the two. The double member of a sentence
seems most neatly parted from the copulative combination forming
a single member, when the pair of members take between them
I
130
PRINCIPLES OF THE HISTORY OF LANGUAGE.
R
a member common to both without the employment of a copu-
lative particle; this is the so-called construction ȧrò кowoû, a
construction common enough in MHG.; cf.: Dô spranc von dem
gesidele her Hagene also sprach. Should we however say, Da
spranc vom sitze Hagen und sprach so, we have at once a transition
stage from a double predicate to a single compound one. Not,
however, that the two predicates are really conceived as one, as is
manifest from the fact that the predicate regularly stands in the
singular after such double subject, as in German: Der mann ist
tot und die frau ('The man is dead and so is the woman'). In
the older stage of the language they were apprehended as one if a
further predicate still was subjoined; cf.: Petrus aber antwortete
und die Apostel und sprachen (LU.), a case in which we now have
to supply a new subject. The instinct of language is much less.
decided when the two members are not thus separated by any
insertion. In that case it is just as possible to assume several
members, successively connected with the different members of
the sentence, as it is to assume a compound sentence united by a
single act. The first-named way of apprehension is less obvious
when the couple of members of the sentence is placed at the
beginning than when they are placed at the end. The vacillation
in the instinct of language manifests itself in the fact that where
several subjects are present, of which the one standing last is a
singular, the predicate can stand as well in the plural as in the
singular. Where the predicate comes at the end we feel ourselves
nowadays compelled to employ the plural, but in Latin the
singular is equally permissible—cf.: Speusippus et Xenocrates et
Polemo et Crantor nihil ab Aristotele dissentit (CIC.); Consules,
praetores, tribuni plebis, senatus, Italia cuncta semper a vobis depre-
cata est (CIC.); filia atque unus e filiis captus est (CAES.); even Et
ego et Cicero meus flagitabit (CIC. ad Att.). So in Italian, Le
THE FUNDAMENTAL FACTS OF SYNTAX.
131
ricchezze, gli honori e la virtù è stimata grande; in French, Le
fer, le bandeau, la flamme est toute prête (RACINE); and thus in
old Modern High-German, Wolken und dunkel ist um ihn her
(LU.); dass ihre steine und kalk zugerichtet würde (LU.).
197. The logical relation, however, of two subjects to the same
predicate may also be of different natures. In this case we have
the foundation laid for the possibility of the differentiation, in
the course of the development of language, of the double subjects
into subject and object. We can best portray this process by one
sentence, like I smell the dinner. We can say equally well, omit-
ting the personal subject, The dinner smells; and we are thus able
to transport ourselves in fancy back to a time in which, owing to
the complete absence of all case-suffixes, and of anything to fix the
position of the words in a sentence like I smell dinner-A dinner
smell I-the words I and dinner fell under the same common
category of the psychological subject. The affinity between sub-
ject and object is also evident from the fact that the object can be
made the subject by turning the verb into the passive.
Object.
198. The object-to take the word in the widest possible Double
sense may again include very different logical relations. Here
again many objects may be attached to the same predicate, either
in the same or in a different logical relation. By this means the
impulse is given to a grammatical differentiation of the object
corresponding to the logical relation (Accusative, Dative, Genitive
of the object, etc.).
199. The object may be conceived of by the side of the subject as
a third member of the sentence of equivalent value with it. It may,
however, equally well enter into a nearer relation to the predicate
than that occupied by the subject, so that a sentence of two mem-
bers takes the place of one of three members, the object forming
one member in unison with the predicate, in such a way that it
132
PRINCIPLES OF THE HISTORY OF LANGUAGE.
There is no
stands in subordination to it and serves to define it.
sharply defined border-line between these two relationships.
200. Just as the predicate may receive a defining word subor-
dinated to itself, so may the subject and the object which is
developed therefrom. Substantival and adjectival attributes com-
monly stand us in the place of such defining words, and genitives
of substantives, though substantives and adverbs connected by
means of prepositions sometimes do the same. The aid of these
different methods of definition renders it possible to express to
some extent even in language the difference of the logical relation
between the defining factor and the defined. A language which
has yet developed no inflexion and no connecting words is not yet
in a position to do this. Such a language has no other resource
but the simple juxtaposition of the defined matter and the
defining word. The fact that the determinant attached to the sub-
ject is not a predicate, can then, in case no fixed order of the
words has become normal, be only discovered by the presence of a
third word, which is detached from the two words which together
make up the subject by a more decided stress, and it may be by a
slight pause. The relation of the determining element to that deter-
mined is analogous to that of the predicate to the subject in the
broad sense in which we have apprehended it above. And in truth
the determinant is nothing but a degraded predicate, not uttered or
spoken for its own sake, but merely in order that a further predi-
cate may be assigned to the subject (or object). Just as the
determination of the predicate has its origin in sentences containing
a double subject, so the determination of the subject—and accord-
ingly that of nouns in general-has its origin in sentences con-
taining a double predicate.
201. We can most readily appreciate the degradation of the
predicate to a mere determination in cases where a finite verb is
THE FUNDAMENTAL FACTS OF SYNTAX.
133
affected by them. We meet here with a process which has grown
up spontaneously in different languages and at different epochs,
and which is, indeed, to some extent to be traced historically.
The starting-point is afforded by the construction ȧrò KOLOÛ
spoken of, p. 130.
202. In this process one of the two predicates may be logically
subordinate to the other, so that it may be replaced by a relative
sentence.¹ This usage is found in OHG. and in MHG.; cf.: mit
zühten si ze hûse bat ein frouwe saz darinne (='a lady who had
her dwelling therein'), wer was ein man lac vorme Grâl? (=who
lay before the Graal), die worhte ein smit hiez Volcân (= named
Vulcan); nist man, thoh er uuolle, thaz gumisgi al irzelle (=there
is no man, who, even if he wished it, could completely count the
number of human beings). A case depending on the main verb
may also serve as the subject of the dependent verb: thus von
einem slangen was gebunden (the title of a fable of Bonar): ich hab
ein sünt ist wider euch (H. Sachs); dar inne sach er glitzen von
kolen rot ein glut wart auf sein fallen (which waited for his fall, ib.).
This construction becomes more frequent towards the close of the
middle ages than before. The corresponding usage in English,
Swedish, and Danish, has gained a far larger area; cf. the follow-
ing examples from Shakespeare:-'There is a devil haunts thee,'
¹ On this phenomenon a considerable literature exists: (cf. especially J. Grimm,
Ueber einige fälle der attraction (Kl. schr. 3. 312 sqq.); Steinthal, Assimilation und
Attraction (Zschr. f. völkerps. i. 93 sqq.=Kl. schr. 107 sqq.; cf. especially p. 173 sqq.;
Tobler, Ueber auslassung und vertretung des pronomen relativum (Germ. xvii. 257 sqq.);
Jolly, Ueber die einfachste form der hypotaxis im idg. (Curtius, Studien vi. 217); Kölbing,
Untersuchungen über den ausfall des relativpronomens in den germanischen sprachen,
Strassburg, 1872; Erdmann, Syntax Otfrids, ii. p. 124 sq.; Behaghel, Asyndetische
parataxe (Germ. xxiv. 167 sqq.); Lohmann, Ueber die auslassung des englischen relativ-
pronomens (Anglia, iii. 115 sqq.). In some of these writings we find conclusions widely
differing from those given above. I have deemed it unnecessary to state my reasons for
disagreeing with these, as it seems to me that the correctness of the views which I have
adopted, those of Jolly and Behaghel, must be sufficiently evident to all who are not
slaves to their own instinct of language and traditional grammar.
134
PRINCIPLES OF THE HISTORY OF LANGUAGE.
}
'it is thy sovereign speaks to thee,' 'here are some will thank you,'
'I have a mind presages me,' 'it is not you I call for.'
203. In the examples hitherto cited, the common clause stood
in the middle. There occur in OHG. cases as well, in which it
takes precedence, or is inserted between the first predicate and its
determinations. In this use it may serve as subject, or as object,
or as any other kind of adverbial determination; and further, it
need not necessarily bear the same relation to both predicates.
Under this head will come cases like these passages quoted from
Otfrid, where the second predicate is subordinated-e.g.: thaz selba
sie imo sagetun sie hiar bifora zelitun (They said the same thing
to him that they had said before'); uuer ist thes hiar thenke ('Who
is there that could here think such a thing'); nist man nihein in
uuorolti thaz saman al irsageti ('There is no one in the world who
could say all that together'). The first predicate is subordinated
in the following case: in selben uuorton er then man thô then êriston
giuuan so uuard er hiar fon thesemo firdamnot ('by the same
words with which he overcame the first man, was he condemned
by the other'). In a sentence like this, so resumes the words in
selben worton, as it may resume any given element of a sentence.
In another case the common element of the sentence is resumed by
a pronoun: allo uuîhi in uuorolti thir gotes boto sageti, sie quement
số gimeinit ubar thân houbit.
1
204. The construction ȧrò κowoû is most frequently found in
OHG., in sentences containing general statements, especially where
these are negative, with a conjunctive dependent verb. This con-
struction is not unknown to the Romance languages; cf.: Old
Italian, non vi rimasse un sol non lacrimassi; prov. una non sai vas
vos non si aclina, anc non vi dona tan mi plagues; Old French, or
n'a baron ne li envoit son fil.
1 Cf. Diez, iii. 381.
THE FUNDAMENTAL FACTS OF SYNTAX.
135
205. If we examine dispassionately the testimony of history,
we shall find the view untenable that this construction, wherever
found, is a traditional inheritance from the IE. original language.
On the contrary, it seems probable that it has likewise sprung
into existence spontaneously in later epochs, although other
Out-
more perfect forms of expression had already been created.
side the IE. family of languages it is found (for instance) in
Arabic, where we meet with expressions like, 'I passed by a man
slept; cf. Steinthal, Haupttyp. 267.
206. If, then, the finite verb could be degraded to the value of
an attributive determinant, how much more easily could a pre-
dicate be so degraded, which hitherto bore no distinguishing
marks of its verbal character? The origin of the attributival
relation is accordingly quite clear.
207. With regard to the function of the determination, we
have to notice certain differences which commonly find no expres-
sion in language, but which are, nevertheless, logically of very
great importance. The determinant may leave unaltered the
extent of the meaning which the word serving as the subject has
independently, or in virtue of a limitation, already imposed by
other means, the determinant itself applying to this whole extent:
cf.: der sterbliche mensch, der allmächtige gott, das starre eis; it may,
however, on the other hand, if it applies to a part only of what
is contained in the usual or already otherwise specialised meaning
of the word, narrow and individualise it; cf.: old houses, an old
house, a (or the) son of the king, the journey to Paris, Charles the
Great in the same way, 'the old house,' when the expression is
used to contrast it with a new one: whereas the combination falls
under another head if it is obvious, without any adjective, what
house is meant. In those cases which come under the second
category, the determination is indispensable, because without it
136
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the predicate has no value. In the first category the following
further distinctions are of importance. In the first place, the
determinant may already be known as applicable to the concep-
tion to which it is appended (as is the case with the repetition of
the standing epithets in the language of Epic poetry): or it may
communicate something new. In the latter case the determinant
has a greater independence, and approaches the value of a true
predicate [as Tullia ferox, Tullia mitis (LIVY)]. We prefer, in
such cases, commonly to use periphrases by means of a relative
sentence; as 'Charles, who was poor;' 'Louis, who was a clever
painter.' In the second place, while the determination has not
necessarily any relation to the predicate, it may yet stand in a
casual relation to it, as in 'The cruel man would not listen to the
victim's prayers.'
208. We have conceived the determinant as a weakened
predicate. There is, further, an intermediate stage at which the
determination has a greater independence still, and is not so
closely connected with the subject. For this reason it seems
best to regard it as a special clause in the sentence. Under
this head comes what we commonly call the predicatival attribute;
for instance, 'he arrived safe and sound.' But prepositional
determinants may equally well stand in the same logical relation,
e.g. 'he begged me on his knees,' for which 'kneeling' would be
a perfect substitute. The relation of the predicatival attribute
to its substantive is not so close, because it denotes, not a pecu-
liarity necessary and enduring, but a merely accidental and trans-
itory situation. It may, therefore, be regarded as an independent
link connected with subject and predicate alike. This independ-
ence shows itself in most languages by the greater freedom of
the position of the words as contrasted with the fixed position
of the pure attribute. In NHG. the nearer affinity to the predicate
THE FUNDAMENTAL FACTS OF SYNTAX.
137
has further found expression in the use of the uninflected form
of the adjective, as in the case of the predicate. When once
the adverbial and adnominal determinations have been developed,
as special varieties, from original subjects or predicates, a further
complication of the sentence becomes possible, since a combina-
tion consisting of a determined and a determining element may
again be determined by a new element, or itself serve as a deter-
minant; and since, further, several determining elements may be
joined to one determinate, or several determinate elements to
one determinant, just as several subjects may be joined to one
predicate, or several predicates to one subject. As examples we
may take (1) alle guten geister, Müllers älteste tochter, er gerät leicht
in zorn (to be understood gerät in zorn+leicht); (2) sehr gute
kinder, alles opfernde liebe, er spricht sehr gut; (3) trübes, regner-
isches (trübes und regnerisches) wetter, er tanzt leicht und zierlich ;
(4) Karls hut und stock, er schlägt weib und kind.
209. The first method of combination is that commonly
described as the relation of inclusion. It is not always possible
to draw a sharp line between it and the third. For instance, if
I say large round hats, it makes no essential difference whether
we classify this combination under (1) or (3). In NHG., where
two adjectives are used together, the use of the strong or of the
weak form affords a means of distinguishing the relation of juxta-
position from that of inclusion, a means which leaves us, however,
in perplexity when both forms phonetically coincide. But the
difficulty of correctly maintaining the difference manifests itself
in many instances where the writers have offended against the
rules of grammar; compare the instances given in Andresen,
Sprachgeschichte, p. 38 sqq.
210. Constructions (3) and (4) may be interpreted in two
fundamentally distinct ways. They may either, as has been stated
138
PRINCIPLES OF THE HISTORY OF LANGUAGE.
above, be conceived as constructions aπò кowoû, or as the com-
bination of one element with two elements united by a copula.
Hence it is that we find in (4), in those languages which have
developed grammatical concord, the same vacillation in the form
of the attribute as we have found above (p. 130) in the form
of the predicate. Cf., on the one hand, the French le bonheur et
le courage constants, la langue et la littérature francaises; Latin,
Gai et Appii Claudiorum; on the other hand, la fille et la mère
offensée (RACINE); the Latin Tiberius et Gajus Gracchus, et
tribunis et plebe incitata in patres (LIVIUS). Not, however, that
all cases of the same grammatical form are in this way open to
two meanings. In the cases cited each of the two substantives.
denotes an independent substance. It may, however, also happen
that the combination denotes merely two aspects of the same
object, e.g. my uncle and foster-father. In this case, where the
combination appears independently as either subject or object,
we must classify only as my uncle and stepfather. When
each word denotes a particular object, the German language in
modern times prefers [as also the English], at least in the case
of singular nouns, to assign to each its special attribute; so that
my uncle and my foster-father bears a different meaning from my
uncle and foster-father. We can refer the former combination to
a person only when it is expressly placed in relation to one
either as predicate or as attribute, or finally as an address. On
the other hand we find, though the grammarians condemn the
usage, frequently enough the simple juxtaposition of the attri-
bute with several substantives, each of which denotes a special
object; cf. the numerous examples given in Andresen, Sprach-
geschichte, p. 125 sqq. Similarly Lessing wrote über die grenzen
der malerei und poesie.
THE FUNDAMENTAL FACTS OF SYNTAX.
139
Sentence as
predicate.
211. The amplifications of the sentence hitherto described pro- Use of a
ceeded from the formula (a+(b)+c),—cf. p. 129-in connexion subject or
with the copulative combination. We turn to the amplifications
according to the formula (a+b)+c. We find these, for instance,
represented by the combination of a verb with the accusative and
infinitive, or with two accusatives, of which the one is a predicate :
memini―me audire, reddo-te beatum. In order, however, to under-
stand the origin of these constructions, we shall be compelled to take
another point of departure. Our best course is to confine ourselves
in the first place to such cases as those in which a composite
member of a sentence (a+b) still plainly exhibits the form of the
independent sentence, and thus contains a finite verb. We here
pass again beyond the borders of the so-called simple sentence, and
encroach on the area of the compound sentence. Actual historical
and psychological observation shows us that this division cannot
be strictly maintained. It depends upon the assumption that the
presence of a finite verb is the strict characteristic of the sentence;
a view which is absolutely inapplicable to many languages and
epochs, and is absolutely applicable to none. In cases where the
definite stamp of a finite verb is wanting, the division too between
the simple and compound sentence in the ordinary sense ceases to
hold good. The so-called composite and the so-called amplified
sentence are accordingly in their essence absolutely identical.
thus appears further to be a mistaken view that the degradation
of a sentence to the position of member of a sentence-the so-
called hypotaxis-did not develop until a comparatively late stage
of language. The existence of the amplified sentence, which is
found even in the most primitive languages, presupposes this
degradation as completely carried out. Further, the ordinary view
that the hypotaxis has consistently arisen from parataxis is mis-
taken. It might be asserted with equal reason that the division of
140
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Union of
Dependence
and Inde-
pendence.
a sentence into subject and predicate has arisen from the copulative
connexion of two words. This view was enabled to arise because
the oldest form of hypotaxis no doubt lacks any special grammatical
denomination, and is merely logical and psychological. But to call
a logical sub-ordination such as this a co-ordination (parataxis)
is quite incorrect.
212. It happens very often in Modern German, and in other
languages in which the structure of the sentence is already richly
developed, that combinations, in form undistinguishable from the
main sentence, are used as objects. Under this head we may
reckon the 'Oratio directa.' Under the same head too come sen-
tences like, I maintain he is a liar; I think you are mad; I see you
tremble; consider, it is dangerous. Demands, too, and questions.
are thrown into the same relation of dependence: Pray, give it me;
cf. Latin quaeso, cogita ac delibera; sage, hast du ihn gesehen ;
sprich, was bekümmert dich; also Latin, videte quantae res his
testimoniis sunt confectae (CIC.); quaero de te, qui possunt esse
beati (CIC.); responde, quis me vendit (PLAUtus). It is more
rare to meet with subjects of this kind, excepting where the
passive is used, as besser ist, du lässt es bleiben; das macht,
sie ist sehr mannigfaltig (LESSING).
213. It is no doubt true that in all these cases the subject or
object sentences possess a certain independence; and, excepting
in the case of the 'Oratio directa,' they cannot be employed without
an independent value being assigned to them. For instance, we
cannot say, I thought you are ill; nor in German either, ich glaubte,
du bist krank, or ich glaubte, du warst krank. It does not, however,
follow from this limited independence that the relation to the main
verb was in its origin paratactical; but, with reference to the
main verb a decided case of hypotaxis arises, and independence.
only in as far as the presence of this is disregarded. The inde-
THE FUNDAMENTAL FACTS OF SYNTAX.
141
pendence is more marked if the governing sentence is placed last
or intercalated, as in this case the dependence is not noticed until
the conclusion-e.g.: He is a liar, I believe; or, he is, I think, a
liar; so in Latin, quid illi locuti inter se? dic mihi (PLAUTUS);
signi, dic, quid est? (PLAUTUS). In the case of intercalation our
grammarians are disposed to treat the intercalated sentence as
subordinate, and they might appeal to the fact that such an
expression as I think is nearly equivalent to as I think, or in my
opinion, or in my belief. In the older NHG. it is quite common to
set down a sentence in the first instance as independent, and then
simultaneously to make it the subject or the object of a succeeding
sentence: cf. the following examples from Hans Sachs-ein evolk
dreissig jar fritlich lebet, verdross den teufel gar; der frauen wart
sein hab vnd gut, geschah nach Christi geburt zware vierhundert vnd
auch funfzig jare; des wirt ein böse letz der lon, deut der schwanz
von dem scorpion; das betrübt weib sich selbst erstach vnd nam ein
kleglich end, beschreibt Boccatius; darum jm jederman vol sprach,
tut Plutarchus beweisen. In these cases it would be quite unjusti-
fiable to assume the ellipsis of a das.
:
214. The use of the persons in such sentences as follow must
also be explained from this union of independence and depen-
dence, e.g. er denkt, er hat was rechtes getan, instead of ich habe;
that is, from the point of view of the speaker, and not after that of
the person to whom the thought is ascribed: just as believe me, you
are in error; he thinks, he can deceive you.
215. It occurs also that some writers and speakers prefer the
emphasised form of the parataxis in spite of the logical dependence.
A common instance of this is the German sei so gut und tue das
('Be so good and keep quiet'). Cf. Hans Sachs' expression, ir seidt
gewonet alle zwen vnd tragt mit euch was nit wil gehn.
examples may be found in ANDR. Sprachg., p. 140.
Other
142
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Indirect
Speech.
Sentence in
apposition to
a Noun,
216. The 'Oratio obliqua' in German must at the present day
be regarded as grammatically dependent, and the sign of its
dependence is the conjunctive mood. If we, however, regard the
origin of the construction, it is clear that, in this case also, a some-
thing half-way between logical dependence and logical indepen-
dence lies at the root. Such a construction as er meint, er könne
dich betrügen originally stood on the same footing as the phrase
cited above, er meint, er kann dich betrügen; only that the asser-
tion is made with less confidence, and that therefore the conjunc-
tive (or optative) is employed in a potential meaning. The fact
that the potential has ceased to be employed in main sentences
has promoted the apprehension of the relation as one of real
grammatical dependence.
217. Now, a combination of the form (a+b)+c may, just like
the simpler formula a+b, be degraded from the position of a sen-
tence to that of a member of a sentence. In this manner a sen-
tence may become determinant to a noun, or be put in apposition
to it. Compare such sentences as er sprach die worte: das tue ich
niemals; eins weiss ich: es geschieht nicht wieder; folgendes ist
mir begegnet: ich traf einen mann; ein sonderbarer zufall hat sich
gestern zugetragen: es begegneten sich zwei freunde, etc.; er hat die
gewohnheit: er erwidert nie einen brief; ich habe die überzeugung :
du wirst dich noch bekehren. A pronoun is especially often used
in this way, to which the sentence stands in the place of an appo-
sition; cf.: das ist sicher, er wird es nicht wagen; es ist besser, du
gehst; Latin, hoc relicuomst: si infitias ibit, testis mecum est anulus
(TER.); hoc capio commodi: neque agri, neque urbis odium me
unquam percipit (TER.). In the same way sentences stand in
apposition to a demonstrative adverb: er ist so lieb, man kann
ihm nicht böse sein.
218. If it is only a pronoun which is defined by the sentence,
THE FUNDAMENTAL FACTS OF SYNTAX.
143
we can omit it in thought without any essential change in the
sense. In this case we have again the form referred to above, in
which the sentence is directly made the subject or the object. Cf.:
es ist gewiss, du bleibst with gewiss ist, du bleibst. Thus the two
forms of expression come into very close contact.
Sentence.
219. Conversely, a noun may become appositive to a sentence; a Noun, to a
cf.: You always squint, a bad habit. This construction is especially
common when a relative sentence is in addition connected with
the noun; as, He means to start, a resolve which has cost him
dear. In this case we again clearly recognise the apposition as
a degradation of the predicate. And it is precisely this degrada-
tion that has prevented the sentence preceding from being itself
degraded to a mere subject.
220. We have thus traced the development of the sentence Parataxis.
from its simplest to its most complicated form. We now turn
to the paratactic conjunction of several sentences. This stands
parallel to the copulative conjunction of co-ordinate members of
a sentence; for which reason even the most highly developed
languages employ the same resources in order to designate both
kinds of connexion. Originally in this case, too, mere juxta-
position had to suffice. And if, as we have seen in the case of
hypotaxis, the one member may have a certain independence, so,
on the other side, we find that a parataxis with full indepen-
dence of the sentences connected together nowhere occurs; that,
in fact, it is impossible to connect sentences together without a
certain kind of hypotaxis. We can only call a sentence inde-
pendent a principal sentence, in the strictest sense, if it is uttered
merely for itself, and not to give a determination to another
sentence. In correspondence with this we should have to define
+
144
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the subordinate sentence as one uttered merely in order to deter-
mine another. Now it is quite obvious that a sentence can at
one and the same time be uttered for its own sake, and may
still serve as determinant to another sentence, and that it is
accordingly sure to exhibit a series of intermediate stages be-
tween the two extremes. It is further obvious that no reason-
able ground could exist for arranging sentences paratactically to
each other, were it not that there were an inner connexion be-
tween them; that is, unless one in some way determined the
other. There is thus no such thing as a purely paratactical
relation between two sentences in the sense that neither is de-
termined by the other; and the only possible conception of
parataxis is this, that instead of one sentence determining the
other, the two reciprocally determine each other.
221. Pure parataxis in this sense exists between parallel
sentences, whether the thoughts conjoined be of analogous or of
contrary import; he is blind, she is deaf; he laughs, she cries.
This is no longer, however, the case with narrative. If any one
tells us: I arrived in London at twelve o'clock; I went to the
nearest hotel; I was told that they were all full; I went on ;—
in such a case the preceding sentence in every case gives to the
following a temporal as well as a causal determination. This is,
however, a function not yet thought of at the moment when
the sentence is spoken. Accordingly we have again a union of
independence and dependence. We might desire a more cir-
cumstantial method of expression in which the sentence should
always occur twice; once as independent, once as dependent.
Instead of such a repetition, which at any rate occurs but ex-
ceptionally in practice, language avails itself of substitution by
means of a pronoun or a demonstrative adverb. It was for the
development of syntax a most significant step by which the
THE FUNDAMENTAL FACTS OF SYNTAX,
145
demonstrative, which originally referred solely to something
immediately before the senses, acquired a reference to some-
thing just uttered. By this means it became possible also to
give grammatical expression to the psychological process in
which a sentence is set down as independent, and at the same
time serves as determinant to a sentence following it. The
demonstrative may have reference to an entire sentence, or
merely to one member of the sentence. In the latter case also
it frequently happens that the entire sentence which contains
this member serves to determine the following. If, for instance,
a German says ich begegnete einem knaben; der fragte mich, the
word der refers to einem knaben; but the entire signification of
der is not exhausted by the general notion knabe; it is 'the
boy whom I met.' Thus the preceding independent sentence
is to some extent changed by the influence of the demonstrative
into a compound member of a sentence; the other parts of the
sentence subordinating themselves to the word referred to by
the demonstrative as its attributional determinant.
Hypotaxis.
222. Now if it is of the essence of all combination of sen- Graduated
Approxima-
tences that even the sentences set down as independent should tion to
maintain a certain element of subordination, it is quite natural
that from this point the possibility is open for a graduated
approximation to entire subordination, the independent value of
a sentence tending more and more to give way to the function
of serving as determinant to another. In the case of narrative,
the logical subordination is indicated in the IE. languages by
the employment of the relative tenses (the imperfect and plu-
perfect). Cf.: Cincta premebantur trucibus Capitolia Gallis; Fece-
rat obsidio jam diuturna famem: Juppiter ad solium superis
regale vocatis Incipe!' ait Marti (Ov. Fast. VI. 351). This usage
is very common in Ovid; it serves to introduce the situation
(
K
146
PRINCIPLES OF THE HISTORY OF LANGUAGE.
from which the narrative starts. Sentences principal in form,
but with decided logical subordination, are especially frequent
in the most various languages, if such words as just, already,
scarcely, still, etc., are added, or in such phrases as es dauerte
nicht lange, etc.; cf.: kaum seh' ich mich auf ebnem plan, flugs
schlagen meine doggen an (SCHILLER); Latin, vix bene desierat,
currus rogat ille paternos (OVID). In Latin also connected by
the copulative particle: vix ea fatus erat senior, subitoque fragore,
intonuit laevum (VERGIL); nec longum tempus et ingens exiit ad
caelum (VERGIL). [Vix prima inciperat aestas Et pater Anchises
dare fatis vela iubebat (VERGIL).] The form of this construction
-one universal in the German of to-day-is where a demon-
strative appears in the ensuing sentence, as ich war noch nicht
eingeschlafen, da hörte ich einen lärm; es dauerte nicht lange, so
kam er wider, etc.
223. In MHG. it occurs not unfrequently that of two sen-
tences placed asyndetically together, the first alone serves merely
as determinant to some member of the second.¹ Cf.: ein marc-
grave der heiz Herman: mit deme er iz reden began (ROTHER);
Josephus hiez ein wiser man: alse schiere er den rât vermam, mit
michelen listen muose er sich vristen (KAISERCHRONIK); ein
wazzer heizet In: dâ vâhten die Beiere mit in (KAISER.).
224. In the case of sentences introduced by either . . . or, the
former may be logically subordinated in such a way that it is
equivalent to a sentence introduced by in so far as . . . not; cf.
MHG., die ir christenlichen anthäiz mit andern gehäizzen habent
gemeret, . . . eintweder diu schrift ist gelogen oder si choment in
ein vil michel nôt (HEINRICH v. MELK); French, ou mon amour
me trompe, ou Zaïre aujourd'hui pour l'élever à soi descendrait
jusqu'à lui (VOLTAIRE).
1 Cf. Behaghel in the Introduction to Veldeke's Eneide, p. xxviii.
THE FUNDAMENTAL FACTS OF SYNTAX.
147
225. Where the sequence in the sentence is inverted, logical
independence and dependence cannot be united in the same
way. If a sentence serves as definition to a preceding one, it
is clear from the outset that it is uttered merely for the sake
of that sentence; cf.: ich kam nach hause, es schlug gerade 12
uhr; I had to tell him everything, he was so curious.
The fact
of the dependence comes out most clearly if the determining
sentence is intercalated in that which it determines. Such
intercalated sentences (or parenthesis) are of course common in
all, even the most highly developed languages, and they occur
indifferently in the most diverse logical relations to the governing
sentence.
from demand
gation to
226. When sentences of demand or interrogation come into Transition
logical dependence, they pass into designations of the condition or or interro-
of the concession. Cf.: geh hin: du wirst sehen or so (dann) wirst Hypotaxis.
du sehen; Latin, cras petito: dabitur (PLAUTUS); sint Maecenates,
non deerunt, Flacce, Marones (MARTIAL); also in combination
by means of a copulative particle, as sage mir, mit wem du
umgehst, und ich will dir sagen, wer du bist; Latin, divide et im-
pera; impinge lapidem et dignum accipies praemium (PHAEDRUS).
From such employment as this of the sentences of demand,
forms of sentences have arisen in different languages which are
felt as independent from the fact that what originally was an
only occasional' conception has received a 'usual' value.
Cf.: ich bin dir nah, du seist auch noch so ferne, or the English
imperatives, such as suppose, say [grant] ('Say you can swim,
'tis but a while,' SHAKESPEARE), which have to a certain degree
passed into the rank of conjunctions. Under this head we
must reckon, too, the Latin conditional sentences with modo
(cf.: ego ista studia non improbo, moderata modo sint), which
must not be apprehended as a governing construction, and
148
PRINCIPLES OF THE HISTORY OF LANGUAGE.
which can actually stand equally well side by side with dum.
In the same way it is well known that from the interrogation
has arisen a form of conditional sentences very common in
German and in English, and not unknown to the Romance
languages (willst du est tun, so beeile dich).
CHAPTER VII.
CHANGE OF MEANING IN THE DEPARTMENT OF SYNTAX.
TH
of Syntac.
Verbal
meaning.
HE most general of the statements which have been made Comparison
about the signification of words and their changes in tical with
Chapter IV. may be equally applied to the signification of syntactic change of
relations. In these, too, we must distinguish between 'usual' and
'occasional' meaning. The 'usual' meaning may be more than a
single one its changes arise from the variations of the occasional
signification, and they consist either in the enrichment or in the
impoverishment of the contents with a corresponding narrowing
or enlargement of the extent. Peculiar circumstances, however,
arise from the fact that we have here to deal with the relations
of several elements to each other (cf.: amo patrem, amor patris),
and that these relations are compounded into narrower or wider
groups (e.g., verb-objective accusative, substantive-genitive of
another substantive). Accordingly, besides the difference between
'usual' and 'occasional' signification, we must make another dis-
tinction, likewise a very important one, namely, that between
the signification of a general relation absolutely, and that of the
relation to some definite word. The signification which the
accusative has in its relation to a single definite verb must be
distinguished from the general signification which it has in its
relation to any word whatever, and also from that which it has
in its relation to any transitive verb whatever. The signification
150
PRINCIPLES OF THE HISTORY OF LANGUAGE.
The Genitive.
it bears in the first-mentioned instance may be of a more special
nature, and more or less isolated with respect to the general
signification. In modern times the view of the elder grammarians
has been much disputed, that a case is actually governed by a
verb or a preposition, a mood governed by a conjunction, etc.;
and it has been sought to derive the use of the case or the mood
from its general signification. But still, in a certain sense, and
with certain limitations, the traditional doctrine may be defended.
These general statements will be supported in what follows by
examples.
228. No simple signification can be laid down for the genitive
from which the functions which it already fulfils in the original
Indo-European language can be directly gathered. For in-
stance, we must from the very outset regard the genitive when
dependent on verbs as in a different category from the same case
when dependent on substantives. If we examine the latter, we
are justified in maintaining that the genitive, as is, generally
speaking, the case in ancient Greek, could be employed in IE.
for the expression of any given relation between two substantives;
we may, therefore, ascribe to this category a simple signification
of very meagre content and very wide extent, which is only in
occasional use. On the other hand, in NHG. the function of
the genitive in connection with substantives is considerably
restricted. Many usages possible still in MHG.-cf., goldes
zein ('staff of gold'), langes lebens wân ('hope of long life')—
are at the present day obsolete. We must nowadays look for
special meanings if we would set forth the usage of the geni-
tive, and in this we are compelled to distinguish several cate-
gories, and to set side by side several independent significations.
These might be most simply represented in this way possessive-
genitive, partitive-genitive, and the genitive denoting that the
CHANGE OF MEANING IN DEPARTMENT OF SYNTAX. 151
governing substantive is what it is, in virtue of that which de-
pends upon it (e.g.: the man's brother, the god of wine, the
writer of the work, the exploit of the hero). The last-named cate-
gory may be divided into two subdivisions in the case of nouns
of action, viz., the subjective and the objective genitive (cf.: the
government of the Czar, or the government of the country). The
statement of such categories has no doubt been lately regarded
as a purely logical division, to be sharply separated from grammar.
This view is, however, hardly correct, assuming that the state-
ment is made accurately and properly. The categories in ques-
tion have gained an independent position with respect to their
original general signification, and only owing to this fact has it
been possible for these alone to survive, while the other ways
of applying them, which would likewise range themselves under
the original signification, have disappeared.
tive.
229. The relation of the accusative to its governing verb is The Accusa-
analogous to that of the genitive to its governing substantive.
If we would give a general statement of the meaning of the
accusative, under which all the single methods of employing it
might be arranged, we must say that it denotes generally every
conceivable kind of relation which a substantive can bear to a
verb, except that of a subject to its predicate. But still we are
unable to employ it in each case in which such a common relation
occurs; and, indeed, as early as in the epoch of the IE. funda-
mental language this was inadmissible, even though the applica-
tion was still much freer and more extended in its range, as we
may see, for example, in Greek. Hence the assumption of one
single all-comprehending meaning is insufficient: we must place
side by side different usages which have gradually become
independent. But in this place the further fact must be taken
into consideration that in its relation to single verbs also a
152
PRINCIPLES OF THE HISTORY OF LANGUAGE.
fixed usage, with regard to the employment or otherwise of the
accusative, and a specialisation of the signification, has established
itself. We must accordingly distinguish between the free accusa-
tive, which is independent of the nature of the verb to which it is
attached, and the attached accusative, which is placed in connexion
with a small number of verbs only, and in each individual case in
a restricted signification.
230. Among the free uses of the accusative dating from the
earliest times, is its employment for the designation of what
extends over space and time (used not merely with verbs): also
the accusative of the contents of substantives etymologically
connected with the verb (as to fight a hard fight); in Latin the
accusative of the names of towns in answer to the question
Whither? A usage which has only arisen in later times is
that with verbs commonly intransitive in connexion with a
predicative adjective; cf.: to cry one's eyes red-to wash one's
forehead cool-to eat oneself full-to dance oneself tired—to cry
oneself hoarse, etc. In these cases, therefore, we should have a
widening of the signification. Still we must take into considera-
tion that special factors have contributed to start this construc-
tion; on the one hand, probably, the feeling-not yet thoroughly
extinct for the general signification of the accusative; on
the other, the analogy of cases like to shoot a man dead-to
buy a man free-to strike a man dumb-to beat black and blue,
etc. The case is similar in constructions like those in the vulgar
phrases to talk one's head off—to worm oneself into another's con-
fidence-to read one's own thoughts into those of one's author-to
laugh a man down, etc.
231. The accusative with compounds occupies a kind of
border-ground between the absolutely 'free' and the attached,'
when the simple verbs are either intransitive or govern quite a
CHANGE OF MEANING IN DEPARTMENT OF SYNTAX. 153
different kind of accusative.
ing that at least a great number of such verbs unite into a
group; while in their formation and transitive application, as
opposed to use, a certain freedom of movement makes itself
felt. Composites with be- in German [and in a less degree in
English] have the quite general function of making an intransitive
verb transitive, or of enabling a transitive verb to adopt a different
kind of object: cf.: befallen, beschreiben, bestreiten; besetzen, bewer-
fen, bezahlen; belabour, begrudge, bewitch.
We say a border-ground, consider-
232. The accusative, when attached to a definite individual
verb, has, as a rule, only a single meaning, limited by use. But
multiplicity of meaning is not quite exceptional, and this is in
such cases partly old-perhaps to be referred to the original
general signification of the accusative-and partly it proves
that originally one signification, and one only, has been ‘usual,'
while the other has grown up by 'occasional' transgression of the
usage; cf. in German, wunden schlagen-den feind schlagen—
das schwert schlagen; einen mit steinen werfen-steine auf einen
werfen; einen mit dem messer stechen-ihm das messer durch das
herz stechen; worte sprechen-einen menschen sprechen; in Latin,
defendere aliquem ab ardore solis-defendere ardorem solis ab
aliquo; prohibere calamitatem a provincia-prohibere provinciam
calamitate [in English, to keep a man from harm-to keep harm from
a man; to stick a knife into a man-to stick a man with a knife].
Undoubtedly a later development, and due partly to 'occasional'
usage, are the following constructions, which are especially com-
mon in poetry: in German, ein kind schenken (=saügen), wasser
in einen eimer füllen; in Latin, vina cadis onerare (VERG., a
variation for cados vinis), liberare obsidionem (LIVY, instead of
liberare urbem obsidione); in Greek, Sárрva тépyew, 'to stain tears'
(instead of 'to stain with tears,' PINDAR); aîµa Seveiv, 'to stain
>
154
PRINCIPLES OF THE HISTORY OF LANGUAGE.
The Preposi-
tions.
Apposition
and Partitive
Genitive.
(
More examples are given by
Since the relation expressed
more than a single one, the
blood' (instead of to stain with blood,' SOPHOCLES). [Cf., The
nightingale pours her throat.']
Madvig, Kleine schriften (p. 337).
by the accusative may in itself be
connexion of one verb with several accusatives is a circumstance
which arises quite naturally.
233. It would be incorrect to say of the IE. prepositions
that they governed this or that particular case. The case in
question was rather directly to be referred to the verb; its
general meaning was still apprehended, and was merely special-
ised by the preposition; whence it comes that different cases
could also stand after the same preposition, each with its own
special meaning. The Greek stands in many respects near to this
original state. But the case has more and more lost its inde-
pendence with respect to the preposition; the connexion of the
preposition with the case has become matter of custom, and the
consciousness of the signification of the case has grown fainter. In
the case of the NHG. prepositions which govern one case only, like
zu, um, or which govern several without any difference in the sense,
like trotz, the meaning of the case has decidedly disappeared; the
employment of the particular case is merely a traditional habit, to
which no value can be attached. Between this absolute fixity and
fast connexion on the one hand, and the original life and freedom
on the other, stands half-way the employment of the dative and
accusative in a different sense after in, auf, über, unter.
234. Appositional construction often appears when we ought,
if strict accuracy of expression were aimed at, to employ a partitive
genitive; not merely where the apposition consists of several
members which, taken together, give the same meaning as the
substantive to which they are appended; e.g.: They went, one this
way, the other that way; Classes populi Romani, alteram naufragio,
CHANGE OF MEANING IN DEPARTMENT OF SYNTAX. 155
alteram a Poenis depressam interire (CICERO); Capti ab Jugurtha
pars in crucem acti pars bestiis objecti sunt (SALLUST); but also
where the whole apposition represents a part only of the sub-
stantive belonging to it-cf.: Volsci maxima pars caesi (Livy);
Cetera multitudo decimus quisque ad supplicium lecti (LIVY); Nostri
ceciderunt tres (CAESAR): and correspondingly in cases where the
subject is expressed by the personal termination of the verb only—
cf.: Plerique meminimus, 'most of us remember' (LIVY); Simoni
adesse me quis nuntiate, 'one or other of you' (PLAUTUS). So in
MHG. si weinten sumelîche, 'many of them;' ja sint iu doch
genuogen diu mære wol bekant, 'to many of you.' In the case of
the designation of materials, which would regularly be expressed
by the partitive genitive, the less accurate appositional relation
appears; cf., in Latin: aliquid id genus, instead of ejus generis
(CICERO), coronamenta omne genus (CATO), arma magnus numerus
(LIVY). This more simple method of construction has gained a
wide hold in NHG. as against MHG.; cf.: ein stück brot (MHG., stücke
brôtes), ein pfund mehl, ein scheffel weizen, ein glas wasser, eine
menge obst, eine art tisch, etc. [cf, the Scotch use of a wee bit
body']. In this case the collective material appellations are inde-
clinable throughout. We cannot, if we analyse correctly the
instinct of language, longer recognise here any nominative or accu-
sative, but merely the simple stem without any designation of case.
Language has returned to its most primitive method of construc-
tion, the only one possible before the appearance of the case, and
as it appears to us in the old compounds.
a Verb.
235. As the object, so even the subject of a verb can lend itself Subject of
to denote a relation in a way varying from previous usage. Cf.
NHG. phrases, such as die bank sitzt voller menschen, ihm hängt der
himmel voller geigen, der eimer läuft voll wasser, or läuft leer [in
English, The cistern is running dry]. In MHG. the employment of
156 PRINCIPLES OF THE HISTORY OF LANGUAGE.
1
Substantive
and Adjecti-
val Predi-
cate.
such combinations with vol is even more frequent; cf.: daz hûs saz
edeler vrouwen vol, ouch gienc der walt wildes vol, daz gevilde was
vollez pavelûne geslagen (cf. HAUPT on the Erec, 2038); even in Hans
Sachs, den (wald) sach er springen vol der wilden tiere, all specerey
voll würme loffen; and a similar use prevails in Danish. Cf.
further, der narren herz ist wie ein topf, der da rinnt (LU.,--at the
present day, rinnen and laufen are used in a precisely similar way);
dass unsere augen mit tränen rinnen, und unsere augenlieder mit
wasser fliessen (LU.); das gefäss fliesst über; in Italian, le vie corre-
vano sangue (MALESPINI); in Spanish, corrieron sangue los rios
(CALDERON, cf. Diez iii. 114); in Latin, culter sanguine manat,
membra sudore fluunt; in English, the hall thick swarming now
with complicated monsters (MILTON); [their eyes run with tears];
NHG., der wald erklingt von gesang; das fenster schliesst schlecht;
just so in French, la fenêtre ne clôt pas bien. The Germans can
say indifferently, die blume riecht and ich rieche die blume, der
wein schmeckt and ich schmecke den wein. Compare the use of the
NHG., stinken; Latin, sapere; French, sentir. The German use
of sehen for aussehen is a parallel [cf., it looks bad]. If we take
the position that the relation between subject and object is to
be fixed and immovable for all time, we are compelled to assign
a double meaning to the verb for cases like those which we have
cited.
236. The corresponding departure from ordinary usage occurs
when a substantive is connected with an adjectival predicate, and
on a still larger scale when the connexion is attributival. Whereas
the adjective ought, properly speaking, to be employed only for the
quality which inheres in the substantive attaching to it, we find it
also employed where the relation is merely indirect. Cf. such
expressions as auf schuldigen wegen (SCHILLER)-i.e. 'ways in
which one incurs guiltiness;' einige gelassene augenblicke (GOETHE)
CHANGE OF MEANING IN DEPARTMENT OF SYNTAX. 157
etc.;
—¿.e. ‘moments in which one is quiet;' der hoffnungsvollen gabe
(GOETHE); bei ihrem unbekannten besuche (LESSING)-i.e. 'one in
which she remains unknown;' des trones, ungewiss, ob ihn mehr
vorsicht schützt, als liebe stützt (LESSING)-i.e. 'in which it is un-
certain.' Many such linguistic licences have become quite 'usual.'
We say quite commonly, a happy event, a joyful surprise, happy hours,
a learned treatise, in an intoxicated condition, in a foolish manner,
and further, we say, he gives us an unhealthy impression, and
a stingy gift, etc. The word sicher in German, like secure and sûr,
refers in the first instance to a person who has no need to be
anxious; in the second place to a thing or a person about whom
or which no one need be anxious: ekel refers on the one hand to
a person who easily feels disgust; on the other to an object at
which disgust is felt. If such freer combinations are apprehended
after the analogy between the substantive and the adjective agree-
ing with it, we arrive at a point where we may lay it down that a
change in word-meaning has occurred.
237. Such a licence is allowed with special frequency in the
case of participles; cf.: einer reuenden träne (LESSING); lächelnde
antwort (GOETHE); in der schaudernden stille der nacht (LESSING);
zum schaudernden concert (SCHILLER); der könig betrachtet ihn
mit nachdenkender stille (SCHILLER); in seiner windenden todesnot
(GOETHE); nach dem kostenden preise (NICOLAI). [We can say in
English a suspicious calmness-a smiling answer-a melancholy
task.¹]
238. Common examples of this are to be found in such usages
as sitzende, liegende stellung-fallende sucht-schwindelnde hohe-
im wachenden traume, etc., and the now condemned phrase bei
nachtschlafender zeit. Many similar expressions are very common
in English, such as, dying day—parting glass-writing materials
1 Further examples are given in Andresen, Sprachgebrauch, p. 82 sqq.
158 PRINCIPLES OF THE HISTORY OF LANGUAGE.
Participial
Construc-
junctions.
-dining room-sleeping apartment-singing lesson-falling sick-
ness. Cf. also such examples in French as, thé dansant, café
chantant. Tacitus has such uses as haec plebi volentia fuere, in-
stead of volenti, etc. (cf. DRAEGER, § 193, 3). We find examples
for the perfect participle in ein längst entwohnter schauer (GOETHE);
in diesen letzten zerstreuten tagen (GOETHE); der beschuldigten
heuchelung (SCHILLER), i.e., 'the hypocrisy of which I am accused ;'
and in English, the ravished hours (PARNELL), for the hours of
ravishing pleasure.' Common examples are ein eingebildeter mensch,
ein bedienter.
239. Parallel to this use is probably the custom of freely
attaching a predicative attribute, which is indeed condemned as
careless, but which still occurs frequently enough in such cases
as seltene taten werden durch jahrhunderte nachahmend zum
gesetze geheiligt (GOETHE); lustig davonfahrend wurden die
eindrücke des abends noch einmal ausgetauscht (RIEHL); zurückge-
kehrt wurde des ermordeten kleidung untersucht (BRACHVOGEL).
Andresen cites other instances, mostly from newspapers (Sprach-
gebrauch, p. 113). In this case we feel ourselves compelled to
supply a subject to the predicative attribute; but it would be
equally possible to fill up the example cited above 'mit nachden-
kender stille' to 'mit stille, während welcher er nachdenkt' without
any thing of this being contained in the expression.
240. In the case of participial constructions the time-relation
tions: Con- alone is expressed in which the state or the event which has
been denoted by the participle stands to the finite verb. It is
possible, however, for numerous relations to subsist at the
same time, so that in resolving the participial construction
into a single sentence we have to employ sometimes one
conjunction and sometimes another. Still we cannot on this
account maintain that the participial construction in itself admits
CHANGE OF MEAning in department of SYNTAX. 159
1
of different meanings-i.e. that it denotes now the reason, now
the condition, now an opposition, etc. These relations remain in
each case only 'occasional' and accidental. The case is otherwise,
however, with dependent sentences introduced by a temporal con-
junction. In this case it is possible for the accidental relation to
the governing sentence to attach itself to the conjunction, and to
become an integral portion of its 'usual' signification. Thus, for
instance, the employment of the word während in German to
denote a contrast must be acknowledged as a peculiar 'usual'
function, side by side with the fundamental meaning. This is
manifest, apart from our feeling for language, from the considera-
tion that this function operates equally where that which is spoken
of as having passed between the dependent and governing sentence
is not contemporaneous-cf.: du beligst mich, während ich dir
immer die wahrheit gesagt habe. In the same way we must grant
the MHG. sit, besides its temporal signification, that of the NHG.
causal da, as an independent one, for it can be used in contradic-
tion with the fundamental meaning in a case of contemporaneous
action between dependent and governing sentence-cf.: sît ich âne
einen vrumen man mîn lant niht bevriden kan, so gewinne ich gerne
einen. The development may then proceed further, as the original
temporal signification disappears entirely, as in the case of NHG.
weil. In precisely a corresponding way prepositions such as
through, by, of local or temporal meaning, pass over to a causal
one.
CHAPTER VIII.
Definition.
Contamina-
tion :
BY
ON CONTAMINATION.
Y contamination' I understand the process by which two
synonymous forms of expression force themselves simul-
taneously into consciousness, so that neither of the two makes its.
influence felt simply and purely a new form arises in which
elements of the one mingle with elements of the other.
process is also of course in the first place individual and momentary.
However, by means of repetition, and the intercourse of different
individuals, it is possible for the individual usage gradually to
become 'usual.'
This
242. Contamination manifests itself partly in the phonetic form
1. Phonetic. of single words, partly in their syntactical combination.
243. A confusion of two words not etymologically connected is
comparatively rare. Schuchardt has indicated one characteristic
example. In the Aemilian dialect there is a word cminzipià,
'to begin,' a contamination arising from the words cominciare and
principiare of the Italian written language. The confusion was
rendered easier in the case of forms which may be amplified into
a paradigma. The older form wis ('sei'), from OHG. wesan, is in
MHG. gradually thrust aside by bis under the influence of bist.
OHG. bim (bin) is probably a contamination of im (Gothic), and
*bium (AS. beóm); a similar occurrence may be noted in the con-
verse way in AS. eóm.
ON CONTAMINATION.
161
244. The confusion of words belonging to the same etymologi-
cal group is more common. Cf. gewohnt from the MHG. adjective
gewon (still found in gewohnheit aud gewöhnlich), and the MHG. par-
ticiple gewent from wenen (gewohnen); doppelt from the adjective
doppel (= French double), and the participle gedoppelt, still usual in
the last century; zu guter letzt from zu guter letz (MHG. letze,
'departure') and zu lettz.
245. Not merely do two single forms suffer reciprocal contami-
nation, but single forms affect each other similarly.
In this way
then there arises not unfrequently a pleonasm of the formative
elements, a form composed in an unusual way, being further
enriched by the suffix of the regular formation. Under this head¹
come forms like NHG. ihrer, ihnen, derer, denen; OHG. inan (from
in influenced by blintan, etc.); NHG. Fritzens, Mariens, from the
older Fritzen, Marien, forms which have been affected by the most
common genitival termination. Further, words like the Latin
jactitare, cantitare, ventitare, instead of jactare, etc., have arisen
under the influence of volitare, etc.; and Spanish adjectives, like
celestial, divinal, humanal (cf. MICHAELIS, p. 38). Especially com-
mon is the multiplication of the suffixes of the comparative and
superlative; cf. NHG. öftrer (common in LESSING); letzteste
(GOETHE); OHG. mêriro, as against the Gothic maiza; Gothic
aftumists, auhumists, frumists by the side of aftuma, auhuma,
fruma; hindumists, spêdumists; late Latin pluriores, minimissimus,
pessimissimus, extremissimus, postremissimus; Greek ȧpelóτepos,
χερειότερος, πρώτιστος, etc. [Cf. in English the forms former,
nearer, lesser, and Most Highest.] In the same way we must
explain the double prefix in gegessen, MHG. gezzen.
Contamina-
246. Contamination plays a great part in the area of syntax. Syntactical
We will cite, in the first instance, a few examples of merely tion.
Cf. Brugman, Morph. Unt. iii. 67 sqq., and Ziemer, Streitz. 146.
L
1. Momen-
tary.
162
PRINCIPLES OF THE HISTORY OF LANGUAGE.
2. Usual.
um
passing anomalies which have no influence on usage. Lessing
uses the phrase, um deines lebens wegen; a confusion between
willen and wegen; a similar confusion is quoted by
Andresen (Sprachgebrauch, 194), from the Kölnische Zeitung, as
between um and halber. Goethe writes: freitags als dem ruhigsten
tage, as if am freitage had been written. Lessing writes: ich
habe nur leugnen wollen, dass ihr alsdann der name malerei
weniger zukomme; a confusion between leugnen... dass
zukomme and behaupten . . . dass weniger zukomme. Hans
Sachs writes: Ein jedes thut, als es dann wolt als jhm von jhem
geschehen solt. In this the two thoughts are confused; wie es
wollte dass ihm von jenem geschehen sollte, and wie ihm geschehen
sollte. Hartmann von Aue writes: er bereite sich dar zuo als er ze
velde wolde komen (from dar zuo daz er ze velde kæme, and als er ze
velde wolde komen). Again the same author writes: des weinens tet
in michel nôt from daz weinen tet in and des weinens was in.
Goethe: im betragen unterschied sich auch hier der gesandte von
Plotho wider vor allen andern; a confusion with zeichnete sich
aus or something similar. Goethe writes: die schicksale meiner
wanderschaft werden dich mehr davon überzeugen, als die wärmsten
versicherungen kaum tun können; here the word kaum belongs
properly to an entirely different manner of expression.
247. We turn to cases in which the contamination has become
usual, or at least makes its appearance as a frequently occurring
licence. The construction das gehört mein is very common (cƒ.
GRIMM, Deutsches Wörterbuch, 4a, 2508); a confusion between
gehört mir and ist mein. In English we say, 'I am friends with
him,' from I am friendly with him and we are friends; the Danish
popular idiom is similar, han er gode venner med hem ('he is good
friends with her'). The Danish popular idiom also has the
expression, jeg fölges med ham (I follow with him '), from jeg
ON CONTAMINATION.
163
1
fölger med ham, and ve fölges ad ('we follow each other,' i.e. ' we
go together ').¹ In Greek we find expressions like o povs
τοῦ χρόνου, τὴν πλείστην τῆς στρατιᾶς, a confusion between
ὁ ἥμισυς χρόνος and τὸ ἥμισυ τοῦ χρόνου, etc., correspondingly
we find in Spanish muchas de virgines, instead of muchas virgines
or mucho de virgines; à pocos de dias, una poca de miel, tantas
de yerbas, la mas de la gente (CERVANTES); in Italian, in poca d'ora,
la più della gente (BOCCACCIO). Similar confusions are found also
in Portuguese, Provençal, and OF. (cf. DIEZ, iii. 152.) We have a
similar contamination in the case of the Latin gerund: poenarum
solvendi tempus (LUCRETIUS), from poenarum solvendarum and
poenas solvendi; exemplorum eligendi potestas (CICERO; cf.
DRAEGER, 597da). Cicero writes: eorum partim
eorum partim in pompa
partim in acie illustres esse voluerunt, in which there is a confu-
sion between eorum pars and ii partem: the corresponding process
is common in older NHG.; cf. theils leute nennen ihn zum spott den
unverstand (CRonegk).
248. Not unfrequently, in referring to what has preceded, an
inaccuracy arises owing to the displacement of a word by the idea
of a word etymologically related with the word actually employed,
where the speaker might equally well have employed either.
Thus, for instance, the idea of the inhabitants displaces that of the
town or of the country ; cf. Θεμιστοκλῆς φεύγει ἐς Κέρκυραν, ὢν
avтŵν EveρYÉTηS (THUC.); Domitius navibus Massiliam pervenit
atque ab iis receptus urbi praeficitur (CAESAR); Sutrium, socios
populi Romani (LIVY); NHG. so waren wir denn an der grenze von
Frankreich alles französischen wesens auf einmal bar und ledig.
Ihre lebensweise fanden wir zu bestimmt und zu vornehm, ihre
dichtung kalt, etc. (GOETHE). Other examples are: innere stärke
kann man der Bodmerischen und Breitingerischen kritik nicht
1 See Madvig, Kl. Schriften, 193³.
164
PRINCIPLES OF THE HISTORY OF LANGUAGE.
absprechen, und man muss den ersten als einen patriarchen ansehn
(HERDER); het ich mich nicht jung thun verweiben, die er mir jetzt
drey jar anhengen thet (the die referring to a wife whom he should
have taken; HANS SACHS);¹ MHG. in dem palas, der wol gekerzet
was, die (i.e. ‘welche kerzen') harte liehte brunnen (WOLFRAM);
entwâpent wart der tôte man und an den lebenden gelegt (diu wâpen
must be supplied as subject; id.); servili tumultu, quos (as if
servorum had preceded; CAESAR).
The most common case is
that the relative refers to a possessive pronoun, as if the personal
pronoun had preceded; cf. laudare fortunas meas, qui gnatum
haberem tali ingenio praeditum (TERENCE); TŶs ẻµîs ẻπtelσódov,
ôv µýt' ỏkveîte (SOPH.); in MHG. it is universal.
249. In Latin there arise from the confusion of the comparative
and superlative manner of expression, combinations like hi cetero-
rum Brittannorum fugacissimi (TACITUS); omnium ante se genito-
rum diligentissimus (PLINY).2 [Cf. Milton's copy of this use: Adam
the goodliest man of men since born; the fairest of her daughters, Eve;
and Shakespeare's Midsummer Night's Dream, V. i. 252: This is
the greatest error of all the rest.] Conversely the superlative occurs
sometimes with the meaning of the comparative; cf.: omni vero
verissimum certoque certissimum (ARNOBIUS). Compare the Old
Norse hæstr borinn hverjun jọfri (GRIPISSPA, 'the highest,' for
'higher than any prince').
250. In Latin we often have joined to the imperative the word
jam dudum; e.g., jam dudum sumite poenas—a confusion between
the thoughts 'pray take' and 'you should long ago have taken,' as
if he had written jam dudum debitas. [Cf. Those dispositions that
of late transform you from what you rightly are (LEAR, I. iv. 242);
and He is ready to cry all the day.]
1 Other examples may be found in Andresen, Sprachg. 252 sqq.
2 Cf. Ziemer; Comp. 55 sqq.
ON CONTAMINATION.
165
251. In MHG. an interrogation is common with the infinitive,
e.g.: do enweste er wie gebâren; we expect a finite verb, and the
construction is probably only to be explained by our assuming the
simultaneous influence of those cases in which the infinitive was
directly dependent on the verb, without any interrogative. The
same thing holds good, of course, of the corresponding Romance
constructions. Cf. in French, je ne sais quel parti prendre; in
Italian, non so che fare (DIEZ, iii. 230); [and in English I do not
know what to do]. We see a similar construction in the Italian
phrase non ho che dire; in the Spanish, non tengo con quien hablar;
in the French, il trouva à qui parler, and la terre fournit de quoi
nourrir ses habitants; in Late Latin, non habent quid respondere
(DIEZ, u.s.); in English, how have I then with whom to hold converse
(MILTON); then sought where to lie hid (ib.); [hath not where to lay
his head], etc.
252. It must further be regarded as a contamination, if an
interrogative sentence be made dependent on a verb, and at
the same time be made the subject of this interrogative sentence.
as nominal object; cf., in Latin, nosti Marcellum quam tardus sit
(CICERO); viden scelestum ut aucupetur (PLAUTUS); observatote
eum quam blande palpatur mulieri (TERENCE); dic modo hominem
qui sit (PLAUTUS); patriam te rogo quae sit (PLAUTUS); in Italian,
tu 'l saprai bene chi è (BOCCACCIO); and a similar use is common
in the older Romance languages. [Cf.: You hear the learned
Bellario what he writes (Merchant of Venice, IV. i. 167); cf. also
Lear, I. i. 272]; cf. Diez, III. 391. Just in the same way a nominal
object stands side by side with an object sentence with dass;
cf. MHG. swenne er sin séle sæhe daz si in tôtsünden wære, die
liset man si wilen waren des wunderlichen Alexandres man, do
hiez in got das er dar in gienge, die wil ich daz siz merken,
NHG. da ihn sahen alle, die ihn vorhin gekannt hatten, dass er
;
166
PRINCIPLES OF THE HISTORY OF LANGUAGE.
mit den propheten weissagete (LUTHER); welchen ihr sprecht,
er sei euer gott (ib.). The object of the governing sentence.
may also be contained in the dependent object, cf.: vier-
hundert taler, die sie nicht wüsste, wie sie sie bezahlen sollte
(LESSING). In the same way it is possible, by the side of a
subject sentence with dass as subject, for the subject or object
of that sentence to appear as subject of the main sentence; cf.:
mich will Antonio von hinnen treiben und will nicht scheinen,
dass er mich vertreibt (GOETHE); and again, nichts, was ihn
gereuen könnte, dass ers gab (ib.).
253. In German instead of der selbe der or der gleiche wie
they sometimes use the phrase der selbe wie and der gleiche der;
just so in Latin, idem ut; e.g., in eadem sunt injustitia, ut si in
suam rem aliena convertant (CIC.). We often meet with phrases
of the following kind: dass sie nichts spricht kommt daher, weil
sie nichts denkt (LESSING); der gedanke wurde dadurch notwendig,
weil man voraussah (WIELAND); wortstreit, der daraus entsteht,
weil ich die sachen unter andern combinationen sentiere (GOETHE);
in dem augenblicke, wenn wir ihn auch seines bogens beraubt sehen
(LESSING); die grösste feinheit eines dramatischen richters zeiget
sich darin, wenn er in jedem falle zu unterscheiden weiss (LES-
SING). In universal, and in some cases even obligatory, use are
combinations like jedesmal wenn or wo (instead of dass) in dem
augenblicke wo (Goethe still says in dem augenblick, dass er
amen sagte); correspondingly we find in French, au temps où;
at an earlier period, au temps que; zu dem zwecke, in der
absicht damit; deshalb, deswegen, aus dem grunde weil; desto
besser weil (MHG. daz); and in English, the rather because, as
well as the rather that.
254. When Cicero says cum accusatus esset, quod contra rem-
publicam sensisse eum dicerent, he makes a confusion between quod
ON CONTAMINATION.
167
sensisse eum dicebant and quod . . . sensisset.
For further
examples cf. Draeger, § 537.
τόδε, ὡς οἶμαι, ἀναγκαιότατον εἶναι, Ziem., 105 [cf.: Marry, that
I think be young Petruchio (Rom. and Jul., I. v. 133), which is a
confusion between 'that, I think, is,' and 'I think that that be.'-
ABBOTT, § 411.]
Plato even uses constructions like
255. A common construction in MHG. seems to have been
in gesehe vil schiere mîn liep (unless I see my love soon'), ich
bin, or so bin ich tôt. The same sense would have been yielded
by the paratactical combination ich gisihe vil schiere mîn liep
oder ich bin tôt. Instead of this the minnesinger Steinmar says.
in gesehe vil schiere mîn lieb alder (=oder) ich bin tot. Another
kind of confusion is still more striking, in which oder appears
before the sentence with ne; ich gelige tôt under mînen van
oder ich nebeherte mîn êre (KAISERCHRONIK).¹
256. A predicatival attribute may have the same function
as a dependent sentence introduced by a conjunction. Con-
sequently many conjunctions may be placed before the simple
adjective, whereby a more exact description of the circumstance
is attained. English is peculiarly rich in such constructions;
cf.: talents angel-bright, IF wanting worth, are shining instruments
(YOUNG); nor ever did. I love thee less, THOUGH mourning o'er
thy wickedness (SHELLEY); Mac Ian, WHILE putting on his
clothes, was shot through the head (MACAULAY).2 In German, too,
it is possible to say: ich tat es, obschon gezwungen, etc. Cor-
respondingly, in Latin, many conjunctions are placed before
the ablative absolute; cf.: quamvis iniqua pace honeste tamen
viverent (CICERO); etsi aliquo accepto detrimento (CÆSAR); etsi
magno aestu (CICERO). The conjunctions quasi and sive, which
3
1 Further examples are given by Dittmar, Zeitschrift für deutsche Philologie, Ergän-
zungsband, p. 211.
2 Cf. Mätzner, iii. p. 72.
3 Cf. Dräger, § 59.
168
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Pleonasms
arising from
originally could merely serve to introduce a sentence, are quite
commonly added to mere dependent clauses.
257. Conversely, coincidence in the function of dependent sen-
tences and prepositional determinants tends to the employment of
prepositions to introduce dependent sentences. English has many
examples of this; cf.: FOR I cannot flatter thee in pride (SHAKE-
SPEARE); AFTER he had begotten Seth (GENESIS); WITHOUT
they were ordered (MARRYAT). [(This use of without is incorrect.)
'I HATE him for he is a Christian, but more FOR THAT . . .
he lends,' etc. (Merch. of Ven. 1. iii. 43)]. Till and until are
specially common in this use. [Indeed the prepositional use
has almost died out in written English, but was common
in the Elizabethan age; cf. Shakespeare, from the first corse TILL
he that died to-day (Hamlet, I. ii. 105), where he should be strictly
speaking him. Other instances are quoted by Abbott, § 184.] It
must, however, be particularly noticed that the constructions for
that, after that, etc., are permissible as alternatives to for, after, etc.,
when used as conjunctions. A preposition also stands before
indirect questions; cf.: at the idea of how sorry she would be
(MARRYAT); the daily quarrels about who shall squander most
(GAY);¹ cf., in Spanish, este capitulo habla de como el rey non
deba consentir: and similar constructions are found in Portuguese
and Old Italian.2
258. The result of contamination in the area of syntax is
Contamina often a pleonasm. Thus, for instance, in Latin we meet with a
tion.
3
multiplication of particles expressing similarity, as pariter hoc fit
atque ut alia facta sunt (PLAUTUS); cf. the common but in-
correct German expression als wie, [and the English like as if].
Similarly we find in Latin quasi si, nisi si.
In English we can
2 Cf. Diez, iii. p. 388.
1 Mätzner, iii. p. 445.
Cf. Dräger, § 516, 14.
4 Ib. § 518, 16.
5 Ib. § 557, fg.
ON CONTAMINATION.
169
connect a preposition either with a substantive or with a governing
verb; the two even occur in combination; cf.: that fair for
which love groaned for (SHAKESPEARE) [cf. also, In what enor-
mity is Marcius poor in? (Coriol. II. i. 18)]. Nay, we often find
such expressions as of our general's (SHAKESPEARE) instead of
of our general or our general's. Not uncommonly a preposition
denoting whence is added to adverbs of place, which of them-
selves denote the same direction; this preposition should, strictly
speaking, be connected with an adverb denoting rest in a place;
cf. Latin deinde, exinde, dehinc, abhinc; NHG. von hinnen, von
dannen, von wannen, von woher, Span. donde from de unde; cf.
the English use of from henceforth. In Latin we often find
in the passive a pleonastic denotation of the pluperfect; e.g.: censa
fuerunt civium capita (LIVY); sicuti praeceptum fuerat (SALLUST).¹
We often meet with expressions like erlauben Sie, dass ich mich
dabei beteiligen darf.²
259. The forms of comparison of the adjective and adverb
present many examples which seem instructive in this connection.
In MHG. we often find a baz added to the comparative, as græzer
baz, etc.; and in the same way in Latin, especially with the
comic poets, we find magis or potius; in Greek, µâλλov;³ thus
also in Gothic, mais vulprizans. We find a similar use with
the superlative; cf. µáλiota µéyiσтov (XENOPHON); die zunächst-
stehendsten (quoted from the Frankfurter Zeitung by Andresen).
We may compare with this such combinations as magis or potius
malle, prius praecipere, πλέον προτιμᾶν (ΧENOPHON), πρότερον
πρoλaµßáveiv (DEMOSTHENES). Lessing writes in the Laokoon :
niemand hatte mehr recht, wegen eines solchen geschwieres bekannter
zu sein [cf.: most unkindest (Jul. Cæs. III. ii. 187); thy most worst
2 For other instances see Andresen, Sprachg. 136, 7.
1 Dräger, § 134.
3 Cf. Ziemer (comp. 154, 5).
17O
PRINCIPLES OF THE HISTORY OF LANGUAGE.
1
(Winter's Tale, III. ii. 180).] The comparative is united with a
preposition denoting preference; such preposition could of course,
strictly speaking, be only united with the positive: olσw ǹ tupavvis
πρὸ ἐλευθερίης ἦν ἀσπαστότερον (HERODOTUS); αἱρετώτερον
εἶναι τὸν καλὸν θάνατον ἀντὶ τοῦ αἰσχροῦ βίου (XENOPHON);
prae illo plenius (GELLIUS); ante alios immanior omnis (VERGIL).¹
[Cf. numerous instances drawn from Old Greek, where the
comparative degree is used where several, or indeed all the objects
belonging to the same class are compared with a single object: e.g.
THEOCRITUS, xv. 139, Εκτωρ, Εκάβας ὁ γεραίτερος εἴκατι παίδων.
Cf. Berliner Phil. Wochenschrift, No. 52, p. 1622. Wolfram von
Eschenbach places the two possible turns side by side: diu
prüevet manegen für in baz dan des mares herren Parzival (in
refers to Parzival).]
0
260. Pleonasm arising from contamination occurs most exten-
sively in the case of negations. It has almost disappeared from
the [English and] German written language; but in the last
century it was still very common in the latter. Thus, we find
after negative expressions, in the dependent sentence intro-
duced by dass, a negation which seems to us illogical; cf.:
es kann nicht fehlen, dass die meisten stimmen itzt nicht gegen
mich sein sollten (LESSING); wird das hindern können, dass
man sie nicht schlachtet? (SCHILLER); der verfasser ver-
bittet sich, dass man seine schrift nicht zu den elenden spöt-
tereien rechne (CLAUDIUS); dir abzuraten, dass du sie nicht
brächtest (SCHILLER); nun will ich zwar nicht läugnen, dass an
diesen büchern nicht manches zu verbessern sein sollte (LESSING);
ich zweifle nicht, dass sie sich nicht beide über diese kränkung
hinwegsetzen werden (LESSING); der lord Shaftesbury erklärt sich
1 Cf. Ziemer (comp. 95 sqq.).
2 Further examples are cited in Dittmar, Zeits. f. d. Philol., Ergänzungsband, 299 sqq.
ON CONTAMINATION.
171
dawider, dass man nicht zu viel wahrheit sagen solle (Translation
of Tom Jones, 1771). We find a corresponding usage in MHG. dar
umbe liez er daz, daz er niht wolte minnen (KUDRUN); ich wil des
haben rât, daz der küene Hartmuot bî mir niht enstât (KUDRUN).
[In Chaucer and Shakespeare the use of the double negative is
common, as, first he denied you had in him no right (Comedy of
Errors, IV. ii. 7); you may deny that you were not the cause
(Richard III., I. iii. 90). See other instances in Abbott's Shak-
spearian Grammar, 406.]
261. As early as in MHG. the negation can be dispensed with.
If the governing sentence is negatived, in MHG. the dependent sen-
tence is not commonly introduced by a conjunction; instead, merely
the negation en with the subjunctive is used; cf.: mîn vrouwe sol
iuch niht erlân irn saget iuwer mære. The origin of the construc-
tions seems to be owing to the thought of the dependent sentence
forcing itself into consciousness, on the one hand, as dependent
on the governing sentence; on the other hand as independent.
When, for instance, we read in the Kudrun das wil ich widerraten,
daz ir mich mit besemen gestrafet nimmer mêr, this is, strictly
speaking, a confusion of the two thoughts, I will counsel you not to
punish me ever again and never punish me again. This explana-
tion is certainly only applicable in cases where the governing
sentence is positive. Not until the application of the negative has
become usual can it be transferred to the cases with a negative
governing sentence. It is possible-nay, it is probable—that the
employment of the negation is traditional, and dates from a time
when no strict grammatical subordination of one sentence to the
others existed at all. Assuming the truth of this, we still have to
deal with a contamination. Kindred peculiarities occur in Latin,
in the Romance languages, and in others.
262. Correspondingly, the negation appears in connexion with
172
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the infinitive, in cases where it is not possible to derive it from an
originally independent position; cf.: freilich hüten wir uns sie nicht
an den gnädigen herrn zu erinnern (GOETHE); ich habe verschworen
nicht mehr an sie zu denken (GOETHE); ich habe es verredet, in meiner
gegenwärtigen lage niemals wieder eine nacht in Braunschweig zu
bleiben (LESSING); der habe ihm verboten, den ring weder der
königin zu geben, noch dem grafen zurück zu senden (LESSING). [Cf.
He waived indifferently 'twixt doing them neither good nor harm
(Coriolanus, II. ii. 19, 20); cf. ABBOTT, 405.]
263. A negation may also be found after an expression not
essentially negative, but negatived; cf.: vnd gentzlich kein hoffnung
mehr handt zu samb zu kummen nimmer meh (HANS SACHS).
264. A pleonastic negation occurs in various languages after
without (cf. MAETZNER, French Grammar, § 268). Cf.: sans nul
égard pour nos scrupules (BÉRANGER); sin fuerza ninguna (CAL-
DERON); senza dir niente; sin hablar palabra ninguna; sans que son
visage n'exprimat la peine (SAINT-PIERRE); sin que nadie le viese
(CERVANTES); NHG. ohne dass wir bei seiner beurteilung weder auf
irgend ein gesetz noch auf irgend einen zweck rücksicht nehmen
(SCHILLER); ohne dass ich weder von dem vorhergehenden noch von
dem nachfolgenden irgend unterrichtet gewesen wäre (GOETHE); ¹
A similar construction is found with ausser: ihr findet widersprüche
überall, ausser da nicht, wo sie wirklich sind (LESSING);² and after
als referring to a preceding nichts, cf.: es mangelt ihm nichts, als
dass es nicht gekläret ist (SCHOCH); es fehlt nichts als dass du nicht
da bist (GOETHE).
265. A regular negative is sometimes prefixed to words which
in themselves have no absolutely negative signification, but merely
acquire it by litotes. Thus in MHG. we find nie attached to selten ;
1 Another example is found in Andresen, p. 145.
2 Cf. Andresen, loc. cit.
1
ON CONTAMINATION.
173
cf.: ein wip, der ich selten nie vergaz (MINNESINGER); daz man nie
deheinen also richen sô senftes willen selten vant (BITEROLF); in the
same way selten nieman = never any one. In NHG. we sometimes
find a negative word after kaum : nichts mag kaum sein so ungelegen
(FISCHART) = 'Scarcely can anything be so difficult;' and after
schwerlich: schwerlich niemals (LESSING).¹
1 Cf. Sanders, 2 b, 1048 b.
CHAPTER IX.
The condi-
tions of new
creation still
present.
WE
ORIGINAL CREATION.
E have hitherto, in considering linguistic processes, made
it a rule to base our views on observations made on
the historic development of language, which is easily traced;
and, taking this as a starting-point, to draw conclusions bearing
on the original history of language. We must endeavour to
extend this principle to our judgments on original creation as well;
though in this case we find ourselves confronted by greater
difficulties. We do not easily meet with opportunities which.
facilitate accurate observations on this subject. For exceptional
cases, such as the capricious invention of the word gas, do not
throw much light on the natural development of language. A
mysterious darkness veils the process, and views are always
being put forward which refer to it as a special heirloom, now
declared to be lost, of the original human race. All such views
must be decidedly rejected. The conditions requisite for primitive
language-making must one and all be inherent in the bodily and
intellectual nature of mankind as man now exists. Indeed, if
our intellectual tendencies have developed into higher perfec-
tion, we shall be actually driven to conclude that these con-
ditions are now present, and in a higher state of perfection than
at the period of the rise of human language. If we, generally
speaking, create no new material of language, this is simply due
1
ORIGINAL CREATION.
175
to the fact that the need for doing so no longer exists. It is
scarcely possible for an idea or sensation to manifest itself in
us without some link of the material of language which we have
inherited from it. This immense mass of material, to which we
are once for all habituated, forbids anything new to spring up by
its side, allowing, as it does, of convenient augmentation by means
of manifold combinations and by transitions of meaning. But if
the experiment were to be made of allowing a number of children
to grow up unacquainted with any language, excluding them
carefully from outside intercourse and limiting them to their own
society, we can hardly doubt as to the result; they would, as
they grew up, form a language of their own out of words origi-
nated by themselves. Something approaching such an experiment
is said to have been actually made. Robert Moffat's report of
the state of language, as evidenced in isolated desert villages
of South Africa, is well known from Max Müller's lectures.
According to that report, the children invent a language for
themselves during the prolonged absence of their parents. But
I should not be inclined to attach too great value to such stories
without ample confirmation.
has never
ceased.
267. We do not, however, require to go so far afield. We Creation
have every right, I think, to maintain that even in the languages wholly
of the European civilised peoples the creation of new material
has never completely ceased. After all the progress made by
IE. etymology in the last ten or twenty years, there still remains
a very large residuum of words which can neither be referred to
roots of the original language, nor yet proved to be borrowed
from foreign languages. Even when we go through the repertory
of the living German dialects we find much there that we are
unable to connect with the MHG. repertory. We must doubt-
less account for this fact in part from the imperfection of
176
PRINCIPLES OF THE HISTORY OF LANGUAGE.
our record as we have received it, and must remember that our
scientific combinations are as yet imperfect. Still, however,
there remain a large number of cases in which it is difficult to
see how, by the aid of sound-development and formation by
analogy, a connexion with older material could ever be possible.
We shall thus have to ascribe to the more and most recent
periods of language not merely the capacity for original creation,
but, what is more, the actual translation of this capacity into
action. We must here enter our objections to the theory that
two periods have to be distinguished in language—in the one of
which the original material of language, the so-called roots, was
created, and a second in which speakers confined themselves to
the formation of combinations out of the material at hand. The
fact is that in the development of popular language no point of
time can be assumed from which original creation is excluded.
On the other hand, it is undoubtedly true that shortly after the
first original creations the same kinds of further development as
we have observed in the later periods have appeared. In this
respect there exists between the various phases of development
no difference in their nature, but solely in their degree. There
is no change save in the proportion of original creation to trans-
mission by tradition of the material created, and to the other
means whereby language is enriched-the enlargement of signi-
fication by apperception, the combination of simple elements,
formation by analogy, etc.
268. The essence of original creation consists, as we have
already seen, in the fact that a sound-group is placed in relation
to a group of ideas, which then comes to constitute its signifi-
cation, and this without the aid of a connected group of ideas
already attached to the sound-group. Such an original creation
is in the first instance the work of the impulse, which may dis-
ORIGINAL CREATION.
177
appear and leave no lasting traces. In order that a real language
may thus arise, it is necessary for such utterances to leave behind
them a psychical after-operation as well, whereby the sound can
be reproduced by memory by means of the signification, and the
signification by means of the sound. Further, the word must
be understood by other individuals, and then be reproduced by
them as well.
to original
the results
departments
history of
269. The experiences which we have made as to the rise of Application
new words by analogical formation, and the apprehension of new creation of
conceptions by the aid of the existing vocabulary may serve won in other
to aid us equally in the judgments we form as to original of the
creation. Hitherto we have always seen that the process of language.
naming the new follows as a result of a perception connecting it
with the already known, whether it be that we simply transfer
the name already existing to the new, or that we form from
it a compound or a derivative. In other words, a connexion of
cause and effect exists between the newly-named object and its
name, and the connecting link is an object named before. This
connexion of cause and effect is necessary, in the first instance,
in order that the name should be called into utterance by who-
ever employs it first, and in order that it should be understood
by others. Such relation of cause and effect only becomes super-
fluous by frequent repetition, as the mere external association is
gradually attached by a link strong enough for the purpose.
The conclusion that original creation, in order to have come
into being and been understood at all, needs such causal con-
nexion, is certainly not to be rejected. Now as there is no.
connecting link, we must look for a direct connexion between
object and name. But besides this, the possibility of under-
standing is originally rendered possible, just as in the process
of adding additional ideas to a word already in existence, by
M
tes
178
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the aid of the intuition given by the situation and by gesture-
language.
270. We have seen that nothing can, generally speaking,
become usual in language unless it is the spontaneous creation
of different individuals. It is a further indispensable condition
that it may be spontaneously created by the same individual at
different times, and this without co-operation on the part of the
memory. If, however, the same combination of sounds is found
connected with the same meaning, at different times and in
different individuals, then this connexion must necessarily be
conditioned by the same cause. And this cause must be rooted
in the nature of sound and of meaning, and not in any for-
tuitously accompanying circumstance. It may be conceded that
occasionally even a connexion created by a single individual on
a single occasion meets with general acceptance. But the possi-
bility of such occurrence is confined within definite limits. If
the first to name an object happen to be the discoverer or in-
ventor of that object, so that all others receive their knowledge
about it from him, then the name given by him enjoys a
certain authority. But such cases naturally occur but seldom.
The common acceptation of the name given can depend on
nothing but its appropriateness: in other words we have again
to deal with the intimate relation between sound and meaning,
which, in the absence of any apparent connecting link, must
depend on the impression made by the sound on the senses of
the hearer, and on the satisfaction accorded to the speaker by
the activity of the motor nerves indispensable to sound-produc-
tion.
271. Now if we closely scrutinise words which may fairly
be suspected of being comparatively new creations, we shall
find them in the main to denote different noises and movements;
ORIGINAL CREATION.
179
cf.: NHG. bambeln, bammeln, bummeln, bimmeln, batzen (NG. to ring
out), bauzen (=batzen- ' bark '), belfen, belfern, blaffen, blarren,
blerren, blatzen, platzen, pletzen, bletschen, pletschen, platschern,
planschen, panschen, plätschern, blodern, plaudern, blubbern, plappern,
blauzen, böller, bollern, bullern, ballern, boldern, poltern, bompern,
bumpern, buff, buffen, puff, puffen, burren, bubbeln, puppeln, puppern,
dudeln, fimmeln, fummeln, flattern, flinder, flindern, flinderling,
flandern, flink, flinken, flinkern, flirren, flarren, flarzen, flartschen,
flismen, flispern, flitter, flodern, flunkern, flüstern, gackeln, gackern,
gautsche, gautschen, glucken, glucksen, grackeln, hampeln, humpen,
humpeln, hätscheln, holpern, hurren, hussen, kabbeln, kichern, kirren,
kischen (zischen), klabastern, klachel or klächel (Bavarian = tongue
of a bell or other dangling object), klatschen, kletzen, kleschen (:
klatschen), klimpern, klirren, klunker, knabbeln, knabbern, knacken,
knacks, knarpeln, knarren, knarzen, knarschen, knirren, knirschen,
knurren, knascheln, knaspeln, knastern,knisten,knistern,knaster(-bart),
knatschen, knetschen, knitschen, knutschen, knattern, knittern, knuffen,
knüffeln, knüllen, knuppern, knuspern, kollern, kullern, krabbeln, krib-
beln, krakeln, kräkeln, kreischen, kuckern, (cucurire), lodern, lullen,
mucken, mucksen, munkeln, nutschen, pfuschen, pimpeln, pimpelig,
pinken, pladdern, plumpen, plumpsen, prasseln, prusten, quabbeln, quab-
belig, quackeln, quaken, quäken, quiken, quitschen, rappeln, rapsen
rascheln, rasseln, räuspern, rempeln, rummel, rumpeln, rüppeln, schlab-
bern, schlampen, schlampampen, schlockern, schlottern, schlürfen,
schmettern, schnack, schnacken, schrill, schummeln, schwabeln, schwap-
pen, stöhnen, stolpern, strullen, summen, surren, tatschen, tätschen,
tätscheln, ticken, torkeln, turzeln (Hessian = torkeln), tuten, wabbeln,
wibbeln,watscheln, wimmeln, wimmern, wudeln, ziepen, zirpen, zischen,
zischeln, zullen and zulpen (' to suck '), züsseln (schütteln), zwitschern.¹
1 [Cf. the numerous similar words cited as occurring in English, and other languages,
in Wedgwood's Introduction to the Dictionary of English Etymology.]
180
PRINCIPLES OF THE HISTORY OF LANGUAGE.
1
Some words denote at once a noise and an explosion, like klack,
klaff; others denote a noise and a stain, such as klacks, klecks,
klatsch. I have purposely confined myself to words not demon-
strably existent till late MHG. times. It would likewise be possible
to collect a rich list of words dating from the older German
dialects, which have nothing corresponding with them in the other
IE. languages, and the same holds good as to Greek and Latin.
We cannot resist the conclusion that, as far as our observations.
reach back, the proper domain of linguistic original creation is
to be found in words like these.
272. Because, in the case of words like these, we recognise an
intimate affinity between the sound and the signification, it does
not follow that they really owe their origin to such affinity.
There exist demonstrably many words which, owing to the
form or the meaning which they have acquired by secondary
development, convey the impression of being onomatopoetic. But
an examination of these words in their entirety excludes the pre-
sumption of the universal play of chance. One circumstance of
great weight has still to be considered: the frequent occurrence of
similar words of similar meaning, mostly differing in their vowels
alone, which still cannot, conformably to the laws of sound, be
deduced from a single original form. Thus we frequently find in
different languages words of this kind, resembling each other in
sound, but which the laws of sound forbid us from regarding as
related. [Cf. avy and the German auge.]
273. The onomatopoetic tendency explains, besides, certain
transformations of words already coined. One of the most in-
structive examples is the MHG. gouch=NHG. kukuk, with the transi-
tion forms guckauch, kuckuch, etc. [cf. our gowk and cuckoo]. These
formations also denote in some cases noises; in some cases restless
movements. Such transformations must not be confused with

ORIGINAL CREATION.
181
sound-change, and must be regarded, in some measure, as new
creations. Indeed, the words previously cited cannot be regarded
as absolutely new creations, as will be explained later. Strictly
speaking, the only absolutely new creations are interjections.
274. This is the place to dwell somewhat more fully upon the Interjections
nature of this part of speech. We cannot but be interested in the
question whether we are right in thinking that we see in interjec-
tions the most primitive utterances of linguistic activity; a theory
maintained by some, denied by others. We understand by inter-
jections involuntary reflex sounds, elicited merely by sudden
emotion, and without any design of communication. But we
must not therefore assume that they are genuine natural
sounds, arising by original necessity from sudden emotion, like
laughing and crying. The fact is that the interjections usually
employed are learned as truly by tradition as the other elements
of language. Association alone transfers them into reflex emotions.
This accounts for the fact that expressions for the same feeling
may result very differently in different languages and dialects, and
also in the case of different individuals speaking the same dialect.
The fact holds good in the most different languages that interjec-
tions owe their origin to other words and groups of words: cf. ach
gott, alle wetter, Gott sei dank, leider, [dear me! Heavens! etc.].
The origin may be so completely obscured by sound-changes as to
be past recognition, even by the aid of reflexion: cf. herrje! (herr
Jesus), jemine Jesu Domine) [Zounds, by Jeminy!]. Thus in deal-
ing with interjections which admit of no analysis, and which seem
perfectly simple, we are from the outset exposed to doubt as to
whether they did not come into being in the same way. But,
on the other hand, we meet among interjections of indisputably
recent origin, and fairly certain etymology, a considerable number
which seem to own affinity to no other words-or only to the
182
PRINCIPLES OF THE HISTORY OF LANGUAGE.
category cited above-and very probably owe their origin im-
mediately to reflexive emotion. Most of these, and those which
bear the strongest stamp of individuality as regards the form taken
by the sound and the tone expressive of the feeling, are produced in
response to sudden excitements of the sense of hearing or of sight.
This, at any rate, we must imagine to have been their origin.
They then come to be employed when we recollect and recount
the causes of such sudden excitement. I mean words like paff,
patsch, bardautz, perdauz, bauz, blauz, blaff, buff, puff, bums, futsch,
hurre, husch, hussa, klacks, klaps, kladderadatsch, knacks, plump,
plumps, ratsch, rutsch, schrumm, schwapp, wupp, etc.
275. Many of these words are substantives as well, or have
corresponding verbs; and in these cases it is often hard to say
which is the original of the two. This is, however, of no great
importance if only the words be apprehended as reactions against
excitement of the feelings. The onomatopoetic character of
such words comes out even more strongly in the case of the
duplication and triplication often employed, and especially when
the elements collected from different letters are differentiated by
ablaut; cf.: fickfack, gickgack, kliffklaff, klippklapp, klitschklatsch,
klimperklamper, kribbeskrabbes, krimskrams, mickmack, pinkepanke,
ripsraps, ritschratsch, schnickschnack, schnippschnapp (schnur), strip-
strap (strull), schwippschwapp, ticktack, lirumlarum, bimbambum,
piffpaffpuff; English, criddle-craddle, widdle-waddle, hankey-pankey,
ding-dong; French, clic-clac, cric-crac, drelin-drelon. These words.
are to some extent used as substantives as well, and indeed sub-
stantives are directly formed in this way, as kringelkrangel, tingel-
tangel [hurly-burly]; and further derivatives are formed from such
formations like fickfacken, fickfacker, wibbelwabbelig. Old language-
material, too, which has no true interjectional character apart from
this is often employed in the process; cf.: klingklang, singsang,

ORIGINAL CREATION.
183
hickhack, mischmasch, wirrwarr, zickzack. We may compare
also such onomatopoetic false formations as klinglingling (possibly
derived from klingklingkling), and hoppsasa. There are other
words due to the same impulse, which confine themselves, how-
ever, to the limits of regular language, such as combinations of
several words echoing the sound, and differing only in their
vowels, such as flimmen und flammen, flimmern und flammern,
kickezen und kackezen, klippen und klappen, klippern und klappern,
klistern und klastern, klitschern und klatschern, knistern und knas-
tern, knittern und knattern, krimmen und krammen, kritzen und
kratzen, gekritz und gekratz, rischeln und rascheln. All these
occur in the works of standard authors.
language.
276. Most words belonging to nursery language are onomato- Nursery
poetic, and reduplication plays an important part in these as
well; cf., in German, wauwau, putput, papa, mama, etc.; [bow-
wow, puff puff, etc.]. This language is not an invention of children,
but is handed down to them just like any other language. Its
value consists in the fact that it serves to aid the purpose of
the teacher. The more intimate relation of the sound to the
meaning which still exists therein, and is in any case constantly
being renewed, facilitates considerably the connexion of both.
This indeed goes so far that actually words of the language of
culture are to some extent learnt in the first instance composed
with words of nursery language; cf., wauwauhund, bäschaf, puthuhn
[moo-cow, baa-sheep, etc.].
277. There is, again, a notable difference between the original
creations whereby a language which has already arrived at a
state of culture is enriched, and those with which language
creation has universally begun. The former, where they are
not pure interjections, simply accommodate themselves to the
existing form system. They appear with the derivation and
184
PRINCIPLES OF THE HISTORY OF LANGUAGE.
The first
creations
no gram-
matical
category;
denoted
complete
intuitions,
flexion syllables common at the time when they were created.
For instance, assuming poltern to be a word of this sort, polt is
the only part due to original creation; ern is formed by analogy.
And thus in such a word as this we can recognise only a partial
original creation. We gather, moreover, from this example that
what we commonly abstract, as the root, from a word, need
certainly never have existed as an independent element, not
even in an older phonetic form; but immediately upon its
appearance may be provided with one or several suffixes, and
indeed must be so provided if the exigencies of the language
of the day require it.
278. Not merely are suffixes created according to the analogy
of the material of language at hand, but the function of words
as substantive, verb, etc., is determined by the same analogy,
and thus an element is imported into the new words which
does not depend upon original creation.
279. We cannot of course suppose that analogy co-operated
belonged to in this manner in the case of the first creations with which lan-
guage began. No trace of any grammatical category is seen in
them. They answer to entire conceptions. They are primitive
sentences of which we may form an idea from such sen-
tences as fire!-thieves! spoken in a single word. They are
thus, like these, strictly speaking, predicates, and a sense-impres-
sion forms their subject. For a human being to arrive at the
utterance of any such sentence, something definite must be
selected from the store of all that falls at the same time
into his perception. Now this selection cannot be made by
any logical operation, and must therefore be due to the outer
world. Something must come to pass which directs the atten-
tion to a definite goal. It is not the world of rest and silence,
but the world of movement and of noise which first comes home
ORIGINAL CREATION.
185
to man's consciousness, and to meet which he creates the first
A movement of his own body may also
sounds of language.
take the place of a movement of his environments, and serve
to direct his eyes to an unexpected sight. The impres-
sion will naturally be the more intense when joy or sorrow,
curiosity or fear, are awakened thereby. Hence the object
which awakes curiosity is represented by the linguistic utterance
no less than what is happening to the object. We approach
this primitive method of speech in exclamations of surprise and
in emotion. We may thus say of the oldest words that they
connect the imperfect expression of a conception, as it is at a
later period reproduced by a sentence, with an interjectional
character.
uttered with
of communi-
280. There is another respect in which the circumstances and were
attending new creations differ from those attending later ones, no thought
In the case of the latter the intention of communication may lend tion.
its aid from the very outset ; not so in the case of the for-
mer. We do not arrive at a point when we intentionally call
an activity into operation to carry out a particular purpose, until
we know by experience that this purpose can be thus attained;
and we gather this experience by seeing that the unintentional
activity, or indeed the activity employed with another object,
has had the success due. Before the creation of language man
is ignorant of the fact that he can communicate anything to
another by the aid of the sounds of language. This reason
would of itself suffice to justify us in rejecting any assumption
of an intentional invention. With regard to the first sounds of
language, we must abide by Steinthal's views that they are
nothing but reflex movements. As such, they satisfy merely a
need of the single individual without taking any account of his
1
1 Cf. his Ursprung der sprache and his Einleitung z. sprachwissenschaft.
186
PRINCIPLES OF THE HISTORY OF LANGUAGE.
living with others. But immediately upon the perception by other
individuals of a reflex-sound like this, together with the feeling
which caused its utterance, each can be set in relation to the
other. The fact that another individual feels this relation may
depend on the true connexion of cause and effect which exists.
between the feeling and the sound by the aid of the excitement
of the nerve. If the different individuals are in essential points
similarly organised, then the same impression of feeling will
produce in them almost the same reflex-sound, and when they
hear this from others they must feel themselves sympathetically
touched. But no doubt the number of the reflex-sounds thus
produced has been, comparatively speaking, small. Conceptions,
widely contrasting with each other, will have called into utter-
ance the same reflex-sound. It is therefore inconceivable that
such a sound, even assuming it to be repeatedly produced by
different individuals in the same way, can call into life the
memory-picture of a definite conception. The utmost that
it can achieve is to excite attention. It is only the con-
ception itself which gives more special details. The fact
that the attention of the other individuals is directed to the
same object which has elicited the reflex-sound in one or in
several may to some extent be due to the accompanying
gestures. We shall have to suppose, on the whole question,
that the language of sounds commenced by developing side by
side with gesture-language, and that its power to dispense
with this aid was a comparatively late acquisition, and grew
with its more perfect development. Gesture-language must of
course owe its origin to involuntary reflex-movement. This
origin is the more easily traceable in its case, as we can
watch it in a more primitive stage of development. If an indi-
vidual has repeatedly succeeded in attracting attention by a
ORIGINAL CREATION.
187
reflex-movement, whether such attention be attracted by his eyes,
his features, his hands, or by the organs of his speech, he is
gradually led to excite attention voluntarily, by the aid of such
movement, as soon as he is compelled by necessity to do so.
281. As soon as the possibility of intentional communication
is recognised, nothing prevents sounds, in whose production the
intention of communication has from the very first assisted, from
attaching themselves to those produced by involuntary reflex-
movement. We lay stress upon the words intention of com-
munication, because there is no intention of creating a lasting
instrument of communication. Any such intention remains.
excluded from original creation as from the regular development
of language. It is the necessity of the moment which produces
a new sound-group. But whether such sound-group disappears
with the first production, or whether it leaves behind a lasting
effect, this depends on its adaptability and on many fortuitous
circumstances.
282. We have yet to take notice of a difficulty which must Incapacity
of primitive
be surmounted before even the first rudiments of language can man to pro-
duce speech
be formed, a difficulty which appears to have been hitherto at will.
appreciated by no one. The original human being, who has as
yet not spoken at all, is as unable as a new-born babe to utter
at will any sound of speech. He has to learn such sound first:
in his case also it is only gradually, owing to manifold activity of
the organs of language, that a motory sensation associated with a
sound-formation can develop, which may then serve as a regulator
for his future speech. We must not therefore suppose that a sound-
group, as it has once been uttered by an individual, could
immediately be imitated by others. Indeed even the same
individual could not voluntarily repeat it. The problem is
harder for the original human being than for the children of our
188
PRINCIPLES OF THE HISTORY OF LANGUAGE.
ever anew.
time. The latter are, commonly speaking, surrounded by a number
of their fellow-beings in whom essentially identical motory
sensations have already developed. They hear, therefore, out of
the whole series of possible sounds a definite and limited number
Thereby a definite tendency is given from the very
outset, in the direction of which their own motory sensations
develop, and to which their attempts at language approach ever
more nearly. There is no fixed rule or authority for human
beings before the creation of language. It seems accordingly
that language must have begun with a confused utterance of the
most various articulations such as we never find combined in any
language. But how out of such a chaos could consistency in
motory sensation develop?
283. These considerations force us once more to the conclu-
sion that certain sound-groups must very frequently be produced
not merely from the same individuals, but from different ones
spontaneously; i.e. without the co-operation of any sort of imi-
tation, and to all intents contemporaneously. In the absence
of a rule for guidance, no motory sensation can take form,
except for such sound-groups as are favoured by their natural
conditions. Among these stand first and foremost the pure
reflex-sounds, and it is in connexion with these that the first
motory sensations will have developed themselves.
We can
hardly fancy the process to have been other than this: the
motory sensations connected with the single sounds must have
developed very slowly one after the other, and the traditional
language must have contented itself at its origin with a minimum.
of sound-signs, even though, during the process of development,
different sounds were on different occasions uttered by the different
individuals.
284. It is plain from what we have laid down that a long
ORIGINAL CREATION.
189
tion neces-
conception
period of linguistic activity must have preceded, before anything Reproduc-
came into being worthy to be called a language, in the sense sary to the
in which we speak, say, of the German and French languages. of language.
This holds good even if the language consist but of a couple of
words. What we have called original creation is of itself in-
sufficient to form a language. Its product must be stored up in
memory by individuals who belong to one linguistic community.
True language does not exist until speech and apprehension depend
upon reproduction.
285. Now if we regard this as sufficient for the recognition
of the existence of a language, no doubt we must ascribe language
to many beasts.
It will hardly be disputed that their calls,
whether of enticement or warning, are traditional and not spon-
taneous. They represent a stage of development which must
have been passed through by human language as well; the stage,
in fact, which we have endeavoured to depict. But yet another
step is.necessary before such a language can come into being as
we find nowadays common among the human race.
It is un-
doubtedly of great significance that the number of the traditional
words, and herewith the number of the differentiated conceptions,
is far greater among mankind than among any species of beasts;
but the strict characteristic which differentiates the language of
men from that of animals, existing language from a previous
linguistic stage, consists in something very different. This
decisive advance consists in the collocation of several words
into one sentence. Only thus does man receive the power to
free himself from simple intuition, and to pronounce judgment
on what is not before him.
CHAPTER X.
Possibility
of a regularly
observed
system of
ON ISOLATION AND THE REACTION AGAINST IT.
HE concentration of linguistic elements into groups depends
TH
upon each individual member of a linguistic community.
Thus these groups are in their nature thoroughly subjective. But
each linguis as the elements of which they are composed are within any definite
grouping in
tic period.
Varieties
in these
systems:
Isolation.
community practically identical, it follows that the formation of
groups by all the individuals who compose that community must be
similar, in virtue of the essential correspondence of their psychical
organisation. Thus, as we are in a position to describe the general
linguistic phenomena which characterise a definite period, we are
equally in a position to propose a system of grouping which shall
essentially hold good for each period of the development of a
language. This common material in the elements composing the
groups gives scientific observation the firm hold it has, while the
individual peculiarities of single speakers-to discard exceptions
which disappear in the great mass-escape observation.
287. Now if we compare our abstractions made at different
times as to the process of grouping, we become aware of con-
siderable differences-differences not merely caused by the loss of
some elements and the appearance of others; but also in cases
where the old elements have maintained themselves, they fall
1 I mean, of course, maintained themselves not in the strict sense, but as we commonly
apply the expression in the history of language. I have shown with sufficient clearness
how the process is to be apprehended in its essential reality.
ON ISOLATION AND THE REACTION AGAINST IT.
191
into different groups, according as a change has passed upon either
their sound-form or their meaning, or both. What at an earlier
period hung closely together now coheres but loosely—or indeed
not at all. We may properly give the first-named process the
name of isolation, since even the loosening of the tie is a partial
isolation. This expression, likewise, is of course due to our
inevitable employment of abstractions. Strictly speaking, it
would be wrong to say that what was before united has become
isolated; we might only maintain that what was in the minds of a
former generation united, is no longer so in the minds of a later
generation.
conditioned
correspon-
form and
288. The formation of groups depends upon the similarity or System
identity of the sound-form and of the meaning. This similarity or solely by
identity depends eventually, in by far the greatest number of cases, dence in
on etymological connexion. But the etymological connexion is not meaning.
to be considered absolutely decisive in producing the union, but
only in as far as it manifests itself at a particular time in a total or
partial identity of sound and meaning; and conversely, each cor-
respondence, as it may casually arise, has the same consequence.
Numerous errors of the old school of philology are due to the non-
recognition of this indisputable fact.
isolation.
289. We have to consider, in this Chapter, in the first place, the Causes of
loosening and dislocation of the groups. This is due to change in
sound and change in meaning; sometimes to analogical formation
as well. No doubt the latter operates, as we shall see further,
chiefly in promoting the reconstruction of the union, now broken;
but since different principles of analogy cancel each other, it may
have the contrary effect.
290. We have already seen in Chapter IV. that the different
significations of a word may tend to isolate themselves more and
more completely from each other. We have seen in the same
192
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Destruction
of Groups:
Chapter that a word regarded as the element of a firm syntactical
connexion may isolate itself as against its ordinary method of
application. In the same way the groups of words and word-
forms described in Chapter v. may be dislocated.
291. The etymologico-phonetic groups are deranged when, from
(1) Etymolo- whatever reason, the conditions which have caused the sound-
gico-phonetic
groups.
(2) Syntac-
tical groups.
change, and by whose guidance, aided by analogy, they have
proceeded, disappear. According to Verner's Law, in the original
German a regular change has set in between the hard and the soft
fricative (h-3, þ-x. ƒ-3, s-≈), and this change depends upon the posi-
tion of the accent according to the original IE. method of accentua-
tion. On the replacement of this method by the younger and speci-
cally German method of accentuation, there was no longer any
appreciable reason for the change, which accordingly could not but
appear a matter of perfect caprice. A general feeling might, no
doubt, grow up, that the sounds in question had a habit of
interchanging, but it was impossible for speakers to accommodate
themselves to the usage of language in any other way than by
mastering each several form specially. The sound-change had in
fact ceased to be a living force it was stiff and dead. In the
second place, it is possible for a more recent sound-change to act
destructively on this kind of groups. MHG. has, instead of the
original German change between hard and soft fricatives, the
change -g (sometimes ck), d-t, f-b (sometimes pp), s-r. Thus the
single method of change has split up into several quite different
ones, and any such split is always equivalent to a weakening.
But the real enemy of the etymological sound-groups is the
assimilating effect of the material-formal proportion-groups, as we
shall have to explain at another opportunity.
292. The cases of isolation which can occur in the department of
syntax have been partly described in Chapter VII. We have here,
ON ISOLATION AND THE REACTION AGAINST IT. 193
in the first instance, the cases of isolation of the different meanings
of a syntactic relation as against each other. By this process the
syntactical proportion-groups are not disarranged as long as each
single function of the relation remains perfectly alive. But each
case of deadening by customary connexion with a definite word
is a breaking off from the common proportional connexion.
Thus, for instance, we can hardly say that the combination zu dir
can stand in any analogous relation to the connexion of any other
preposition with the dative, far less that a more general function.
of the dative is placed in an analogical relation to it by the
operation of the instinct of language. This connexion, however,
still keeps its place within the limits of a narrower proportion-
group, and one in which the same member passes through all
the proportions: thus, zu: dir=zu: dem vater=zu: allen, etc.
293. In this case it is possible to vary at will the word in
which the syntactical relation is formally expressed. There is
still another kind of isolation, in which this word is fixed, while the
other, in which the relation finds no expression, may change at will.
This isolation consists in the fact that methods of construction
commonly disappear, but maintain themselves in certain survivals
which have left their impress specially strong owing to their frequent
use, so that they need no aid from analogous proportions, and
hence are enabled to survive, even after the disappearance of these.
294. Thus in MHG. there exist several functions of the
genitive once fully alive, but at the present day confined to
the genitives of some few words, which now stand entirely by
themselves, or crystallise into comparatively small groups, capable
of only a very small analogical extension, or indeed of none at all.
To indicate a point of time (if we disregard the isolated formulæ
derzeit, jener zeit, dieser tage), nächster tage, no case can be employed
except the genitive singular of masculine and neuter substantives.
N
194
PRINCIPLES OF THE HISTORY OF LANGUAGE.
1
German admits of such phrases as des morgens, eines morgens, abends,
tages, jahres; but not of saying der stunde, einer stunde, etc., and
further, not of des monats. Moreover, these genitives cannot be con-
nected with any adjective at will; they appear in standing formulae
alone, such as eines schönen tages, morgens, &c. The function of
denoting the point of time is in this case bound up, not with the
genitive as such, but with the suffix (e)s, whose original identity
with the genitival suffix is scarcely felt any more at the present
day. We see this even more distinctly in the case of forms
without an article such as abends, morgens, tags; particularly in
the archaic form (des) nachts, which is now differentiated in
sound likewise from the form which performs the function of
the regular genitive. Certain genitives which denote a relation of
space are even more isolated than such definitions of time; cf., des
weges, gerades weges, rechter hand, linker hand, allerorten, aller
wegen. Further, there are certain causal genitives, such as hungers
sterben, todes verblichen; and again, such formulæ as der hoffnung
and des glaubens leben, if these are not to be taken in a different
way. Those which express a modal relation are more numerous
and not less isolated. Different applications have to be dis-
tinguished in this case. One group of related genitives is used
predicatively. German admits of ich bin der ansicht, meinung,
hoffnung, zuversicht, des sinnes, des glaubens; willens is similarly
used without any article, as also anderer ansicht, guter hoffnung ;
compare also such expressions as er ging fort, der meinung, dass
etc. The expressions guten mutes, guter dinge are somewhat
different [cf., be of good cheer]. The expressions reinen sinnes,
göttlicher natur, etc., seem already antiquated. The following are
immediately joined to the substantive like an adjective, and they
appear to be no more felt as genitives, allerhand, mancherhand,
einerhand, keinerhand, allerlei, aller art, etc. The expression is
ON ISOLATION AND THE REACTION AGAINST IT. 195
also common, es ist einerlei. Other formulæ, again, are adverbially
joined to the verb, as meines bedünkens, meines erachtens, alles
ernstes, stehenden fusses, eilenden schrittes, kurser hand, leichten
kaufes, unverrichteter sache, vorsichtiger weise, törichter w., vernünf-
tiger w., etc., vorkommenden falls, besten f., keines f., etc., keineswegs,
einigermassen, gewissermassen, etc., dergestalt, solchergestalt. Some
of these formulæ, as the ordinary way of writing them shows, arc
simply regarded as adverbs. The same holds good of flugs, sporn-
streichs, augenblicks, teils, grössten teils, etc., and of the following
adjectival derivatives, anders, rechts, links, stets, straks, bereits,
besonders, blindlings, etc.
295. The formula es sei denn dass is a survival of an old way of
construction, usual in MHG.; cf. Genesis xxxii. 26, ich lasse dich
nicht, du segnest mich denn; the same was even more common in
MHG. with the negative en and also without denne. We have an
unrecognisable survival of this in the adverb nur enwaere.
296. The process of isolation may finally proceed even further,
since none of the members connected is able to change freely,
so that then each single formula is passed on by memory alone
without producing any new combination.
297. It is no longer possible in MHG. to connect prepositions
with a singular substantive without prefixing the article. For
instance, it is not permissible to say an hause, vor tür, zu see, etc.,
but only am hause, vor der tür, zur see. But within certain definite
limits it is still possible to create combinations without any article;
e.g., vor liebe, besorgniss, kummer, etc. [to explain the preventing
cause: thus, Ich konnte vor kummer nicht weiter lesen]; auf ehre,
gewinn, weisheit, geld gerichtet (thus auf may be connected with
every abstract or collective to denote the goal of the effort): su
gelde, weine, wasser werden, machen, and thus with all collectives,
but die arbeit wird ihm zur erholung, zum genuss, der knabe wird
196
PRINCIPLES OF THE HISTORY OF LANGUAGE.
zum mann, das mädchen zur frau. Other combinations, however,
belong no longer to any creative group; and nothing perfectly
analogous to them can be created. The fórmula with zu are
probably the most common, as zu hause¹ (but not zu dorfe, zu stadt),
zu wasser, zu lande (this last is only used in contrast to the former,
and is no longer employed like the MHG. ze lande, analogously to
zu hause), zu schiffe, wagen, fusse, pferde, zu anfang, ende, zu tische,
bette, markte, zu leide, liebe, gute, zurück, zurecht, zunichte; other
usages are now limited to the connexion with definite verbs,
while in older NHG. in many cases a more free usage prevails: zu
grunde gehen, zu rande sein mit etwas, zu berge stehen, zu kopfe
steigen, mir ist zu mute, zu sinne, einem zu gemüte führen, zu
schaden kommen (but zum schaden gereichen), zu tode kommen,
quälen, zu statten kommen, zu wege bringen, zu gesichte kommen,
einem etwas zu danke machen, einem zu willen sein, zu rate gehen,
halten, zu abend, zu nacht, zu mittag speisen, zu tage bringen, fördern,
but not zu tage in the sense of 'on the day' or 'on this day,' though
it is so used in heut zu tage. Such parallel connexions as zu nutz
und frommen are noteworthy, though zum frommen and zum nutzen
are correct; not to speak of the phrase sich etwas zu nutze machen;
zu spiel und tanz, but zum spiel, zum tanz; in freud und leid, but
in der freude, im leide; in krieg und frieden, but im kriege, im
frieden (in frieden has a different meaning); in, or durch feld und
wald, but im felde, im walde, durch das feld, durch den wald; in
dorf und stadt, but im dorfe, in der stadt, etc.
298. Another example under this head is the following:-In
MHG. the adjective may still be used, when placed attributively,
especially after the indefinite article, in the nominative singular of all
genders, and in the accusative singular neuter, in the so-called unin-
1 In several of these formulæ, zu is still used to denote rest in one place, a use only
possible in quitè definite combinations.
ON ISOLATION AND THE REACTION AGAINST IT.
197
flected form, as ein guot (schane) man, frouwe, kint. On the other
hand, in NHG. we can only employ the inflected form: ein guter mann,
eine gute frau, ein gutes kind. But the old method of composition has
left numerous traces in the improper compounds which have arisen
through the growth of an adjective into a substantive, as altmeister,
bösewicht, kurzweil, Neumann, Schönbrunn, etc. And further, the
uninflected form still appears in certain standing combinations, as
gut wetter, schlecht w., ander w., ein gut stück, ein gut teil, ein ander
mal, manch mal, ein ander bild (ander occurs in this use down to
the eighteenth century), gut ding will weile haben. Such phrases
as jung Roland, schön Suschen, lieb mütterchen, are antiquated.
299. Isolated survivals are seen in zweifelsohne (in MHG. dne
can be suffixed to any given genitive), mutterseelenallein (in MHG.
aleine is commonly used with the genitive in the sense of 'deprived
of'), vergissmeinnicht (vergessen was once ordinarily constructed
with the genitive), dass es gott erbarme (MHG. mich erbarmet ein
dinc='something grieves me').
and material
change of
300. The instances of syntactic isolation are, to some extent, at (3) Formal
the same time isolations in the department of formal grouping, for groups (a) by
this too depends in great measure on the syntactical function; cf. meaning
especially the genitives cited above. But the formal isolation again
stands in close relation to the isolation of the material element, as far
as the latter is a result of change of meaning. A separation of
etymologically united forms is avoided as long as the development
of meaning of the single forms moves in parallel lines. This will
be found to be more and more true the more they are referred
anew to each other. The relation, however, is the most strongly
felt, when it does not merely hand on each by itself by an effort of
memory, but when one is perpetually being created after the other,
on extraneous analogies. Seeing that in the case of every new
creation of any form, two groups, a material and a formal, co-
198
PRINCIPLES OF THE HISTORY OF LANGUAGE.
operate, both of these reciprocally suit themselves to the conditions
of their creative force. A formal isolation is almost always at the
same time a material one as well. As soon as rechts ceases to be
felt as a genitive, it ceases to stand in the same intimate relation
to the nominative recht. Kunst does not stand in the same intimate
relationship with können as führung with führen; for -ung is still
a living suffix, and by its aid we can always form new substantives
from verbs not so with -st. Indeed we may assert that regierung
in the sense of 'a governing body,' mischung in that of 'a mixture,'
kleidung in that of 'a way to clothe,' etc., do not stand in such inti-
mate relation with the verbs that answer to them as regierung in
the sense of the act of ruling,' etc. For the living function of the
suffix -ung is confined to the designation of an activity; and in this
a substantive can be placed by the side of every transitive verb.
(
301. The groups which depend upon inflexion hang naturally
closer together than those which depend on word-formation. On
the one hand the measure of the common element is greater; on
the other the feeling for the method of the formation is most vivid.
In this respect the position of the nominal forms of the verb is
characteristic. As long as these are used as real nouns, the
infinitive provided with the article, the participle employed to
denote a lasting characteristic, the connexion with the remaining
verbal forms is loosened, and thereby the possibility is created for
a further departure in the meaning.
302. An enlargement of the meaning of the root-word, or of
the word which commends itself as such to the instinct of lan-
guage, communicates itself more readily to the derivation, than
an enlargement of the derivation communicates itself to the root-
word. In fact, while in the case of the derivation we remember
the root-word more readily than in the opposite case, we connect
the derivation more readily with all the meanings of the root-word
ON ISOLATION AND THE REACTION AGAINST IT. 199
than we connect the root-word with all the meanings of the deriva-
tion. And hence the impulse to isolation commonly proceeds from
a change in the meaning of the derivation. The relation of the
simple to the compound is as that of the root-word to the derivation.
303. The reason for the discrepancy in the development of
meaning of words etymologically connected is, as far as it is not
conditioned by some other process of isolation, to be sought for in
the originally existing difference in their function. A noun may
develop itself in directions in which a verb cannot follow it. It is
only the proper nouns of the agent and nouns of action which
stand in true correspondence with the verb. As soon as the noun
of the agent has passed into the denomination of a permanent
quality or of the owner of a permanent quality, and the noun of
action has come to denote a standing condition, or a product, or a
tool, it is possible for an enlarged meaning to attach thereto, such
as cannot attach itself to a verb. Thus the NHG. ritter is the noun
of the agent connected with reiten, but at a later period comes merely
to signify a man who commonly pursues riding as a profession.
Up to this stage it is closely connected with the verb. But as the
word is specially applied to mounted warriors, and as a privileged
class or order developed from these, it passed into a meaning
which has no verbal meaning to correspond to it. And thus it has
proceeded to adopt a sense which has nothing to do with the
original one. Thus, many developments of meaning are possible
to the adverb which are impossible to the adjective. Take as an
example the common adverbs, whether strengthening or limiting,
such as NHG. sehr MHG. sêre, derived from an adjective ser
('wounded'), OHG. harto and drato (valde') from the adjectives
herti ('hard') and drâti (swift'), NHG. in the language of common
intercourse schrecklich, furchtbar, entsetzlich, fast as related to fest;
and also such as schon as against schön.
200
PRINCIPLES OF THE HISTORY OF LANGUAGE.
(b) By change
of sound.
304. The etymological groups, and the forms which correspond
in sound, and thus the proportion-groups to be compounded of
both, undergo operations due to sound-change, which either notably
limit the power of cohesion or entirely cancel it. A quantity of
purposeless differences are produced thereby. For the most com-
mon conditions of sound-change prove that a sound very rarely
undergoes consistent changes each time that it appears in language.
Even such a spontaneous sound-change as the original German
sound-shifting process has found certain opposing barriers which
have prevented a systematic and regular carrying out of the
process; for instance, in the case of the combinations sk, st, sp, the
shifting-process has not operated. A much stronger disposition to
the differentiation of sounds originally identical is given where the
change is conditioned by the surrounding sounds or by the system
of accentuation. There thus arise, almost in the case of each
sound-change, purposeless differences between the different deriva-
tions from the same root, and between the different inflexion-
forms of the same word; (cf. e.g. Greek στίζω-στίξω-στικτός
Oтíyμa; NHG. sitze—sass, heiss-heitze—hitze; schneide—schnitt ;
friere-frost, etc.); the same derivative and flexional suffixes sepa-
rate into different forms; (cf. e.g. the different shapes taken by the
IE. suffix -ti- in Lat. hostis, messis, pars, in Goth. ansts—gabaurps
-qiss; the different treatment of the nominatival termination in
ON. sonr―steinn (from *steinr)-heill-îss-fugl (from *fuglr, etc.):
nay, the same word takes different forms according to its position
in the sentence; (cf. the different forms of Greek prepositions
such as έν—ἐμέγ, συν-συμ-συγ). An unnecessary burden to
the memory for future generations is the result. At the same
time the inevitable consequence is that the single forms, owing to
the diminished amount of sound-correspondence, draw now. less
easily and less firmly into groups. The result of this is that a
ON ISOLATION AND THE REACTION AGAINST IT. 201
change of meaning transfers itself too easily from one related
word to another. Thus the cancelling of the agreement in the
shape taken by the sounds favours the cancelling of the agreement
in meaning.
305. The disappearance and death of living modes of forma-
tion dates, generally speaking, from an isolation of sound which
is frequently material as well as formal; the isolation of meaning is
a later matter. We may, for instance, assume in German a period
in which possibly a weak causative might have been formed from
every intransitive strong verb. This was from the IE. period
differentiated in the root-vowel from the present of the root-word,
but as it agreed with the singular indicative preterite (brinna—brann
-brannjan, etc.), a new sound-relation was maintained.
But a
separation had already set in in the original German period, owing
to the operation of Verner's Law, in consequence of which in
many cases a consonantal deviation of the causative—not merely
from the present, but from the singular preterite of the root-word,
ensued. This deviation entails in OHG. further deviations in
the vowels.
The causative, deviating in this from the singular
preterite, takes the umlaut where possible. Thus, in MHG. relations
appear like springen—spranc—sprengen, varen—vuor—vüeren, sîhen
-séch-seigen, ziehen-zôch-zöugen, genesen-genas-neren. Under
such circumstances it was natural that the root-word and the
causative should now proceed on their respective ways, so that in
the case of NHG. no connexion is any longer felt between genesen
and nähren. But the sound-changes mentioned have the effect
of attacking the uniformity of the method of formation, and there-
by the connexion of the causative verbs suffers on the side of
meaning as well, and is finally wholly destroyed.
306. The disappearance of the IE. derivative suffixes in the
Teutonic languages is due, in the first instance, mainly to a
202
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Reaction by
aid of Unifi-
cation.
sound-change.
Thus, for instance, the t of the suffixes -tei, -teu,
-to, etc., appears, after the Lautverschiebung, in five different forms :
t (GOTH. paurfts, 'need,' connected with paurban; gaskafts, ' crea-
tion,' with skapjan; mahts, 'might,' with magan; fravaurhts, to
pass away,' with vaurkjan); p (gaqumps 'meeting,' with qiman ;
gabaurps, ' birth,' with bairan); d (-deds, 'deed,' with O. SAXON dôn ;
gamunds, 'memory,' with munan); st (ansts, 'pardon,' with unnan) ;
alabrunsts, ‘burnt-offering,' with brinnan); s (-qis-s, 'speech,' with
qipan; -stass,' step,' with standan; gaviss, 'connexion,' with gavi-
dan). Of course there can be no such thing as a consciousness
of the original identity of the different shapes taken by these
sounds. The great group separates into five smaller groups. Of
the five suffixes none can be universally employed. Further, the
connexion with the root-word is frequently loosened by changes
in the auslaut of the root, as in the examples given. Hence
the inevitable consequence has been that the old suffixes were
deprived of their capacity of again serving for the creation of new
words; that henceforward only the old formations were trans-
mitted by memory, though only so far as through constant
use they were able to dispense with the support of the root-word.
Thus, again, the suffix no has died out, because it had in many
cases become unrecognisable in consequence of the assimilation
of the n to the preceding consonants; cf. fulls = IE. plnos, etc.
307. Thus the symmetry of any system of forms meets in
sound-change an incessant and aggressive foe. It is hard to
realise how disconnected, confused, and unintelligible language
would gradually become if it had patiently to endure all the
devastations of sound-change without any possibility of reaction.
Such reaction is, however, rendered possible by analogical for-
mation. By the aid of analogy language gradually works its
way into more satisfactory circumstances, to a firmer cohesion
ON ISOLATION AND THE REACTION AGAINST IT. 203
and a more convenient method of grouping in inflexion and
word-formation. And thus it is that we see in the history of
language a perpetual oscillation between two opposite tendencies.
Each disorganisation is followed by a reorganisation. The more
violent the attack to which the groups are exposed by sound-
change, the more vigorous is the activity in new creation.
308. Where an unnecessary and inconvenient difference, due
to sound-change, has arisen, it may be cancelled by the aid of
analogy; a form marked by such difference may be gradually
thrust aside by a new creation which does not contain the dif-
ference in question. We may call this process unification; only
we must clearly recognise that this term does not satisfactorily
describe the actual course of procedure, which, strictly speaking,
consists in a complicated system of single processes as described
in Chapter V.
309. Unification is hindered by the material-sound propor-
tions. A sound-change yet living and supported by such pro-
portions resists unification for a long period, without, however,
setting any insurmountable obstacle to eventual unification. If
the material-sound proportions are once broken through, the sound-
change loses much of its power of resistance.
•
of differences
from differ-
ence of posi-
tion in the
sentence.
310. We will now endeavour to observe the different modes Unification
of unification a little more closely. Where one and the same which result
form has parted into several different forms, owing to the differ-
ence of their position in the sentence, the original difference in
the employment of these forms is lost when the one form is used
also in that position in the sentence in which phonetic develop-
ment has led to the production of the others.
311. G. Curtius has shown, in his Studien, x. 205 sqq., that
the final sound of the Greek preposition, as well as that of the
accusative singular of the article, depended at a former linguistic date.
1
204
PRINCIPLES OF THE HISTORY OF LANGUAGE.
on the opening sound of the following word; cf. κàd dè — KàK
κεφαλὴν—καγ γόνυ-κάπ πεδίον—κὰν νόμον—κὰμ μὲν—κὰρ ῥόον
κάλ λαπάρην, τομ βέλτιστον—τὸν κράτιστον-τὸν θρασύτατον—
Tòλ Aŵσtov, etc., while at a later date a single form only of these
manifold varieties, or the adverbial form,¹ which differs again
from it, became the regular form in use.²
312. In the Teutonic languages we find the following process
repeating itself at various periods. Words used as adverbs and
as prepositions indifferently, as they receive the full stress or
are mere enclitics, and as, in their position of enclitics, they receive
a collateral stress, or have none at all, separate into two different
forms or, it may be, more, whose original difference of function
however is not always maintained, since one form intrudes into
the place of the other; cf. Beiträge z. Gesch. d. deutschen Sprache,
vi. 144, 191 sqq., 199 sqq., 207 sqq., 248 sqq., 1372. To cite one
example only. The original Teutonic tô (to'), in cases where
it received the full stress and remained unweakened in adverbial
use, was shortened as a proclitic to *to. From this the forms.
in OHG. za, ze, zi arise under the influence of accentual differences.
These forms are employed in some of the oldest records indif-
ferently side by side; in more recent times one takes firm root
in each dialect. All three appear in MHG. as ze. But side by
side with this appears the form zuo regularly developed as a
preposition, and in NHG. this form alone holds the field. The
same holds good of the forms of the pronouns and the article;
cf. Beiträge, vi. 1372, 144 sqq.
313. In the period of the transition from OHG. to MHG. r in
1 For as such we must regard ἀνά, κατά, παρά in contrast to άν, κατ, παρ, with their
lifferent collateral forms ; in the same way ἐνί, περί, ποτί, προτί, as against ev, περ, ποτ
οι πος, προτ οι προς.
2 How far this applies merely to writing, and how far to actual usage in language,
› remains still doubtful in many cases.
ON ISOLATION AND THE REACTION AGAINST IT. 205
auslaut after a long vowel disappears in dâ from dâr, hie from hier,
etc., though it survives, and is maintained in close connexion with
a following word, because it is then drawn over to the following
syllable, thus dar an, hier an, etc. In NHG. hier often takes the
place of hie, and eventually completely supplants it in the written
language, to say nothing of the phrase hie und da. Conversely in
MHG. we also find the combinations hie inne, hie ûze, and the con-
tracted forms hinne, hûze, which remain South German forms to
this day.
314. The process of differentiation and of unification may repeat
itself several times in succession. In OHG. the form ana parted
into ana the adverb, and an the preposition; the first form then
displaced the second. In MHG. ana again parted into ane and
an, and the second then displaced the first. Aba (ab) has gone
through a similar development.
315. The influence of the component parts of the sentence
upon the sound-development is explained, as we have seen, by
the fact that a word-group, like a single word, is apprehended
as a unity, which does not require to be previously analysed by the
hearer into its several parts, nor to be made up by the speaker
from the elements which compose it. Thus the circumstances are
the same as in a compound; and indeed there is no sharp line, as
will be afterwards more fully explained, to be drawn between a
compound and a word-group. Least of all is it to be supposed
that a distinction originally existed between the connexion of
the preposition with a noun, and that with a verb. Thus, in the
case we have chosen, a newly created method of composition takes.
the place of the traditional configuration of the group.
316. There are two different ways of development possible in
this process. Either one form may encroach on the function of
the others, or the encroachment may be reciprocal. The latter case
206
PRINCIPLES OF THE HISTORY OF LANGUAGE.
will naturally occur when the different forms occur with fairly
equal frequency; the former, if one form occurs more frequently
than the other. As a result, in both cases in the first instance,
double, or it may be triple, forms run parallel to each other for a
long period; in the one case, however, only with restricted limits,
while elsewhere uniformity is maintained; in the other case they
run parallel with unlimited application. A general uniformity
results in the course of a further development, owing to the
disappearance of one form. In cases where plurality of form
in one department is met by uniformity in the other, it cannot,
of course, be long doubtful which form will eventually triumph.
But where plurality of form has once become general, the forces
are not so unevenly distributed, the struggle is not so easily de-
cided; the issue depends upon chance circumstances not always
appreciable by us. The more dissimilar the relation, the shorte
is the struggle, and the earlier the commencement of the attack.
317. The severance of one form into several different forms
may take place in such a way that a change occurs under every
circumstance; but it may also take place in such a way that the
root-form remains preserved side by side with a single or several
changed forms.
In the last mentioned case, in the course of further
development, the root-form, as such, has no advantage over the
derived forms, for it is not recognised as such. No doubt, how-
ever, this form has an advantage over the others in which the
word appears, if it is independent of any influence from the syntax
of the sentence, whether it be the root-form or not. A French-
man, who has no scientific knowledge of his mother-tongue, hears
with surprise that the phrase a-t-il represents a more original form
than ila, and that in un ami the n represents a more original pro-
nunciation than in un fils. If he thinks about the matter at all,
he will be rather inclined to take the t for an insertion, and to
ON ISOLATION AND THE REACTION AGAINST IT. 207
assume the pronunciation of the n in un ami to be a departuré
from the regular pronunciation.
318. These remarks are, mutatis mutandis, equally applicable
to every other kind of unification due to formation by analogy.
of phoneti-
entiated
belong to the
319. Essentially similar is the unification of forms phonetically Unification
differentiated which proceed from the same stem, or of words cally differ-
formed from the same root. We may call this material in con- forms which
trast to formal unification, the latter being concerned with the
corresponding formations from different roots, or different systems
of formation of inflexions or word-composition. It often happens,
however, that material unification is at the same time formal.
320. Examples of this are not far to seek. The most in-
structive are certain systematic differentiations which make their
appearance at a very early period. The following generations
are often occupied with the reaction against these for many
centuries after, during which one case after another falls a victim.
to unification, though eventually there still remain not unfrequently
certain traces of differentiation. The development is more mani-
fold and instructive as well, if, after the commencement of the
differentiation of sounds, the language has parted into many
different dialectic shapes. The example of this kind known to
me as occurring on the largest scale is the graduation of vowels
in the IE. original language, the traces of which the surviving
languages are still endeavouring to efface. In the Teutonic
domain the operations of Verner's Law are the most striking,
according to which, in the original German, the hard fricatives
h, þ, f, s have maintained themselves after an originally accented
syllable, but after an originally unaccented syllable have passed
into the corresponding soft fricatives (Gothic g, d, b, z). The
impulse created thereby is specially adapted for methodic study,
especially as we find ourselves here upon safe and generally
same stem.
1
208
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Inequalities
of unification
owing to
favouring
and hostile
conditions.
recognised ground. The philologist who has been at the pains
to follow the instances of reaction against any such sound-law
into every detail, cannot conceivably harbour such perverse asser-
tions and objections as to formation by analogy as are unfortu-
nately so widely in vogue. And as with one phonetic law, so is it
with the rest. There is in fact no phonetic law which, having once
in a number of cases differentiated in sound what holds closely
together in etymology, fails to produce a reaction against this
differentiation, unless in the case where the sound-change left
be continuously supported by analogy (see p. 107). This must
be recognised as a fundamental maxim of the historical investi-
gation of language. We may examine every language whose
development admits of a continuous investigation to find a single
sound-law which has failed to show, as a result, some centuries
after it has operated, some kind of reaction. However great the
prize offered for such discovery, no one would win it.
321. Whoever has followed such a development in all its con-
nexions will not, as has been recently done, claim by an explana-
tion of forms, resting upon the supposition of unifications, that
the unification in all the forms affected by the sound-law must
have come about uniformly, and in the same direction. This
would be to demand a development directly contradicting the
experience which we can abstract from the facts which we have
actually to observe. Besides, any such demand depends on a
manifest confusion of ideas. In the case of sound-change we
certainly must demand that it should make its appearance in the
same way in every case in which the same phonetic conditions are
present. But identity or non-identity of the phonetic circum-
stances has nothing to do with unification. Either does such
group, connected by material relationship, proceed on its own
course of development, or, if several such groups operate on each
ON ISOLATION AND THE-REACTION AGAINST IT. 209
other, it happens owing to the fact that formal unification is simul-
taneously at work; but the fact of being affected by the same
sound-law does not in itself yield any reason for a reciprocal
influence in the unification. On the other hand, there are many
circumstances, serving either to foster or to check the process,
which cause it to have very different results in different cases.
322. To these must be added as well a phonetic impulse.
Those forms which are differentiated by the operation of several
sound-laws are less favourable to unification than those in which
only one of them has operated towards differentiation.
=
=
323. The well-known NHG. lengthening of the vowels never,
except in certain definite connexions, appears before double
consonants, in which case, on the contrary, a syllable origi-
nally long is actually shortened (cf., brachte MHG. brâhte,
acht MHG. ahte, etc.). In accordance with this, the second and
third persons singular and the second person plural of the
present indicative are shortened in cases where the terminal
vowel is syncopated, and this holds good of cases where the
other forms of the present have allowed the lengthening to set
in. But in by far the greatest number of cases unification has
set in; this is regularly the case in the weak verb (e.g., lebe-lebst,
lebt), where the quality of the vowel was always uniformly
identical from the earliest times; and further, in the strong
verbs with radical a: trage-trägst, trägt (in Low German with
shortened vowel, dröchst, dröcht). On the other hand, the short
syllable of the second and third persons singular is maintained
in those verbs in which the root-vowel has fluctuated from
the earliest times between e and i: in nehme-nimmst, nimmt,
trete-trittst, tritt, this change is general; and it appears like-
wise in lese-list, gebe-gibst, gibt; at least according to the
usual pronunciation in Lower Germany. The reason for the more
210 PRINCIPLES OF THE HISTORY OF LANGUAGE,
་
¿
successful resistance opposed by these verbs than the others to
the unification affecting the quantity must certainly be sought
in the fact that the quality is at the same time different. This
assumption is confirmed by the fact that they have not, in the
case of the second person plural, withdrawn from the influence
of unification. The difference between a and ä is not felt in this
way, because the umlaut seems natural to our instinct of language.
324. In OHG. the participles of the verbs lesan, ginesan, uuesan
had, according to Verner's Law, to be sounded gileran, gineran,
giuueran; but, with the exception of a few traces in the oldest
records, the forms gilesan, ginesan, giuuesan—approaching the
present form-actually appear. On the other hand, the parti-
ciples of kiesen, friesen, verliesen appear even in NHG., maintaining
the change gekoren, gefroren, verloren. The identity of the vowel
in the one case, and the difference in the last, serves as a pre-
cedent for the condition of the consonants.
325. The strong verbs which have the same vowel in the
singular and plural of the preterite, likewise cancelled at an early
period the consonantal difference which rose according to Verner's
Law; cf. OHG., sluog-sluogun, hieng-hiengun, huob-huobun, hluod
-hluodun against zôh-zugun, meid-mitun. [Cƒ. OE. slôg, 'I
struck'-slôgon,
-slôgon, with têah, 'I drew'—tugon]. We can see how
in this way even such forms as are not merely affected by the
same phonetic law, but are connected by their function and by
their method of formation in other respects, are differently treated.
326. This phenomenon demands a psychological explanation.
We should be inclined at first sight to suppose that since what
we call unification proceeds by analogy from a new creation,
the phonetic configuration of the form expelled by the new
creation did not come into consideration at all. If the picture
of the traditional phonetically differentiated form comes into con-
ON ISOLATION AND THE REACTION AGAINST IT. 211
sciousness, no new creation is possible; if, however, it does not
so come into consciousness, the new creation is given free play.
But no real reason appears why one form should more easily
come into consciousness because, phonetically speaking, it differs
more strongly than another form from a related one. The diffi-
culty can only be explained by the assumption of the co-operation
of simple reproduction by the aid of memory and creative com-
bination in the creation of new forms, just as we were compelled
to recognise this in the daily utterance of the forms already usual
in language. There is a state in which the picture of the tradi-
tional form is not powerful enough to enter into consciousness
under any circumstances more easily than a new creation due
to analogy; but still not so weak as to retire without opposition
before such new creation. There thus appear in opposition to
each other two ideas contending as to whether of the two shall
first enter into consciousness, and thus drive the other back. It
is only where a relation like this exists that the degree of
divergence between the traditional form and the eventual new
creation comes into consideration. If the latter is on the point
of pressing first to the front, the former may yet, even without
being plainly apprehended in consciousness, exercise a retarding
influence upon it, which, in the case of the latter, the linguistic
sense prevents from acquiring the instinctive sureness requisite,
and thus prompts to reflexion upon it. The idea of the tradi-
tional form, however, co-operates as a stronger retarding agent,
the more widely its contents differ from the new combination.
And it is with the hearer as with the speaker. A new creation
operates the more bewilderingly upon him, and is the less readily
accepted and imitated, the more many-sided the opposition which
it presents to the traditional form, so far as the memory of it is
still in some degree operative in his mind.
212
PRINCIPLES OF THE HISTORY OF LANGUAGE.
/
(2.) Varying
cohesion of
logical
groups.
327. There are two other circumstances which play a much
more important part in the process of furthering or impeding
unification than phonetic difference, namely, the greater or less
stability of the connexion of the etymological groups and the
greater or less intensity with which the single forms are stamped
on the memory.
328. The first depends on the degree of correspondence in
the Etymo- the meaning, and on the degree of living capacity for formation
in the several forms. Both stand, as we have already seen, in
reciprocal relations to each other. The more or less intimate
nature of the connexion may be at once given with the mere
function of the forms; as, for instance, the forms of the present
are more closely connected with each other than with those
of the preterite, the forms of the same word more closely con
nected with each other than with the forms of words derived
from the same root. The tie between them may, however, be
weakened by secondary development. Every kind of isolation
which affects the function increases the difficulty of reaction against
the isolation by which the sounds are affected, and renders it
actually impossible as soon as it has reached a certain degree.
329. A few examples may serve to illustrate these sentences
The numerous consonantal differentiations, which took their rise
by the operation of Verner's Law, are completely obliterated in
noun inflexions, as seen in the earliest records which have come
down to us. We see, however, their traces in many double
forms which remain parallel to each other without any difference
of meaning. In the verb, however, the differentiation has main-
tained itself better, obviously aided by the differentiation of vowels
(ablaut) which coincides with it; cf. MHG. ziuhe-zôch―zugen-
gezogen. We are able now in several ways to observe how the
process of unification which sets in later begins by cancelling
Į
ON ISOLATION AND THE REACTION AGAINST IT.
213
he difference between the singular and plural of the preterite,
and cancelling it so as to differentiate the singular by this pro-
cess for the first time from the present. This has happened in
he case of the West Teutonic languages in almost all those
cases in which no variation of the vowels interposed to retard
the process; thus OHG. slahu-slùog—sluogun, instead of *sluoh—
sluogun, fâhu—fiang—fiangun, instead of *fiah—fiangun, etc. We
see an example in which this development is not prevented even
by the difference of the vowel arrangements in the OS. fîthan.
This ought, according to strict phonetic development, to form
he preterite fôth-fundun. We find, however, fand-fundun
only; and though it be true that in the present we find the
form findan, yet this is only by the side of fîthan. The few
MHG. traces of this old change all display a departure from the
older circumstances still existent in MHG., where the singular of
the preterite is assimilated to the plural: ziehe-zog (OHG. zôh)—
Bogen, leide-litt (OHG. leid)-litten, schneide-schnitt (OHG. sneid)-
schnitten, siede-sott (CHG. sôd)-sotten, erkiese-erkor (OHG. irkos)
—erkoren. In the same way the ablaut has certainly maintained
itself in MHG., but a correspondence has been re-established
between the singular and plural of the preterite.
330. We are often able to observe that phonetic differentiations
which are disregarded, either completely, or with the exception of
scanty traces, maintain their place between words etymologically
connected, or are only cancelled when their relationship to each
other is very close. In the Teutonic languages a change dating
from the oldest times has been in operation between the sound of
German ½ and our ch,-the sound of h appearing at the beginning
of a syllable, that of ch at the end of a syllable and before con-
sonants: cf. MHG., rûch, 'raw,' gen. rûhes, ich sihe—er siht (pro-
nounced like the MHG. sicht)—er sach-wir sâhen. In the case of

1

214
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the written language of to-day, this change in the inflexion is
disregarded, except in the case of hoch; while, besides this, the
comparative and superlative have been approximated to the posi-
tive, except in the case of höher-höchste and näher-nächste. In
other cases, however, it is maintained: cf. sehen-gesicht, geschehen
-geschichte, fliehen—flucht, ziehen-zucht, schmach-schmähen. An
interchange prevailing in many instances in the case of the vowels,
appeared in the OG. dialects under the influence of the vowel of the
following syllable, viz., between e and i and between u and o. This
interchange has been almost disregarded in the noun-inflexion
before the commencement of our tradition. In MHG. it is consis-
tently maintained within groups etymologically connected, except-
ing in cases of feminine formations of nouns of the agent; cf. go
—gotinne (OHG. gutinna), but cf. also the present use of birin by the
side of berinne, and wolf by the side of wülpinne and the diminu-
tives (cf. vogel-vögelîn [OHG. fugilî]). In NHG. the unification only
appears in the case of close relationship. Thus it appears regularly
in the case of names of materials (cf. leder-ledern, MHG. liderin,
gold-golden [MHG. guldîn]; holz-hölzern [hulzîn]; besides e.g., in
wort-antwort, antworten [MHG. antwürte, antwürten]; gold, ver-
golden, still in the old-fashioned style vergülden). On the other
hand, we find at the present day recht-richten, richtig, gericht,
berg-gebirge; feld-gefilde; herde-hirt; hold-huld; foll—fül-
len; koch-küche, etc.
331. As a matter of course, no unification sets in where the
feeling for the etymological connexion has completely disappeared
owing to a divergence in the development of meaning, or even
where it is so far faded that it does not come into con-
sciousness without some reflexion. This explains, for instance,
why the differences in sound just described are maintained in
the following cases: rauhrauchwerk, rauchwaare, rauchhandel,

•

ON ISOLATION AND THE REACTION AGAINST IT. 215
nach (MHG. nach)—nahe; erde-irden, irdisch; gold-gulden (an
adjective used substantively). In MHG. the contracted forms du
ireist, er treit from tragen actually exist; these are in NHG., again,
replaced by trägst, trägt; but the contraction is maintained in the
derivation getreide. MHG. gar shows in its inflected form a w
(garwe, etc.), which ought, according to the laws of sound-change,
to develop into b; but an inflexion, gar-garber, could not be
maintained indefinitely, and the inflected followed the example
of the uninflected forms: on the other hand, the b of the verb
gerben remained, owing to the different development of mean-
ing. Every language, in every possible stage of development,
affords a plentiful supply of proof of this phenomenon.
332. The intensity of the impression received by the memory
determines, in the first instance, the relative strength of the factors
opposed to each other, in which respect the remarks made above
(p. 206) equally apply. If, e.g., in ON. the first person singular of
the conjunctive mood ends, in both present and preterite, in a (gefa,
gafa), while in all other forms an i appears (gefir, gefi, gefim, gefið,
gefi and gæfir, gæfi, etc.) the chances are, of course, highly un-
favourable to the former, and thus gefi and gæfi appear in the more
recent sources likewise. But, of course, under certain circumstances,
one isolated form may carry the day against several corresponding
forms-that is, if they are more frequently used for themselves than
the others taken together. If, for instance, in NHG. ziemen the i
runs generally through the present, from which then, instead of the
old strong preterite, a further new weak preterite is formed, while
in MHG. most forms still maintain e, this depends on the fact that
the third person singular es ziemt occurred, then, as now, more
frequently than all the others.
333. But most of the irregularities in the treatment of etymo-
logical groups, which otherwise stand perfectly parallel to each

216
PRINCIPLES OF THE HISTORY OF LANGUAGE.
other, proceed from the fact that the individual groups differ very
markedly from each other with respect to the frequency of their
occurrence, and thus with respect to the facility with which the
individual forms can be reproduced by an effort of memory
with their traditional differences. The rarest words are those
which, assuming circumstances in other respects to be similar,
are earliest exposed to unification; the most commonly occurring
are the latest to be so exposed, or indeed not exposed at all.
This assertion can be proved inductively as well as deductively.
334. But besides this, the course of the movement is influenced
by a quantity of casual processes in the mental activity of the
single individuals and their reciprocal effect-processes which
lie beyond the sphere of our reasoning, as of our observation.
Especially do factors withdrawn from our knowledge play a great
part in the contest which the double forms, which have taken
their rise owing to unification, have to go through with respect
to each other. Omniscience alone could explain in every case
why the decision has been in certain cases thus, and in other
cases otherwise. And we cannot close our eyes to the fact that
very frequently analogous cases result differently in the same
dialect, and one and the same case have different issues in different
dialects. Thus, to cite only a quite certain example: while the
Gothic has in other cases unified the so-called grammatical
change, by rendering general the use of the consonant of the
present and of the singular preterite, the verbs hvairban, svairban,
skaidan have taken the opposite course, and have generalised the
use of the consonant of the plural preterite and of the participle;
and especially in the case of the last-mentioned verb in HG., which
carries the consonant of the plural preterite through more com-
monly than the Gothic, the consonant of the present has succeeded
in carrying the day.
ON ISOLATION AND THE REACTION AGAINST IT.
217
indivi- Influence of
the formal
formal grouping.
and e (2) regularly carried
Thus in the Gothic we find
335. Of course, however, the development in the
dual material groups is not wholly independent of the
grouping. When a phonetic differentiation affects all the
etymological parallel groups belonging to a formal group, a
co-operation of the material and the formal grouping is thereby
caused. This co-operation sometimes serves to decide the direction
of the unification. In the original German there occurred in
the numerous noun-formations with the suffix -no, a change of
the vowel preceding the n between (afterwards further de-
veloped into o-a) and e (i), so that both distributed themselves
among the different cases according to a regular rule.¹ Next,
at a later period, we find u (a)
through all the cases of a word.
forms like piudans, 'king,' standing beside forms like maurgins,
'to-morrow; and in ON. forms like Jormunn beside forms like
Odinn, and morgunn and morginn standing together. But the
participles falling under this head have, in contrast with the
lawless caprice evidenced in the other forms, invariably an in
Gothic, and invariably in in ON. The decisive part played by
the formal grouping is especially seen from the fact that such
participles as have been completely transformed into adjectives
or substantives, have in some cases struck out a new path; cf.
Gothic folgins, 'hidden,' as against fulhans, the genuine participle
of filhan, 'to hide;' aigin, 'property' (eigenthum), the participle of
aigan, ‘to have,' used substantivally; further, the ON. jqtunn, 'giant,'
the old participle of eta, ' to eat,' with an active signification.
336. The formal grouping may, however, not merely determine
the direction of the unification, but also whether it shall occur at all.
The less phonetic differentiation disturbs the formal parallelism of
the single groups among themselves, the more capable are they of
1 Cf. Beiträge, vi. 238 sqq.
218
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Conversion
of a merely
formal dif-
ference into
of meaning.
withstanding the tendencies towards unification. Thus, for instance,
the maintenance of the ablaut series in Teutonic for so long a
period would have been impossible if each verb had had its own.
special kind of ablaut, and had there not existed larger groups
of verbs with the same scheme. And it can be demonstrated
that the schemes preserved to us represent merely a selection
from those which were in existence before the commencement of
our tradition, all those which were represented either by a few or
by only a single example having perished, save for a few traces.
The disappearance of others may be traced historically; e.g. Gothic
truda - trad trê du trudans. It is the same with the umlaut
in the second and third persons of the singular indicative present
of the strong verbs; OHG. faru - ferist - ferit, and thus in MHG.
fahre-fährst - fährt.
337. Another circumstance which contributes to the preserva-
tion of a phonetic difference is when this happens to coincide with a
a difference difference of function. If, for instance, all the cases of the singular
are arranged so as to correspond with all the cases of the plural,
the relation they bear to each other is more easily and more
indelibly stamped on the memory than if certain forms of the
singular, coupled with certain forms of the plural are arranged so
as to correspond with certain other forms of the singular and plural.
And thus it is also natural that where, in the majority of cases, the
phonetic differentiation corresponds with the difference of function,
the unification should be confined to the groups which are more
immediately connected, and this unification thus perfects the
correspondence between phonetic and functional difference. In
Old Danish the plural of barn, 'bairn,' in accordance with a general
Scandinavian sound-law is børn, barna, børnum, børn ; while in
the singular, a prevails throughout. Modern Danish has admitted
børna also instead of barna. In another word, lagh, 'law,' o is
ON ISOLATION AND THE REACTION AGAINST IT.
219
The
carried even in Old Danish throughout the whole plural.
unification within the narrower groups is frequently enough the
preliminary to wider unification. Thus, in the case of lagh, the o is
actually making its appearance sometimes in Old Danish, and in
Modern Danish lov is the prevailing form. The coincidence with
a functional difference may, however, be the reason for the lasting
retention of a phonetic difference, and especially when it is ren-
dered capable by formal analogy of resistance in the way already
described.
338. The coincidence of these two circumstances enables the
idea of the phonetic variation to connect itself so closely with the
functional variation that the two may appear inseparable to the
instinct of language. Thus it is that the non-significant variation
originated by accident became significant. It will become more
significant still, the less plainly that variation in meaning is
accentuated by other phonetic variations. Language thus claims.
compensation for the loss of the characteristic signs of the
functional variation which set in as a consequence of phonetic loss.
339. The ablaut in the German verb depends upon a vowel
differentiation already manifest in the IE. original language. This
is a mechanical result of the shifting accent, and has nothing to do
with the original difference in the function of the single forms,
This differentiation was even for the original language superfluous,
except as regards the distinction between the present and imper-
fect on the one hand, and the aorist upon the other (cf. GK. λeiπw,
ἔλειπον, λείποιμι — ἔλιπον, λίποιμι). The perfect stem, in parti-
cular, was clearly differentiated from the present stem by re-
duplication. Hence we see, even in Greek, distinct signs of decay
in the vowel-change between present and perfect; we have, on
the one hand, λείπω - λέλοιπα, but πλέκω - πέπλεχα and not
*πéπλoxа. And only a few traces remains of the original change
220
PRINCIPLES OF THE HISTORY OF LANGUAGE.
L
between the singular and the plural of the perfect (olda
"oμev). The ablaut disappears, in fact, simply because it is
superfluous; and it was superfluous because the old characteristic
sign of the perfect stem, the reduplication, survived, and was con-
sistently maintained, while the present stem was also, in many
respects, specially characterised. In Teutonic, on the other hand,
the loss of the reduplicate went hand in hand with the confir-
mation of the ablaut. We cannot indeed assert that one was the
cause of the other. Rather, the primary impulse to the loss of the
reduplication was given by the phonetic development, in conse-
quence of which certain forms were no longer recognisable as re-
duplicated (cf. the type bêrum); and the preservation of the ablaut
is due, in the first instance, to the existence of parallel series. But
in the further course of development, a reciprocal causal relation
has become manifest. For instance, it is instructive for our purpose
that, in Gothic, chiefly those verbs have maintained the reduplica-
tion in which the IE. vowel variation between present and perfect
(preterite) has disappeared on phonetic grounds; and this holds
good of every case; cf., halda - haihald, skaida-skaiskaid, stauta
staitaut. Still, even in OHG., there was as yet no absolute
necessity of differentiating the root syllable of the present and
preterite, because the variation was marked in the termination as
well in the case of each single person of the indicative as of the
conjunctive. The case is different in MHG., where in the first and
second persons plural of the indicative, and throughout the whole of
the conjunctive the variation between the present and the preterite
depends entirely upon the form taken by the root-syllable, cf. geben
= gåben, gebet = gåbet, gebe = gabe, etc. In NHG. the second
person singular and the third person plural of the indicative must
be added to the list. Thus the ablaut has become a character-
istic, whose necessity has been increasingly felt. But it is only
ON ISOLATION AND THE REACTION AGAINST IT.
221
the distinction between the present and the preterite that is of im-
portance, and not that between the singular indicative preterite, or
merely the first and third persons of it, and the other forms of the
preterite. The last-mentioned distinction, passing as it did like-
wise from the original language, was maintained solely owing
to the frequency of the occurrence of certain verbs, and the exis-
tence of parallel series. It was thus, in the case of certain
classes, cast aside at an early date (Gothic fôr - fôrum, faifâh
faifâhum, OHG. flang fiangum) [OE. scôc — scôcon, etc.]. In the
case of other classes it has maintained its ground even into NHG.,
but has finally been cast aside entirely save for a few traces. It
is merely in respect of phonetic convenience a gain to say, as we
do now, sprang - sprangen, flog-flogen, and not as in OHG.,
spranc― sprungen, floug — flugen. Thus the ablaut received no
functional value till the NHG. period. And in this connexion
there is one noteworthy circumstance. The variation between
the singular and plural is (except in the past-present verbs) re-
tained by the modern written language solely in the common verb
werden [ward-wurdon], and in this case also by-forms prepon-
derate, in which the variation is lost; on the other hand, there is
still a number of verbs in which, though the vowel of the singular
has penetrated into their plural, yet the conjunctive has main-
tained its peculiar vowel arrangement: starb-stürbe, schwamm —
schwömme (with a by-form schwämme), etc. In this case a phonetic
contrast is maintained within narrower limits, but again in virtue
of the coincidence with a functional one. Since, however, the
umlaut alone would suffice to express the latter (schwammen
schwämmen), it would be superfluous to maintain the old vowel.
But precisely in the case of those verbs in which the umlaut is
maintained most strictly (verdürbe, stürbe, würbe, würfe, hülfe)
another consideration comes in, namely, the need for distinction
222
PRINCIPLES OF THE HISTORY OF LANGUAGE.
from the conjunctive present: helfe and hälfe, which form occurs
as a variant of hülfe, are no doubt differently written, but phone-
tically they correspond with each other. On the other hand, there
is no verb with i going through all the persons of its present
tense which still forms a subjunctive preterite with ü (e.g., singe
sünge), because in this case it is precisely the old form, which,
according to the pronunciation prevalent in most German dialects,
would fall into correspondence with the conjunctive present. And
this explains the fact why the verbs with mm and nn show double
forms still (schwämme-schwömme, sänne - sönne; cf. geschwom-
men, gesonnen as contrasted with gesungen.)
340. A part similar to that played by the ablaut is that played
by the umlaut produced by an i orj of the following syllable.
In the masculine ¿-declension in OHG. it had happened to become
the rule that the whole singular was free from, and the whole
plural affected by, its umlaut (gast-gesti, etc.), and it is hence
that the difference is maintained. The circumstance is best ex-
plained if we compare the history of the change between e and
i, u and o, caused by the following vowel. The 7-declension
must in original German have run much as follows :—
1
Sing.
N. meduz
Pl.
Sing.
midiviz
SUNUZ
Pl.
suniviz
G. medaus
medevô
sonauz
sonevô
D. midiu
medumiz
suniu
sunum
A. medu
meduns
sunu
sununs
So purposeless a change was not likely to be long maintained.
OHG. has
And thus we find traces of it in Old Norse alone.
from the very earliest times made the u of sunu uniform through-
The terminations are comparatively unimportant for our purpose.
1
ON ISOLATION AND THE REACTION AGAINST IT.
223
out; in metu, ehu, eru, the e; and in situ, quirn, the i The
umlaut in the i-declension has not yet come to be a necessity
in OHG., as the cases of the plural are independently of its aid
clearly differentiated from those of the singular; indeed it is not
a necessity in MHG. either as long as the e of the flexional
terminations is retained; for the nominative, accusative, and
genitive plural geste could not easily be confused with the dative
singular gaste, even should they dispense with the aid of the
umlaut. But directly the e disappears - and this is the case
notably in the South German dialects-the umlaut remains in the
nominative and accusative as the only mark of distinction be-
tween the singular and the plural. At this stage of development
the i-declension has a decided advantage over the a-declension,
and the purely dynamical value of the umlaut is complete.
This is very clearly manifest in the fact that it spreads beyond
its original domain. This disposition to spread stands in the
closest connexion with the absence or presence of a differen-
tiating e. In the same way in South German the umlaut has
affected almost all the substantives of the old a-declension capable
of being affected by it (cf. SCHMELLER, Mundarten Baierns, § 796,
and WINTELER, Kerenzer mundart, p. 170 sqq.). Thus they say
tag-täg, arm-ärm, etc. The Middle and Low German dialects.
have this tendency in a much less degree, and notably only in
the case of words of more than one syllable, such as sattel, wagen,
in which they too reject the e of the plural. At an early period
the umlaut succeeded in affecting words expressive of relationship,
which originally were inflected with consonants, and thus could
dispense with a termination in the nominative and accusative plural;
MHG. vater-veter, muoter-müeter, etc.
1 Cf. Beiträge, z. g. d. d. spr. vi. 80. The possible alternative remains of a unifica-
tion in different directions, assuming that for OHG. a phonetic transitive of e into i before
" is to be supposed.
224
PRINCIPLES OF THE HISTORY OF LANGUAGE.
341. The formal unification, which we had already occasion
repeatedly to call into consideration, is also often a reaction
against an objectless phonetic differentiation. The process is then
the following. Within a class which hitherto adopted similar for-
mations, phonetic differences arose in one form or in several. For
instance, the genitive in certain words formed itself in one way,
and in others differently, while in other cases the uniformity was
not disturbed. Then the tendency makes itself felt of restoring the
same uniformity in the one or the few differentiated forms as well,
and of changing the partial agreement of the method of formation
again into a complete agreement. This method of unification is
specially seen in connexion with the material method, as the
instances cited may serve to show. But even independently of
this it occurs frequently enough. Under this head falls, for
instance, the unification of the hard and soft fricative in the ter-
minations of the cases and persons of the Old Teutonic dialects.¹
According to Verner's Law p=IE. t was divided into p and (d), s
intos (hard) and (soft). Accordingly in the original German we
had * trdési (du trittst,' thou treadest'), trdépi (er tritt), * trdépe (ihr
tretet)* trdónpi (sie treten), as against* bérezi, 'thou bearest,' * béreði,
* béreðe, * bérondi, while in the first persons singular and plural no
differentiation had yet appeared; further, in the o-declension, the
nominative singular was * stigós (steg), but *éhwoz ('horse'); the
nominative plural *stigôs, but éhwôz; accusative-plural * stigóns
but éhwons, while the other case-endings remained the same; and
similarly in other classes of inflexion. The unification which
ensued terminated in nearly every case in favour of the soft
sound; on which we have to remark that ≈ appears as r in the ON.
and in the West Teutonic dialects, but disappears from the latter.
in the case of original auslaut. Still, in certain cases, too, the hard
1 Cf. Beiträge, vi. 548 sqq.
ON ISOLATION AND THE REACTION AGAINST IT.
225
s has carried the day. Thus, in the nominative plural of the a-
declension, we find in AS. and Old Friesic dagas by the side of ON.
dagar; in OS. the Heliand exhibits -os, only in solitary cases o or
a (grurio, slutila); while in the Freckenhorst Roll a appears more
frequently than os and as. OHG. recognises in appellatives a only;
on the other hand, in genealogical designations which have passed
into names of towns, it displays -as as well, supposing that these
are not to be taken in any other way.¹
342. An example, taken from modern times, is the restoration
of the inflexion e in NHG. in cases where it had already disappeared
in MHG.
The derivatives terminating in -en, -er, -el are par-
ticularly instructive. In the case of substantives the MHG. rejec-
tion of the e is maintained; cf., des morgens, dem wagen, die wagen,
der wagen, den wagen, as against tages, tage, tagen, and in the same.
way schüssel, schüsseln, as against schule, schulen. On the other
hand, in the case of adjectives which are connected more firmly,
owing to the prevailing uniformity which characterises them, the e
has been restored after the analogy of the monosyllabic words:
gefangenes like langes, gefangene, gefangenen (MHG. gevangen),
andere, anderes, andere (= MHG. ander, anders, ander). The NHG.
forms occur even in MHG. by the side of the syncopated forms.
We are able in this case to make observations again on the pro-
cess of isolation. We write invariably die and den eltern, beside
die, den älteren; der jünger, den jüngern (substantive), beside
der jüngere, den jüngeren (adjective); einzeln, dative plural of the
MHG. adjective einzel; anderseits, unserseits, as against andere seite,
unsere seite; vorderseite, hinterseite, oberarm, unterarm, edelmann,
innerhalb, ausserhalb, oberhalb, unterhalb (all spurious composites
sprung from the union of the adjective and substantive), as against
die vordere seite, etc.; anders, as against anderes.
1 Cf. Kögel, Zeitschrift für deutsches altertum, xxviii. p. 110 sqq.
Р
226
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Conversion
of elements
of the stem
into in-
flexions.
343. Another case in which the umlaut has become dynamical
from analogous reasons is the subjunctive mood of strong pre-
terites, and of the weak ones formed with no intervening vowel,
MHG. fuor-füere, sang, pl. sungen—süngen, mohte-möhte, brâhte-
bræhte, etc. In this case the umlaut is either uniformly prevalent,
or, at least for the plural, is the only method of distinction.
The dynamic acceptation of umlaut by the linguistic instinct
announces itself in NHG. by the fact that, in spite of all other
unification of the vowel-system, the umlaut still remains, as sang,
sangen—sänge, for sungen, sünge; and further, even more decidedly
in MHG., in the transference of the umlaut from the preterites
which had originally no vowels to the syncopated preterites (brante
-brente, instead of brante after the analogy of brâhte-bræhte).¹
344. A third case is the umlaut in the present as contrasted with
the suppression of the umlaut in the preterite and the participle:
OHG. brennu-branta-gibrantêr. In the participle a change has
developed phonetically: gibrennit―gibrant-. The first result,
however, of unification under these circumstances is that the unin-
flected form gibrennit is rejected in favour of gibrant. But then
the contrast is maintained in the root syllable between the present
and preterite participle for centuries unintermittently, although it
is not necessary to the characterisation of the forms.
345. In this way it is possible for even elements of the stem of
the word to be changed into flexional terminations. This is the
case in the German weak declension. In this the n (as in words
like namen, frauen, herzen) belongs to the original stem. As, how-
ever, every trace of the original flexional termination has been
wiped out by phonetic decay, and as, on the other side, the n in
the nominative (in the case of the neuter, in the accusative as well)
singular has disappeared, as in the case of name, frau, herz, it has
1 Cf. Bech, Germania, xv. p 129 sqq.
ON ISOLATION AND THE REACTION AGAINST IT. 227
come to be a characteristic of the oblique cases in contrast with
he nominative singular. Another case-suffix, which took its rise
n a similar manner, is the syllable -er, which serves to form
plurals (rad-räder, mann-männer). The method of formation
cook its origin in certain neuter s-stems (cf. Latin genus—generis),
n which s had, obeying the law of sound-change, turned into r.
In the nominative singular this same s had naturally, according to
the same laws, to disappear together with the preceding vowel.
Thus, under the partial influence of the vowel-declension, there
arose in the first place in OHG. the following scheme :-
Sing.
N. kalp
Pl.
kalbir
G. kalbir-es
kalbir-o
D. kalbir-e
kalbir-um
A. kalp
kalbir
}
In the genitive and dative singular, at least, the -ir- was useless
and puzzling. And thus the forms in question have already dis-
appeared at the time from which our oldest sources date, leaving but
isolated traces, and are replaced by kalbes, kalbe, which are formed
after the model of the normal inflexion from the nominative
accusative. Then the -ir necessarily appeared as the characteristic
of the plural, and the more so since there was in the nominative
and accusative no other differentiating symbol. The functional
character then of the -ir (MHG., NHG. -er) is attested by the fact
that it is gradually transferred to a quantity of words with which
it had originally no connexion.
changes in-
346. These examples will suffice to make it plain how a varia- All phonetic
tion which sprang up without any idea of purpose, if favoured by voluntary.
the casual coincidence of different circumstances, may unperceived,
and unintentionally, be made to subserve a purpose, causing it
5
1
228
PRINCIPLES OF THE HISTORY OF LANGUAGE.
to appear as though the variation were designedly made to suit
this very purpose. This appearance grows actually stronger the
more perfectly the differences which arose at the same time un-
intentionally are abolished. We may generalise our experience
drawn from historical development as far as we can trace it, in
the proposition that there is no such thing as variation of sound
created of set purpose with a view to denote a difference of
function. The difference of function only attaches itself to the
variation of sound by secondary development, and this through a
development which the individual speaker neither designs nor per-
ceives by means of a natural association of ideas.
:
*
1
CHAPTER XI.
}
IF
FORMATION OF NEW GROUPS.
of distinc-
sound-
F sound-change has, generally speaking, the effect of producing Obliteration
differences where none existed previously, it serves, on the tions by
other hand, not unusually to obliterate existing differences. This change.
is under certain circumstances a salutary arrangement; but in most
cases it does more harm than good. For by this process differences
essential to the denotation of the function disappear; and in
addition, the clear division of the single groups from each other is
rendered an impossibility. Hence the effect of sound-change is
commonly to produce further results, and is especially operative
in producing many analogical new-creations.
348. The simplest process under this head is, when words
etymologically unconnected, and equally unconnected in significa-
tion, come to coincide by the action of secondary development ;
e.g., enkel (talus) = MHG. enkel- enkel (nepos) = MHG. enenkel,
garbe (manipulus) = MHG. garbe garbe (schafgarbe) = MHG.
garwe, kiel (carina)
kiel (caulis pennae) = MHG. kil,
mähre (equa) = MHG. merhe, tor
=
MHG. kiel
mähre (narratio) = MHG. mære
(porta) = MHG. tor tor (stultus) = MHG. tôre, los (solutus) = MHG.
lôs-los (sors) MHG. lôz, ohm (amphora) = MHG. âme ohm
(avunculus) = oheim, schnur (linea) - MHG. snuor schnur (nurus)
-MHG. snur. English is specially rich in instances of this process.
[Cf., such cases as shed (asceádan, ' to part'), shed, ' a hut' (doublet of
230
1
PRINCIPLES OF THE HISTORY OF LANGUAGE.
shade); sheer (allied to Icelandic skærr, 'bright'); sheer (a doublet
of schear, akin to Dutch scheren); sallow (a willow'), sallow,
('yellowish'); rouse, row, rock, quarry, purl, punch, pump, box, etc.]
349. It sometimes happens that two such words, in spite
of their different meaning, become, for the instinct of language,
one. No one without special philological knowledge will suspect
that two entirely distinct words have coalesced in the German
word unter, namely the Latin word inter and another con-
nected with the Latin infra (SKT. andhari). Few would
suspect that the phrase ein schiff lichten is of a different
origin (MHG. lîhten, 'to lighten') from die anker lichten (Low
German form for lüften, English lift). Schlingen (devorare) is the
MHG. form of an elder slinden (cf., schlund), and the word has
perhaps taken firm root in the written language because it has
coalesced with schlingen = MHG. slingen. In the phrase in die
schanze schlagen few are aware that we are dealing with the French
word chance, and not with the German word schanz. Certain cases
are more instructive still in which formal influence has been at
work. It may be a mere conjecture that mahlen (MHG. main) has
passed from the strong into the weak conjugation under the
influence of malen (MHG. mâlen). But it is less questionable that
the transference of laden (einladen, OHG. ladôn) into the strong con-
jugation has been caused by contact with laden (aufladen, OHG.
hladan, English lade); while conversely weak forms of the latter
appear as well-e.g., überladete occurs in LESSING, and ladest and
ladet are in actual use. It is certain that a strong form er befährt
occurring in JEAN PAUL, in connexion with befahren (else conjugated
as a weak verb = MHG. vâren) is due to a confusion with the strong
befahren (MHG. varn). In Austria, kennen and können are confused
one hears, for instance, der schauspieler hat seine rolle nicht gekannt.
In the case of the two last instances, we have words etymologically
FORMATION OF NEW GROUPS.
231
connected, but still essentially different, confused. In MHG. we
find two etymologically different particles wan; one with an
adversative, the other with a causal meaning used in the sense of
NHG. denn. The latter has a fuller by-form, wande, which comes
to be used sometimes in an adversative sense, where such sense
is not justified by its origin (cf. MHG. Wörterbuch, iii. 479 b). In
OHG. the prepositions int- and in in composition with a verb,
have often coalesced into the form in-, the t disappearing by assimi-
lation with the following consonants. It thus happens that the
form int-, being taken for a doublet of in-, has passed into cases.
where in should, strictly speaking, be used; cf., NHG. entbrennen,
entzünden, etc. The German prefix zer- had in olden times a by-
form ze-, which developed into zer- before a vowel, and ze- before
a consonant. This form was, phonetically speaking, identical with
the preposition ze-, which is of wholly different origin. The
adverbial form zuo (MHG. zu) appeared side by side with these,
which gradually thrust completely aside the form ze. This zu is
also found for ze- = zer-, e.g., in LUTHER. The AS. tô-, used in the
sense of zer-, is to be similarly explained.
unrelated
material
(simplest
350. Unconnected words coalesce into material-groups owing Union of
to accidental partial unification. This is the simplest variety of words in
so-called popular etymology, consisting solely in a change of groups;
meaning effected by the impulse of language without the sound- form of
form undergoing any change. It is a necessary condition of this mology).
that the true etymology of one of the words has been obscured,
so that it has no other more natural connexion.
=
351. The component parts of a compound word are specially
subject to such changes; thus the German word erwähnen is
apprehended as compounded with wähnen MHG. wanen, while
it really and truly contains rather the MHG. (ge)wehenen: in
the case of freitag we are tempted to think of the adjective frei.
popular ety.
232
PRINCIPLES OF THE HISTORY OF LANGUAGE.
[In sovereign we think that we see the word reign, whereas the
old English form was soveraine, French souverain: in island
we fancy that we have a word connected with the Latin insula,
whereas it represents in fact the AS. ealand.] Proper names are
more exposed than any others to change of significance; cf.
Reinwald, Bärwald, Braunwald, in which the latter half was
not originally equivalent to silva, but was the noun of the agent
from walten; Glaub-recht and Lieb-recht are originally probably
compounded with brecht OHG. beraht; Sauerlant is a High
German form of Sûerland = Süderlant. Other instances may be
seen in Andresen's Volksetymologie [and A. S. Palmer's Folk-Ety-
mology, 1882]. In this case the change of meaning has ensued
without any support derived at the very outset from a connexion
in the meaning. Nothing is in fact operative but the natural
expectation of finding, in a word which looks like a compound,
familiar elements.
=
352. Proper names hold out least successfully against such
secondary relationship attaching to the sound alone, simply be-
cause here no contradiction, any more than agreement, of mean-
ing is possible. There are, however, cases as well in which
it is possible to restore a relation between the significations of
the word in question; cf. MHG. endekrist, phonetically developed
from antikrist; NHG. lanzknecht from landes knecht; wahnwitz,
wahnsinn, wahnschaffen, falsely connected with wahn (MHG. wân),
while the NHG. wan ('empty' [seen in wan-hope]) lies at their
real root; friedhof, from MHG. frîthof; vormund is referred to
mund, 'a protection;' verweisen does not belong to weisen (= MHG.
wisen), but comes from the MHG. verwizen. Umringen, as its
weak inflexion proves, is in its origin no compound from ringen,
but is a derivative from the MHG. úmberinc, now obsolete. But
the fact that it is accentuated umringen shows that it is appre-
FORMATION OF NEW GROUPS.
233
hended in meaning as a compound from um and ringen. A
further result of the change of meaning was the formation of a
participle umrungen, and even of a preterite umrang (see Sanders,
ii. 764). Words, too, which are really and truly no compounds,
but which give the effect of being so owing to the full conforma-
tion of their sounds, are thus fashioned with true compounds;
cf. leumund, which is appreciated as leutemund, but is really
derived from the Gothic hliuma (auris); weissagen, MHG. wîssagen
=OHG. wîzagon, which is really a derivation from wîzago, 'the
knowing one,' 'the prophet;' trübsälig, armsälig, etc., which are
mere derivatives from trübsal, armsal, etc., by the derivative.
suffix -sal.
353. It is a rarer case when a word is apprehended as a derivative
of another with which it is wholly unconnected.¹ The NHG. sucht
is apprehended by the instinct of language as connected with
suchen. It really comes from the MHG. suht (= Gothic_sauhts),
which has nothing to do with the MHG. suochen (Gothic sôkjan).
The NHG. inclination to connect the word with suchen depends
on such compounds as wassersucht, mondsucht, gelbsucht, schwind-
sucht, eifersucht, sehnsucht, ehrsucht, etc., which were apprehended
as 'a longing for water,' 'for the moon,' etc.
Hans Sachs appre-
hends -suht still as an illness,' as he shows by writing wann er
hat auch die eifersucht.
On the other hand we may compare
the well-known proverb eifersucht ist eine leidenschaft, die mit
eifer sucht, was leiden schafft. The word laute is regarded as
connected with laut, but is really a loan word from the Arabic.
[In the same way in English, posthumous is popularly connected
with post humum, dismal with dies malus; the sixteenth century
connected abominable (written abhominable) with ab-homo, as if the
1
¹ [Cf. Outlines of a History of the German Language (last Chapter), by Strong and
Meyer.]
234
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Phonetic
transforma-
tion (com-
plex kind of
popular
etymology).
word signified inhuman.] In the case of the word hantieren, which
comes from hanter, we think of hand; in the case of fallieren
(from the French) of fallen; in the case of beschwichtigen, the
Low German form of the MHG. swiften, of schweigen; in that
of schmälen (properly 'to make schmal, to impair') we think
of schmähen. The words Herrschaft, herrlich, and herrschen are
derived from hehr (hence MHG. form herschaft, etc.), but are now
referred to herr, with which word they were originally but in-
directly connected.
354. There is a more complex kind of popular etymology
which must be distinguished from those described. This consists
in a phonetic transformation whereby a word which reminds us of
another by the similarity of its sound is brought into nearer corre-
spondence with it still. Such transformation may be made of set
purpose, and with the full consciousness that those who employ
the word are deliberately altering its form. Such linguistic con-
tortions are often handy tools for humorous writers: Fischart
employs them on a large scale in German [and Dickens and
Thackeray in English]. Many of these are handed down as tradi-
tional jokes, especially in academical circles. This intentional
humorous phonetic transformation offers no problem to the philo-
loger. It merely concerns him so far as it is not apprehended as a
contortion by the simple minds of children and of the uneducated,
but is accepted and propagated as the genuine form. There is,
however, doubtless such a thing as an unintentional and uncon-
scious transformation, which proves itself to be such by the absence
of any humour.¹ Foreign words, proper names, and other words
whose etymology has become obscured are specially exposed to
this; and almost exclusively compounds, or words which in virtue
¹ We must remark that this must not be confused with the phonetic substitution to
be discussed in Chapter XXII. The operations of the two processes are not always,
however, clearly separated.
FORMATION OF NEW GROUPS.
235
of their fuller sound-arrangements give the impression of being
such. In this process either the first element is exposed to a
change, as in the case of jubeljahr (Hebrew jobel), dienstag, Hul-
dreich (from MHG. Uolrich), maulwurf (from MHG. moltwurf), Latin
aurichalcum, from the Greek opeixaλkos: or it may be the second
only; cf., hagestolz, Reinhold, Gotthold, Weinhold, etc., from -olt=
walt,¹ abspannen, from MHG. spanen, 'to allure;' abstreifen, from
MHG. ströufen;2 einöde, from MHG. einæte (œte being a mere suffix):
or both parts may be affected; cf., armbrust, from the Latin
arcubalista; liebstöckel, from the Latin ligusticum; felleisen, from
French valise; ehrenhold, from herolt; Greek σvvédpiov, from the
Hebrew sanhedrin [cf. the vulgar English sparrow-grass for
asparagus]. One part changes its form, while the other merely
changes its meaning, in the case of abseite, formerly apside, from
the Greek ais; Küssnacht from Cussiniacum; probably also
in Mailand from MHG. Mîlân. These examples help to show
that an additional impulse is given to unification if the meaning
of the transformed word permits of being connected with that
of its model; but such impulse is not absolutely necessary for
its production. To explain the process we have to notice in
the first instance that we do not as a rule apprehend the words
and sentences which we hear with absolute exactness according
to their several factors. We commonly to some extent divine
them by guesswork, and our guess depends on the sense which
the context leads us to expect. In the-process we come naturally
upon sound-groups with which we are quite familiar; and it may
thus happen that the very first time that the word is heard,
instead of an intrinsically meaningless portion of a long
word, a word in ordinary use with a similar sound may
1 The has probably never been pronounced, and in that case we have only a change
of meaning, which has found expression in the orthography.
2 In this case the dialectic transition from cu into ei comes into consideration.
236
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Union in
formal
groups: (a)
Where func-
tion is the
same.
substitute itself. And further, a portion of a word, which
otherwise has no hold on language, has, even when correctly
caught, no firm hold on the memory, and it is therefore possible
that on the attempt at reproduction a familiar element may sub-
stitute itself as an independent word. And when such substitution
has once taken place either in the process of hearing or in that
of speaking, the substitution has the advantage over the genuine
of imprinting itself more lastingly on the memory. It is natural
to expect that this process should be, as a rule, confined to words
of more than ordinary length. Shorter words are more readily
caught and more readily retained. But further, we are accus-
tomed to expect to find that there are a number of simple words
standing isolated, or at least grouped with derivatives generally
familiar, and admitting of being found at will, while we expect
besides, of a word which gives the impression of a compound,
that its single elements should admit of connexion with simple
words.
355. Phonetic coincidence is a more active efficient on the
formal than on the material side of language.
The processes
which range themselves under this head fall under two main
groups i.e. according as forms functionally identical or func-
tionally different come to coincide.
356. The disappearance of phonetic differences in cases of
identity of function may act beneficially by simplifying the forma-
tion of the formal-groups. Sometimes in the process only the
phonetic differentiation described in the last chapter is again
cancelled. Thus, for instance, the OHG. formative syllables -ul,
-al, -il, which lie on the same lines, converge into the form el;
and similarly with -un, -an, -in, which converge in the form -en,
etc. But such differences as the double formation of the com-
parative and superlative in OHG. -iro, -ist--ôro, -ôst, or the two
FORMATION OF NEW GROUPS.
237
synonymous methods of forming the adjective in -ag and -g-in
the case of the isolated instance of einag, einîg, we have a difference
of signification—serve no purpose, and it is thus a clear gain that
we have at present only -er, -[e]st, and -ig. The convergence,
again, of two entire classes of inflexions like the OHG. words in
-ôn, -ên, into the MHG. -en is merely a useful simplification.
1
357. But it does not always happen that phonetic convergence
affects so regularly entire systems of material-formal groups.
Commonly speaking it affects only a single part of forms con-
nected with each other. Then it contributes not to the simplifica-
tion of the relations, but often to their confusion.
(a.) The phonetic divergence affects consistently the entire
forms of a system of inflexion; but it affects in one class of
inflexions or in several only one part of the words which originally
belonged to it. While, as we have seen, of the three OHG. classes
of the weak verbs in MHG. two have completely converged, only
those of the third class which have the root-syllable short (Gothic
ending in -jan) have fully united with them; those with long
root-syllable remain still differentiated by the old syncope of the
central vowel in the preterite and perfect participle, and it may
be by the Rück-umlaut'; cf., manete, lebete, wenete, from manota,
lebeta, wenita, connected with manen, leben, wenen, by the side
of neicte, brante, connected with neigen, brennen. The OHG. 2-
declension has completely coalesced with the o- declension with
respect to its terminations, but with respect to the arrangement
of its stem only when the root-vowel admits of no umlaut. Thus
in this case there is always a separation closely connected with
the convergence, or it may be that a separation has preceded
the convergence.
(b.) The convergence affects all the words of several classes
of inflexion similarly, but not all the forms of a system of in-
238
PRINCIPLES OF THE HISTORY OF LANGUAGE.
flexions. This is a very common case.
Thus the second declension
in Latin has converged with the fourth in the nominative and
accusative singular only; and so with the o- and i-declensions
in Gothic (fisks, fisk—gasts, gast).
(c.) The convergence affects only one part of the words of
several inflexional classes, and only one part of the forms of the
inflexional system. Thus in OHG. the nominative and accusative
of the i-, u-, and o- stems, in the case of words ending with either
a long syllable or of words which are polysyllabic, while these
cases have remained separate in the case of those terminating
in a short syllable; cf., gast, wald, arm, from *gasti(z), *waldu (z),
*armo(z), as against wini, sunu, and a form *goto which may at
least be assumed.
358. When the case (a) has occurred, the convergence, as well
as the divergence, of inflexional classes is definite, and admits of
no reaction. The lasting result is a displacement in the relative
forces of the groups in question, the one receiving an accession
at the expense of the others. The cases of (b) and (c), on the
other hand, produce a confusion in the grouping. Where pho-
netic modifications once different are employed for the same
function, it is most convenient for the phonetic difference to go
through all the forms of a system, so that a sharp line of demarca-
tion may be drawn between the single classes of inflexions, so that
each single form betrays at a glance the class to which it belongs.
359. Now supposing that certain forms in two classes coincide,
while some diverge, then we shall find that a word is easily
arranged wrongly on the basis of the forms that agree, and ana-
logical formations arise in the place of the traditional forms of
the one class which rightly belong to the other. It is then possible
for language to gradually work its way out of the vacillation, and the
confusion which is its result, to simpler and more stable conditions.
FORMATION OF NEW GROUPS.
239
+
360. Examples are easy to find in numbers. I would refer
particularly to the reciprocal influences exercised by the different
classes of the declensions of the IE. language on the single lan-
guages-influences which have almost always followed as a result
of phonetic convergence in several cases, especially in the nomi-
native and accusative singular. In most cases the classes which
thus converge had at an earlier period an identical or approxi-
mately identical method of formation; and this original identity
has only become obscured owing to secondary sound-develop-
ment, against which an immediate reaction is not possible, because
the differentiation was so consistent. Thus, for instance, the
unity of the IE. declension had been mainly destroyed owing to
the vowel separation which has set in under the influence of the
accent, and the contraction of the final syllable of the stem with
the regular inflexional termination. These changes were so wide-
reaching that many further changes - and notably changes in
the direction of weakening-were necessary in order to partially
unite again on an entirely different basis what was separated.
361. In this kind of convergence the result commonly is that
words of a class formed in one way pass into a class formed in
a different way, and they may do this either entirely or only in
some cases; either in all forms or only in certain ones. The
following may serve as an example of the latter. In Gothic the
masculines of the i- declension in the singular have passed into
the a- declension on account of the phonetic convergence in the
nominative and accusative, and the case is similar in OHG. The
plural, however, remains in both dialects differently inflected.
The fact of the unification being arrested at this particular point
is a result of the never-failing co-operation of the etymological
grouping, and so far the maxim is confirmed; the closer the
bond, the more easy the power of influence.
240
PRINCIPLES OF THE HISTORY OF LANGUAGE.
362. It may happen that either only the one group is active
while the other contents itself with a passive rôle; or that both
groups are active and passive at once. In NHG. a quantity of
weak masculines have taken the inflexion of the strong ones in
-en, from which they differed even in MHG. only in their nomina-
tive and genitive singular; cf., bogen (=MHG. boge), garten, kragen,
schaden, etc. There are, however, certain cases in which conversely
a strong masculine in n has passed over into the weak inflexion;
cf., heide (=MHG. heiden), krist(e) (=MHG. kristen), rabe (=MHG.
raben).
+22
363. If such reciprocal influencing and counter-influencing of
two groups appears in the same words, it may also happen that'
after a long period of vacillation a completely new system of
flexional formation comes into existence. Thus a mixed class
has arisen by contamination of the two classes just referred
to: der glaube—des glaubens, der gedankě-des gedankens, etc.
The rise of this mixed class is easily explained if we remark
that once double forms existed in the nominative just as in the
genitive der glaube-der glauben, des glauben-des glaubens. Then
the nominative of one class and the genitive of the other class
became established in the written language. Thus, again, a mixed
class has arisen from the reciprocal influence exercised by the
weak masculines with rejected final vowel on the one hand, and
by the strong masculines on the other, which inflect the singular
strong and the plural weak: schmerz, -es, -e--schmerzen. Similarly
in the case of neuters: bett, -es, -e-betten. The most extensive
instance of the kind in NHG. is the regular inflexion of the femi-
nines in -è, which is a fusion of the old a-declension with the
In MHG. the inflexion still runs
n- (or weak) declension.
follows :-
as
FORMATION OF NEW GROUPS.
241
SG. N. vröude
zunge.
G. vröude
zungen.
D. vröude
zungen.
A.
vröude
zungen.
PL. N.
vröude
zungen.
G. vröuden
zungen.
D. vröuden
zungen.
A. vröude
zungen.
364. In NHG. the word freude or zunge holds good for the
whole singular throughout, and the forms freuden and zungen
run through the whole plural. This is again a characteristic
example of a useful transformation attained without any con-
sciousness of its result. The greater advantage of the NHG.
scheme consists not merely in the fact that the memory is thereby
considerably relieved; but that the two terminations which are
alone common to both are distributed in the most convenient fashion
possible. It is more important to discriminate numbers than
cases, because the latter are also characterised by the article which
is in most instances attached to them. In MHG. die vroude and
die zungen may be the accusative singular and the nominative
accusative plural, while der zungen may be the genitive singular
and plural. These uncertainties are now no longer possible; on
the other hand, in the case of sunge, only the power of marking
the difference between the nominative and accusative singular is
cancelled. But if we examine how these circumstances have
developed, we shall find that as a preliminary process a general
encroachment has taken place of each of the two classes into the
department of the other. And it was natural that this should
be the result as soon as once phonetic convergence had declared
itself in the case of three forms (nominative singular, and genitive
and dative plural). The circumstance was thus produced that
Q
242
PRINCIPLES OF THE HISTORY OF LANGUAGE.
each form, with the sole exception of the nominative singular,
might terminate in -e as well as in -en. In all this no single
form was created with the special object of serving a purpose;
but the maintenance or disappearance of every form has been
determined by its utility.
365. Reciprocal influence in the case of two groups always.
assumes that their reciprocal forces are not too unevenly balanced.
In any other case the influence will be one-sided, and thus more
prevailing and more rapid in attaining its end. Those classes
of course are always exposed to especial risk which are not re-
presented by numerous examples that is, supposing that
these are not protected by the special frequency with which
they occur. The limited extent of certain classes as contrasted
with others may have existed from the very earliest times,
through no more words having been formed in the way in
question, but it is, commonly speaking, a result of secondary
development. It either happens that many words originally
belonging to the class die out, in which process the case
may occur that one method of creating forms, originally a
living method, dies out, and is merely handed on as a 'usual'
formation, in a few cases in common use. Or it may be that
the class splits up into several sub-divisions owing to phonetic
variations, which lose all bond of connexion as no immediate
reaction against them is made. Thus the greatest possible sub-
division of one is sometimes the best possible way of finally
uniting two different methods of formation with each other.
Observations in this sense may be made, for instance, in the
history of the gradual disappearance of the consonantal and the
u-declension in German.
366. Supposing that one class has gained any decisive advan-
tage over one or several others with which it has certain points
FORMATION OF NEW GROUPS.
243
of contact, the latter are infallibly predestined to perish.
It
is only the fact of frequent occurrence which is able to impart
force enough to enable certain words to escape for any long period
the otherwise overwhelming influence. These then exist for the
future in their isolation under the name of anomalous forms.
367. Every language is incessantly engaged in eliminating all
useless irregularities, and securing identity between functional and
phonetic expression. All languages are not equally successful.
We find single languages, and single stages of development of these
languages, anything but equidistant from this goal. But even
the language which succeeds in approximating most nearly to its
aim, fails signally in attaining it. In spite of all the transforma-
tions which operate to this end, it remains always unattainable.
368. The reasons why it is thus unattainable are easily appre-
hended from the remarks previously made. In the first place,
forms and words which have been isolated, by whatever means,
remain untouched by the normalising process. For instance,
a case formed after an ancient model survives as
an adverb
or as a member of some compound, or a participle formed
after an older model survives as a purely nominal form. This
certainly does not interfere with the regularity of the actually
living methods of formation. However, in the second place, it
is a matter of the merest accident whether a partial abolition
of the class-differences by phonetic means, which is so often the
condition precedent to entire unification, appears or not. In
the third place, the capacity for resistance shown by the single
words of the same formation differs considerably according to the
degree of force with which they are impressed on the memory;
and this is why, as a rule, precisely the most indispensable elements
of daily language survive as anomalies. In the fourth place, the
inevitable prevalence of a single class is always the result of a

244
PRINCIPLES OF THE History OF LANGUAGE.
chance convergence of circumstances. As long as this prevalence
does not exist, it is possible for the single words to be impelled
now to one side and now to another, and thus it is possible, owing
to the operation of analogy, for a chaotic confusion to be created,
until the very excess of such confusion conduces to the cure of the
evils in question. In the face of so many adverse circumstances,
it is natural enough that the process under the most favourable
circumstances should advance so slowly that before it has even
approximately come to a termination, new phonetic differences
should have arisen, requiring unification. The same ever-variable
character of the sounds which is indispensable as a spur to the
work of unification, proves also the destroyer of the work set
on foot by it, before that work is accomplished.
369. We can see the truth of this in the relations of the declen-
sions in the NHG. written language. In the feminine the three
main classes of MHG., ¿.e., the ancient i-, a-, and n-declension are
reduced to two (cf. supra, p. 241). As the remains of the con-
sonantal and of the u-declensions (cf., e.g., MHG., hant, plural
hende, hande, handen, hende) have also gradually intruded into
the i-class, we should have two simple and easily-distinguish-
able schemes: (1) the singular without -e, the plural with -e,
and it may be with the umlaut (bank-bänke, hinderniss -
hindernisse); (2) the singular with -e, the plural with -en (zunge
—zungen). However, the non-monosyllabic stems in -er and
-el do not at once accommodate themselves to these schemes
(mutter-mütter, achsel-achseln), having, according to a com-
mon rule already established in MHG., throughout lost their e
(where it was present at all). These would not be very effective as
disturbing factors. But there are, besides, many feminines which
have lost their terminal -e in the singular; all the non-monosyllabic
stems in -inn and -ung, and many monosyllabic, as frau, huld, kost,
FORMATION OF NEW GROUPS.
245
etc. MHG. frouwe, hulde, koste, etc. The process of development
in the case of the latter has probably been this, that originally
in the case of all bisyllabic feminines in -e, duplicate forms
made their appearance, according to their different positions in
the sentence, and that the unification which again set in after
this has had a different result. Besides this, the strife between
Upper German and Middle German for the supremacy in the
written language comes into consideration. However this may
be, a new separation appears: zunge-zungen, but frau-frauen.
And at the same time the clear distinction between the two
classes is an impossibility for the future; frau corresponds to
bank in the singular, in the plural to sunge. This new confusion
was certainly of advantage to the further unification. The contact
between the formation frau and the formation bank entailed the
attraction of a great number of words, indeed of the majority, from
the former into the latter; cf., burg (plural burgen=MHG. bürge),
flut, welt, tugend, etc., and all the words in -heit, -keit, -schaft. By
this method we might have arrived at a uniform way of forming
the plural in -en (n), and the difference between words with
and without e would have subsisted in the singular only. But the
impulse did not hold to the end, and considerable traces of the old
¿-declension remain in antagonism to it.
370. Similar observations hold good for the case of the mas-
culine and neuter; only that, in the case of these, still more
confusing circumstances are combined. In this case, too, the
relations were calculated to produce a sharp inflexional division
between the substantives without -e and those which have -e in the
nominative singular (arm-arme, wort-worte; but funke-funken,
auge-augen) were it not that the rejection of the e in one part of
the word had interfered (mensch-menschen, herz-herzen).
371. The phonetic convergence of forms functionally different,
t
246
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Union in
formal
Where the
function is
different.
comes about within the etymological groups. Thus in OHG. the
groups: (b) transition from final unstressed m to n causes the coincidence of
the secondary termination of the first and third person plural: in
the older sources gâbum-gâbun, gâbîm—gâbîn, in the later for
two persons gâbun, gâbîn. Such coincidence is occasioned on the
largest scale by the weakening of the full final vowels of the OHG.
to a consistent e. Thus, MHG. tage=OHG. tage (dative singular)-
taga (nominative plural)—tago (genitive plural); MHG. hanen, OHG.
hanin (genitive and dative singular)—hanun (accusative singular,
and nominative and accusative plural)-hanôno (genitive plural)—
hanom (dative plural) and in the OHG. forms we find already
evidence of the convergence of forms which differed before. The
convergence does not prevail always through an entire inflexional
class; it need not necessarily affect more than one part of the
words which originally belonged to that class; cf. e.g., tag-tage-
tagen with sessel―sessel - sesseln, winter-winter-wintern, and
wagen—wagen-wagen. This convergence in the case of deriva-
tives from the same base is rarer than in the case of inflexional
forms. Since such derivatives may of themselves merely make up
an entire system of forms, it is possible for the convergence to be a
merely partial one in two directions. It is possible, on the one
hand, that in several classes of words originally differing pho-
netically, only a part of the words may coincide. Thus, for
instance, it is possible in OHG. for two weak verbs to be formed
out of each adjective, viz., an intransitive in -ôn and a transitive in
-en (=Gothic -jan). In MHG. both classes come to coincide in their
terminations, but in the form of the root-syllable this coinci-
dence is only partial, as most of them remain separated by the
presence or absence of the umlaut; cf. on the one side leiden from
leidôn, 'to be unpleasant,' and leiden from leiden, 'to render un-
pleasant;' richen, 'to become rich,' and 'to render rich;' niuwen, 'to
FORMATION OF NEW GROUPS.
247
become new' and 'to render new;' on the other hand, armen, ‘to
become poor'—ermen, 'to render poor,' swâren, 'to become heavy'
-swæren, ‘to make heavy.' On the other hand, there is no neces-
sity for phonetic convergence to spread over all the forms of two
related words. In the case of NHG. schmelzen, two words absolutely
different have converged, viz., smëlzen (with open e) strong and
intransitive, and smelzen with closed e, weak and transitive. The
convergence, however, extends only to the forms of the present,
and even from these the second and third persons of the indicative
singular, and the second person singular of the imperfect are
excluded; cf., schmilzt, schmilz-schmelzt, schmelze.
372. The phonetic convergence of functionally different forms
has further results. One such result is this, that persons grow
so accustomed to phonetic regularity, that they transfer it to
cases where it is not yet warranted by the natural process of
sound-development. In the case of the OHG. verb the first person
plural has taken the same form as the third person plural; this is
owing to the transition of m final into n (gâbun from gâbum—
gabun), with the exception of the case of the indicative present
where the difference has passed, even into the MHG. period; geben
-gebent. This difference disappears in the first place in Middle
German, and then in Upper German as well, as has been already
remarked on p. 107, by the identification of the form of the third
person plural with that of the third person plural of the preterite and
subjunctive. It may be that the process of becoming gradually
accustomed to the coincidence of the first and third person plural
has co-operated to this end. It is, no doubt, the effect of such
process when, in the Alemannic, forms in -ent have been used from
the fourteenth century for the first person plural as well. The
coincidence between the first and the third persons plural appears
even in the written language of to-day in sind = MHG. sîn—sint;
248
PRINCIPLES OF THE HISTORY OF LANGUAGE,
in Upper Saxon, conversely, the third person pl. is sein. Another
example is given us by the coincidence between the nominative
and accusative in German. In original Teutonic both cases were
in the masculine and feminine for the most part still different. It
seems probable that identity merely existed in the plural of the
feminine a-stems (GOT. gibôs, ON. giafar). In OHG., as in the
other West-Teutonic dialects, the nominative singular of the o-,
i-, and u-stems, and of the consonantal stems with the exception
of the so-called weak declension, has become identified with the
accusative by the disappearance of the final s (fisc, balg, sunu, man
= GOT. fisks — fisk, balgs balg, sunus sunu and ON. fiskr
fisk, belgr - belg, sonr—son, maðr — mann); further, phonetic
coincidence has made its appearance in the nominative and accu-
sative plural of the weak declension (hanun, zungûn, OG. probably
*hanoniz - *hanonz). This gives a tendency to further converg-
ence. The form of the nominative plural of the o-, i-, and u- stems
and of the consonantal stem has penetrated to the accusative, and
has thus restored the same agreement as in the singular: taga,
balgi (belgi), suni Gothic dagôs dagans, balgeis balgins,
sunjus sununs, and ON. dagar daga, belgir belgi, synir —
sunu (sonu). The forms which regular sound-change in OHG.
would have led us to expect in the accusative, would have been
*tagun, *balgin, *sunun. In the case of the consonantal stems,
unification has also set in in Gothic and ON.; we may assume that
in original Teutonic such forms as manniz*-mannung* existed =
OHG. man—mannun, which latter form has been driven out by the
former. In the case of the adjective as well, and the pronoun which
marks gender, the nominatival form has affected the accusative:
blinte (-a), die (dia) = GOT. blindai—blindans, þai-pans. In the case
of the feminine a- sterns on the other hand, the phonetic identity of
both cases has brought about an identification in the singular. In
J
FORMATION OF NEW GROUPS.
249
the first place both forms, the nominative and accusative form, were
used indifferently; next the accusative form established itself as a
rule, while the nominative form was confined to definite cases, and
disappeared more and more. While AS. differentiates giefu and ziefe,
ârâre, we have in OHG. only the accusative forms geba and êra,
and side by side as nominative and accusative halba and halp, wîsa
and wis, etc. In NHG. further, in the feminine of the weak adjective,
the accusative form is ousted by the nominatival form: lange
MHG. lange langen; further the feminine nominative form of the
article is ousted by the accusatival form: die = MHG. diu
as early as in MHG. the nominative siu is ousted by sie. In
Rheno-Franconian and Alemannic, on the contrary, we find also
the nominative form of the article der used accusativally.
=
die;
373. If, in any language, a convergence of originally different
case-forms occurs on a large scale, the cause of this may be that
the remains which were spared from this convergence are entirely
or mostly abolished, as has happened in the case of English and
the Romance languages. Thus pure stem-forms again take their
rise as they existed before the case-formation which we incorrectly
call nominative or accusative.
374. When the forms of cognate words partially converge, the
feeling for the difference of these words is rendered less keen, and
thus the forms which have not yet converged are easily confused.
The partial coincidence, touched on above, of MHG. smëlzen and
smelzen has entailed the result that the strong orms schmilzt,
schmolz, geschmolzen are employed transitively as well: the weak
forms have at the present time fallen completely into disuse. In
the same way, the weak forms of verderben, which acquired an
exclusively transitive signification, are ousted by the strong forms,
which were originally transitive only, and can be used at the
present day only in a moral sense. In the case of quellen,
250
PRINCIPLES OF THE HISTORY OF LANGUAGE.
schwellen, löschen, the difference is maintained in what passes for
correct language now-a-days; but from löschen weak forms some-
times come in an intransitive form, e.g., es löscht das licht der sterne
(Schiller); in the case of quellen and schwellen, we find a confusion
in both directions, e.g., dem das frischeste leben entquellt (GOE.) -
gleichwie ein born sein wasser quillt (LU.); schwelle, brust (GOE.)
die haare schwellten (TIECK) die ehrsucht schwillt die brust
(GÜNTHER), seifenblasen, die mein hauch geschwollen (CHAMISSO).
CHAPTER XII.
ON THE INFLUENCE OF FUNCTIONAL CHANGE ON FORMATION
BY ANALOGY.
TH
into a new
changes the
analogical
HE arrangement of the single words and forms, and of the Entrance
syntactical connections between the linguistic groups group
depends in every case upon their function. Thus a functional direction of
change may be the ultimate cause of a transition into another formation.
group. Participation, however, in this group entails participation
in its creative power. Thus there arise new analogical creations
which proceed in a direction different from that which the origin
of the word-form or method of construction would have led us to
believe. The following examples may serve to exemplify this.
the change
lative into
name;
376. The change of an appellative into a proper name causes a Results of
corresponding change of declension; cf. the accusatives and of an appel-
datives of such words as Müllern, Schneidern, Beckern, etc. It was a proper
a consequence of the monotheism introduced by Christianity, that
from got an accusative gotan was formed in OHG. after the analogy
of proper names. We may compare with this such datives and
accusatives as vatern and muttern which we commonly hear in
Berlin.
into an
377. The Greek adverbs in -ws were originally cases of the of a case
o-declension. But when they had once become detached from the adverb;
inflexional system, and -ws felt as a formative word-suffix, it
was able to attach itself to other stems which had received no
252
PRINCIPLES OF THE HISTORY OF LANGUAGE.
of a syntac-
tical group
word.
influence from the o-stems; cf. such cases as ndéws, owopóvws, etc.
The case is similar with the adverbial suffix -0 in OHG., which has
been likewise transferred from the o-stems to the old - and μ-
stems: kleino and harto, following liobo, etc.
378. There are in NHG. many adverbs which are in their origin
genitives singular of nouns; such are falls, rings, rechts, stracks,
blindlings. But the s has long since ceased to be felt as a sign of
the genitive; it must now appear as an adverbial suffix. Conse-
quently it has since the seventeenth century been transferred to
other adverbs, which are in their origin cases of nouns, or unions of
a preposition with a case, but are likewise not felt as such, but
have passed into the general category of adverbs; cf. allerdings
(from aller dinge, gen. pl.) schlechterdings, jenseits, disseits (MHG.
jensît, acc. sing.), abseits (from ab seite), hinterrücks, in the seven-
teenth century also hinterrückens (from an old form hinterrück,
hinterrücken), unterwegs, unterwegens (from unter wege, unter
wegen), vollends (earlier vollen, vollend); erstens, zweitens, etc.
The change of the s from a case-suffix into a formative element of
words has also made it possible for it to be adopted in derivations :
desfallsig, allenfallsig.
379. Hans Sachs forms a comparative flüchser from flugs. This
results from the fact that the substantival case has come to stand
on the same line with the adjectival adverbs which alone originally
admit of comparison.
380. If a syntactic combination has passed into a verbal unit,
into a single then this new unit is treated after the analogy of the simple word,
and the possibilities affecting this are transferred to it. It occurs
in many languages that an inseparable particle attaches itself to a
pronoun. The result may be that the inflexion is transferred after
the model of the simple words from the middle to the end.
Plautus employs still the accusatives eumpse, eampse, and the
INFLUENCE OF FUNCTIONAL CHANGE BY ANALOGY. 253
ablatives eopse, eapse, from i-pse, which forms are later replaced by
ipsum, etc. A similar development has been undergone by the
German pronoun diser, as may be proved especially by the ON.
Runic forms. This word is a compound of the article and the
particle se. Language is notably enriched by the fact that out of
such compounds, which owe their origin to secondary fusion, the
same derivations are formed as from simple words, and that they
are able to serve like these, as the member of a compound: cf.
überwinder, überwindung, ergiebig, befahrbar, gedeihlich, betrübniss,
gevangenschaft, befangenheit; edelmännisch, hochmütig, jungfräulich,
landesherrlich, landsmannschaft, grossherzogtum, bärenhäuter, kinder-
gärtnerinn; sofortig, bissherig, jenseitig; rotweinflasche, gänseleber-
pastete; überhandnahme, vorwegnahme, zurücknahme.
tion.'
381. Not unfrequently an inflexional form becomes fixed, when 'Crystallisa-
it is transferred to cases with which it has strictly speaking nothing
to do.¹ 1 The German selber 2 is the nominative singular masculine,
and at the same time, the genitive and dative singular feminine
and genitive plural, of an older adjective selb, which is preserved as
an adjective still only in the word der selbe. The word selbst =
older selbes, of the same signification, is the nominative and
accusative singular neuter, and at the same time the genitive
singular masculine and neuter, of the same word. In MHG. the
adjective is inflected sometimes strong, sometimes weak, and
follows the noun with which it is connected in gender, number, and
case; cf. im selbem, ir selber, sîn selbes, etc. Now when the forms.
received in MHG. have forced themselves into places where others
properly belong, this must result from the fact that the word is no
longer felt as an adjective. Since, in the word selber, nothing more
was felt than the function of an energetic identification, the form
came to be employed in every case where such identification had
1 Cf. Brugmann, Ein problem der homerischen text-critik, p. 119 sqq.
2 Cf. Voller, lauter, etc.
254
PRINCIPLES OF THE HISTORY OF LANGUAGE.
to be affirmed. The case of the dialectic halber is parallel: die
nacht ist halber hin, es ist halber eins; and it is the same with
einander, instead of which we find in OHG. regularly inflected: ein
anderan, ein andermo, etc. In MHG. it was still possible to say
beider des vater und des sunes in which phrase des vater und des
sunes stands strictly speaking in an appositional relation to beider.
More commonly, however, we find beide, des vater und des sunes.
We thus find the nominatival form beide crystallized, as the origin
of the construction presents itself no more to our consciousness and
the function of beide — und approaches the modern German sowohl
- als auch. In Latin the nominative quisque coupled with the
reflexive pronoun and the possessive which corresponds to it has
passed beyond its proper area, cf. e.g., multis sibi quisque imperium
petentibus. In Plautus we find praesente testibus for præsentibus, in
Afranius absente nobis: from this we must acknowledge that the
participial forms in question have approached the characteristics of
prepositions. Combinations like agedum conferte, agedum creemus,
etc., result from the fact that age was no longer felt as the second
person singular imperative, but merely as a general cry of exhorta-
tion. Correspondingly we find in Greek aye used before a plural
as we do eiπé, pépe, idoú;¹ further, in Latin we find cave dirumpatis
(Plautus), etc.; in ordinary German conversation we hear some-
times such expressions as warte mal, even where the words are
addressed to several persons, or to one only who is commonly
addressed by Sie. In the older stages of NHG. we find siehe
employed even where more than one are accosted; the French
voici and voilà are completely crystallized. In late Greek peλov
ὠφελον
and wþeλe were employed, without any consideration of number or
person as simple conjunctions. The German nur took its origin
from enware (es wäre denn). Thus this enware has forced its way
in the place of enwærest, enwæren, enî, ensîn, etc.
1 Brugmann, u.s., p. 124.
INFLUENCE OF FUNCTIONAL CHANGE BY ANALOGY. 255
1
t
E
1
382. It is a similar process when in late MHG. the word sich
dependent on a preposition, makes its way also into sentences
in which the subject is the first or the second person.¹ This
results from the fact that an expression like über sich or unter
sich is no longer analysed, but is apprehended as = upwards
or downwards; cf. the NHG. expressions vor sich gehn and an und
für sich. And hence these combinations are also employed in
cases where they cannot be referred to the subject, but only to
an oblique case; cf. heb hinten über sich das glas ('raise your glass
high,' Uhland, Volkslieder). The same crystallization has occurred
in such expressions as seiner zeit; cf. e.g., die jugend ist unterneh-
mend, wir sind es seiner zeit auch gewesen (HACKLÄNDER). Cor-
respondingly we find in Latin suo loco, sua sponte, suo nomine. In
Roman jurists we find combinations like si sui juris sumus. In
ON. a middle and passive is formed by the aid of the reflexive.
In this case the -sk, later -z, which goes back to -sik, and could
originally be applied to the third person only, is transferred in
the first place to the second, and afterwards to the first person;
e.g. lúkoms, instead of an older lukomk (=*luko-mik); the z was
no longer apprehended in its original meaning, but merely as a
sign of the middle or passive. In very many Upper and Middle
German dialects the word sich is also employed as a reflexive
for the third person plural; occasionally indeed for the second
person as well. The ordinary practice of confining its use to the
first person plural is probably to be explained by the fact that
by this restriction the process of transference is rendered more
easy,2 owing to the formal correspondence of the verbal form with
1 Cf. Brugmann, u.s.
2 The view put forth by Brugmann, u.s., p. 123, that this sich has risen from a
form unsich seems to me untenable, because the form unsich had already disappeared
before this application of sich appears at all. The wide area covered by the phenomenon
forbids our assuming with Weinhold, Bair. Gram. § 359, and Schuchardt, Slawodeutsches,
p. 107, any Slavonic influence.
256
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Influence of
change of
meaning
upon con-
struction.
the third person plural.
In the Bavarian dialect the possessive
sein is made to refer to the feminine, and to the plural as well;
cf. Schmeller, p. 198.
383. Plautus [Catullus, and other poets], employ the words
perire, deperire, and demori in the sense of to be mortally enamoured
of, with the accusative; in the same way Vergil, Horace, and other
authors employ ardere in the sense of to be inflamed by love for.
It is plain that the construction of these words is influenced by
that of amare, because their metaphorical use approximates to the
proper use of amo. We may fairly conclude that this signification
of these words was already somewhat obsolete, at least in the
language of the poets. For if their full significance had been
still completely appreciated, it is probable that no such change
in the construction would have set in. Still, in a case like this, we
have always to set something down to the account of an intentional
audacity of language on the part of the poet. It is different with
the ordinary language of prose. In this case, too, it frequently
occurs that a word changes the method of its construction which
should characterise it according to its fundamental meaning for
another which does not suit this, since it is influenced either by a
particular single word or by a group of words to which it has
in the course of time approximated its meaning. In this case the
change of construction is an infallible mark of the disappearance
of the fundamental meaning. Especially often do we find thus
indicated the detachment from the concrete perception which
originally lay at its root.
384. We find many compounds with adverbs of place which
For
are especially instructive for this process of detachment.
instance, the preposition in belongs originally to the words.
einwirken and einwirkung; and this use was common up to the
last century; cf. sobald kunst und wissenschaft in das leben einwirkt
INFLUENCE OF FUNCTIONAL CHANGE BY ANALOGY. 257
(GOETHE); durch die einwirkung in gewisse werkzeuge (GARVE).
These words are at the present day connected with auf, and this
proves that the feeling for the sensuous perception, to which the
ein points, has disappeared. The same change has made its
appearance in the case of einfluss; cf. gesundheit ist ein gut, welches
in alles einfluss hat (GARVE); and this was the general usage in the
last century (in and auf were formerly connected with einfliessen 'to
have influence,' as well); einschränken, cf. es hat längst aufgehört
in die engen grenzen eingeschränkt zu sein (LESSING), etc.; eindruck,
cf. die nähe des schönen kindes musste wol in die seele des jungen
More
mannes einen so lebhaften eindruck machen (GOETHE).
sensuous still is um durch das grosse dieses todes einen unauslösch-
lichen eindruck seiner selbst in das herz seiner Spartaner zu graben
(SCHILLER); though it already appears with auf in Lessing.
Abneigung gegen, or, as the older writers also express it, vor, can-
not have been an original usage, but von only, which Sanders cites
Heine alone to prove. The earliest instance of nachdenken über
seems to occur in Schiller's Don Karlos; in other cases the simple
dative (properly dependent on nach), is common; e.g. um ihren
briefen nachzudenken (NICOLAI).
385. If we now say in German sei mir willkommen in meinem
hause, it is evident that the last-named component part of the word
is no longer apprehended as a participle of kommen. As long as
it was so apprehended the denotation of a certain direction was
also understood; e.g. willekomen her in Guntheres lant (NIBE-
LUNGENLIED).
386. Such an expression as quin conscendimus equos is, properly
speaking, why do we not mount our horses? but is understood as
let us mount our horses; and hence it follows that we are able
to employ even after quin an imperative or so-called adhortative
subjunctive; e.g. quin age istud, quin experiamur. Correspondingly
R
<
258
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the MHG. wan fürchtent si den stap (warum fürchten sie nicht den
stab), approximates to the sense of mögen sie den stab fürchten;
as a consequence of this we find after wan the ordinary subjunctive
preterite employed in optative clauses without any introductory
conjunction; cf. wan hæte ich iuwer kunst. It seems probable
that the OF. usage of car (=quare) connected with the con-
ditional and the imperative is to be thus explained; cf. DIEZ,
iii. 214.
387. The Greek oỦкoûv is originally = 'not therefore,' and
serves to introduce a question to which an affirmative answer
is expected. The sentences introduced by oùкoûν come, however,
gradually to be apprehended as direct positive assertions. And
thus the particle has retained merely the function of marking
consequence, and it is introduced in sentences which can no
longer be apprehended as interrogative sentences, e.g. coupled
with the imperative; cf. οὐκοῦν ἀπάγαγέ με αὖθις ἐς τὸν βίον
(LUCIAN).¹ The Sanskrit na-nu shows a precisely parallel develop-
ment. It serves, in the first instance, like nonne, to introduce
interrogative sentences; but then, as such interrogative sentences
are transformed into affirming sentences, it may be translated
by surely, and it occurs next in petition sentences; cf. nanu
ucyatam = it seems to be said.
388. The accusative with an infinitive could originally stand
only in connexion with a transitive verb, as long as the accusative
of the subject was felt as directly dependent from the finite verb;
cf. Chapter XVI. After, however, the interpretation of it had so
far changed, that the accusative and infinitive was looked upon
as a dependent sentence, and the accusative as its subject, it was
possible to extend the construction far beyond its original limits.
Thus in Latin, too, verbs are construed with the accusative and
¹ Cf. Kühner's Griech. Gram., II. i. p. 717.
INFLUENCE OF FUNCTIONAL CHANGE BY ANALOGY. 259
infinitive which cannot have any accusative of the object connected
with them, as gaudere, dolere; and further, we find combinations
like magna in spe sum, spem habeo, etc. In very many cases it
then happens that the accusative and infinitive is employed as the
subject; thus after such words as licet, accidit, constat, etc., after fas,
jus est, etc., and in the case of passives with the nominative and
infinitive; cf. non mihi videtur ad beate vivendum satis posse virtutem
(CICERO); Volscos et Aequos extra fines exisse affertur (LIVY).
Then the accusative and infinitive construction passes into sentences
which depend on another accusative and infinitive. Thus, in the
first place, it passes into relative sentences loosely connected; e.g.
mundum censent regi numine deorum, ex quo illud natura consequi
(CICERO); cf. Draeger, § 447, I. Further, it passes into sentences
of comparison: e.g. ut feras quasdam nulla mitescere arte, sic
immitem ejus viri animum esse (LIVY); addit etiam se prius
occisum iri ab eo quam me violatum iri (CICERO); cf. Draeger,
448, I; 453, 2. It passes also into indirect questions: e.g. quid sese
inter pacatos facere, cur in Italiam non revehi (LIVY); cf. Draeger,
450. It even passes into temporal and causal sentences: crimina
vitanda esse, quia vitari metus non posse (SENECA); cf. Draeger,
448, 2, 3. A corresponding extension is met with in Greek. The
custom of having the subject of the infinitive in the form of the
accusative conduces in this instance, too, to the employment of
this case by the side of the infinitive where the latter is turned
into a substantive by the article, in whatever case this may be;
cf. αἴτιος τοῦ νικηθῆναι τοὺς Λακεδαιμονίους, διὰ τὸ τὴν πόλιν
ᾑρῆσθαι, ὑπὲρ τοῦ ταῦτα μὴ γίγνεσθαι.
differently
389. If two methods of construction partially cover each other Construction
in their functional use, it is possible, in the case of many tradi- understood
tional syntactic combinations, for an uncertainty to arise as to influence of
which of the two is the fundamental one. In this way there
through
a synonyme.
260
PRINCIPLES OF THE HISTORY OF LANGUAGE.
arises a new interpretation of the combination, which turns the
activity of analogy into another channel.
390. A genitive dependent on a substantive has a function
similar to that of the attributive adjective. In combinations like
Hamburger rauchfleisch or Kieler sprotten, the genitive of the
designation of the inhabitants is at the base as the first member;
but it is more in accordance with the instinct of language to
apprehend it as an adjective derived from the name of the place;
in any case we refer it directly to the place and not to the
inhabitants. No doubt the absence of inflection proves that we
have no true adjective before us. On the other hand, the way in
which the article is employed in the connexion (das Hamburger
rauchfleisch) shows that the genitive is not felt any longer as
such; for the position of a genitive between the article and the
substantive has now become an impossibility. OHG. had no
possessive pronoun of the feminine and the plural sie.
It em-
ployed, instead of this, the genitive of this pronoun, ira, iro.
In MHG., too, we have the genitive ir; but here and there the
usage has set in of apprehending it as an adjective, and declining
it adjectivally. This use has become general in NHG., and this
is the origin of the German possessive pronoun ihr. It seems
probable that the contact of the genitive with the attributival
adjective was the occasion of employing it after the model of
the adjective predicatively; cf. er ist des todes, reines herzens, so
sind wir des herrn (LUTHER), etc.. This use belongs most
probably to the IE. fundamental language.
CHAPTER XIII.
DISPLACEMENTS IN THE GROUPING OF WORDS ETYMOLOGICALLY
CONNECTED.
IF
ing of ety-
connected
diverges, in
time, from
again that required
by their
itself formation,
hence
creation on a
F we collect and arrange all the words and forms which contain The group-'
the same root according to the original laws of formation, mologically
as they are found by the dissecting process proper to the older words
comparative grammar, we obtain a system compounded of many the course of
members, or a larger system of smaller systems which may
on their side consist of systems. Even a single IE. verb in
represents a highly complicated system. From the verbal stem analogical
are developed different stems marking tense; from each tense-stem new basis.
different moods; and then, and not till then, are the different
persons in both genders developed. It is the endeavour of
analytical grammar to separate what by its origin is closely re-
lated from what is related only in a further degree; always to
distinguish between root-word and derivative; to avoid all sudden
leaps and bounds, and not to apprehend anything as a direct
derivation which is only a derivation from a derivation. But that
which, judged from its point of view, is an error in the judgment
formed as to word- and form-composition, is a thing to which
linguistic consciousness is extraordinarily liable. It is unavoid-
able that the way in which the etymologically connected forms
group themselves together in the mind of the members of any
linguistic community should, at a later period, prove quite other
262
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Examples.
than when the forms first took shape, and the result of this is
that the formation by analogy as well, which reposes upon irregular
grouping of this kind, diverges from the original laws of formation.
Secondary coincidence of sound and signification also often adds.
its influence. The importance of the part played by this process
in the history of language may be shown by a list of examples.
-
392. There are in NHG. a large number of masculine names of
action traditionally bequeathed to us side by side with strong verbs
corresponding therewith; cf. fall-fallen, fang-fangen, schlag —
schlagen,streit — streiten,lauf— laufen, befehl (OHG. bífelh) — befehlen.
If we go back to the original principle of the formation, we shall be
compelled to say that neither the noun is derived from the verb nor
the verb from the noun, but both are immediately derived from the
root. We have further certain cases in which, side by side with a noun
of the agent, there stands a weak verb derived from it; cf. hass
hassen, krach- krachen, schall schallen, rauch rauchen, zil -
zilen, mord-morden, hunger- hungern. In NHG. these two classes.
cannot be kept apart, and chiefly for this reason, that the difference.
of the verbal terminations in the present has entirely disappeared.
The words schlag and hass seem to bear a precisely similar relation-
ship to schlagen and hassen; and the usage is now to form other
nouns similarly connected with other verbs, irrespective of the class
of conjugation to which they may belong, simply by omission
of the termination; cf. betrag, ertrag, vortrag, betreff, verbleib,
begehr, erfolg, verfolg, belang, betracht, brauch, gebrauch, verbrauch,
besuch, versuch, verkehr, vergleich, bereich, schick, bericht, ärger, etc.
In MHG. we have side by side with the substantive gît a verb gîtesen
derived from it. The latter in late MHG. develops regularly into
geitzen, geizen; and from this is formed the substantive geiz, which
thrusts aside the older form geit.
393. Where a noun and a verb of similar signification stand.
DISPLACEMENTS IN ETYMOLOGICAL GROUps.
263
side by side, it is inevitable that the derivative formed from the one
should proceed to form relations with the other as well, so that it
may seem to the instinct of language to be formed from the latter
as well as from the former; and this relationship deviating from
the original condition may then give an impulse to new formations.
The German suffix -ig (OHG. -ag and -îg) serves originally merely
to mark derivatives from nouns. But words like gläubig, streitig,
geläufig stand, as far as their form and signification goes, in just as
near relation to glauben, streiten, laufen as to glaube, streit, lauf;
while others, like irrig, stand in even a nearer relation to the verb.
in question, because the substantive irre in the development of its
meaning has not proceeded in a parallel path with the ad-
jective; in the case of others like gehörig, abwendig, the substantive
which lies at the base of the word (MHG. hôre) has been lost, or at
any rate is no longer in ordinary use. In the same way we have
next a quantity of adjectives formed straight from verbs; cf.
erbietig (as against the noun-derivative erbötig), ehrerbietig, freigebig,
ergiebig, ausfindig (this word seems, however, formed after the MHG.
fündec), zulässig, rührig, wackelig, dämmerig, stotterig: the word
abhängig too must, according to its signification, be referred not to
hang, abhang, but only to abhangen. The circumstances are similar
in the case of the adjectives in isch, of which at least the words.
neckisch, mürrisch, and wetterwendisch must be apprehended as
derivatives from verbs, after the model of such words as neidisch,
spöttisch, argwöhnisch, etc. The modern German suffix -er (OHG.
-âri, -eri, MHG. -ære, -er), which at the present day serves as the
general method of forming nouns of the agent from verbs, was
originally applied only to such formations as we have still in
bürger, müller, schüler, and many other words. In Gothic the words.
bokareis, 'a scribe,' from boka (in the plural 'book'); daimonareis,
'one possessed,' from Saíµwv; motareis, 'publican,' from mota, ‘toll;'
264
PRINCIPLES OF THE HISTORY OF LANGUAGE.
vullareis, ' cloth-fuller,' from vulla, 'wool;' liupareis, ' singer,' from
an assumed word *liup OHG. leod, NHG. lied. In the same way
we shall have to derive the words laisareis, 'teacher,' and sokareis,
'inquirer,' not from the verbs laisjan and sokjan, ‘to seek' (suchen),
but from assumed substantives *laisa, = OHG. lera, NHG. lehre,
[English 'lore'], and *soka = MHG. suoche. These two last-men-
tioned verbs, however, exhibit already the possibility of bringing
the formation into connexion with a verb. By the side of liuþareis
stands also liupon, 'to sing.' Thus, from association with instances
like these, the derivations from verbs begin in OHG. already. We
can see that the nominal derivation is the original, especially from
such instances as zuhtâri'educator,' from suht, and not from
ziuhan; and again notnumftâri, 'robber.' In the cases in which the
root-vowel of the nominal derivative does not agree with the
present of the verb, it often happens that a new verbal formation
appears beside it; and both formations have maintained themselves
side by side down to the NHG. time; cf. ritter reiter, schnitter -
schneider, nähter -näher, mähder — mäher, sänger — singer (OHG.
only sangari), Schilter (as a proper name) = MHG. schiltære
('painter') — schilderer. The abstracts in OHG. ending in -ida
Gothic -ipa) seem to have been originally formed from adjectives
only, and from verbs as a result of secondary relation only; cf.
kisuohhida with kisuohhen, pihaltida with pihaltan after chundida
chunden-chund, etc.
394. As it is in derivation so is it in composition. The gradual
process of change of the first portion of a nominal compound into
a verbal, and the new formations called into existence thereby, have
been treated in great detail by OSTHOFF. Thus for instance, OHG.
waltpoto 'procurator,' sceltwort, betohus, spiloman, fastatag, wartman,
I
¹ Das verbum in der nominalcomposition im deutschen, griechischen, slavischen und
romanischen. Jena, 1878.
DISPLACEMENTS IN ETYMOLOGICAL GROUPS.
265
spurihunt, erbereht, which must be allowed to contain the nouns
walt (giwalt), scelta, beta, spil, fasta, warta, spuri, erbi enter into
direct relationship with the verbs waltan, sceltan, betôn, spilôn, fastên,
wartên, spurien, erben; and from these and similar components
arises the class which has become so common in German of com-
pounds with a verbal noun, forming the first portion of the com-
pound, as esslust, trinksucht, schreibfeder, schreibfaul, etc. Under
this head come especially many compounds with -bar, -lich, -sam,
-haft),¹. which, however, from the standpoint of linguistic instinct
are rather to be regarded as derivations, and to be ranged parallel
to the formations with -ig and isch mentioned above; cf. words.
like wählbar, unvertilgbar, unbeschreiblich, empfindlich, empfindsam,
naschhaft. The transition shows itself with special clearness in
the case of words like streitbar, wandelbar, vereinbar. Streitbar
may just as well be referred to streit as to streiten; but unbestreit-
bar can only be referred to bestreiten. In MHG. wandelbære is
referred straight to wandel; and as this word signifies commonly
'a spot,' the word commonly signifies accordingly 'marked by a
spot;' in NHG. on the contrary we find wandelbar, unwandelbar
wholly attracted to the meaning of the verb wandeln. In MHG.
we have an adjective einbære, (einträchtig 'concordant '), with no
relation to the verb.
two suffixes.
395. The case is very frequent that a derivative from another de- Fusion of
rivative is placed in direct relationship to the root-word, by which
process it then happens that really direct derivatives are produced
with the fusion of two suffixes into a single one. This explains,
for instance, the rise of the NHG. suffixes -niss, -ner, -ling. In
Gothic there still survived a distinct suffix -assus, ufar-assus (over-
ness abundance). This is, however, most commonly employed
for formations from verbs in -inon; e.g. gudjinassus (priesteramt,
=
1 Cf. Osthoff, p. 116 sqq.
266
PRINCIPLES OF THE HISTORY OF LANGUAGE.
'post of priest) from gudjinon, ' to perform priestly functions.' As
soon as this word was referred directly to gudja, a priest,' -nassus
must have been felt as a suffix. Further, an n was found in such
formations as ibnassus from ibns (eben, 'even '), and in derivations
from participles such as OHG. farloran-issa (forlorn-ness). It has
thus come to pass that the West Teutonic dialects always, with the
exception of a few antiquated fragments, exhibit an n coalescing
with the suffix. The formations in -ner start from noun-items which
contain an n; cf. gärtner (MHG. gartenære), lügner (MHG. lügenære
from ligene by the side of lüge), hafner (MHG. havenære), wagner,
or from verbs terminating in OHG. in -inôn; cf. redner (OHG.
redinûri from redinôn), gleissner (MHG. gelîchsenære from gelîchsenen).
When then lügner, for instance, is referred to lüge, and redner to
rede, reden, the suffix -ner arises, which we find in bildner (as early
as the fourteenth century, bildenære, but earlier still bildære)
harfner (MHG. harpfære), söldner (late MHG. soldenære, earlier soldier.
In künstler (MHG. kunster) -ler also appears as a suffix, for we refer
it directly to kunst, because the verb künsteln, from which it strictly
speaking comes, is confined to a more special meaning. The suffix
-ling (in pflegling, zögling, etc.), proceeds from formations like the
OHG. ediling 'the noble' from edili or adal; chumiling (NHG. in
abkömmling, ankömmling) connected with (uo-)chumilo. Thus be-
tween jung and jungilinc there must probably have stood once a
diminutive form or *jungilo.
396. The NHG. verbs in -igen originated in derivatives from
adjectives in -ig. MHG. einegen, huldegen, leidegen, nôtegen, manec-
valtegen, schedegen, schuldegen doubtless arise from einec, huldec,
leidec, nôtec, schadec, schuldec, etc.; but the NHG. vereinigen,
beleidigen, beschuldigen are rather to be directly referred to ein, leid,
and schuld: and in the case of huldigen and schädigen, no other
reference is possible than to huld and schade, because the adjectives
DISPLACEMENTS IN ETYMOLOGICAL GROUPS.
267
that forwarded the transition have disappeared, and nôtigen in the
same way, because nôtig no longer corresponds in meaning. Thus
then others appear directly derived from the substantive, such as
vereidigen, befehligen, befriedigen, einhändigen, beherzigen, sündigen,
beschäftigen, or from simple adjectives, as beschönigen, senftigen,
genehmigen. The verbs ending in -ern and -eln arose from a
nucleus of derivations from nouns ending in OHG. in -ar and -al
(-ul, -il), the OHG. spurilôn, (investigare) for instance, going back
not to the verb spurien, but to a presumable adjective *spuril
= OHN. spurall); but they are at the present day derived simply
from more simple verbs, cf. folgern, räuchern (late MHG. rouchern,
more anciently rouchen), erschüttern (MHG. even in the sixteenth
century erschütten), zögern (from MHG. zogen), schütteln, lächeln,
schmeicheln (from MHG. smeichen, etc.) In the same way
derivatives from nouns like äugeln, frösteln, näseln, frömmeln
klügeln, kränkeln, etc., have been formed.
(
397. In MHG. many adjectives form an adverb in -lîche, cf.
frôlîche, grôzliche, lûterlîche, eigenlîche, vermezzenliche, sinneclîche,
einvalteclîche. Forms like this are of course derived in the first
instance from adjectival compounds. But as the adverb of the
simple adjective becomes obsolete, a direct relation sets in between
the adverb of the compound and the simple adjective. In fact, the
development proceeds even further, as after the analogy of
grimmeclîche, stæteclîche, etc., which are referred directly to grim
or grimme and state, armecliche, miltecliche, snellecliche, etc. are
formed, although no such word as armec, etc. is in existence.
English abverbs in -ly have the same origin.
The
398. Similar processes clearly came into operation at a period
so early that we are unable to trace their gradual development.
We find in the different IE. languages, in the earliest period of their
development known to us, a large number of suffixes whose
1
I
1
Displace-
ment of
compounds.
268
PRINCIPLES OF THE HISTORY OF LANGUAGE.
phonetic arrangement proves to us that they are complications of
several simple suffixes, and which probably one and all took their
rise in such a way, that in the way indicated a derivative of the
first has passed into a derivative of the second degree.
399. Further, the relation of compounds to each other gives
relations in occasion to many displacements of the relations. If two related
words enter into composition with a similar element, it is almost
inevitable that a direct relation between the two compounds must
arise, and the consequence follows that one of them is apprehended
no longer as a compound, but as the derivative of a compound. Con-
versely it is possible for a derivative from a compound to be placed
in direct relation to the corresponding derivative from the simple
word, and the result is that it is apprehended as a compound.
400. The history of composition in German gives rich material
for the illustration of these processes. Originally a sharp distinc-
tion was made between verbal and noun composition. In verbal
composition prepositions alone are employed, as the first members.
of a compound in noun-composition, noun-stems, and adverbs-
at first only such as were identical with prepositions; though at a
later period others as well. In the case of the verbal compounds,
the accent falls upon the second component portion of the word;
in case of the noun-compounds it falls upon the first portion. It is
thus the accent which, in the case of composition with particles, is
the distinguishing sign. The case often occurs that a verb and a
noun of action connected therewith are compounded with the same
particle. There are numerous cases of such in which the old
relationship is maintained down to the present day in spite of the
parallelism of meaning between the two compounds; cf. durch-
1
1 However, generally speaking, the noun compounds have a tendency to follow the
attraction of the improper verbal compounds, and this precisely because of the similarity
of accent, whereas from the proper ones substantives in -ung are derived, cf. durchfahren
dúrchfahrt
durchfahrung, etc.
—
durchfahren
=
1
11
1
I
F
DISPLACEMENTS IN ETYMOLOGICAL GROUPS.
269
"
bréchen — durchbruch, durchschnéiden — durchschnitt, durchstéchen
-durchstich, überblicken - überblick, überfallen - ü'berfall, über-
gében — übergabe,- übernehmen - i'bernahme, überschauen
1
schau, überschlagen
—
überschlag, übersehen
über-
übersicht, überziehen
überzug, umgehen - limgang, unterhalten — unterhalt, unter-
scheiden- unterschied, unterschreiben unterschrift, widerspréchen
widerspruch, etc. In other cases the difference of accentuation
has produced a different arrangement of the sounds of the particles,
by which process the verbal and nominal compound stand out in
even sharper contrast to each other. In this case the old condition
is maintained only in some few cases, where the development of
meaning has not been parallel, as in the case of erlauben — urlaub,
erteilen urteil. In MHG. we have besides empfangen — ámpfanc,
enthéizen-ántheiz, entlä' zen-ántlâz, entságen-antsage, begraben
-bígraft, besprechen bisprache, bevahen- íbvanc, erheben —
úrhap, ersten - úrstende, verbieten vi'rbot 'citation before
Justice,' versetzen — virsaz (versetzung, 'pledge'), verziehen, —–—
pii'rzoc, etc. In all these cases when the words have maintained
themselves at all, we find this discrepancy cancelled by the
attraction of the noun-compound to the verb: empfang, verzug,
etc. In other cases the unification set in as early as in MHG.,
and the particle ga- (NHG. ge-) is always unaccented, at least
as early as in OHG., if not at as early a date as in original-
Teutonic. It is clear that in this process the relation of the
verbal compounds to the noun-derivation thence formed has
aided in this operation (MHG. erlasen - erlæsære, erlæsunge, etc.),
which on their side are merely analogical formations from simple
verbs. There is the infinitive and participle as well, which
in
many cases develop into simple nouns (cf. NHG. behagen,
belieben, erbarmen, vergnügen; bescheiden, erfahren, verschieden,
etc., and the substantives compounded of the latter (cf. gewis-
270
PRINCIPLES OF THE HISTORY OF LANGUAGE.
sen, bescheidenheit, bekanntschaft, verwandtschaft, erkenntniss, etc.)
co-operate.
401. On the other side, the principle that a verbal compound
cannot contain a name, is severely damaged, to the instinct of
language, by the fact that derivatives like handhaben, lustwandeln,
mutmassen, nottaufen, radebrechen (proved by its weak inflexion to
be a derivative of MHG. -breche), ratschlagen, wetteifern, argwöhnen,
notzüchtigen, rechtfertigen, verwahrlosen from handhabe, notzucht,
rechtfertig, etc., as well as the word weissagen misinterpreted by
popular etymology (OHG. wîzagon from the adjective wîzag, made
into a substantive wîzago 'the prophet'), can also be apprehended
as compounds. Perhaps the coalition of syntactical groups into
compounds (lobsingen, wahrsagen) is thereby aided.
1
1
402. Another curious instance of the displacement of the
relations in composition is found in numerous examples in Late
and Middle Latin, and in the Romance languages. We have here
a large number of verbs which are either actually derived from the
connexion of a preposition with its case, or at least appear accord-
ing to their meaning to be so derived, cf. accorporare (ad corpus),
incorporare, accordare, excommunicare (ex communione), extemporare
(extemporalis is found as early as the first century after Christ ;)
emballer, déballer, embarquer, débarquer, enrager, affronter, achever
(ad caput), s'endimancher 'to dress one's-self in Sunday costume,'
s'enorgueillir.¹ The words formed from adjectives are connected
with these, which signify' to place oneself in the condition implied*
by the word;' such are affiner, enivrer, adoucir, affaiblir, ennoblir, etc.
The original basis for these formations was twofold. On the one
hand, there were derivatives from compound nouns, cf. assimilis
assimilare, concors concordare, deformis deformare (with the
¹ More examples are given by Arsène Darmesteter, Traité de la formation des mots
composés dans la langue française (Bibliothèque de l'école des hautes études. Sciences
philologiques et historiques 19) Paris 1875, p. 80 sqq.
DISPLACEMENTS IN ETYMOLOGICAL GROUPS.
271
meaning of 'to deform),' degener
exanimis
exanimare, exheres
exsucus exsucare, demens
degenerare, depilis depilare,
exheredare, exossis - exossare,
dementire, insignis insignire;
which stand to each other like sanus sanare; further, dedecus
dedecorare. On the other side there are compounds from denomi-
native verbs like accelerare (celerare is poetical), adaequare, addensare,
aggravare, aggregare, appropinquare, assiccare, attenuare, adumbrare,
dearmare, decalvare, dehonorare, depopulari, despoliare, detruncare,
exhonorare, exonerare, innodare, inumbrare, investire. Both classes
had gradually to be fused with each other, and especially in the
case where in the first class the noun which lay at the foundation,
in the second the simple noun, became obsolete.
CHAPTER XIV.
Causes of
superfluity
in language.
Tendency to
WE
ON THE DIFFERENTIATION OF MEANING.
E have seen that it is an essential characteristic of the
development of language that each single language
is incessantly engaged in developing a plurality of words, forms,
and constructions of similar meaning.¹ One source of this pheno-
menon we have found in formation by analogy, a second in
the converging development of meaning from different sides;
we may add as a third the acceptation of a foreign word for a
conception which is already represented by a native word (cf.
vetter cousin, base-cousine); and under this head we must of
course range words borrowed from a kindred dialect.2
404. Inevitable, however, as the rise of such superfluity is,
eliminate it. it is as unable to maintain itself permanently. Language abhors
luxury. The objection will not hold good that, if this were so,
she would not permit it to arise. It is not the habit of language
to adopt precautions against the occurrence of possible disasters,
but merely to take measures against such as have actually occurred.
The individual speakers who add new linguistic forms to those
of equivalent meaning already existing take no account, at the
moment when they do so, of the latter, since these are either
unknown to them, or at least do not enter their consciousness at
¹ [Excellent illustrations of this chapter from the French are found in Darmesteter's
La Vie des mots, Chap. iv., Synonymie.]
2 [In English the case is somewhat different, at any rate as regards Norman-French.]

ON THE DIFFERENTIATION OF MEANING.
273
the moment in question. It is, as a rule, only other persons who,
hearing the new form from one and the old from another of their
interlocutors, use both indiscriminately.
405. Our assertion applies in any case absolutely to the lan-
guage of common life. It applies somewhat less to the language
of literature, and less to the language of poetry than to that of
prose. But the deviation merely confirms our fundamental view,
that a need, and the means of satisfying that need, are always
striving to adjust their relations; a process which equally involves
the rejection of superfluities and the filling up, as far as may be,
of lacunae. But the conception of a need must not be pressed
so as to comprise merely the elements absolutely necessary for
social life. On the contrary, we have to take account of the whole
sum of intellectual interests, and of all the impulses of poetry and
rhetoric. A cultivated style, which forbids the too frequent
repetition of the same expression, demands of course that where
possible several phrases shall be available for the same thought.
In a still higher degree, rhythm, rhyme, alliteration, and similar
devices demand the possibility of choosing among several phonetic
groups of similar meaning, if the constraint which they impose
is not to be exceedingly burdensome. The result is that the
language of poetry avails itself of the groups of synonyms which
have gradually grown up, using them indiscriminately, where the
language of conversation attaches the use of each to special con-
ditions, and maintaining them where the language of conversa-
tion gradually restricts itself to a single use. This is precisely
one of the most essential factors in the differentiation of poetical
expression from that of prose. It is easy from the poetical lan-
guage of every people and age to show that its relative luxuriance
depends in the closest way upon the character of its poetic
technique; most easily, it may be, in the language of the Old
>
S
274
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Elimination
and utilisa-
tion.
Teutonic alliterative poems, which is remarkable for its special
wealth of synonyms.
406. But to assume, for the common language, the co-existence
through many centuries of double forms, or double words with
the same meaning, is opposed to all experience, and must decidedly
be designated as a fault in method, a fault which has very fre-
quently been committed in the construction of IE. original forms.
407. We must not of course connect the retrenchment of this
superfluity with any conscious design. The purposeless over-
burdening of the memory brings its own remedy.
408. The simplest case of retrenchment is the disappearance
of all among a group of similar forms and phrases, save one. It
is easy to see that the superfluity belonging to the language at
large belongs only in a small degree to the individual. The
characteristic peculiarity of individual language consists mainly
in a certain consistency in the choice made by the speaker of the
different possible forms of expression at his disposal. For if one
has for whatever reason become more usual than another-i.e.
if its capacity for forcing its way into consciousness under given
circumstances is greater-the tendency is also at work by which,
where no special influences draw in the opposite direction, this
preponderance will be reinforced on every fresh occasion. Now
as soon as the great majority of a moderate-sized linguistic com-
munity coincides in its selection from any group of forms, the
natural result is again that the correspondence becomes more and
more confirmed, and, after the disappearance of several generations,
absolute. Thus the different possibilities of choice form a main.
source of the distinctions of dialect. It often happens also, of
course, that the choice leads to the same result over the whole
linguistic area, and especially in cases where conditions peculiarly
favourable for one form occur.
}
ON THE DIFFERENTIATION OF MEANING.
275
1
409. In addition, however, to this merely negative process of
retrenchment, there is a positive process: the utilisation of super-
fluity by means of differentiation in the meaning. This process
is no more the result of conscious purpose than the other. We
have seen that the different senses of a word, an inflexion, a
particle, etc., have to be learnt separately and successively. Now
when a plurality of synonymous expressions is in use, each of
which has several senses and applications, it is quite obvious that
the different meanings cannot appear to each individual in a
linguistic community evenly divided among the different expres-
sions. On the contrary, it will often happen that he hears one
expression earlier, or more frequently connected with one meaning,
another expression earlier or more frequently connected with an-
other signification. Should it, however, happen that of the different
expressions each one is familiar to him in a special sense, he
will abide by these, unless he be drifted by specially powerful
impulses to the opposite side.
Phonetic
differentia-
tion for the
purpose of
only
410. In cases where we cannot trace historically the single fac-
tors of the development, but only perceive their total result, it often
appears as if a differentiation of sound had set in for the purpose differentiat-
of differentiating meaning. And even now most philologists do ing meaning
not shrink from assuming something of the kind. If it were only apparent.
to definitely show the unreasonableness of such assumptions, it is
very important to collect the cases which fall under this head taken
from modern languages as fully as possible.
Doublets.
411. What has been hitherto done in this subject relates chiefly Works on
to the Romance languages. As early as 1683, Nicolas Catherinot
published a work, bearing the title of Les Doublets de la Langue
Françoyse, which collected material to our purpose. In fact the
matter has always excited interest since the foundation of the
scientific grammar of the Romance languages. Rich material
15
276
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Cases of
apparent
tion.
from French has been gathered by A. Brachet, Dictionnaire des
doublets de la langue française, Paris 1868, Supplement, 1871;
from Portuguese by Coelho in the 'Romania,' ii. 281 sqq.;
from Spanish, and, by the way, from other Romance languages
by Caroline Michaelis, Romanische wortschöpfung, Leipzig 1876.
M. Bréal has given us a collection of Latin doublets in the
Mémoires de la société de linguistique de Paris, i. 162 sqq. (1868).¹
For German may be cited O. Behagel, Die neuhochdeutschen zwil-
lingswörter, Germania 23, 257 sqq. A small collection of English
doublets is to be found in Mätzner, Englische grammatik, i. 221
sqq. [Cf. also Skeat, Etym. Dict.; Appendix]. C. Michaelis has
some penetrating observations on differentiation (cf. especially
p. 41 sqq.). She decidedly supports our view that the differentia-
tion in sound and in meaning stand originally in no causal con-
nexion with each other. Behagel (p. 292) speaks yet more positively
on the subject: 'In living language there is no such thing as a
voluntary conscious differentiation of form for the purpose of
differentiation of meaning.' His own work, however, deals almost
exclusively with the phonetic side.
412. The material collected in the above-mentioned works is
differentia- far from being wholly comprised in the category with which we
have to deal. It stands to reason that all cases must be excluded
in which a loan-word is from the first accepted in a different sense
from a word either native or borrowed at an earlier period, or
from another source, even where the two words, when traced back
far enough, lead to a common source. The French chose and cause
owe their origin alike to the Latin causa; but the difference in
their meaning did not arise from any differentiation on French
soil; but cause was borrowed as a law-term when chose had lòng
since developed into the general meaning of 'thing.'
¹ [Cf. Regnaud Mélanges, p. 299. Paris, 1886.]
It is the
1
1
ON THE DIFFERENTIATION OF MEANING.
277
same with by far the majority of doublets in the Romance [and
English] languages, which, for this reason, do not affect us here.¹
It is the same again with such NHG. words as legal-loyal, pfalz-
palast, pulver-puder, spital-hôtel, etc. [regal-royal, orison-ora-
tion, penance-penitence. For a fuller list see Morris's Historical
Outlines of English Grammar, p. 32, § 28.] But we must further
exclude all those cases in which the differentiation of meaning is
the result of a grammatical isolation. If, for instance, the old
participle bescheiden is still employed as an adjective with the
meaning of 'modestus,' while, on the other hand, as a proper parti-
ciple, we use beschieden, there is no doubt that the two words.
bescheiden and beschieden were for a time used indifferently in this
latter function, but beschieden can never have been used for
modestus.
real differen-
413. On the other hand, no notice is taken, in the works cited, Examples of
of our second class, in which the equivalence of meaning is only tiation.
the result of a secondary development. We are thus after all
without a clearly arranged collection of certain instances of the
undoubted differentiation of equivalent expressions. It will there-
fore be well to illustrate the process somewhat at large. These
instances are chosen for the most part from NHG.
414. The forms knabe and knappe are in MHIG. absolutely
equivalent, each possessing the divergent NHG. significations.
Similarly raben (=NHG. rabe) and rappe are both applied to the
bird, whereas in modern times, in the written language, rappe is
confined to the metaphorical use as a term for a black horse.2 A
third form, rappen, with an n transferred from the oblique cases,
has established itself as a name for the coin (originally with a
¹ C. Michaelis is certainly on the whole wrong (p. 42 sqq.) in apprehending the more
nearly Latin sense of the more nearly Latin form, as the result of a differentiation.
2 I cannot, it must be owned, adduce any example of rabe in the transferred sense.
278
PRINCIPLES OF THE HISTORY OF LANGUAGE.
black head of a bird), which was at first known as rappe, rapp, and
further, as rabenheller, rabenpfennig, rabenbatzen, rabenvierer (cf
Adelung). The MHG. backe ('hinterbacken,'' schinken ') bears the
same relation to backe (original Teutonic bako-bakko) as knabe to
knappe; and it is hence probable enough that we have to deal
with a differentiation of meaning developed, as in the previous
instance, secondarily, but of much older date. The difference
between reiter (MHG. riter) and ritter, scheuen and scheuchen,
belongs entirely to NHG., as does the different shade of meaning
between jungfrau and jungfer. Hain is a contraction from hagen,
and in MHG. the same meaning attaches to both (as we see still in
the case of such compounds as hagebuche-hainbuche, hagebutte
hainbutte, etc.); hagen in the derived sense, now restricted to hain,
appears in Burkhardt Waldis.
·
415. Double forms arising from the confusion of different
methods of declension are often differentiated; thus Franke-
franken, tropf-tropfen (cf. for their identical usage the examples
given by Sanders, e.g. Haller: Du bist der weisheit meer, wir sind
davon nur tröpfe; and conversely Wieland: dem armen tropfen),
fleck-flecken, fahrt-fährte, stadt-stätte (MHG. nom. vart, stat-
gen. verte, stete); at the same time, with difference of gender, der
lump-die lumpe, der trupp-die truppe, der karren-die karre, der
possen die posse. The difference of gender, with the same
nominative form, is utilised in der band, das band (cf. examples of
der band used in the sense of 'fascia,' 'vinculum,' in the Deutsches
w.bh.), der die flur (the former only in the signification of 'haus-
flur, though the form die flur occurs also with the same significa-
tion; der-die haft (in MHG. the separation of meanings is already
fairly established); der-das mensch (the latter form was still
employed in the seventeenth century without any suggestion of
contempt); der-das schild (the separation is hardly, even yet,
ON THE DIFFERENTIATION OF MEANING.
279
completely carried out, cf. Sanders, s.v.); der-das verdienst, der—
die see, der die schwulst (examples for both genders, alike in the
proper and metaphorical signification, are to be seen in Sanders);
die-das erkenntniss (the latter form still used very frequently
by Kant in the sense of 'cognitio') [la critique, le critique: la
manœuvre d'un navire, le manœuvre : la statuaire des anciens, un
statuaire: le crêpe, la crêpe: similarly livre, manche, période, poêle].
To these must be added the cases in which different plural forma-
tions are differentiated, bande-bänder, dinge-dinger (as opposed
to the modern usage, e.g. cf. LUTHER, Luke xxi. 26, für warten
der dinger die kommen sollen auf erden); gesichte--gesichter
(examples of cases where the distinction is not observed are given
by Sanders); lichte-lichter (the difference is not consistently
maintained); orte-örter (the same remark applies); tuche-tücher,
worte-wörter (examples are given in Sanders iii. 1662b), in which
the former is still employed like the latter); säue—sauen (cf. for
an older stage passages like von den zahmen sauen entsprossen
or wilde säue und bären, etc., given in Sanders); effecte-effecten,
[l'aïeul, les aïeux, and aïeuls: le travail, travaux, and travails (a
minister's 'reports ')]. In an older stage of MHG. we have
from druck the plural drucke, as well as drücke: at the present day
we have the plural drucke employed only in the sense of printed
works, for which Goethe still uses drücke: though, on the other
hand, we have abdrücke, eindrücke, ausdrücke. The differentiation
of tor—tür goes back further (cf. Sievers, Beiträge V. III¹) and
buch—buche (OHG. buoh, still frequently feminine, is the old nomi-
natival form, buocha the accusative form); the old nominative
forms buoz, wîs, halp are restricted to the employment in certain
formulae (mir wirdit buoz, managa wîs, einhalp, etc., even yet we
use anderthalb, drittehalb), while the accusative forms buoza, wîsa,
halba have become usual with no restrictions.
280
PRINCIPLES OF THE HISTORY OF LANGUAGE.
(
416. This utilisation of different inflexional forms meets us
in nearly all inflected languages. In English we might cite many
instances of duplicate plural formations: cloths, clothes; brothers,
brethren, while in the older language both methods of formation
were employed indifferently as in the case of most similar words:
pennies, pence [dies, dice]. [In Russian the accusative plural is the
same as the nominative in the case of inanimate beings; the same
as the genitive in the case of animate beings.] In Dutch the plurals
in -en and -s are employed in the case of some words indifferently
(vogelen, vogels); in the case of others one alone is commonly used
(engelen, but pachters); again, however, in the case of others both are
employed, but with different meanings; cf. hemelen ('heaven,' in
its proper sense), hemels (betthimmel), 'canopy of a bed'; letteren,
'letter,' or 'literature', letters, 'letters of the alphabet'; middelen,
'means', middels, waists'; tafelen, 'law-tables,' etc.; tafels
'tables'; vaderen, ancestors', vaders, 'fathers'; wateren, 'waters,'
waters, 'streams.' In the same way the forms in -en and -eren
stand side by side: kleeden, tablecloths,' kleederen, 'clothes';
beenen, 'legs,' beenderen, 'bones'; bladen, 'leaves' in a book,
bladeren, 'leaves,' in the proper sense. From the Danish we may
cite skatte (schätze), ‘treasures,' skatter, 'taxes,' vaaben, 'weapons,'
vaabener, ‘armorial bearings.' Where in O. Norse a in the root-
syllable of the nouns changed with g (the -u umlaut), according to
the form of the inflexion; e.g. sqk(u), sakar, etc., we find that
in later Norwegian, in the first place, double forms arose, one in
a and one in o, of which it commonly occurs that either the
first or the second has disappeared. In some cases, however, both
have maintained themselves with a differentiation of meaning :
gata (gasse), 'street,' 'gate,' gota, 'road'; grav, 'grave,' grov,
'ditch'; mark, 'field,' mork, 'wood'; tram, 'height,' trom,
'border.'
ON THE DIFFERENTIATION OF MEANING.
281
417. The difference observed at the present day in the use
of the shorter and more lengthened forms in the inflexion of der
was of gradual development. The forms der in the genitive
singular feminine, and in the genitive plural of all genders, and
den in the dative plural, which are now confined to the adjectival
use, still occur frequently in the seventeenth century, and occa-
sionally in the eighteenth, in the substantival sense; e.g., in Goethe :
die krone, der mein fürst mich würdig achtet. On the other hand,
we find conversely derer and denen used adjectivally-nay, even
as a pure article; cf. e.g. derer dinge, derer leute (LOGAU); derer
gesetze (KLOPSTOCK); zu denen dingen, zu denen stunden (HEIN-
RICH von WITTENWEILER, fifteenth century). Even as late as
the eighteenth century denen in this application is frequent in
the written language; and even at the present day dene, with
the usual apocope of the n, is the generally prevailing form in
Alemannic and South Franconian dialects. Further, the usage
of the present day confines deren to the genitival function only,
while der, on the contrary, is exclusively employed in the dative;
likewise a purely secondary formation: cf. von deren ich reden,
in deren die schmeichler seind (GAILER von KAISERSBERG), o
fürstin, deren sich ein solcher fürst verbunden (WECKHERLIN).
Finally, the curious difference made in NHG. in the employment
of the forms derer and deren is of purely recent development;
cf. wie viel seind deren die da haben (PAULI); and conversely,
mit mancher kunst, derer sichs gar nit schemem thar (MELISSUS).
418. Schaffen, as a strong verb, and schöpfen, have arisen
from the same original form; cf. Gothic skapjan, pret. skôp.
To the preterite scuof a new regular present scaffan was formed in
OHG. beside the old form scepfen; in MHG. a preterite schepfete
and a participle geschepfet was then formed for schepfen.
MHG. schuof, geschaffen, and schepfete, geschepfete, are synonymous,
In
282
PRINCIPLES OF THE HISTORY OF LANGUAGE.
uniting in themselves the signification of both the NHG. words.
The same conjunction is found in the present schepfen. The
present schaffen certainly appears from the earliest times confined
to the signification of to create, shape.
419. Zücken and zucken are originally duplicate forms with
the same meaning; cf. der schon das schwert zucket (LESSING);
den anblick eines zuckenden (HERDER). It is the same with
drücken and drucken.
420. The conjunction als is derived from also through alse.
[So the English word as, like also, took its rise from AS. ealswâ.]
In MHG. the pair are synonymous; both alike can be used
demonstratively or relatively. In the same way there is no
difference of signification between danne and denne, wanne and
wenne. The modern distinction in usage has been developed
by a very slow process; and the fortuitous nature of its origin
shows itself at the present day in the want of a logical principle
of differentiation. The modern difference between warum and
worum is likewise secondary.
421. The participle of the intransitive verdorben and that of
the corresponding transitive verderbt have separated; the latter
being employed in a moral sense only. [Cf. wrung, wrong.] The
difference of meaning is also secondary in the case of bewegt and
bewogen; cf. e.g. das meer... vom winde bewogen (PRÄTORIUS);
der hat im tanze nicht die beine recht bewogen (RACHEL); on the
other hand, dass er dardurch bewegt ward, solches in eigener person
zu erfahren (Buch der Liebe) [cf. the difference between aged
and agéd.]
422. The words in -heit, -schaft, -tum, were formerly essentially
identical in meaning. They are all capable of denoting a pro-
perty; many have in addition developed a collective meaning;
and words, too, in -niss, and simpler formations like höhe, tiefe
ON THE DIFFERENTIATION OF MEANING.
283
often covered nearly the same ground. This state of things has,
on the whole, continued; but in single cases we find that where
several of these formations stood side by side they have mostly
in one way or other suffered differentiation. Cases in which the
different usages now distributed over several such formations were
once completely united in each are certainly not very common ;
still, compare gemein(d)e and gemeinschaft, from which gemeinheit
also did not originally differ in meaning. The two following pairs
are also worth remarking: kleinheit, kleinigkeit; neuheit, neuigkeit.
Examples of the earlier undifferentiated application of the former
pair are cited in the Deutsches Wörterbuch; cf. so verhält es sich
auch mit gewissen kleinheiten, die es im haushalt nicht sind (Goethe-
Zelter Correspondence), die ausnehmende kleinigkeit der masse
(KANT). As regards the latter pair Adelung teaches us that
neuheit was used as a concrete, a thing hitherto not experienced
or known, for which, however, neuigkeit is more common'; on the
other hand, 'die neuigkeit einer nachricht, einer empfindung, eines
gedankens etc., for which neuheit is at the present day more
common in polite conversation.'
423. It is much the same with the adjectives in -ig, -isch, -lich,
-sam, -haft, -bar, in which the existing differences of meaning
do not depend upon any difference of meaning in the suffixes
themselves. An example in point is ernstlich, ernsthaft;
cf., for the older usage, die stets gar ernstlich und sauer sieht
(AYRER); der ernsthaft fleisz (FISCHART).
424. In MHG. so and als (also, alse) are absolutely synonymous ;
both serving equally as demonstratives and as relatives. In NHG.
they are differentiated, so being in general used as a demon-
strative, and als as a relative; cf. e.g. so wol als auch (MHG. so wol
sô or als wol als), so bald als. Still a trace of the demonstrative
als survives in alsbald [cf. English as soon as]. In MHG. lihte and
284
PRINCIPLES OF THE HISTORY OF LANGUAGE.
vil lihte have equally the meaning of NHG. leicht and vielleicht.
The restriction of the form ehe to the conjunction is secondary.
Even Gleim writes ehe als Klopfstock; GOETHE, er soll eh gewonnen
als verloren haben.
425. In MHG. sichern may bear the same meaning as NHG.
versichern, and, conversely, versichern may bear the same meaning
as NHG. sichern (e.g. die stat mit mûren und mit graben versichern).
The differentiation of sammeln, sammlung, and versammeln,
versammlung, is unknown to older NHG.; cf. Moses und Aaron .
sameleten auch die ganze gemeinde, Gott ist fast mächtig in der
samlunge der heiligen (LUTHER); Des festlichen tages, an dem die
gegend mit jubel trauben lieset und tritt und den most in die fässer
versammelt (GOETHE); Die linsen sind gleichsam eine versammlung
unendlicher prismen (GOETHE); Dass sie (the Jews in their dis-
persion) keiner versammlung mehr hoffen dürfen (LUTHER). The
simple form öffnen is at an earlier period employed in the meta-
phorical sense of offenbaren (to reveal'), like eröffnen at the
present day. A similar relation often exists between simple and
compound, or between different compounds which have a simple
verb in common.
426. Certain processes must also be noticed here which, with-
out being strictly speaking differentiations, yet arise from the
same fundamental processes as these, and are hence important
as aiding us to form a judgment on them. The starting-point is
here a partial, instead of complete equivalence of meaning.
427. It may be that that partial equivalence has been
preceded by a complete one, the immediate ground of the
change being that the one received an enlargement of meaning
in which the other did not participate. Then follows very
frequently the further consequence, that the former is completely
ejected from its original meaning by the latter, and confined to
ON THE DIFFERENTIATION OF MEANING.
285
the new.
Kristentuom and Kristenheit are certainly used even
by Walther v. d. Vogelweide in their distinct modern senses;
but the latter is still employed in MHG. in the fundamental
meaning of christentum; cf. e.g. Tristan 1968 (of a child to be
christened), durch daz ez sine kristenheit in gotes namen empfienge.
MHG. wîstuom denotes the same as wîsheit, but side by side of
this the derived meaning 'instruction in law' appears, and then
the NHG. weistum is restricted to this. MHG. gelichnisse can
still be employed in the same sense as gelîchheit, the NHG. gleich-
niss has renounced this original meaning. Indessen (indes) has
originally a purely temporal meaning; cf. ich bin indess krank
gewesen (LESSING); it has been ejected from this meaning by
unterdessen.
428. A more common occurrence is that a word, which at
an earlier date was quite different in meaning from another,
occupies one portion of the domain of the latter, and then gradu-
ally claims it for itself alone. Thus base is restricted to the area
of moral signification (MHG. also basiu kleit, etc.) by the encroach-
ment of schlecht (originally 'smooth,' 'straight,') [English 'slight'].
Similar restrictions have been imposed upon siech (originally [like
English sick] the common word for 'ill'), seuche, sucht, by krank,
krankheit (originally 'weak,' 'weakness'); arg (MHG. also in the
signification of 'avaricious'), by karg (originally 'shrewd '); als
by wie (originally an interrogative word; then a relative, in the
first instance only of the generalising kind); ob by wenn.
429. Finally, it is very common to find that a newly-formed
word, or one borrowed from a foreign language, ejects an older
word from a portion of its domain. Thus the MHG. ritterschaft
has also the signification of rittertum; but, on the formation
of the latter word, it loses this meaning. In the same way freund-
lich is menaced by freundschaftlich, wesentlich by wesenhaft,
286
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Syntactical
differentia-
tion.
empfindlich by empfindsam, einig by einzig, gemein by gemeinsam
and allgemein, lehen by darlehen, stegreif by steigbügel, künstlich
by kunstvoll and kunstreich, bein by knochen (originally Middle
German).
430. These various processes may appear in manifold com-
binations with each other and with the process of differentiation
in the stricter sense. Should the history of the development of
meanings ever grow into a science, it will be of primary import-
ance to take the most scrupulous account of these phenomena.
Our principle that every detail must be judged with constant
reference to the entire mass of the linguistic material is confirmed
in this respect as well; in no other way is appreciation of the
causal connexion possible. As the hints above given lead us
to suppose, it is exactly the absence of consistently applied
logical principles which is here characteristic. The accidental
and involuntary are everywhere evident.
431. We have previously touched on several occasions on the
domain of syntax. The processes above discussed manifest
themselves equally in the case of purely syntactical phenomena.
432. In OHG. duplicate forms had arisen in the strong declen-
sion of the adjective for the nominative singular as well as for the
accusative singular; cf. the neuter guot, guotêr; guotiu, guotaz.
No difference in the usage of these forms meets us at the outset.
On the one hand, the so-called uninflected attributive is employed
before the substantive,-this usage is still universally prevalent
in MHG.-while at the present day the inflected one has established
itself universally with the exception of a few isolated traces;
on the other hand, the inflected form is used as well in places
where the uninflected has at a later date established itself; it
is thus used attributively after the substantive; e.g. Krist guater,
thas himilrîchi hôhaz (OTFRID); even in NHG. der knappe guoter
ON THE DIFFERENTIATION OF MEANING.
287
(PARZIVAL); ein wolken so triebez (HEINRICH von MORUNGEN),
by the side of the more usual der knappe guot, etc.; further, as
predicate: ist iuuar mieta mihhilu (TATIAN); uuird thu stummêr
(OTFRID); occasionally still in MHG., e.g. daz daz wîte velt vollez
frouwen wære (PARZIVAL, 671, 19); thus also ih habetiz io giuuissaz
('I always held it as certain,' OTFRID); also naszer muose ich
scheiden (WALTHER v. d. VOGELWEIDE). In the case of ein and
the possessive pronoun the uninflected form has established
itself even before the substantive; formerly the two forms stood
side by side: cf. sîner sâmo, sînaz korn, einaz fisgizzi (OTFRID).
433. The duplicate forms ward and wurde have parted
company: the former being confined to the signification of the
aorist, while the latter alone can be employed in the sense of
the imperfect. The separation is, however, not completely carried
out, because wurde can be used in both ways. We may assume
with tolerable certainty that in IE. no original difference of
meaning existed between the indicative imperfect and indicative
aorist, nor, again, between the different moods of the present
and those of the aorist. For the double sense probably arose
from a single paradigm, where a discrepancy due to the shifting
accent was removed by two divergent processes of correction.
In Sanskrit, even in the form in which we have received it, the
forms are not distinguished in all classes of verbs. Whether we
prefer to call the Gothic viljau ('I will') an optative present or
aorist is absolutely indifferent. Generally speaking, it may be
laid down that the tense and mood system of the IE. language
must have proceeded to its development by means of a number
of differentiations of meaning-a process accompanied at every
step by the opposite one, viz., the coincidence in meaning of
different formations.
The original
harmony
between
psychological
and gram-
matical
categories
is, in course
of time,
disturbed :
remove the
E
CHAPTER XV.
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
VERY grammatical category is produced on the basis of a
psychological one. The former is originally nothing but
the transition of the latter into outward manifestation. As soon as
the agency of the psychological category can be recognised in the
use of language, it becomes a grammatical category. Its agency,
tendency to however, by no means ends with the creation of the latter. It is
discrepancy. itself independent of language. As it existed before the gram-
matical category, so it does not cease to operate when this
comes into being. In this way the original harmony between the
two may be in the course of time disturbed. The grammatical
category is to some extent a petrifaction of the psychological. It
adheres to a fixed tradition. The psychological, on the other
hand, remains always a free, living agent, capable of taking mani-
fold and changing forms according to individual apprehension. In
addition to this, signification-change operates in many ways to
prevent the grammatical category covering the same ground as the
psychological. Then, as a tendency to convergence makes itself
felt, the grammatical category suffers a displacement, whence may
arise ambiguous phenomena which admit of no simple adjustment
to the categories thus far existing. The consideration of these
processes, which we can observe more accurately, throws light at
the same time upon the origin, which our observation cannot reach,
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
289
of the grammatical categories.
We proceed accordingly to
consider some of the most important grammatical categories from
the points of view indicated.
GENDER.¹
435. The foundation of grammatical gender is the natural Gender.
difference of sex in men and animals. If a masculine or feminine
gender is ascribed to other objects-nay, even to the names of
qualities and activities-this is the operation of fancy which
apprehends these after the analogy of human personality. But
neither natural sex nor that ascribed by fancy has in itself any-
thing to do with grammar. The speaker might think of anything
as a male or female individual without the least trace of any
such thought appearing in language. The linguistic instruments
whereby we now recognise the grammatical gender of a substantive
are the concord in which, on the one hand, attribute and predicate,
on the other hand, a substitutory pronoun stands therewith. Thus
the rise of the grammatical gender stands in the closest corre-
spondence with the rise of a variable adjective and pronoun.
The variability with regard to gender of the adjective presup-
poses that the difference in gender has become attached to a
special stem-ending. This phenomenon might be explained by
supposing that the stem-ending in question was originally an inde-
pendent word, a pronoun which, while still independent, had
acquired a reference to a male or female individual. Still this
assumption is not absolutely necessary. It might conceivably
happen that, by pure accident, an overwhelming majority had
pronounced for the masculine in the case of one stem-ending, and
1 For the subject of this section cf. especially Grimm, Gr. iii. 311-563; Kleine
schrift. iii. 349 sqq.; Diez, iii. 92-8; Miklosich, iv. 17-37; Schroeder, p. 89; Brug-
mann, Z. f. spr. xxiv. 34 sqq.; Delbrück, iv. 4—13; W. Meyer, Die schicksale des
lateinischen neutrums im romanischen, Halle 1883; Lange, De substantivis Graecis
feminini generis secundae declinationis capita tria, Lipsiae 1885 (Diss.). [See Sayce,
Principles of Comparative Philology, p. 264 sqq.].
T
290
PRINCIPLES OF THE HISTORY OF LANGUAGE.
an equal majority for the feminine in the case of the other.
In the pronoun, as in the adjective, the distinction of gender may
appear in the stem-ending; it may, however, also be expressed
by specific roots. It seems probable that grammatical gender
developed earliest in the case of the substitutory pronoun, just as
it is there that it has maintained itself longest in languages, such
as English, where it has partially disappeared.
436. When it first
it first arose, grammatical gender doubtless
coincided throughout with natural sex. Departures from this rule
gradually came about, especially through changes in meaning, as
well as merely 'occasional' modifications of meaning. As a result
of this, the natural sex again asserts its claims, in the first instance,
by causing a violation of the grammatical concord; cf. such cases
as eines frauenzimmers, die sich am artigsten gegen mich erwiesen
hatte (GOE.); die hässlichste meiner kammermädchen (WIELAND);
Lat. duo importuna prodigia, quos egestas addixerat (CIC.); capita
conjurationis virgis caesi ac securi percussi (LIVY); septem milia
hominum in naves impositos (LIVY); Greek, & píxтar', & TEρiσσà
τιμηθεὶς τέκνον (EUR.); φίλτατ' Αἰγίσθου βία (AESCH.). From this
stage we arrive next at a complete change of gender. Thus in
Greek we find that masculine designations of persons and animals
are made also feminine by the simple process of referring them to
feminine objects. For instance, we find side by side the duplicate
forms ὁ — ἡ ἄγγελος, διδάσκαλος, ἰατρός, τύραννος, ἔλαφος, ἵππος,
etc. Conversely, in Christian times a form ó Taρévos 2 was con-
structed. The diminutives, originally neuter, readily assume the
masculine or feminine gender when the diminutive meaning has
been obscured. Just so in German die fräulein is common in the
dialects, even with older writers. If collectives or designations of
qualities become designations of persons, the result may be a
2 Ibid. p. 28.
1 Cf. Lange, u.s. p. 27 sqq.
1
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
291
change of gender. The French le guide, originally 'guidance,'
answers to the Italian la guida; the French le garde, 'the watch-
man,' is originally identical with la garde, 'the guard;' cf. further,
the Spanish el cura, 'the priest,' el justicia, 'the magistrate;' the
Old-Bulgarian junota, 'youth,' as a masculine means 'a youth;'
starosta, ‘age,' as a masculine 'the elder of a village;' the Russian
golova, which in the feminine means 'a head,' and in the masculine
'a conductor.' Of specially common occurrence are feminine sur-
names given to masculine personal names; cf. Latin Alauda,
Capella, Stella; It. Colonna, Rosa, Barbarossa, Malaspina, etc.
437. It often happens that the fact that a word belongs to a
special category decides its gender. This depends partly upon the
fact that the gender of the common designation of the species
fixes the gender of the more special designation. Thus it is easy
for a change of gender to follow in connexion with words connected
in idea.
438. Here therefore we have cases of analogy. Thus the
word mittwoch, earlier mitte woche (media hebdomas) used dialecti-
cally even at the present day as a feminine, has come, like the
French dimanche, to be used as a masculine after the model of the
other names for the week-days. The foreign names Tiber and
Rhône have joined the majority of German river-names. In Greek
many names of trees and plants have become feminine; the words.
Spûs and Boτávn, as class designations, causing the feminine gender
to become the normal one.¹ This process manifests itself most
clearly in words which in their proper meaning exhibit a different
gender, and have only become feminines in their transference
to plants:2 cf. ò Kúavos, 'steel' — Kúαvos, 'the corn flower,'
so called from its resemblance in colour to that metal. In the
same way the names of towns exhibit a tendency to the feminine
¹ Cf. Lange, u.s. p. 35 sqq.
2 Cf. Lange, u.s. pp. II, 42 sqq.
1
292
PRINCIPLES OF THE HISTORY OF LANGUAGE.
1
gender; cf. ἡ Κέραμος from ὁ κέραμος, clay; ἡ Κισσός from
ὁ κισσός, ‘ivy;᾽ ἡ Μάραθος from ὁ μάραθος, fennel; ἡ Ιπνος
from ὁ ἰπνός, oven ; ἡ Ιαλυσός, the town, from ὁ Ιάλυσος, the
name of a person,' [cf. Mätzner, FR. GR. p. 380.]
439. In other cases formal reasons have brought about a
change in gender. Thus in Latin it was customary for words in -a,
so far as they were not designations for male persons, to be of
feminine gender. In consequence we find that the Greek neuters
in -μa in ante- and post- classical writers appear, at least in
popular Latin, as feminines: cf. schema, dogma, diadema, and for this
reason they often appear in the Romance languages as feminines.2
The Italian word ago, answering to the Latin acus, is masculine.
The old Greek feminines in -os are in modern Greek mostly dis-
carded, in some cases having passed into the lists of masculines:
ó Ô
e.g. ὁ πλάτανος, ὁ κυπάρισσος. Even the natural sex has in some
cases not prevented the change of gender; cf. prov. papa and
profeta used as feminines.¹
440. The contrast between the traditional gender of the single
word and that which we expect from its termination may be
cancelled in another way, namely, by a change not of gender,
but of termination—the new termination being, of course, one
characteristic of the gender in question.
441. Thus we find that in Latin peristromum appears by the
side of peristroma. The Latin word socrus produced the Spanish
and Provençal word suegra, Portuguese sogra: the Latin nurus, the
It. nuora, Spanish nuera; Port. and Prov. nora: Old French nore.
Modern Greek, again, has employed this method in order to
1 Cf. Lange, u.s. p. 42 sqq.
2 Cf. W. Meyer, p. 93 sqq., where many other examples are found of change of gender
for formal reasons.
3 Cf. Hatzidakis, Zschr. f. vgl. spr. 27, 82; Lange u.s. p. 9.
4 Cf. W. Meyer, p. 9.
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
293
displace the feminines in -os: hence we get forms like
πaρléva,
ἡ πλατάνη, etc. Even in Old Greek we find forms like ή μίνθη by
ἡ μίνθος, ἡ ἐβένη by ἡ ἔβενος, etc.1
442. In some cases the traditional was at the same time the
natural gender,—an additional reason for its modifying the termina-
tion, instead of being modified by it. To this head too belongs
the fact that in Greek the a-stems which have become masculines
have adapted the s of the nominative (e.g. veavías).²
443. Thus far we are moving on fairly safe ground. But it
is a hard matter to decide how far the natural gender, as viewed
by imagination, has affected the change of grammatical gender.
The subjective views of separate individuals may take very dif-
ferent forms in connexion with the same object. In modern
English this subjectivity is capable of asserting itself unchecked to
a certain point; and we are thus able to form an idea of the way
in which originally the transference of the masculine and feminine
genders to objects which possess no natural sex proceeded. In
other languages the free activity of the imagination is held in
check by the traditional sex, and cannot assert itself so long as
the memory of the latter continues vigorous. A certain hesitation
in regard to the tradition must therefore in every case give the
impulse which sets the imagination to work in this direction. If
however, the traditional sex is not impressed on the speaker, or
only insufficiently so, it needs no very violent effort of fancy to
impel him to attribute any gender at pleasure to the word in ques-
tion. For the difference in gender has so thoroughly permeated
language that it is in many cases impossible to leave the gender
undetermined, and therefore needful to decide for one or another.
Under these circumstances the issue is often decided by chance—
that is, by some trifling circumstance, not necessarily connected in
1 Cf. Hatzidakis and Lange, u.S.
• Cf. J. Grimm, Kl. schrift. p. 357.
294
PRINCIPLES OF THE HISTORY OF LANGUAGE.
any way with the original grounds of grammatical gender. We
have merely to think of the errors we make in a foreign language.
444. Now whatever the positive causes for a change of gender
may be, the negative cause ought in no case to be passed over;
indeed it proves often of more decisive importance than the posi-
tive. The part that it plays may be historically demonstrated
from the fact that those words have been especially exposed to
change of gender, the gender of which is, in connected speech,
most frequently without any special mark, and thus leaves the
slightest impress on the memory. German has no longer any
mark of gender in the plural, not even in the article. Hence it is
natural that precisely those words which are most commonly used
in the plural have changed their gender, partly in connexion with
a change in their phonetic form, which is rendered likewise pos-
sible by the fact that the singular was less firmly rooted than the
plural: cf. wange (MHG. neuter), woge (MHG. der wác), locke (MHG.
der loc), trähne (MHG. der trahen), zähre (MHG. der zaher), wolke
(MHG. daz wolken), waffe (MHG. daz wafen), ähre (MHG. daz eher),
binse (MHG. der binez). Further, if many weak masculines have
become feminine (cf. Paul's MHG. Gr. § 130, note 4), this will corre-
spond with the fact that the declension of the weak masculines
and feminines was absolutely identical in MHG. It may be
affirmed that no word will adopt a grammatical gender not habi-
tually associated with the inflexions attached to it, excepting only
those cases in which natural sex operates. This negative function
of the formal element in causing change of gender must not be
confused with its positive agency already discussed, though the
distinction may not admit in every single case of being sharply
drawn.
445. The neuter is in its origin nothing but the 'sexless,' as its
name rightly declares. The masculine and feminine existed as
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
295
psychological categories before they passed into grammatical
ones; the neuter, on the other hand, has, solely as a consequence
of the formal abolition of the two natural genders, and as a conse-
quence of the consistent observance of concord, taken its place as
a third grammatical gender.
NUMBER.
446. Number also passes into a grammatical category solely by Number.
the development of concord. Even in inflexional languages the
看
​plural is not in all cases indispensable where a plurality has to be
designated. Every plurality may be conceived by the speaker as
a unity. And thus there are designations for a definite number
which are singular, such as score, dozen, lakh, just as originally
thousand, hundred, and probably also other numerals also were
throughout.
447. Thus, further, the so-called collectives are comprehensive
singular designations of plurality. Now as the conception of an
aggregate as a unity or as a plurality depends so much on the sub-
jective caprice of the speaker, his conception may also conflict with
that which is exhibited by the grammatical form of the expression
chosen, and this variation of subjective conception attests itself by
the fact that it, and not the grammatical number, determines the
concord,—a result followed, in some cases, by variations in gender.
448. The most common case is where a plural follows a singu-
lar collective. This phenomenon is much restricted in modern
literary German, which in this, as in other ways, has been very
powerfully influenced by the pedantry of grammatical logic. But it
was of common occurrence during the last century, as it is in Greek
and Latin, and is at the present day in English. Cf. ich habe mich
offenbaret deines vaters hause, da sie noch in Egypten waren
296
PRINCIPLES OF THE HISTORY OF LANGUAGE.
(LUTHER); im vollen kreise des volks entsprungen, unter ihnen
lebend (HERDER); civitati persuadet ut exirent (CAESAR); ex eo
numero, qui per eos annos consules fuerunt (CICERO); ängstlich im
schlafe liegt das betäubte volk und träumt von rettung, träumt ihres
ohnmächtigen wunsches erfüllung (GOETHE); das junge paar hatte
sich nach ihrer verbindung nach engagement umgesehen (GOETHE);
the whole nation seems to be running out of their wits (SMOLLETT);
Israel aber zog aus in den streit und lagerten sich (LUTHER); alle
menge deines hauses sollen sterben, wenn sie männer worden sind
(LUTHER); dass der rest von ihnen sich durch Libyen nach Cyrene
retteten und von da in ihr vaterland zurückkamen (LE.); the
army of the queen mean to besiege us (SHAKESPEARE); pars saxa
jactant (PLAUTUS); concursus populi, mirantium quid rei esset
(LIVY); ὁ ὄχλος ἠθροίσθη, θαυμάζοντες καὶ ἰδεῖν βουλόμενοι
(XENOPHON).
449. In the case of many words the combination with the
plural is so common that we may actually apprehend them as
plural—that is, supposing that no formal element points to the
singular. This is, for instance, the case with the English word
people. The development admits even of a further step, in which
the discrepancy between the grammatical and psychological
number is overcome by the assimilation of the former to the latter.
Thus in the case of OHG. the plural liute (people-leute), has
replaced the singular liut (people=volk); perfectly analogous
instances are the Fr. gens (Old Fr. ja furent venu la gent), Italian
genti (beside gente); late Latin populi (Appuleius, Augustinus), and
the English folks. In AS. -waru denotes a 'state;' the plural -ware,
citizens.' The German die geschwister, 'brothers and sisters,' has
arisen from the collective das geschwister, which is still commonly
found in the last century. In Gothic there is a collective neuter
fadrein found in the sense of 'parents.' This is found not merely
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
297
connected with the plural of the predicate, but it turns the article
which belongs thereto into the plural: pai fadrein, pans fadrein.
At the same time it appears as well in the plural form: ni skulun
barna fadreinam huzdjan, ak fadreina barnam.
450. It happens, too, conversely, that a plural expression
receives the function of a singular by the fact that the parts which
are thereby indicated are comprehended into one united whole.
Thus in German one can say, ein zehn mark; in English, I would
do it for another five shillings; even, there's not another two such
women (WARREN). Further MHG. ze einen pfingesten; Latin una,
bina castra, etc.; English if a gallows were on land; there's some
good news; that crystal scales (SHAKESPEARE). Finally, such
plurals also receive a singular form. In German the names of
festivals, ostern, pfingsten, weihnachten, properly dative plurals, are
used as singulars. The German word buch is in Gothic a plural :
bôkôs, properly 'letters;' in OHG. the plural is still used for a book.
The Latin word castra is sometimes apprehended as a singular
feminine, and forms a genitive castrae; in the same way festa in
the Romance languages has passed into a singular feminine. The
Latin litterae, in the sense of an 'epistle,' has passed into the
Italian lettera; French lettre [the Latin vela into the French voile];
minaciae has become in French menace; in Italian minaccia;
nuptiae has become in French noce and noces; tenebrae has become
in Spanish tiniebla as well as tinieblas.
451. When a word is used as an abstract it is, strictly speaking,
incapable of any distinction of number. Since, however, in respect
of the external form, a particular number has to be chosen, it
is a matter of indifference which this is. The sentences man is
mortal and men are mortal, used in an abstract sense, express
the same meaning. Hence a change of numbers is common in
the different languages. Otfrid employs the combination engilon
298
PRINCIPLES OF THE HISTORY OF LANGUAGE.
joh manne. A pronoun which has reference to an abstract ex-
pression stands sometimes in the plural: nicht als ob in ihm kein
einziges punkt wäre, die hat er (HERDER); ein echter deutscher
mann mag keinen Franzen leiden, doch ihre weine trinkt er gern
(GOETHE); nobody knows what it is to lose a friend, till they have
lost him (FIELDING); MHG. swer gesiht die minneclichen, dem
muoz si wol behagen, daz si ir tugent prîsent; jedes triftige beiwort,
an denen er glücklich ist (HERDER). The predicate can stand
in the plural: MHG. daz ieslicher recke in den satel saz und in schar
schihten; Lat. ubi quisque vident, eunt obviam (PLAUT.); uterque
sumus defessi (ib.); uter meruistis culpam (ib.); neuter ad me
iretis (ib.); It. come ogni uomo desinato ebbero; Engl. neither of
them are remarkable (BLAIR). Most IE. languages possess, for
denoting generality, a pair of pronouns, one singular and one.
plural (every, all; jeder, alle). These admit of being readily
interchanged. Thus, even in Latin, the singular is found by the
side of omnes; e.g. militat omnis amans (OVID); in Italian the
singular ogni has become the exclusive form. In Greek, ȧupóтepos
and aμpóтepo stand side by side. Singular forms have grown
out of German beide. The neuter beides is common; it is occa-
sionally found even in MHG. We likewise meet with ze beider sît
in MHG; cf. beiderseits. In older NHG. we meet with other
singular applications of the word: beider baum (MATHESIUS);
mit beidem arm (LOHENSTEIN); auf beyde weise (LESSING).
Conversely the plural jede is common, especially in the previous
century (cf. DWb. 42, 2290).
452. The category of number is further inapplicable to the
simple designations of materials. For individuals, as such, and
the contrast between individual objects and pluralities, arise only
when we contemplate form as well as matter. Hence the desig-
nations of materials are mostly employed only in the singular,
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
299
<
which has to serve the place of that neuter number which we
do not possess. But a transition is easily made from the desig-
nation of a material to the designation for an individual object,
and vice versa, since the individualising form is readily added,
or removed, in imagination; cf. hair, grass, bloom, fruit, weed,
grain, cloth, stone, wood, field, meadow, marsh, heath, earth, land,
bread, cake, etc. Under this head, too, will fall fowl, for the
meat of fowl;' and in German, schwein for schweinefleisch; cf. in
Latin, leporem et gallinam et anserem (CAESAR), and fagum atque
abietem (CAESAR) for beech and firewood. In the same way we
must explain the singular in cases like the enemy is approaching,
the Russian (i.e. 'the Russian army ') is drawing near us.
Livy uses the singular: Romanus, Poenus, eques, pedes, etc., and
even ventures on the combination Hispani milites et funditor
Balearis. [In Hor. Sat. 1, i. we have miles nautaeque.] In Seneca
we even find multo hoste. With this may be compared mit
willkürlicher beliebung des ganzen kaufmanns (MICRÄLIUS, u.a.,
cf. DWb. 5, 337).
Similarly
453. The singular of many words, again, in the function of
an absolute form on which the category of numbers has not yet
been fully impressed, stands in NHG. after numerals. This method
of construction no doubt arose from cases in which a true plural
form, which has merely phonetically come to coincide with the
singular form, lies at the base; thus in the case of mann-pfund,
buch. If, however, the archaic forms have maintained them-
selves immediately after numerals, and other words such as
fuss, zoll, mark, have followed their analogy, this fact must
depend upon special causes. The instinct of language is as
little conscious of a plural form in the archaic combinations as
in those analogically constructed after them. The fact is that,
immediately following a number, there is no need for any special
300
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Tense.
expression of plurality, this being already sufficiently denoted
by the number itself. We thus arrive at a form which in respect
of number is neutral or absolute; and so, once more, at a point of
view similar to that occupied before the rise of grammatical number.
TENSE¹
454. Various attempts have been made to reduce the tenses
of the IE. languages, to a logical system, and neither arbitrary
judgment nor misplaced acumen has been wanting in the process.
We must, here as elsewhere, guard ourselves, when dealing with
logical definitions, against considering solely the grammatical
phenomena, which we have before us, and, when forming a judg-
ment of the latter, against an exclusive regard to purely logical
divisions. There is no such thing as a perfect concord between
the logical and the grammatical categories.
455. The category of tense depends on the temporal relation
in which an event stands to a definite point of time. As such
may be taken, in the first place, the actual moment of speech;
and thus arises the difference between past time, present and
future, to which the grammatical categories of perfect, present
and future, tense correspond. I assume the perfect as the just
expression for this relation, and not the aorist, though the latter
certainly does also occur in this function. The common defini-
tion, according to which the perfect denotes 'completed' and the
aorist 'past' action, is a mere verbal explanation with which no
clear conception can be associated. The characteristic feature
of the perfect in contrast to the aorist and the imperfect lies
in the fact that it expresses the relation of an event to the
present time.
1 Cf. for this section, Brugmann, Ber. der phil.-hist. class. der sächs. gesellsch. d.
wissenschaften 1883, p. 169 sqq.
1
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
301
456. Instead, however, of the present, a point may be taken
which lies in the past or the future; and to this, again, corre-
spondingly a threefold relationship is possible.
An event may
be contemporaneous, previous, or in prospect. Contemporaneous-
ness, with a moment in past time, has found its expression in
the imperfect; that which has preceded it is denoted by the
pluperfect; for what is in prospect in past time no special tense
exists; we have to content ourselves with periphrases. That
which has preceded a point of the future is denoted by the
futurum exactum; that which is in prospect, from this point,
can be expressed by periphrasis only; the contemporaneous is
rendered by the simple future. In this scheme the aorist, and
the substitutes for it which have grown up in single languages,
has as yet found no place. It is the narrative tense-i.e. it denotes
a process which falls in the past; not, however, in its relation
to the present, but in relation to another, but earlier, point of
past time. Herein, however, the process in question is not regarded
as impending, but as already fulfilled. The point of time at which
we place ourselves is continually changed and moved forwards.
457. What I have said of the relation of the tenses which
actually occur to those which must be ideally constructed holds
good absolutely for the indicative only. For the infinitive and
participle, the point of time with reference to which we record
our movements is defined by the finite verb to which they are
attached. A threefold distinction of tense therefore suffices.
The same tenses which serve to express the relation to a present
moment of time are likewise employed to denote the relation
to a point of the past or of the future.¹ This is also the reason
why the participles, in connexion with a finite verb, are so well
adapted to replace such tenses as happen to be defective. The
1 Cf. Brugmann, u.s. p. 174.
302
PRINCIPLES OF THE HISTORY OF LANGUAGE.
imperative is in its nature always future, and similarly the con-
junctive and the optative, so far, that is, as they denote that
anything is to happen or is wished for.
458. Before the formation of grammatical tenses one and the
same form had to do duty in their place, and the time-relation
had to be either denoted by special words, or else guessed at
by the context. No special form undistinguished by any mark
of tense any longer occurs. But the function of such a form
Present as the least characteristic
tense side by side with its strictly Present function. We are
accordingly in a position to realise the circumstances which
preceded the formation of grammatical tenses.
is partially fulfilled by the
459. The present fulfils the function of an absolute tense, in
the first instance, in all abstract sentences (cf. p. 117). A sentence
like the ape is a mammal applies to the past and the future as
well as to the present. If another sentence is subordinated to
the abstract one, the action of this may be conceived as preceding
that of the main sentence, and thus the perfect may be employed;
as we shut the door when the horse is stolen. What is foreign,
then, to the abstract sentence is not all tense-distinction as such,
but the fixation of a point of departure.
j
460. The concrete-abstract sentence (cf. p. 117) has this in
common with the pure abstract sentence, that no definite single
point of time is decisive; that it rather holds good for a number
of different points of time; whence it comes that in it the present
includes in itself both past and future. Its time, however, is
not an absolute one. It is bounded before and behind in narrow
limits, and within these limits it is possible for interruptions to
occur. It is likewise possible for all the points of time to fall
into the past or the future, and hence it comes that the imperfect
or perfect, and the future, can stand.
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
303
461. The present, in very many languages, performs the
functions of the future. This is specially the case when another
word sufficiently indicates that the action in question is in the
future; cf. ich reise morgen ab, das nächstens erscheinende buch;
and in English such phrases as I start to-morrow; the next day
that dawns; but in other cases also, where the situation admits
of no misunderstanding. Further, the future character of the
main sentence is extended to the by-sentence, so that the present
and perfect receive a future sense; cf. wenn er kommt, werde ich
dich rufen; wenn ich die arbeit beendigt habe, werde ich es dir
sagen. Conversely we find in Greek the present of the main
sentence after the future of the by-sentence; cf. ei avтη ǹ TÓλIS
αὕτη πόλις
ληφθήσεται καὶ ἡ πᾶσα Σικελία (EUR). In OHG the present
ǹ
is likewise used in a future sense without any further support.
1
462. We are not accustomed to use the present instead of
the preterite except in the case of the historical present, where,
however, we have to assume an actual displacement of the point
of view in the imagination. But in Sanskrit the word pură, in
Greek Tápos, is found with the present in the sense of the
preterite ; cf. πάρος γε μὲν οὔ τι θαμίζεις = formerly thou camest
not often' (HOMER).2
πάρος,
463. There occur also cases in which the present refers alike to
past and future; cf. ich weiss das schon lange = I know this now
and have long known it: er ist seit 20 jahren verheiratet; so lange
ich ihn kenne, habe ich das noch nie an ihm bemerkt; seitdem er in
Rom ist, hat er mir nicht geschrieben.
464. The relative time-relation of two events which take place
in the past or in the future remains in many cases undefined. The
Germans say als ich ihn erreichte as well as erreicht hatte, wenn ich
ihn finde, as well as gefunden habe. It is well known that in
1 Cf. Brugmann, u.s. p. 170.
2 Ibid. p. 170.
304
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Greek the aorist stands in by-sentences instead of the Latin
pluperfect, in Latin the perfect even after postquam; in MHG. the
simple preterite is quite common, where we now employ the
periphrasis which has to take the place of the pluperfect. This
inaccurate employment of the tenses is the more primitive. The
pluperfect is merely a secondary formation. More commonly still
the relative time-relation is neglected in the case of the participle;
and in this the lack of forms strictly needful co-operates. Cf. In
Zug ans land steigend, kehrten wir im Ochsen ein (GOETHE, and
more examples may be found in Andr. Sprachg. 112); haec Maurus
secum ipse diu volvens tandem promittit (SALLUST; for further
examples vid. Draeger, § 572).
perfect participle with a present meaning: moritur uxore gravida
relicta (LIV., cf. Draeger, § 582). The participle in -ndus is not
used only in a future sense but in certain cases also as a present,
as in volvenda dies, volvendis mensibus (VERGIL); alienos fundos
signis inferendis petebat (CICERO; Draeger, § 599). The German
participle present unites the signification of the present and the
future, cf. e.g. der noch immer betrauerte, früh verstorbene vater.
Conversely we find in Latin the
465. It is possible too for many factors of a secondary nature to
come into consideration for the signification of grammatical tenses.
Since, for instance, an event which has occurred commonly leaves
traces of itself, it is possible for the statement that an event has
occurred, to carry with it the further suggestion of its subsequent
result; and this, strictly speaking, accidental portion of the
meaning may become its principal element. The result being
now, however, regarded as the essence of the signification, the
signification of the perfect necessarily appears as present. Thus
the disappearance of the regular present conduces to the result
which is called in the German grammars präteritopräsens.
466. The same logical relation as that of the present to the
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
305
perfect, denoting result, may subsist also between distinct verbs:
-cf. treten stehen, fallen liegen, verstummen schweigen,
erwachen wachen, entbrennen brennen, sich setzen sitzen.
Since in this case the passing into a given condition, and the
subsistence in such condition, are rendered by two different
linguistic expressions, there are also cases in which the same verb
may denote both. In MHG. it is possible for sitzen, stân, ligen,
swigen to bear the sense of sich setzen, treten, sich legen or fallen,
verstummen; cf. NHG. aufsitzen, aufstehn, abstehn, etc.; and the
Modern Upper German use of sitzen. Consequently the MHG. ich
bin gesezzen and ich sitze may bear an identical signification.
Similarly the Greek deuyo may mean 'I am exiled,' and adıxw, ‘I
am sorry.' Here too may be cited cases in which events belong-
ing to the past are denoted by the present because their effect is
still operative; cf. er lässt dich grüssen; der herr schickt mich; ich
höre, dass er zurückgekehrt ist; er schreibt mir, dass alles gut steht
etc. In the same way the Greek ἀκούω, πυνθάνομαι, αἰσθάνομαι,
μavОávo, etc. are used; and other languages employ the same
usage.
467. We have seen above (p. 303) that the modal and temporal
circumstances are not independent of each other. As it is
characteristic of the imperative to denote a process to be realised
in the future, it is easy to understand that the future can, with the
aid of the situation and of the stress laid on the word be under-
stood as an imperative; e.g. you will do it at once. In the same
way the future can be optative, cf. sic me di amabunt, ut me tua-
rum miseritumst fortunarum (TERENCE). In the interrogative
sentences of demand ('frageaufforderungssätzen') cf. p. 125? the
conjunctive and the future have the same function; cf. Latin quid
faciamus with the Greek Tí Tоinσoμev. Nay, the future may be
employed even as a potential, cf. das wird sich so verhalten, that
U
306
PRINCIPLES OF THE HISTORY OF LANGUAGE.
will be so' and similarly in Latin popular language, e.g. haec erit
bono genere nata (PLAUTUS), cf. Draeger, § 136. The same use is
common in the Romance languages, cf. Diez, iii. 282: Mätzn. Fr.
Gr. p. 72, 3, 4; 75, 2. It is possible to seek two different explana-
tions for this phenomenon. In the first place, since all that takes
place in the future is more or less uncertain, the signification of the
future might have developed in such a way that nothing but the
factor of uncertainty remained over. In the second place, however,
we might apprehend a sentence like he will be at home as 'it will
turn out that he is at home.' The French conditional is a preterite
to this potential mood. It denotes originally the future process
from a point of time of the past, as, e.g., in such a sentence as nous
convînmes que nous partirions le lendemain. As a strict conditional
it may, but need not necessarily, bear a future sense. In German
as well, a future periphras is employed which has not necessarily a
future sense; but for this the conjunctive is employed, as ich
würde zufrieden sein. As the future has been transferred to a
modal signification, conversely in Latin the conjunctive has passed
into a future.
VOICE.
468. The tenses and moods have intrinsically nothing to do
with syntax, and only come to express syntactic relations owing to
their relationship to each other, and consequently not till we have
to deal with compound sentences. The distinction between Active
and Passive, on the other hand, is of essentially syntactical nature,
since it expresses nothing else but a differing relation of the verb
of the predicate to the subject. That which in the active is the
object, is in the passive the subject. Hence the employment of
the passive renders it possible to make a psychological subject,
which otherwise would necessarily take the grammatical form of
the object, into the grammatical subject as well; and this is a
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
307
In
principal reason for the use of the passive construction.
impersonal sentences it is a matter of indifference whether the
active or the passive is employed. The custom of language has
ruled that words which under normal circumstances are construed
personally, are employed passively when they are by an exceptional
usage employed impersonally (cf. the German es wird gesungen,
getanzt, etc.), while in the case of verbs which normally speaking
are used impersonally, the more simple active is employed (es
regnet, ‘it rains, hails,' etc.). There occur, however, points of
contact between the active and passive constructions, cf. der wald
rauscht es rauscht, das haus brenntes brennt. In the ON.
Sagas, we often find in the introductions to a division the formula
hér segir 'here it says,' 'here is treated of.' [The same usage is
common in Russian.] In Middle Latin dicit is equivalent to dicitur
of classical Latin. In a superscription of the OHG. Isidor we find
hear quidit umbi dhea bauhnunga hier wird gehandelt von der
vorbildlichen darstellung: and similarly in other cases.
of skal in ON. in the sense of man soll (wird), etc. is similar.
The use
469. The contrast between active and passive could not be
formed until the separation between subject and object had been
completed. Before this, in any case, the simple juxtaposition of
subject and predicate must have denoted the passive relation as
well as the active. We can still observe the change between the
active and passive signification in the nominal forms of the verbs,
which in their manner of formation have nothing to point to one
or to the other.
470. The participle present appears in archaic NHG. frequently
in a passive sense, cf. seine dabei hegende verräterische absicht
(THUMMEL), dem in petto habenden gedicht (SCHILLER);¹ especially
common are expressions like vorhabende reise. In English we say
1 Cf. Grimm, Gr. iv. 66. Andr. Sprg. 82.
308
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the horses are putting to, the casinos are filling, etc.¹ This passive
application is to be apprehended precisely as the free addition of
the participle discussed supra, p. 157.
471. In the case of the so-called perfect participle, it is seen
quite plainly that the difference between active and passive cannot
consist in anything already attaching to the formation, since the
participles of the transitive verbs are employed passively, and those
of the intransitives in part actively. Even these limits, however,
are not invariably preserved. There arise turns of expression like
das den grafen befallene unglück (GOETHE), des den erwartungen
nicht entsprochenen anfenthalts (GUTZKOW); stattgefunden, and
stattgehabt are tolerably common. Notably, however, a number
of participles of transitive verbs used in a transitive meaning have
passed into pure adjectives, cf. erfahren, verdient, geschworen,
gereist, gelernt, studiert, etc. [in English 'learned,' 'aged,' etc.]
2
472. In Latin the passive sense does not originally attach to
the participles in -endus, -undus, cf. oriundus, [volvenda dies,
VIRGIL, etc.], by the side of which in older writers others like
pereundus, ' dying out,' placendus, ' pleasing,' etc., range themselves.
Other observations of a similar nature may be made in Latin as
in other languages.
473. The verbal distinction of voice was originally as foreign to
the infinitive as to the verbal noun. The former, however, makes
a first approximation to the voice-character, on the one hand,
when an object-case is put in dependence from it; and, on the
other, when it shares with the governing verb the reference to the
subject of the latter (he is able to read); or further, to another word
contained in the sentence, to which it stands in no direct gram-
matical relation (befehlen steht ihm übel an, durch fliehen kann er
sich retten, etc.) Such reference is not in itself absolutely necessary.
2 Cf. Andr. Sprg. p. 83 sqq.
1 Cf. Mätzner, ii. 56.
CATEGORIES, PSYCHOLOGICAL AND GRAMMATICAL.
309
=
None such, for instance, is to be found in such a sentence as er
befiehlt zu schweigen or not lehrt beten. In this case the infinitive
is fundamentally neither active nor passive, but voiceless. In
Gothic we find not unfrequently the simple infinitive standing in the
place of the Greek infinitive passive in cases where the modern
Germans also employ the periphrastic passive infinitive; e.g. warþ
pan gaswiltan þamma unledin jah briggan fram aggilum ἐγένετο
δὲ ἀποθανεῖν τὸν πτωχὸν καὶ ἀνενεχθῆναι ὑπὸ τῶν ἀγγέλων, cf.
Gramm. iv. 57 sqq. This is natural when we take notice of the
originally neutral nature of the infinitive. But on the other side
it is equally intelligible that the want felt in the individual IE.
languages led by gradual but inevitable steps to the creation of
a passive infinitive. The necessity for the employment of such
naturally made itself most felt in the case of those languages in
which the accusative has developed into the subject-case of the
infinitive.
474. A grammatical passive exists only in cases where such
passive has been formed from the same stem as the active, and has
been separated therefrom by a special method of formation. The
relation of an intransitive to the corresponding causative is
approximately analogous to the relation of the passive to the
active, cf. fallen - fällen, hangen hängen, and the pairs from
unrelated roots, werden machen, sterben - tödten, (hin) fallen
(hin)werfen. The difference, however, consists in this, that in the
case of the intransitive, an operative agency is not contemplated as
normally as in the case of the passive. This difference is therefore
easily cancelled. In Greek the phrase is allowed, åπoðvýokeι úπÓ
τινος. In Latin fio is employed in the present simply as the
passive of facio. In no other way than this are the periphrases for
the passive by means of werden and sein intelligible. On the other
hand, the so-called deponents show us the transition from the
310
PRINCIPLES OF THE HISTORY OF LANGUAGE.
passive to the active. These cannot justifiably be placed in a
separate category from the proper passive or middle voice by the
circumstance that they have to be translated by the active in a
foreign language. On the other hand, we have to take into con-
sideration the entire loss of the active which originally belongs to
them, and more decisively still, a method of construction which
commonly falls only to the active; thus especially the connexion
with an accusative of the object.
475. One of the most ordinary ways in which the passive takes
its origin is from the middle voice, which on its side is capable of
being formed from the composition of the active with the reflexive
pronoun. [Cf. French se marier, the Scand. -sk, etc., the Russian
way of forming the passive in -sa, and possibly the Latin form.
in -ier]. The exact process is that one element in the significa-
tion of the middle disappears. The middle involves at once the
origin of an action in the subject and its return thither: the passive,
the latter only. In the case of many reflexive combinations in
NHG., the consciousness of an activity of the subject has likewise
disappeared; but they approximate more to the simple intransitive
by means of the relationship which exists between this and the
passive; cf. sich regen, ausdehnen, drehen, teilen; sich freuen,
schämen, verwundern, irren, etc. Still more completely is every
active operation of the subjects excluded in sich finden, befinden, in
expressions like das lässt sich hören, es lässt sich da gut leben, das
hört sich gut an, hier tanzt es sich sehr leicht.
CHAPTER XVI.
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION.
WE
gence of
gical and
distribution.
WE have already seen, in Chapter VI., that the distribution of The diver-
a sentence, the manner in which we combine elements psycholo-
in larger or smaller groups,-admits of being readily modified.
It grammatical
was likewise hinted there that the psychological (or logical) relation
of the elements among themselves, and their purely grammatical
relation, may be in absolute conflict. The syntactical forms (e.g.
the cases) arise, in the first instance, in connexion with definite
elements of the sentence,-such as subject, object, determination
of a substantive, etc. But they express at the same time a more
specific mutual relation than can be expressed by the mere juxta-
position of the several words. Now the use of this method of mere
specific expression, while the old free combination of notions,
which can never be wholly abolished, still prevails, produces a
contradiction, which in its turn, if it becomes 'usual,' gives rise to
new constructions. The departure from the external form of
grammar here consists partly in a different way of grouping and
detaching the single elements, partly in a different psychological
arrangement of them, by which subject, predicate, object, etc.,
change functions.
477. Duality of elements is, as we have seen, the primitive form Duality and
of the sentence.
Multiplicity
Even the fullest sentences admit of being so of elements.
constructed, the whole of the supplementary matter being thrown
352
PRINCIPLES OF THE HISTORY OF LANGUAGE.
The Psycho-
logical Pre-
dicate.
into extensions of the two elements. But it is possible for the
determinants of the predicate, the nearer and remoter object, the
determinants of place and time, to likewise obtain the value of
independent members, so that a multiplicity of members arises.
Conversely, this multiplicity may itself give rise to a simple pair,
several members being grouped in one without regard to the
distribution which would be demanded by the historical develop-
ment of the mode of sentence in question. If we indicate the
subject by a, the predicate by b, the determinants of the latter by
Greek letters, and the single groups by brackets, we should have
as our fundamental scheme (a)(baßyd), and beside it (a)(b)(a)(B)
(7)(8). Hence can develope schemes such as (abaßy)(8) or
(abaßd)(y), etc., or again (a)(baßy)(d), (a)(baßƐ)(y), etc., and others
beside. The violation of the original distribution can go even yet
further, determinants of the subject being likewise detached from
it and combined with other elements; and those of the object in
the same way, etc.
478. Multiplicity of elements, resulting from an approximate
equality in the value of each, occurs especially in calm, connected
exposition. Ordinary conversation always tends to sentences of
two or three elements only.
479. The element most sharply distinguished from the rest is,
in the first place, the psychological predicate, as being the most
important of all, containing as it does that which it is the final aim
of the sentence to communicate, and on which therefore the
strongest emphasis is laid. The sentence Charles drives to-morrow
to London can be conceived as of four members in case it is spoken
to a wholly unprepared hearer, to whom all its several elements are
therefore alike new. We can then say equally well that three dis-
tinct predicates are attached to the subject Charles; or, perhaps
more correctly, that to the subject Charles is attached the predi-
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION.
313
cate drives, to the subject Charles travels the predicate to-morrow
and to the subject Charles drives to-morrow the predicate to
London. Here the last determinant is no doubt somewhat, but
only slightly, more emphasised than the rest. On the other hand,
where the mental disposition of the hearer is previously defined in
a manner known to the speaker, each of the four members may
become a sharply emphasised predicate. If Charles' journey to-
morrow has already been discussed, and only its goal remains
uncertain, then to London is the predicate. We might then ex-
press it otherwise-the goal of the journey which Charles makes
to-morrow is London. If his journey to London has been men-
tioned, and only the time was indefinite, then to-morrow is
predicate, and we might express it—Charles' journey to London
takes place to-morrow. If it is known that Charles will travel to
London to-morrow, but not whether he will walk or drive there,
the predicate lies in drives. But we could not exactly say that
drives is psychological predicate in accordance with the gram-
matical form; it is rather to be split into two elements, a general
verb of motion and a determinant to this verb indicating the
species of motion; and only the latter is predicate. If, finally, it
is known that some one is driving to London to-morrow, it is only
doubtful who, then the grammatical subject Charles is the psycho-
logical predicate, and we can say equally well-the person who is
driving to-morrow to London is Charles.
cal subject
480. As opposed to the psychological predicate, all the other Psychologi-
elements of the sentence can be conceived as subject, as is clear and copula.
from the above examples. But it is also possible for a single
element to stand prominently out as subject, in which case it
approaches most nearly in importance, and therefore in emphasis,
to the predicate. The other elements then appear as copulæ,
which serve to connect the subject and predicate, and to define
314
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Elements
which are
regularly
cal Subject
the precise nature of their connexion. Thus, in psychological
analysis, the predicate of the sentence Mary has toothache is not
has but toothache, while has is only a copula ; in the sentence John
is accustomed to walk very fast, in the same way, the predicate is
very fast, is accustomed to walk being a copula; in he gesticulated
like one possessed, like one possessed is predicate, gesticulated copula.
481. Every member of a sentence, in whatever grammatical
form it may appear, is capable, from a psychological point of
view, of being subject, predicate, or copula, or a constituent of any
of them. Subject and predicate can be indicated by position as
well as by emphasis. In German, wherever the place at the head
of the sentence, normally assigned to the subject, is occupied by
another element, this is either a logical subject or a logical pre-
dicate, the former more often than the latter. In the latter case
this part of the sentence is also the most strongly emphasised,
but not in the former. The view, frequently met with, that the
position at the head of the sentence always serves to emphasise
above all others the element which occupies that place, is there-
fore wholly wrong.
482. When a demonstrative, referring back to a previously
expressed substantive, occurs at the beginning of a sentence, it
Psychologi is as a rule the psychological subject, or a portion thereof. For
or Predicate, precisely because of this backward reference, it represents the
idea from which the thought of both speaker and hearer proceeds,
and to which what follows is attached as new information; cf. ich
traf einen knaben, den fragte ich;-dem sagte ich;-bei dem erkun-
digte ich mich ;—darüber war ich erfreut; or ich ging nach hause, da
fand ich einen brief; ich sah ihn am sonntag zum letzten male,
damals sagte er mir; or Fritz war gestern bei mir; diesen menschen
möchte ich immer zum hause hinaus werfen; aber ich muss rück-
sicht auf seine familie nehmen; aus diesem grunde kann ich es
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION.
315
nicht. In the same way the relative is regularly psychological
subject. The interrogative, on the other hand, is regularly a
predicate, or a portion of it. For the indefinite use of it in the
question, the answer substitutes a definite one. Hence, if Cicero
says, quam utilitatem aut quem fructum petentes scire cupimus illa?
or tu vero quibus rebus gestis, quo hoste superato contionem convocare
ausus es?—the psychological predicate lies not in the finite verb
but in the participle and its appendages. Further, every element
of a sentence, the connexion of which with the rest is denied
by means of a negative particle, is invariably a psychological
predicate; cf. nicht ihn habe ich gerufen = 'the person, whom I
have called, is not he;' nicht ihm habe ich das geld gegeben = 'the
person, to whom I have given money, is not he,' etc.
The nega-
tion belongs, therefore, though not invariably, to the grammatical,
yet invariably to the psychological predicate, or, to speak more
properly, it invariably relates to the connexion of the psycho-
logical subject with the psychological predicate. It follows, of
course, that the adversative clause which is placed in line with the
negated element is also a predicate; e.g. nicht am morgen, sondern
am mittag will ich verreisen. Further, every element which is
emphasised by a nur, allein, ausschliesslich, only, exclusively, or the
like, for they can be replaced by not another; but besonders, vor
allem, especially, above all, mostly, etc., are also marks of the
predicate.
Discrepancy
Periphrasis.
483. The discrepancy between grammatical and psychological The
predicate may be avoided by a more circumstantial form of avoided by
expression, which in many languages is abundantly resorted to;
Christen sind es, die das getan haben, or von denen man das verlangt;
English, 'tis thou that robbst me of my lord; French, c'est moi qui, etc.;
c'est à vous que je m'adresse; English, it is to you, young people, that
I speak; German, was ihn am meisten ärgerte, war ihre gleich-
316
PRINCIPLES OF THE HISTORY OF LANGUAGE.
The
Discrepancy
gültigkeit; English, what I most prize in woman, is her affections,
not her intellect.
484. A familiar expedient in German for converting into a
subject what would otherwise be the grammatical predicate, is the
periphrasis with tun; cf. verbieten tut es niemand.
485. In many languages we find the conflict of grammatical
eliminated. and psychological subject solved in the following interesting way:
the psychological subject is put first in the nominative, ¿.e. in the
form proper to the grammatical subject, and is then resumed by
a pronoun, the form of which is determined solely by its gram-
matical function; cf. ein eichkranz, ewig jung belaubt, den setzt die
nachwelt ihm aufs haupt (GOETHE); French, cette confiance, il
l'avait exprimée; Italian, gli amici vostri non gli conosco; MHG.
rüemære unde lügenare, swd die sin, den verbiute ich mînen sanc;
Spanish, claro é virtuoso principe, tanto esta sciencia le plugo;
Greek, ἐκεῖνος δὲ οὐ δώσω αὐτῷ οὐδέν; MHG. die Hiunen durch
ir haz der garte sich zwei tûsent; French, tous ces crimes d'état
qu'on fait pour la couronne, le ciel nous en absout; Italian, quelli che
hanno costituita una republica, tra le cose ordinate da loro è stato
(MACHIAVELLI); Greek, τὸ μηδὲν ἄκοντά τινα ἐξαπατῆσαι
Tò
μέγα μέρος εἰς τοῦτο ἡ τῶν χρημάτων κτῆσις ξυμβάλλεται
(PLATO); ach, der heiligste von unsern trieben, warum quillt aus
ihm die grimme pein? (SCHILLER). The possessive pronoun here
takes the place of a genitive: MHG. Parzival der valschheitswant
sîn triuwe in lêrte; English, 'tis certain, that every man that dies
ill, the ill is upon his own head (John, iv. 1.); Spanish, la villa sin
regidores, su triunfo sera breve; French, les soudans, qu'à genoux cet
univers contemple, leurs usages, leurs droits ne sont point mon
exemple (VOLTAIRE). A similar phenomenon occurs when an
attribute to the psychological subject appears in the nominative
cf. Greek, διασκοπῶν καὶ διαλεγόμενος αὐτῷ ἔδοξέ μοι οὗτος ὁ
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION. 317
ἀνήρ (PLATO); ἔδοξεν αὐτοῖς ἀποκτεῖναι τοὺς Μυτεληναίους
ἐπικαλοῦντες τὴν ἀπόστασιν (THUC.); παθοῦσα οὕτω δεινὰ πρὸς
τῶν φιλτάτων οὐδεὶς ὑπέρ μοῦ δαιμόνων μηνίεται (AESCH.) ; French,
depuis deux jours, Fatime, absent de ce palais, enfin son tendre amour
le rend à mes souhaits (VOLTAIRE).
486. A still more complete elimination of the discrepancy
consists in simply giving to the psychological the form of the
grammatical subject, i.e. in putting it in the nominative. On the
Rhine they say, according to Andresen, Sprachgebrauch 80, es
geben dies jahr nicht viele äpfel. The nominative is used in the
same way, according to Hildebrand, DWb. 4, 1a, 1404, in Strass-
burg, Osterlande, Thüringen, and Hesse. Andresen adduces from
literary German, es gibt nichts lächerlicheres als ein verliebter mann
(BÖRNE). Even GOETHE (Der junge Goethe, ii. 465) says: müssen
es hier menschen geben; and HERDER, giebts aber keine andere emp-
findbarkeit zu tränen als körperlicher schmerz? In the last case,
therefore, at least, the comparison is treated as if it belonged
to a grammatical subject.
logical treat-
487. Adverbial determinants, which commonly, as the name Psycho-
shows, are simply attached to the predicative verb, play in reality ment of
very various parts in the structure of the sentence.
On the one Deter-
hand they are really verb-determinants; e.g. Charles eats slowly;
das kind zappelt mit händen und füssen. If, then, the essential
point of the communication lies in the adverbial determinant, this
can be treated as the predicate, and the verb as copula between.
it and the subject. But the distribution may also be of this
nature, that the adverb serves to define the combination of the
other members of the sentence. No sharp boundary between this
and the first-described distribution can be laid down. All tem-
poral, local, and causal determinants can be referred to this class.
These constitute, in antithesis to the other elements of the sen-
Adverbial
minants.
318
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Rarity of
Discrepancy
of slight
formal
Develop-
ment.
7
tence, as a rule the psychological subject, sometimes the predicate
cf. to-morrow evening I will visit you; on the table lie two books
the books lie not on the table but in the box. But here the vert
is always subordinated in such a way that it can be equally
well conceived as a copula. On the other hand, there are certair
cases in which the adverb can only be conceived as a predicate
attached to an already completed sentence. To this head belong
all expressions of modality, such as certainly, assuredly, truly
in any case, probably, doubtless, perhaps, hardly. He will certainly
come is equivalent to it is certain that he will come. Further, such
expressions as unfortunately, otherwise, under these circumstances
on this condition, German vorkommenden falls, leider, sonst, etc.;
also, in German, torichterweise, and all other forms in -weise, which
are distinguished just in this point from the simple adverbs toricht,
etc.; the latter refer to the predicate, the former to the relation
between subject and predicate. The attempt to give clear gram-
matical expression to the logical relation has led to such German
phrases as kaum, dass er mich ansieht; vielleicht, dass eine träne
dann von seinem auge fällt (frequent in the last century); glück-
licherweise, dass die gemälde so hoch stehen (GOETHE). Expres-
sions of assurance, occurring alone in the first place-e.g. gewiss,
er wird es tun-are obviously predicates to the independent
sentences which follow.
488. In languages of slight formal development the discrepancy
in Languages between psychological and grammatical subject or predicate is
far more rare; for this is occasioned precisely by the growth of a
manifold special form of expression for the various logical relations
of notions among themselves. The remarkable, and to us very
singular, forms of expression in Dajak, adduced by Steinthal,
Typen, p. 172, 3, seem to me to rest on the essentially following
processes. The psychological subject or predicate is made the
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION. 319
=
(
grammatical subject or predicate likewise; whence either the
former or the latter is put in the first place; and these two
main members are thus, if originally compound, distributed on
the same principle. Compare especially, in Steinthal's translation,
the following: boat this boat of his choice this is the boat
he has chosen'; witness two these which thy desire = 'which
of these two witnesses desirest thou?' thou place of my giving =
it is to you that I have given it'; too much to it moved be
bench by thee (zu sehr ihr geschoben sein bank durch dich) 'thou
hast moved the bench too much' (where too much is the psycho-
logical predicate). Cf. with this the Arabic construction Omar dead
his father=' Omar's father is dead' (Steinthal, Typen, 271), which
further corresponds with the IE. constructions adduced above (p. 316).
Functions
Deter-
the Deter-
489. The same inversion which marks the relation of the Change of
psychological, in comparison with the grammatical, subject and between the
predicate, may occur also in the relation of the determinant to minant and
the determinate. The case where it is least easy to decide which minate.
of two members is the true determinant, and which the true
determinate, is that of two substantives in opposition. I can
say, e.g., Totila, a king of the Ostrogoths, or a king of the Ostro-
goths, Totila. Such a change of rôles is, however, only possible
when the relation of the two members is a somewhat loose one,
a condition of which is, that it is communicated as a piece of
new information. Then the whole combination approximates
to the nature of a sentence, and the first member is related to
that which follows as subject to predicate. If, on the other hand,
the relation is assumed to be already known, no arbitrary exchange
of rôles is possible, and the order decides nothing. If, for instance,
there is question of a certain Mendelssohn, and some one asks:
What Mendelssohn is meant ?' in the reply, 'the composer
Mendelssohn,' Mendelssohn is undoubtedly the determinate, not
י
320
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the determinant, in spite of its not standing first. Similarly, in
Queen Anne, Mr. Smith, brother John, the proper nouns are the
determinates, the titles and other defining epithets the deter-
minants. There also occurs, however, where no previous know-
ledge of the relation can be assumed, a closer combination of
the two members with addition of the definite articles; e.g. the
master-tailor Thomson. Here the article belongs, not to the
first member, but to the whole, and by this very means binds
it together in a unity. For we cannot say, instead of this,
Thomson the master-tailor, but, at most, Thomson a master-tailor;
or, in German, omitting the article, e.g. Schulze schneidermeister,
when a further determinant follows, e.g. in Berlin. But this
alteration would relax the closeness of the combination, and
the whole expression would accordingly have a different force.
In this construction, then, neither element is distinctly determinant
or determinate. To the more closely-connected appositional
relations belongs also the combination of Christian and surname.
It is indubitable that, at the present day, in John Smith, Peter
Robinson, Henry English, etc., the Christian name is the deter-
minant, the family name the determinate; it is equally certain
that originally the converse was the case. There has occurred
here, therefore, a displacement of the distribution.
490. An adjectival attribute cannot thus simply change rôle,
with its substantive. We must here, however, refer to a very
frequent construction, in which, undoubtedly, the principal notion
is deposited in the adjective. When GRIMM says jenes
heranzuziehen untersagt die mangelnde lautverschiebung, we must
if we would make the grammatical form logical, transpose the
distribution, but at the same time introduce a further change of
construction der mangel der lautverschiebung. Cf. the list of
examples in Andresen, Spr. p. 122, 3.
1
1
1
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION.
321
491. A displacement of quite another kind appears in such
phrases as ein sein wollendes original (HERDER), so viele sein wollende
kenner (EBERT to LESSING), sein sollende griechische simplicität
(IFFLAND); ein gewesener soldat, ein sogenanter vetter, and the like;
FR. un nommé Richard. Here the substantives, which are properly
predicates of unnamed subjects, have taken the place of these
subjects, and accordingly determined the form of the participle
also. In cases, again, like sein früherer (ehemaliger) herr, seine
spätere (zukünftige) frau, der angebliche baron, the substantives are
really predicates.
elements
cally con-
492. When the separation of elements which grammatically Separation of
belong closely together, has become 'usual,' new constructions grammati-
arise, of which we are no longer at liberty to say that they still nected.
exhibit the discrepancy between the grammatical and the logical
distribution to which they certainly owe their origin. A relation
in origin purely psychological has developed into a grammatical
relation.
adjective.
493. Thus the genitive frequently detaches itself from immediate Genitive and
connexion with the word on which it primarily depended; where it
is dependent on a predicative adjective, the combination is never
very close, and it is immaterial whether we treat it as dependent
on the adjective alone or on the adjective together with its copula.
It enjoys therefore a similar degree of independence to that of an
object depending on a verb, and the same freedom of varying its
position. Cf. des erfolges bin ich sicher (of success I am certain ').
Now in German the genitive es, which is frequently dependent on
combinations of this kind, has become phonetically coincident with
the accusative (MHG. ez), and accordingly has been accepted as
such by the linguistic sense: cf. ich bin es zufrieden. Furthermore,
in some cases the genitive nichts of the MHG. niht has been
traditionally preserved, which was also inevitably treated as an
X
322
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Genitive and
substantive.
accusative: cf. ich bin mir nichts böses bewusst. These circum-
stances have facilitated, but hardly been the sole cause of, the
the further substitution in many cases of the accusative-the
objective case κат' ¿§о×ýv—for the genitive conceived as an
objective case; a substitution precisely parallel to that which has
occurred in so many verbs,—e.g. erwähnen, vergessen, etc.´ Cf. was
ich mir kaum noch bewusst war (WIELAND); sind sie das zufrieden?
(GOETHE), frequently; wir sind die probe zufrieden (RÜCKERT);
das bin ich vollkommen überzeugt (LESSING); so viel bin ich versichert
(LESSING); ingedenk zu sein die bescheen fragen (WEISTÜMER).
The accusative is frequent with habhaft werden, universal with
gewahr werden, gewohnt, los, überdrüssig, schuldig sein or werden.
What is true of the adjective applies naturally also to the
predicative adverb, hence inne werden is now used with accu-
sative. The substitution of the accusative is in any case favoured
by the fact that sentences in dass can be made dependent on such
combinations (ich bin [es] zufrieden, dass du ihn besuchst), and
treated as the object. . In many of these combinations we have
only evidence of the accusative of a pronoun. This shows the
iufluence of the es. But it is clear from analogous cases in Greek
that the process is possible even where it is not thus facilitated,
e.g. ἐπιστήμονες ἦσαν τὰ προσήκοντα (ΧΕΝ.), ἐξαρνός εἰμι τὰ
ἐρωτώμενα (PLATO).
494. The naturally closer combination of genitive with a sub-
stantive is likewise in many cases relaxed, through its being made
logically dependent no longer on the substantive alone, but on the
combination of the substantive with a verb, and thus becoming an
independent element of the sentence. This is very common in
MHG.; e.g. des wirdet mir buoz ('I have a remedy thence'); des hân
ich guoten willen; des sît âne sorge; si wurden des ze rate; ich kume
eines dinges an ein ende ('I learn something with perfect precision ').
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION.
323
Cf. NHG. des lärmens ist kein ende; aller guten dinge sind drei; lass,
vater, genug sein des grausamen spiels (SCHILLER); nun will ich des
briefs ein ende machen (SCHILLER); des ich ein diener worden bin
(LUTHER); dieses gerechten, welches ihr nun verräter und mörder
geworden seid (LUTHER); ein schiff, dessen man, so es vorüber ist,
keine spur finden kann (LUTHER); den leichten erwähnungen, die
seiner einige alte grammätiker tun (LESSING). For the most part
we have now to use a preposition for the MHG. genitive. But
here also the genitive es was misinterpreted and treated as
nominative or accusative, so that the logical subject became
unreservedly the grammatical also. Cf. es ist genug (MHG. genuoc
as substantive taking the genitive), es ist not, es ist zeit, etc.; er
will es nicht wort haben; er weiss es ihm dank. But the displace-
ment of distribution has had this further consequence, that the
genitive has been replaced by the nomination or accusative, a
change to which the dependent sentences in dass, conceived as
subject or object, doubtless contributed. We now say das nimmt
mich wunder as well as das wundert mich; MHG. has des nimet mich
wunder 'wonder about it seizes me.' Examples for the accu-
sative are wer wird ihm diese kleine üppigkeit nicht vielmehr dank
wissen? (LESSING); was er mir schuld gibt (LESSING, and similar
phrases elsewhere); in ansehung der stärke wird niemand diese
assertion in abrede sein (LESSING, cf. Blümner's note in his edition.
of the Laokoon, 2d edition, p. 588). Wahrnehmen (MHG. war =
'observation') is now treated as a simple conception, and invariably
constructed with an accusative. Cf. constructions in Latin such as
quid tibi nos tactiost (PLAUT), quid tibi hanc curatiost rem (ib.), in
which the accusative cannot be conceived as depending solely on
the substantive; further infitias ire, auctorem esse aliquid. Also
GR. ềv µèv πρŵτa σoì µoµpǹv exw (EURIP.), and the like.
495. In the languages which use a word originally substantive
324
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Verb and
Adverb.
to express, or to strengthen the expression of, negation, we find in
connexion with it a genitive which, originally dependent on it,
has gradually passed into an independent member of the sentence,
and now serves as subject or object, while the word on which it
originally depended has lost its substantival nature. Cf. FR. il n'a
pas (point) d'argent ; properly, 'he has no step (point) of money.'
That the linguistic instinct is no longer sensible of any de-
pendence upon pas or point is evident from this, among other facts,
that de is transferred by analogy to other negative sentences which
contain no word originally substantive (cf. il n'y a jamais de lois
observées), and also to sentences negative only in meaning, not in
form (cf. sans laisser d'espérance; doit-il avoir d'autre volonté).
Similarly in MHG., cf. des enmac niht gesîn; mîn vrouwe bîzet iuwer
niht and so again also grôzer krefte nie mêr reeke gewan. Cf.
further NHG. hier ist meines bleibens nicht.
496. The German Adverbs, which are at the same time pre-
positions, enter into a closer union with the verb, in consequence
of which the case, which is strictly dependent upon them alone,
appears to depend on the combination of verb and adverb; cf.
einem abgewinnen, anliegen, aufdrängen, überwerfen, unterlegen,
vorstellen, zusprechen; einen anreden, anklagen. That the case was
at first really dependent on the adverb is shown by the fact that
in the early period we regularly find the same case employed
which the adverb governed when used as a preposition; and by
the special fact, more particularly, that the verbs which taken by
themselves are transitive can take a double accusative when
combined with an adverb; thus in MHG. very frequently the verbs
combined with ane (er nam ze kînde sich den weisen an), sometimes
those with if, in OS., also those with umbi, cf. stôd ine uuerod umbi.
In English we can clearly trace the process by which a case de-
pending on a preposition is detached from direct connexion with it,
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION. 325
f
and attracted to the verb. In the great majority of cases, this
detachment is caused by the effort to put the psychological subject
at the head of the sentence. Cf. And this rich fair town we make
him lord of (SHAKESPEARE); washes of all kind I had an antipathy
to (GOLDSMITH); further examples in Mätzn. ii. 518. The two
principal cases of this phenomenon are the relative sentences (cf. a
place which we have long heard and read of, cf. ib. 519), and passive
sentences (the tailor was seldom talked of, cf. ib. 65 ff.), where the
passive construction serves, as elsewhere, to make the psychological
also the grammatical subject. This kind of passive construction
is used even with transitive words with an object attached to them
(they were never taken notice of, SHERIDAN, cf. ib. 67). This de-
tachment is further usual in interrogative sentences, where the
precedence of the predicate accordingly comes into question (what
humour is the prince of, cf. ib. 519).
and an
dependent
497. An element grammatically depending on an infinitive may Infinitive
pass into psychological dependence on the combination of this element
infinitive with its governing word; cf. dies buch werde ich dich nie on it.
lesen lassen; das ding selbst bin ich weit entfernt zu sehen
(LESSING); mit welchen sie sich erinnern, gegen mich glücklich
gewesen zu sein (LESSING). In consequence of this the linguistic
instinct may become uncertain whether the element in question
ought to be directly connected with the infinitive or with its
government. Hence these cases come to bear a close resemblance
to others in which the dependence on the finite verb is really the
more primitive—cf. was ich zu besorgen habe. Thus it happens
that a genuine transference of the government from the infinitive
to the finite verb takes place, which is attested unmistakably by
the use of the passive; cf. hier ist sie (Minna v. Barnhelm) auf
ansuchen des herrn von Hecht zu spielen verboten (LESSING); die
anklage ist fallen gelassen worden (Allgemeine Zeitung); die
326
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Origin of
Connecting-
Words.
stellung des fürsten Hohenlohe wird zu untergraben versucht (ib.).
With these compare the Greek examples: χιλίων δραχμών
ὁμολογηθεισῶν ἀπολαβεῖν (‘it having been agreed that I should
receive 1000 drachmas,' DEM.); và ηµîv è§ àρxîs πapayyeλÕévтa
διεξελθεῖν (PLATO); τῶν προειρημένων ἡμερῶν τὰ ἐπιτήδεια ἔχειν
('of the days for which it had been ordered to have provisions,'
XENOPHON). To the same displacement is also due the passive
use of Latin coepi, desino, jubeo, prohibeo (liber legi coeptus est, jubeor
interfici), only that here the infinitive also is thrown into the pas-
sive, involving a double reference to the element which is made
the subject. A similar passive use occurs in archaic Latin with
possum and queo also, e.g. quod tamen expleri nulla ratione potestur
(LUCRETIUS); cf. Draeger, § 93. [So caletur (PLAUT.), captivi.]
Here belongs, further, the conversion in thought of an object
depending on an infinitive into the subject of a governing, but
absolutely impersonal, verb, as in ἦν γάρ τι ἐν αὐτοῖς προσῆκον
ideîv ('what it was becoming to see,' PLATO); Móyov Tivà
προσήκοντα ῥηθήναι (ib.).1
498. We have seen that the most heterogeneous elements of a
sentence are capable of being psychologically conceived as mere
copulas, when two others in their neighbourhood are emphasised
as its real essence. Certain words being regularly used thus, the
psychological category becomes a grammatical one, and the words
in question become connecting-words. By a connecting-word I
mean one which serves to indicate the relation subsisting between
two notions; which can, accordingly, only be used in association
with two notions, and is incapable of any independent meaning
when used either alone or with a single notion only. The copula
is is a connecting-word between subject and predicate. It has
latterly been denied that we are justified in setting up such a class,
1 The above exposition is almost entirely based upon Madvig, Kl. schr. p. 362.
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION.
327
and contended that the copula, like any other finite verb, is to be
treated as predicate, and the predicative substantive or adjective
as the determinant of the predicate. This view seems to me an
example of that misunderstanding of what is involved in the
demand for a separation between grammar and logic which I
have touched on at p. 16 above-an example of one-sided regard
for the outer grammatical form and neglect of its functional
value. We must not, however, ignore that sentences like träume
sind schäume, glücklich ist der mann, are equivalent in value to sen-
tences without copula, such as träume schäume, glücklich der mann,
and that sentences of the simpler form were obviously abundant in
the earliest period, and were only by degrees thrust into the back-
ground by sentences with copula. Were we to concede to the is
an independent value beside the substantival or adjectival pre-
dicate, all sentences of this class would become assertions of
existence, which, on the testimony of the linguistic sense, they are
obviously not. What nonsense would result from the interpreta-
tion of the sentence that is impossible, as 'that exists as something
impossible'!
499. The reluctance to recognise the copula as a connecting-
word arises from its preserving, by means of its inflexion, the
character of a verb. In the case of crystallised forms which are
devoid of flexional change, there is less difficulty in perceiving the
transition from an independent to a connecting-word. This trans-
ition is always effected by means of a displacement of the distri-
bution, as will be shown by a series of examples below.
Reference
Direct.
500. A special variety of this displacement consists in giving to Indirect
two elements, which properly have only an indirect relation to each becomes
other due to their common dependence upon a third, a direct rela-
tion. It is thus that we ought, probably, to conceive the origin of
the predicative accusative. We can now say equally well ich
328
PRINCIPLES OF THE HISTORY OF LANGUAGE.
mache ihn zum narren and ich mache einen narren aus ihm. A
double mode of accusative is therefore possible with machen, one
indicating the object affected by the action, the other the result of
this action. If we attach both directly to the verb, as is still pos-
sible, in certain phrases, in MHG., e.g. ich mache in ritter, the idea
'he becomes a knight,' of its equivalent, must enter the conscious-
ness at the same time, and thus both accusatives are set in mutual
relation on the analogy of subject and predicate. This explana-
tion is applicable to all the cases in various languages, in which a
substantive is used as predicative accusative. The transferred use
of the adjective as a predicative object might then be conceived as
modelled by analogy on the similar use of the substantive. But it
must further be considered that, in addition to ich mache einen
menschen glücklich, we can say ich mache einen glücklichen men-
schen. The same explanation serves for the accusative and infini-
tive. The infinitive is originally a second object to the governing
verb. This is still the case with our ich heisse ihn aufstehen, ich
lasse ihn arbeiten, etc. The infinitive may even stand without
another accusative as object (ich lasse arbeiten). Er lehrt mich
französich sprechen is not essentially different in construction from
er lehrt mich die französische sprache. So we can say in Latin
quod te jubet as well as jubet te facere. In the same way the
nominative and infinitive has an analogy in the passive construc-
tion of verbs which can take a double accusative. Bibulus nondum
audiebatur esse in Syria is identical in construction with Cicero per
legatos cuncta edoctus; quod jussi sunt. The treatment of the sub-
stantival accusative as a subject of the infinitive is then a natural
result of the actual situation.
501. Another not infrequent mode of displacement occurs
when an element, which properly belongs to two co-ordinate or
adversative elements, is conceived as belonging solely to the first,
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION.
329
The An Element
equally
two other
attracted to
and placed in relation to a particle which connects the two.
German entweder-oder is now thought of as a pair of correlative related to
particles. But entweder arose from eindeweder, and means, pro- Elements is
perly, 'one of the two'; hence entweder das auge oder das herz is the First.
literally one of the two, the eye or the heart.' The displacement
has allowed the form to crystallise, so that entweder can be used in
every case, and with every part of speech at will. Where entweder
-oder serves to combine sentences, the attraction of the former
into the first sentence is shown also by the inversion (entweder ist
er tot beside er ist tot). It is just the same with weder-noch, with
MHG. ´weder-oder = Latin utrum-an, MHG. beide-und= English
both—and, etc. We translate Latin aeque ac by 'even as,' 'just as.'
But a phrase like hic mihi aeque placet atque ille is properly 'this
man and that please me alike.' That a real displacement of dis-
tribution has nevertheless taken place, and that the ac of com-
parison is, to a certain extent, detached from the copulative ac, is
shown by the regular use of a singular predicate where the ac is
attached to a singular subject; also by the order; and finally,
by cases in which it is no longer possible to render ac by 'and,'
-e.g. aeque a te peto ac si mea negotia essent. Instructive,
too, are analogous constructions which have failed to become
perfectly normal-where the displacement has either not appeared
at all or not yet become 'usual.' Aeque et sometimes stands for
aeque ac: aeque promptum est mihi et adversario meo (CICERO);
cf. Draeger, § 311, 18. We find further ac or et after par, similis,
idem, alius, etc. (cf. ib.): pariter patribus ac plebi carus; pariter
corpore et animo (TERENCE); simul consul ex multis de hostium
adventu cognovit et ipsi hostes aderant (SALLUST); solet alia
sentire et loqui (CAELIUS); viae pariter et pugnae (TACITUS);
omnia fuisse in Themistocle paria et Coriolano (CICERO); haec
eodem tempore Caesari mandata referebantur et in Licori vadum
330
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Compound
sentences.
reperiebatur (Caesar). This displacement has also affected the
ON. ok.
502. The same displacements which occur within the simple
sentence are naturally also found in the compound, since the
two kinds of sentence are separated by no essential and pervading
distinction. The subordinate sentence has the same function as
an element of the sentence, and is therefore subject to the same
laws as any other element in respect of the distribution of the
entire period. It is therefore wrong to divide, as is usually done,
every period at once into a principal and a subordinate sentence
(or sentences). In the first place, it is to be considered that the
subordinate sentence can represent an indispensable element such
as the subject (e.g. that he does not come, annoys me); and, further,
that what we commonly call the principal sentence is in truth
no sentence at all, but only an element or a complex of elements.
If the subordinate sentence contains an unessential portion of the
period—e.g. a time-determinant-it is doubtless possible to dis-
tinguish the principal sentence from it as an independent whole;
but this distribution is grammatically wrong, and even psycho-
logically not always right. To divide the period I asked him after
his health when I met him at once into principal and subordinate
sentence, is no more justified than it would be to divide the
sentence I asked him yesterday after his health into I asked him
after his health+ yesterday. We can of course just as easily
put the subordinate sentence as the adverb between the other
members. Finally, the subordinate sentence does not always
contain an independent element, but frequently only a portion,
a determinant, of one: this is the case with all relative-sentences
which refer to a word in the principal sentence. Now the sub-
ordinate sentence may, just like any other element, demand, from
a psychological point of view, a distribution not in accordance
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTIÓN.
331
with pure grammar; and it may, just like any other element,
share in the displacement of the distribution. Hence such a
displacement is often the condition which renders possible the
division into principal and subordinate sentence. The subordi-
nate sentence is here always psychological subject, the principal
sentence predicate-of course in the wide sense laid down in
Chapter VI.
To
503. If we apply the distinction drawn on p. 117 between
abstract, concrete, and concrete-abstract sentences to the com-
pound sentence, we find that the hypothetical sentences (in the
widest sense) comprise the abstract and abstract-concrete.
the abstract group belong, e.g., if it rains, it is wet; who touches
pitch will be defiled; to the abstract-concrete, if you do not yet know
it, I will tell you; as often as he meets me, he asks me; whoever
among you is not content, let him say so. The meaning of every
abstract or abstract-concrete sentence may therefore be expressed
by a hypothetical period.
from
to Inde-
504. Just as the grammatically independent sentence admits Transition
of a gradual transition to dependence, so one which is gram- Dependence
matically denoted as dependent may approximate by successive pendence.
steps to independence. In the intermediate stage described
above, p. 145, between logical dependence and independence, the
grammatical form may belong to either class. A preference for
one or for the other serves to distinguish different languages
and styles. Thus it is a familiar characteristic of Latin in the
historical period to communicate facts which are in themselves.
new and have an independent value, but which at the same
time serve as temporal or causal determinants of some other fact, in
the form of a dependent sentence or a participial phrase, while
in German the independent sentence is preferred. In various
languages it is not uncommon to attach by a relative a sentence
332
PRINCIPLES OF THE HISTORY OF LANGUAGE.
1
the relation
which in no way defines or modifies what precedes, but gives
new information, and therefore has the force of a co-ordinate
sentence; cf. he went to Paris, whence later on he travelled to Lyons
(= and thence); I met your father yesterday, with whom I had
a long talk (in contrast with I met to-day the gentleman with whom
I talked yesterday). This is of course especially frequent in Latin,
and we are there accustomed to treat long sentences introduced
by relatives as independent sentences. A similar loosely-attached
relative appears also in conjunctive sentences, such as quod Tiberius
quum fieri animadvertit, sinu pugionem eduxit (BELL. HISP.); quae
si dubia aut procul essent, tamen omnes bonos republicae subvenire
decebat (SALL.).¹ A test of the complete independence of the
relative sentence is the use of the imperative in it. This I find
occurring in the Greek New Testament, 2 Tim. iv. 15: dv kaì où
φυλάσσου, and Heb. xiii. 7: ὧν ἀναθεωροῦντες τὴν ἔκβασιν τῆς
ἀναστροφῆς μιμεῖσθε τὴν πίστιν ; in both passages likewise in
Luther's translation: vor welchem hüte du dich auch and welcher
ende schauet an und folget ihrem glauben nach. The use of quam-
quam and etsi = 'however' corresponds. Especially obvious
is the dissolution of the dependence in a case like do poenas temeri-
tatis meae; etsi quae fuit illa temeritas? (CIC.). The same usage
is found with the German wiewol, obgleich, where the dissolution
of the dependence is attested by the order of subject and
predicate; cf. Wie darfst du dich doch meinen augen weisen ?
wiewol du kommst mir recht (HAGEDORN); obgleich das weissbrod
schmeckt auch in dem schloss nicht übel (HEBEL).
Inversion of 505. Cases likewise occur in which the logical dependence is
of principal precisely the converse of the grammatical. The most familiar class.
and sub-
ordinate
sentence.
¹ In itself, no doubt, the use of the relative in a conjunctive sentence does not imply
any loosening of the dependence. Cf. Luther, Apost. xv. 29, dass ihr euch enthaltet vom
götzenopfer, etc., von welchen so ihr euch enthaltet, tut ihr recht (¿§ ☎v diaтηpoûvтes
éavroùs eû πpáğere. Here the relative is used, as elsewhere, as a portion of an element.
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION.
333
of instances, and one common to many languages, is formed by time-
determinants, mostly eben, gerade, noch, kaum,=just, scarcely, etc.,—
upon which the logical principal-sentence may follow not only, as
we saw (p. 145) in the form of the principal-sentence, but also in
that of the dependent-sentence; cf. kaum war ich angekommen, als
ich befehl erhielt; FR. je n'eus pas mis pied à terre, que l'hôte vint me
saluer. Other examples are: FR. le dernier des Bourbons serait tué,
que la France n'en aurait pas moins un roi (MIGNET) 'though
the last of the Bourbons were killed, France would none the less
have a king;' MHG. jane gêt er nie so balde, erne benahte in dem
walde, though he walk never so fast, night will overtake him in
the wood.'
506. The psychological distribution may also break down the
division between principal and subordinate sentence. A frequent case
of this is where a particle properly belonging to the principal
sentence coalesces with one in relation to it which introduces the
subordinate sentence, the whole group being then regarded by the
linguistic instinct as introducing the subordinate sentence. Cf.
sowie (GOTH. swaswe, OHG. sôso), so dass, sobald als, auch wenn;
LAT. sicut, simulac, postquam, antequam, priusquam, etsi, etiamsi,
tam(en)-etsi. It is more important that certain words, especially
pronouns or particles, which originally belong to the principal
sentence, become connecting elements between the latter and a
psychologically subordinate sentence which was previously intro-
duced by no particle, or was even wholly devoid of any gram-
matical work of dependence. These words are then commonly
treated as a part of the subordinate sentence. In this way a
quantity of conjunctions have arisen which serve to introduce sub-
ordinate sentences, and this simple process of displacement of
distribution has been one of the most effective means of creating a
grammatical denotation for the dependence of sentences. For the
2
334
PRINCIPLES OF THE HISTORY OF LANGUAGE.
most part these words referred originally to the logically dependent
sentence which followed (cf. p. 141). Here belongs the most
important German particle daz, English that, originally nominative
and accusative of the demonstrative pronoun. I see that he is
satisfied comes from I see that; he is satisfied. When the that had
been drawn into the subordinate sentence and so converted into a
conjunction, this construction could be extended, like the accusative
and infinitive (cf. p. 259), to cases in which a nominative or accusative
of the pronoun could not be used; e.g. I am convinced (of this), that
you are guilty; he was (so) amazed, that he could not answer a word.
In many cases daz has passed into the subordinate sentence even
when accompanied by a preposition. Cf. MHG. durch daz er videlen
kunde, ‘because he knew how to fiddle,' literally 'for this reason:
he knew how to fiddle.' Similarly umbe daz, âne daz, für daz, ûf daz
(rare), bedaz ('while'). NHG. has preserved ohne dass and auf
dass; ausser dass, während dass and anstatt dass are probably to be
conceived as analogically modelled on them, since these preposi-
tions do not govern the accusative. On the other hand, some
prepositions with the dative of the demonstrative pronoun have
only in the NHG. period passed, through displacements, into con-
junctions: nachdem, seitdem, indem, währenddem. Here and there
we find darum; darum ich es auch nicht länger vertragen, habe ich
ausgesandt (LUTHER, I Thess. iii. 5). The same is the case in
English with for that, OE. for pâm, ær pam. Also, so in OHG. and
early MHG. so dass; and so in assurances and protestations: so
wahr mir gott helfe, so wahr ich hier stehe, for which may be equally
well said so wahr wie ich hier stehe. So= 'however much'; so
gutmütig er (auch) ist, das wird er nicht tun; cf. MHG. sô vil ze
Salcerne von arzenien meister ist, but also with a second relative.
sô: sô manec wert leben so liebe frumt; cf. English Nature, as green
as she looks, rests everywhere on dread foundations (CARLYLE), a
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION. 335
construction frequent in the older language, while the modern
mostly uses only the second relative as; cf. also OF. si com,
NFR. si que. In the last cases, besides the so, a further element
not properly belonging to it is always thrust into the subordinate
sentence. It is thus with NHG. sobald (als, wie), so lange (als, wie),
(in) sofern, (in) soweit, sowie. This so is often conceived wrongly
as an original relative. Substantives too, in part with, in part
without, an article, in part depending on a preposition, have
frequently entered into a subordinate sentence which served to
explain them (cf. p. 141). E.g. MHG. die wile ich weiz drî hove,
NHG. dieweil, alldieweil, derweil, weil = English (the while; NHG.
falls, im falle, sintemal = sint dem mâle; seit der zeit er aufer-
standen ist (LUTHER); English on (upon) condition, in case (both
also with following that), because.
507. A similar process is, in German at least, a partial ground
of the transition from the demonstrative to the relative. This
transition is a consequence of the variety of amò KOLOû construc-
tion described above, p. 133. The common member may consist
of the demonstrative pronoun der, or a demonstrative adverb, cf.
liefun thie nan minnotun (OTFRID); thâr ther sin friunt uuas iu êr
lag fiardon dag bigrabanêr ('where he, who had formerly been his
friend, lay the fourth day buried,' ib.); ni mag diufal ingegin sîn
thâr ir ginennet namon min ('the devil cannot resist there, where
you utter my name,' ib.); thu giangi thara thu uuoltos (you go
thither, whither you will,' ib.); der mich liebt und kennt ist in der
weite (GOETHE). We should here, if we followed our linguistic
sense, interpret the pronoun or adverb as a relative, and as belong-
ing to the subordinate sentence, and this interpretation has
actually effected the substitution for the old demonstrative of the
other relative, coincident with the interrogative, which in universal
sentences is now alone used; wer wagt, gewinnt; wo nichts ist,
336
PRINCIPLES OF THE HISTORY OF LANGUAGE.
"
da hat der kaiser sein recht verloren. But that the pronoun (and
hence also the adverb) originally belonged alike to principal and
subordinate sentence, appears from the following reasons: Firstly,
the pronoun can be combined with a substantive which must
necessarily belong to the principal sentence also; in droume sie in
zelitun then uueg sie faran scoltun ('in dream they told him the
way which they were to travel,' OTFRID); der möhte mich ergetzen
niht des mæres mir iuwer munt vergiht ('he could give me no com-
fort for the tidings which your mouth announces to me,' WOLFRAM);
er sâr in thô gisageta thia salida in thô gaganta (OTFRID); diu
sich gelîchen kunde der grôzen sûl da zwischen stuont (WOLFRAM).
Secondly, the case of the pronoun is in OHG. and MHG., and even
in early NHG. determined usually by the principal sentence when-
ever this requires a genitive or dative, and the subordinate a
nominative or accusative; uue demo in vinstrî scal sîno virina
stuen ('woe to him who in darkness shall do penance for his sins,'
MUSPILLI); owwê des dã nâch geschiht (WOLFRAM); mit all dem
ich kan vnd vermag (HANS SACHS). Thirdly, the pronoun can be
dependent on a preposition, and this must likewise be drawn into
the principal and subordinate sentences: waz ich baser handelunge
erliten hân von den ichs wol erlazen möhte sin ('from those by
whom I might well have been spared it,' MINNESINGER). Fourthly,
a class distinct from these, but equally tending to prove that the
pronoun belonged originally to the principal sentence also, is that
in which it is dependent on a preposition which belongs solely to
the principal sentence. Cf. waz sol trûren für daz nieman kan
erwenden (MINNESINGER); daz ich singe owe von der ich iemer
dienen sol (HEINR. v. MORUNGEN); or the case may even be
determined solely by the principal sentence; der suerit bi demo
temple, suerit in demo dâr inne artôt (swears by him who
dwells therein,' Fragmenta theotisca); den vater erit da zi himili
DISPLACEMENT OF THE SYNTACTICAL DISTRIBUTION.
337
der sun mid den er hât hî in erdi giwunnun (Summa Theologiae).
If the subordinate sentence precedes, the common element may be
resumed by a pronoun or adverb, cf. ther man thaz giagaleizit thaz
sih kuning heizit, der uuidarot in alauuâr themo keisore sâr ('the
man who undertakes to call himself king, he assuredly opposes the
emperor,' OTFRID); daz erbe ûch ûwere vorderen an brâchten unt
mit herscilte ervâchten, welt ir da von entrinnen (Rolandslied);
den schaden he uns to donde plecht, dar vor kricht he nun sin recht
(Reineke vos).
508. For such cases as those adduced it is clear from the above
reasons that the element which introduces the sentence must really
be conceived as originally common to both divisions of it, and that
its repetition stands originally in the same place with such cases as
den schatz den hiez er füeren; beide schouwen unde grüezen swaz ich
mich daran versûmet hân (WALTHER). We are therefore also
entitled to assume the same origin for sentences like ther brût
habet, ther scal ther brûtigomo sin (OTFRID). This is not, however,
intended to exclude the supposition that relative sentences have
also arisen from a primitive doubling of the demonstrative.
of drawing a
fast line
between
principal and
sentences.
509. Principal and subordinate sentence may also be so intri- Impossibility
cately blended that it is no longer possible to separate the ele- hard and
ments of the one from those of the other, as is shown also by the
order of the words. In many languages the principal sentence is subordinate
logically so subordinate that it can be treated as a connecting
element, and inserted in the dependent sentence. The part of the
latter which precedes is then the psychological subject or predi-
cate. This is especially common in interrogative and relative
sentences. Cf. Italian, mio padre e mio fratello dimmi ove sono
Latin tu nos fac ames (CICERO); verbum cave faxis (PLAUTUS);
matrem jubeo requiras (OVID); ducas volo hodie uxorem (TERENCE);
quid vis curem? (PLAUTUS); quid tibi vis dicam? (ib.); English,
Y
338
PRINCIPLES OF THE HISTORY OF LANGUAGE.
something that I believe will make you smile (GOLDSMITH);
whereof I gave thee charge thou shouldst not eat (MILTON); whose
fellowship therefore unmeet for thee good reason was thou freely
shouldst dislike (MILTON); NHG. zuo Amelolt und Nêren nu hæret
wie er sprach (ALPHART); die enweiz ich war ich tuo; MHG. eine
sammlung, an deren existenz ich nicht sehe warum Nik. Antonio
zweifeln wollen (LESSING). English, but with me I see not who
partakes; which we would know whence learned (MILTON). NHG.
auf diese veralteten wörter haben wir geglaubt, dass wir unser
augenmerk vornehmlich richten müssten; MHG. tiefe mantel wît sach
man daz si truogen; zuo sinem brûtloufte bat er daz si quæmen;
Italian, questi mercati giudico io che fossero la cagione (MACH.);
Spanish, los forzados del rey quiere que le dexemos (CERVANTES);
Provençal, cosselh m'es ops qu'ieu en prenda (it is necessary that I
take a resolve in regard to it'; Latin, hanc domum jam multos
annos est quom possideo (PLAUTUS); MHG. swie sie wil, so wil ich
daz min fröude stê; Italian, solo Tancredi avvien che lei connosca
(TASSO); er hat alles, was man will das ein mann haben soll; MHG.
daz ich ie wande daz iht ware; French, voilà des raisons qu'il a cru
que j'approuverais; Italian, le opere che pajono che abbino in se qualche
virtù (MACH.); NHG. was wollen sie denn dass aus mir werde (LESS-
ING); wie wollt ihr, dass das geschehe? woher befehlt ihr denn dass
er das geld nehmen soll? womit wollt ihr dass ich mich beschäftige?
die mischung, mit welcher ich glaube, dass die moral in heftigen situa-
tionen gesprochen sein will (LESSING). Hence in many cases it be-
comes uncertain whether the first part of the sentence is to be held
as still dependent on the verb of the grammatical dependent sen-
tence or rather on that of the grammatical principal sentence. The
difficulty is now solved in German by a repetition of it with varying
construction, by which the overlapping of principal and dependent
sentence is avoided : wovon er wusste, dass er es nie erlangen würde.
CHAPTER XVII.
IN
ON CONCORD.
started from
which one
to agree
without
the latter,
N inflexional languages there exists a tendency to place Concord
words related in a way for which no specific means of cases in
expression exists, as far as possible, in formal correspondence with word came
each other. Thus is explained the concord in gender, number, with another
case, and person, which subsists between a substantive and its regard for
predicate or attribute, or a pronoun or adjective representing the
latter. We may couple with these, as kindred phenomena, the
correspondence in tense and mood within the same period. Such
concord is by no means to be in all cases considered as arising
naturally and inevitably from the nature of the logical relation.
For instance, there is no logical reason assignable why the
adjective should be placed in the same gender, number, and case
as the substantive. We have rather to think of the matter in this
way. The starting-point for the origin of concord was afforded by
cases in which the formal correspondence of a word with another
was produced not by any regard for the latter, but merely by the
identity of their relation. When, however, the concord once began
to be felt as such, it extended its area by analogical transference to and was then
analogically
other cases. We shall most easily recognise that this is the real extended to
course of its development if we contemplate in the first place those
cases in which the extension of concord can still be historically
followed.
other cases.
FJ
340
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Cases in
which the
secondary
growth of
concord is
historically
traceable.
511. The correspondence in gender and number seems to be
illogically spread beyond the area which rightly belongs to it, in
cases where attention is directed by the subject to something as
yet unknown, which only receives a definite content by means of
the predicate. The pronoun which composes the subject ought
then always to stand in the neuter singular; and it actually does
so habitually in NHG.; as in cases like das ist der mann; das sind
die richtigen; and similarly the French say ce sont mes frères. On
the other hand, it appears in English brought into concord with
the predicate; thus, these are thy glorious deeds (MILTON); [but
þæt sindon, etc.. in OE.]; in Italian, è questa la vostra figlia? in
Spanish, esta es la espada; in ancient Greek, aűтη Toɩ díkη čσTì
Deŵv (HOMER); in Latin, ea demum firma amicitia est (CIC.);
haec morum vitia sunt (CIC.); Athenae istae sunto (PLAUT.); quae
apud alios iracundia dicitur, ea in imperio superbia atque crudelitas
appellatur (SALLUST); though we also find id tranquillitas erit
(SENECA); and this use is common in negative and conditional
sentences. We shall probably best explain this phenomenon by
supposing that the subject has here conformed to the predicate, as
the predicate elsewhere conforms to the subject.
512. We often find words which commonly occur in the
singular only, placed by Latin writers in the plural when connected
by a copula with words that have no singular, and which have in
the plural a special meaning; summis opibus atque industriis
(PLAUT.); neque vigiliis neque quietibus (SALLUST); paupertates
divitiae (VARRO); cf. Draeger, § 7, 4.
513. In a sentence like man nennt (heisst) ihn Friedrich, the
name strictly speaking ought to have no case; the simple stem
should stand; nay, we may apprehend Friedrich and other proper
names which contain no case-signs as the stem, or absolute case.
Further, so far as there is a reference to naming in an address we
ON CONCORD.
341
might expect the vocative, and we actually find this in Greek;
τί
µе KAλеîтE KÚρie? (St. Luke vi. 46), translated in the Vulgate
quid vocatis me domine?¹ In default of a pure stem, the nominative
has to be used, which, in most cases, is undistinguishable in form
from the vocative. In Gothic the passage cited above is trans-
lated hwa mik haitid frauja? Correspondingly, Luther further
translates was heisst ihr mich aber herr, herr? And the nominative
or vocative is thus used in other places in MHG. and NHG.; cf.
daz man in hiez der bâruc (WOLFRAM), ich hiess ihn mein Montan
(GELLERT); den ich herr Stolle nennen hörte (INSEl Felsenburg).
The most common usage at the present day is that of the accu-
sative; and as early as in Gothic we find panzei jah apaustuluns
namnida. This accusative is owing to the customary concord
occurring in such cases as the Gothic izei piudan sik silban taujiþ
('who makes himself king').
514. In like manner, in the case of phrases like er hat den
namen Max, the pure stem, or, in default of the existence of such,
the nominative, should stand; and this is the case in German. In
Latin, however, such a construction as lactea nomen habet (OVID)
is purely poetical and post-classical. In classical Latin the nomi-
native stands side by side with nomen only when this word is itself
in the nominative, so that concord is observed, e.g., cui nomen
Arethusa est (CICERO). At the same time, however, we sometimes
find the name made to agree with the person to whom it is
attributed, e.g., nomen Mercurio est mihi (PLAUTUS).
A corre-
sponding vacillation in the use of concord is seen where nomen is
in the accusative case, cf. filiis duobus Philippum et Alexandrum et
filiae Apamam nomina imposuerat (LIV.) — cui Superbo cognomen
facta indiderunt (ib.). This vacillation is the best proof that
the concord in this case did not spring from the nature of the case,
1 Cf. Ziemer, p. 71.
342
PRINCIPLES OF THE HISTORY OF LANGUAGE.
but rather from a certain perplexity felt by the speakers, who in
default of an absolute form were obliged to choose a case, and
in the process sought for any basis of usage which might seem
in harmony with the principle of concord already prevailing in
language.
515. A similar perplexity arises in the case of the predicatival
or predicatival-attributive noun with an infinitive. The NHG. is
well off in this respect, for it possesses an absolute form of the
adjective: es glückte ihm unbekannt zu bleiben. The substantive
appears, when it is necessary to use a particular case clearly
marked as such, invariably in the nominative; we find not merely
er strebt danach berühmt zu werden, but also es steht dir frei als
verständiger mann zu handeln. In Latin the nominative stands if
a connexion with the subject of the governing verb is possible: e.g.
pater esse disce, omitto iratus esse; in poetry we find such expres-
sions as ait fuisse navium celerrimus (CATULL.); rettulit Ajax esse
Jovis pronepos (OVID); and similarly in Greek, even in the case of
the infinitive used substantivally, in whatever case this may stand;
ὀρέγονται τοῦ πρῶτος ἕκαστος γίγνεσθαι (THUCYDIDES); ἔδοξε
πάσσοφος εἶναι διὰ τὸ αὐτὸς μὴ οἱός τ᾽ εἶναι (PLATO). In Greek
such connexion is found even with a genitive or dative depending
on the governing sentence; as in ἅπασιν ἀνάγκη τῷ τυράννῳ
πολεμίῳ εἶναι (PLATO); οἱ Λακεδαιμόνιοι Κύρου ἐδέοντο ὡς
προθυμοτάτου πρὸς τὸν πόλεμον γενέσθαι (ΧENOPHON). In Latin
too we find the connexion with a dative, though to a more limited
extent; cf. animo otioso esse impero (TERENCE); da mihi fallere, da
justo sanctoque videri (HOR.); nec fortibus illic profuit armentis nec
equis velocibus esse (OVID); and commonly in the case of licet. At
the same time after licet mihi we sometimes meet with the accu-
sative (e.g. si civi Romano licet esse Gaditanum, CICERO); this
construction is to be explained by the fact that the accusative is
as
T
ON CONCORD.
343
the ordinary case of the subject with the infinitive.¹ Cf. Ziemer,
p. 96.
516. I proceed to cite a few cases in which no concord is carried
out, and indeed is to some extent incapable of being carried out.
In these cases, in default of the pure stem, which is the only form.
really justifiable, its place has been supplied by the nominative
In German, for instance, such expressions are common as dem als
eine schreiende ungerechtigkeit bezeichenten befehle, mein beruf als
lehrer, even such as die stellung des königs als erster bürger des
staates; in èiner lage wie die seinige side by side with der seinigen.
In Latin we find such constructions as Sempronius causa ipse pro
se dicta damnatur; flumen Albim transcendit, longius penetrata
Germania quam quisquam priorum (TACITUS). In these cases no
doubt ipse and quisquam depend upon the subject of the finite
verb, but belong strictly speaking to the ablative absolute only,
which offers them no immediate connecting link.¹
1
517. The speaker is specially apt to feel perplexity in cases
where a grammatical concord is from the sense impossible, and a
third clause comes in which custom has led us to expect to agree
with both. We have to decide in favour of one or the other, and
in the case of such decision, usage may establish itself differently
in different languages, while it may vary in one and the same.
of concord
parts of a
518. In the case of subject, predicate, and copula, the original Variation
and normal rule unquestionably is that the copula, like every other between two
verb, follows the number of the subject; and accordingly we find sentence.
in English such cases as it was my orders, what is six winters? in
French c'est eux, c'était les petites îles; in Latin nequam pax est
induciae (GELLIUS). In German, however, when the predicate is
plural, the copula is used in the same number, as das sind zwei
verschiedene dinge. Other languages have similar usages; thus
¹ See on this point Madvig, Kl. schr., sqq.
344
PRINCIPLES OF THE HISTORY OF LANGUAGE.
in Greek we find τὸ χωρίον τοῦτο, ὅπερ πρότερον Εννέα ὁδοὶ
ẻkaλoûvтo (THUCYDIDES), and in French we find such expressions
as ce sont là des vertus de roi. This usage seems due to the fact
that the plural makes itself more characteristically felt than the
singular. In several languages, however, the converse usage is
possible; viz., the copula in the singular is added to a plural
subject, and to a singular predicate; cf. such Greek sentences as
aï xopnyíaı ikavòv evdaiµovelas onµeîov ẻoti (ANT.); such Latin
ones as loca quae Numidia appellatur (SALLUST); such English
ones as two paces in the vilest earth is room enough (SHAKESPEARE);
the Spanish los encamisados era gente medrosa (CERVANTES); and
the NHG. falsche wege ist dem herrn ein greuel (LU.). We find a
corresponding use in the person of the verb,-cf. the English it
was you, is that you? the French c'est moi, c'est nous, c'est vous, in
the older form of the language we actually find c'est eux. On
the other hand, we find in NHG. das waren Sie, sind Sie das?
O. FR. ce ne suis je pas, c'estez vous.
519. In the case of the anticipatory undefined subject with
logical subject and predicate, we find that in French the use is
to write such sentences as rarement il arrive des révolutions, il est
des gens de bien. On the other hand, we find in German, es
geschehen umwälzungen.
520. A participle used as predicate or as copula may follow in
gender and number the predicatival substantive instead of the
subject. Cf. the Greek πάντα διήγησις οὖσα τυγχάνει (PLATO); the
Latin paupertas mihi onus visum (TERENCE); nisi honos ignominia
putanda est (CICERO); on the other hand, we find Semiramis puer
esse credita est (JUSTIN). A similar usage prevails in the case of the
predicatival accusative; cf. the Greek τὴν ἡδονὴν διώκετε ὡς ἀγαθὸν ὂν
(PLATO) ; used attributively as the Greek τὰς θυγατέρας, παιδία ὄντα
(DEMOSTHENES); the Latin ludi fuere, Megalesia appellata (LIVY).
ON CONCORD,
345
521. The predicate, instead of following the subject, may follow
some apposition belonging to it; cf. Θῆβαι, πόλις ἀστυγείτων, ἐκ
μέσης τῆς Ἑλλάδος ἀνήρπασται (AESCHINES); Latin, Corinthum
totius Graeciae lumen extinctum esse voluerunt (CICERO); Volsinii
oppidum Tuscorum concrematum est; NHG. die Aegypter aber, dies
harte und gesetzmässige volk, schlug gleich die form der regel und
der gewohnheit auf ihre versuche (HERDER). This holds good
even when the sentence is turned into the ablative absolute; omni
ornatu orationis tamquam veste detracta (CICERO). In connexion
with a distributival apposition we find the singular, in spite of the
fact that the subject is plural; cf., αἱ τέχναι τὸ αὑτῆς ἑκάστη ἔργον
épуáleтaι (PLATO); die sich nach des meisters tode sogleich ent-
zweiten und offenbar jeder nur eine beschränkte sinnesart für das
rechte erkannte (GOETHE); dâ die Kahedîne und die sarjande von
Semblidac ieslîcher sîner künste pflac (WOLFRAM).
522. More striking is the construction whereby the predicate is
made to agree with a noun compared with the subject (1) in gender,
magis pedes quam arma tuta sunt (SALLUST); (2) in number, me non
tantum literae quantum longinquitas temporis mitigavit (CICERO); ei
cariora semper omnia quam decus fuit (SALLUST); (3) in gender
and number, as quand on est jeunes, riches et jolies, comme vous,
mesdames, on n'en est pas réduites à l'artifice (DIDEROT); (4) in
person, ὅσοι ὥσπερ ἡμεῖς ἐπιβουλευόμεθα (THUCYDIDES); (5)
in person and number, as ἡ τύχη ἀεὶ βέλτιον ἢ ἡμεῖς ἡμῶν
αὐτῶν ἐπιμελούμεθα (DEMOSTHENES). The concord of the
predicate is also curious with a second subject connected by
the words and not, as heaven and not we have safely fought to-day
(SHAKESPEARE).
523. In Greek an apposition, if it is separated from the noun
to which it belongs by a relative sentence, may follow the relative
pronoun in case; as, Κύκλωπος κεχόλωται, ὅν ὀφθαλμοῦ ἀλάωσεν,
346
PRINCIPLES OF THE HISTORY OF LANGUAGE.
ἀντίθεον Πολύφημον (Odyssey); οἱ παλαιοὶ ἐκεῖνοι, ὧν ὀνόματα μεγάλα
λέγεται Πιττακοῦ τε καὶ Βίαντος (PLATO).
524. A demonstrative or relative, instead of following the
substantive to which it refers, may follow a noun predicated of it,
as in Latin Leucade sunt haec decreta; id caput Arcadiae erat
(LIV.); quod si non hominis summum bonum quaereremus, sed
cujusdam animantis, is autem esset nihil aliud nisi animus (CIC.);
animal hoc quem vocamus hominem (CIC.); ii sunt, quam tu
nationem appellasti (CIC.); in pratis Flaminiis, quem nunc cir-
cum Flaminium appellant (LIV.); Greek póßos, ǹv aidò eiπoμev
(PLA.). The predicate of the main sentence may then follow
the relative pronoun; cf. Carmonenses, quae est longe firmissima
totius provinciae civitas, per se cohortes ejecit.
525. A relative pronoun which logically refers to an undefined
subject usually follows the defined predicate which belongs to it,
and of course the predicate of the pronoun as well. Thus the
Germans have to say, es war ein mann, der es mir gesagt hat; es
sind die besten menschen, die dir das raten. In the same way in
French, c'est eux qui ont bâti. In French it is further to be observed
that the person of the verb in the relative sentence follows the
defined predicate, as c'est moi seul qui suis coupable. On the other
hand, the NHG. use is to say, du bist es, der mich gerettet hat.
526. In a relative sentence the verb goes into the first or second
person in connexion with the subject of the governing sentence,
although the relative pronoun refers to the predicate, and the third
person would consequently in strictness be demanded; cf. in Latin
non sum ego is consul, qui nefas arbitrer Gracchos laudare (CIC.);
neque tu is es, qui nescias (ib.); English, if thou beest he, who in the
happy realms of light didst outshine myriads (MILTON); I am the
person, that have had (GOLDSMITH). This kind of construction
might certainly be regarded as contamination as well; in which
ON CONCORD.
347
case, in the last example given, the thoughts, I am the person who
has had, and I have had, must have become confused. The same
holds good of a combination like eine der penibelsten aufgaben,
die meiner tätigkeit auferlegt werden konnte instead of konnten
(GOE.). With this we may compare allaro barno betsta thero the io
giboran uurdi (HELIAND) and secga ânegum para pe tîrleáses trode
sceawode, 'to one of the men who looked at the inglorious track
(BEOWULF); and so commonly in old Saxon and Anglo-Saxon.
527. The predicate or attribute, instead of agreeing with the
subject or the word which it defines, may agree with a genitive
depending on it, cf. ἦλθε δ' ἐπί ψυχή Θηβαίου Τειρεσίαο χρύσεον
σкŷπтρоv exwv (HOM.). Stranger still is the English use, there
are eleven days' journey from Horeb unto Kadesh-barnea (Deut. i. 2).
In French, the idiom is, to say, la plupart de ses amis l'abandon-
nèrent, but la plupart du peuple voulait. If it often happens that
after a collective with a plural partitive genitive the plural stands
(in such an instance as eine anzahl soldaten sind angekommen), the
genitive certainly does not need to be regarded as the only reason
for the plural, since such an usage is possible in itself after the
collective: see Chap. XV., section 447.
528. In isolated cases we find in Latin an attribute referring to
a person addressed, placed in the vocative, as quibus, Hector,
ab oris exspectate venis? (VERG.)
529. We may gather, then, from the examples given, the way
in which concord has spread beyond the area which strictly speak-
ing belongs to it. We are able, accordingly, to form some idea of
the way in which this process grew up at a period so early as to
reach back far beyond all our tradition. No doubt we have to
note the fact that concord was not so inevitable in the oldest
stages of language, because absolute forms without inflexional
suffixes were then the rule.
>
348
PRINCIPLES OF THE HISTORY OF LANGUAGE.
First rudi-
ments from
which
concord
proceeded.
530. Let us now consider the first rudiments from which con-
cord proceeded. This process bears a certain analogy to the
concord of the verb in number and person. Verbal forms seem
mostly to owe their origin to the coalition of a personal pronoun
with the tense stem. We must in any case suppose a period in
which substantives coalesced in the same way with the stem, and
in which pronouns could precede the stem as well. We must
therefore suppose, to illustrate by an instance, that it was possible
to say gehen vater, vater gehen and ich gehen, just as much as to
say gehen ich, gehen du, gehen er, etc. There are various non-Indo-
European languages, such as the Hungarian, in which the third
person singular differs from the other persons of the same number
by dispensing with a suffix. In these languages, then, the original
plan maintains itself of coalition according to the formula gehen
vater or vater gehen. The further development then proceeds from
a reduplication of the subject, a process not without analogies at
certain stages in the life of modern languages; cf. der kirchhof er
liegt wie am tage, die glocke sie donnert ein mächtiges eins; freilich
ist er zu preisen, der mann (cf. supra, p. 116); je le sais, moi, il ne
voulut pas, lui; toi, tu vivras vil et malheureux. [This use is of
course very common in English and German ballad poetry.] We
must here mention the anticipation of the subject by means of an
indefinite es, as es genügt ein wort. The pronoun originally was
doubled only in cases where it had to be specially emphasised.
But how such pronominal reduplication is able gradually to spread,
and especially as it is favoured by the phonetic reduction of the pro-
nominal forms, is shown by Bavarian dialects in which we find such.
curious amalgams as the following: mîr hammer (=wir haben wir)
or hammer mîr; ess lebts='ihr lebt ihr' or lebts ess.
The process
has thus repeated itself in the case of the verbal forms, when
already made and finished, which at an earlier period was in
ON CONCORD.
349
operation on the tense stems. The pronouns which coalesce
enclitically have become fused with the verb, and have shown an
increasing tendency to lose their original character of the subject
of the verbs with which they are connected. In the IE. original
language the development must already have gone so far that the
formula vater gehen was entirely replaced by the formula vater
gehen er. The suffixed pronoun however maintains, in the first
place, still a double function. In certain cases it still serves as the
subject (as in the Latin lego, legit); in other cases it is merely by
the concord that it shows its relationship with the subject (as pater
legit, ego scribo). In most IE. languages of the present day the
second function alone has survived. The main reason which has
conduced to render the employment of a second subject-pronoun
general, is this, that the suffixes were no longer sufficient for the
characterisation of the forms. For the rest, the concord of the
verbal predicate with the subject has no value in itself. Thus our
personal endings would merely be so much superfluous ballast, did
they not, on the one hand, serve to mark the verb as such, and, on
the other hand, in certain cases to express the difference between
different moods; though such service is, indeed, but very imperfectly
performed, and in an unnecessarily complicated way.
531. As for the concord of nouns, that of gender and number
at any rate is first formed in the pronoun to which reference is
made, to which grammatical gender also owes its origin. Concord
in case first appears in the case of apposition. Here, likewise, no
absolute necessity exists for employing the case-sign twice.¹ At
the same time we are tempted to regard the apposition to one part
1 We see this best by observing that at a more recent epoch, when the connexion is
very close, the principle of concord is again given up, and the inflexion of the first com-
ponent part omitted; cf. MHG. des künic Guntheres lip, an künec Artuses hove; NHG.
Friedrich Schillers, des herrn Müller; even in Goethe we find des herrn Carlyle's, etc.
Hans Sachs even says, herr Achilli, dem ritter.
350
PRINCIPLES OF THE HISTORY OF LANGUAGE.
of the sentence as a repeated employment of this part. A concord
in gender and number occurs even at the present day only where
it is demanded by the nature of the case. The concord of the
attributive and predicative adjective can have grown only out of
the substantive used as an apposition or as a predicate: in other
words, their origin reaches back to an epoch in which the adjec-
tive had not yet freed itself from the category of the substantive,
and assumed a position in a category of its own. The starting-
point was afforded by those substantives which in the Latin
grammars are called mobilia, such as coquus—coqua; rex—regina,
etc. As such substantives passed into adjectives (cf. below, chap-
ter XX.), they maintain the concord, and it came to be regarded as
of the essence of the adjective.
532. The concord of tense, the so-called consecutio temporum,
has, generally speaking, failed to extend beyond the area originally
assigned to it. The exceptions to the rules laid down on this sub-
ject show that the tense in dependent sentences does not, strictly
speaking, follow that of the clause on which such sentence
depends, but that it settles itself independently on principles of
its own.
The concord of the mood, which sometimes further
affects that of the tense, is somewhat more extended.
Latin tantum voluit error, ut, corpora cremata cum scirent, tamen ea
fieri apud inferos fingerent, quae sine corporibus nec fieri possent nec
intelligi (instead of possunt, CIC.); invitus feci, ut fortissimi viri
T. Flaminii fratrem e senatu ejicerem septem annis postquam consul
fuisset (fuerat, CIC.); cum timidius ageret, quam superioribus
diebus consuesset (CAES.).¹ The assimilation of mood is toler-
ably regular in MHG.
1 Cf. Draeger, 151, 5.
Cf. the
CHAPTER XVIII.
ECONOMY OF EXPRESSION.
TH
economy of
depends on
HE more economical or more abundant use of linguistic Relative
means of expressing a thought is determined by the need, expression
It cannot indeed be denied that these means are often employed the need.
in luxurious superfluity. But, on the whole, our linguistic activity
is characterised by a certain trait of parsimony. Everywhere we
find modes of expression forced into existence which contain
only just so much as is requisite to their being understood. The
amount of linguistic material employed varies in each case with
the situation, with the previous conversation, with the relative
approximation of the speakers to a common state of mind.
Under some conditions a word may speak as plainly as a whole
sentence under others. If we take as a standard the form of
expression which will convey a thought under all possible condi-
tions to any possible hearer, the other forms in use appear to be
imperfect.
534. It is natural, therefore, that what is called ellipse should Ellipse
have played a great part with our grammarians.
When the assumed in a
either to be

minimum of
cases, or else
to be recog-
the
nised as part
of the essence
of expression.
led
terser expression is invariably referred to a more circumstantial
equivalent as the standard, there is scarcely any limit to
possible assumption of ellipses. The abuses to which this
in the sixteenth and seventeenth centuries are well known. They
were, however, only an extreme result of conceptions still repre-
"
.
.
352
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Words or
phrases
supplied.
sented in our grammars. This standard must be given up, and
every form of expression be apprehended in accordance with its
origin without any intrusion of alien matter. The assumption
of ellipse will then be reduced to a minimum. Or else, the notion
of ellipse will have to be given a far wider application than it
has hitherto had; it will have to be conceded that it is of the
essence of linguistic expression to be elliptical, to contain less
than the full contents of what is thought, so that, in regard to
ellipse, the various modes of expression differ only in degree.
535. We will consider first the cases in which a word or a
phrase is supplied from what precedes or from what follows. Here
the question immediately presents itself, whether and how far
we are justified in the expressions supplied. We saw above (p. 129)
that an element of a sentence may be posited more than once.
The other elements then assume a similar relation to both
positions. It will hardly be contended for all cases that these
other elements would in the normal sentence be also doubly
posited, that they are actually posited once, and a second (third,
fourth) time have to be supplied. Least of all is the notion
of supplying applicable in the construction άπò кowоû. But in
such a sentence as he saw me and grew pale it will likewise not
be thought necessary to supply he with grew pale; nor, in such
a combination as in fear and hope, will any one think of 'supplying'
the preposition before hope, because we can also say in fear and in
hope. It is a question, however, whether the notion of 'supplying
cannot be wholly dropped, and replaced by that of single positing
with plural reference. Only, in this case, we must further cease
to conceive what is commonly called a sentence as a closed
and independent unity, and regard it rather as a link in a con-
tinuous series.
536. It is customary to assume ellipse in cases such as die
ECONOMY OF EXPRESSION.
353
deutsche und die französische sprache, and still more decidedly in
the form die deutsche sprache und die französische. But that
we have here merely a pair of elements standing in the same
relation to a third, is shown by the fact that in other languages,
though not in German, similar modes of expression can be
replaced by others in which the two elements are treated as a
unity, and attached as such to the third, which now becomes,
properly speaking, the second. This is shown by the use of
the plural. We say, for instance, quarta et Martia legiones (beside
legio Martia quartaque, both in Cicero), Falernum et Capanum
agros (for agrum, LIV.); Italian, le lingue greca e latina (beside
la lingua greca e latina); French, les langues française et allemande,
les onzième et douzième siècles; English, the German and French
languages.
537. Similar is the case where a single common element is
accompanied by a plurality of mutually corresponding elements;
e.g. John writes well, James badly. That the current assumption
of an ellipse is here likewise superfluous, nay inadmissible, is
shown again by the use in many languages of the plural predicate ;
cf. Latin, Palatium Romulus, Remus Aventinum ad inaugurandum
templa capiunt (LIV.); and similarly with the ablative absolute:
ille Antiocho, hic Mithridate pulsis (TAC.). Even in the case
of disjoined subjects the plural predicate is in many languages
as current as the singular; cf. Latin, si quid Socrates aut Aristippus
contra morem consuetudinemque civilem fecerint locutive sint (CIC.);
haec si neque ego neque tu fecimus (CIC.); Roma an Carthago jura
gentibus darent (LIV.); French, ou la honte ou l'occasion le détrom-
peront; ni la douceur, ni la force n'y peuvent rien; English, nor
wood, nor tree, nor bush are there (SCOTT). This plural has in
any case originated from instances in which the copulative con-
nexion could be substituted without essential alteration of meaning,
Z
354
PRINCIPLES OF THE HISTORY OF LANGUAGE.
and thence been extended by analogy to cases where no such
substitution was possible. It proves that for the linguistic
instinct the predicate has been posited once and not twice.
538. Instances of an element common to both principal and
subordinate sentence (or, if it be preferred, to be supplied in
one of them) occur in the variety of dò Kоwоû discussed on
p. 133, and also in relative sentences, which have a different
origin; e.g. the Latin (qui tacet consentit). Further, in MHG., when
a subordinate sentence without conjunction stands in the relation
of object to the principal sentence: da wânde ich stæte fünde
(Minnesinger); her sprach were intrunnin (ROTHER). Rarer are
other cases: nune weiz ich wie es beginne (TRISTAN); wes er
im gedâhte daz elliu diu wolde bedwingen (JUDITH); mitthiu ther
heilant gisah thio menigi steig ufan berg (Fragm. theot.); kem
einer her mit dem opfer, brecht auch vil golts darvon (H. SAchs);
da ihn die schöne fraw erblicket, winckt ihm (ib.); was ich da
träumend jauchzt und litt, muss wachend nun erfahren (GOE.);
dais, indem er ihn gesegnete, ihm gebot und sprach (LU.).
539. It occurs very commonly in dialogue that words of one
speaker are not repeated by the other. But this will not serve
to justify the assumption that the words are necessarily supplied.
For dialogue itself must, not less than the single speech, be
regarded as a continuous and connected whole.
540. It strikes us now as a singular anomaly when an element
belongs in common to two sentences which, instead of being
continuous, are separated by a third sentence; cf. swaz er den
künic ê geschalt, des wart ir zehenstunt mêr, und (er) jach, si wære
gar ze her (WOLFRAM); wer mit wölfen wil geulen, der muss auch
mit in heulen, sunst tun sie sich bald meulen und (er) ist bei in
unwert (H. SACHS). Similarly, when the sentences to which the
element is jointly referable are only formally continuous, without
ECONOMY OF EXPRESSION.
355
any direct relation to one another; cf. sô ist geschehen des ir dâ
gert und wænent ('ye ween'), mir sî wol geschehen (HARTMANN
VON AUE).
541. The common clement may either stand between those
which are not common, so that it is referable with equal ease to
either (aπò κowoû), or it stands at the outset or at the close of
the whole sentence; it is then nearer no doubt to the one, but
still without difficulty referable also to the other; or, finally, it
is inserted in one of the groups to which it is jointly referable,
in which case it appears, in the first instance, to belong to this
group alone. In modern German such insertions are only familiar
in the first group. It is here that the hypothesis of supplying the
second (third, etc.) group has the most in its favour. In MHG.
insertion in the second group is not very rare: mâge und mîne man
('my relatives and subordinates'); gelücke und Sifrides heil; daz
ich muoz und sterben sol. Instances in MHG. are: nicht sonne, mond
und sternenschein, mir glänzte nur mein kind (BÜRGER); es belť
und wiite, wie der hund auch immer will (HEINR. ALBERTS arien).
Cf. Italian, il mar tranquillo e l'aura era soave (PETRARCA); non
pur per l'aria gemiti e sospiri, ma volan braccia e spalle (ARIOSTO);
Old French, Breton l'ensaigne lor signor (their lord's battle-cry'),
e li Romain crient la lor ; Greek, οὔτε βωμὸς οὔτ᾽ Απόλλωνος δόμος
σwσel σe (EURIPIDES). In this construction, again, there can be
no question of 'supplying.' The case is rather that the first
group remains incomplete until the common element is uttered,
which, in the moment of utterance, serves to complete the first
and the second group at once.
542. The function of a common element is often not perfectly
identical in its several relations. Hence arises a grammatical
discord, since the grammatical form of the element can only
indicate a single function. The dislike of this discord, evidenced
356
PRINCIPLES OF THE HISTORY OF LANGUAGE.
by the repetition of the common element, varies greatly in dif-
ferent languages and periods.
543. Such a discord is everywhere least offensive in the person
and number of the verb. Cf. er hat mich eben so lieb wie du; du
glaubst es, ich nicht; sie reisen morgen ab ich auch. It strikes
us, however, as abnormal, when the common element is made
to conform to the second group; cf. avτòs µèv vdwp, èyà dè oîvov
Tívw (DEM.); dass ich im vater und der vater in mir ist (LU.);
non socii in fide, non exercitus in officio mansit (LIV.). Discord
of tense is ignored in the following: ἡμεῖς ὁμοῖοι καὶ τότε καὶ νῦν
ἐσμεν (THUC.); ἄλλα μὲν πρότερον, ἄλλα δὲ νῦν πειρᾷ λέγειν (ΧΕΝ.).
Difference of tense and mood at once is ignored in een ov
τότε, ἀλλὰ νῦν δεῖξον (DEM.). It is further of fairly common
occurrence that the infinitive has to be borrowed from a finite
verb: er hat gehandelt, wie er musste; still in MHG. nach der mîn
herze ie vanc und iemer muos ; Greek, πάνυ χαλεπὼς ἔχω, οἶμαι
δὲ καὶ ὑμῶν τοὺς πολλούς (PLA.). It is rarer for a participle
to be thus borrowed, as in MHG.: daz diu minn dich verleitet, als si
manegen hat. The same form sometimes serves in German as
both infinitive and participle: ich habe es nicht und werde es nicht
vergessen (KLOPSTOCK); cf. further examples in Andresen, Sprachg.
p. 133. H. Sachs says zu ehren sein wir zu euch kumen, ein histori
vns für genumen, although the perfect of the second verb would
have had to be written in full with haben, not sein.
544. In the case of nouns such discords are in modern
German almost without exception discarded; but they occur
frequently in the older language, especially in the sixteenth
century, and also in other languages. Thus the adjective agrees
with the nearer only of two copulatively connected substantives :
aus meinem grossen kummer und traurigkeit (LU.); von eurer saat
und weinbergen (LU.); sein sonstiger ernst und trockenheit (GOE.);
ECONOMY OF EXPRESSION.
357
seiner gewöhnlichen trockenheit und ernst (ib.); many examples
are given in Andresen, Sprachg. 127 ff.; French, un homme ou une
femme noyée; Italian, in publica utilità ed onore, le città ed i villagi
magnifichi; Spanish, toda sa parentela y criados, la multitud y
dolor, los pensiamentos y memorias, un pabellon o tienda; Latin,
urbem ac portum validum (LIV.). A word governed by two or
more prepositions which require different cases may be posited
only once without difficulty where its cases are phonetically
identical, as in mit und ohne kost; but also where they phonetically
diverge, as in um und neben dem hochaltare (GOE.); durch und
mittelst der sprache (HERDER); further examples in Andresen,
Sprachg. p. 128. Similarly, a single form, standing in relation
to two or more verbs, may represent two or more cases, as in
Latin, quod tactum est et ille adjunxit (CIC.); quae neque ego teneo
neque sunt ejus generis (ib.); NHG. was geschieht und ich nicht
hindern hann (LE.); eine dose, die er mit 80 gulden bezahlt hätte und
nur 40 wert wäre (GOE.);¹ womit uns für die zukunft der himmel
schmeicheln und bedrohen kann (GOE.); bei dessen gebrauch wir ein-
ander mehr schmeicheln als verletzen (GOE.);2 leidlicher wer mir vnd
het auch lieber das drey oder vierteglich fieber (H. SACHS); with
insertion, vnd wissen nit jr widervart mag offt lang haben nit mehr
fug (H. SACHS). Even a word dependent on a preposition may
be at the same time made the subject of the following verb:
dan leszt er uns fürtragen schon das heilig euangelion durch sein
heilige junger, deuten all christlich prediger (H. SACHS); von ritter
Cainis ich lasz het lieb fraw Gardeleye (ib.). The licence is also
extended to cases in which phonetically divergent forms would
properly be required. Especially frequent is the use of an oblique
case, as at the same time the subject of a following verb. Thus
with asyndetic juxtaposition: liess der bischoff die seinen über das
2 Cf. ibid. p. 133.
Cf. Andresen, Sprachg. pp. 129, 130.
358
PRINCIPLES OF THE HISTORY OF LANGUAGE.
In con-
her laufen, erstachen der etlich (WILTWOLT VON SCHAUMBURG,
1507); with insertion, ich war selb bei dieser handlung, gschach e du
warst geborn (H. SACHS). Similarly in constructions with und;
very common in MHG., as ez möhte uns wol gelingen und bræhten
dir die frouwen; but also still in NHG., as in er setzte sich auf einen
jeglichen unter ihnen und wurden alle voll des heiligen geistes (LU.);
den es krenke meinethalben und meinen ohren offenbare (LU.); auch
dem, der sie verfolgt, und fleht und schenkt und schwöret, wird kaum
ein blick gegönnt, und wird nur halb gehöret (LES.).
structions with wan (='for'): thut euch bedenken, wan wisset
selber ge gar wol (H. SACHS). Incongruence may further accom-
pany the anomaly noticed on p. 354, as in belîbe ich ane man bi
iu zwei jâr oder driu, sô ist mîn herre lîhte tot und kument ('you
come') in sô grôze not (HARTMANN VON AUE). Examples in the
case of aπò KOLOû with logical subordination have been given
ἀπὸ κοινοῦ
above, pp. 133 and 134. In Latin even a nominative may replace
an accusative: qui fatetur . et . non timeo (CIC.); a dative
an accusative : cui fidem habent et bene rebus suis consulere arbitran-
tur (ib.). A possessive pronoun may also represent the corre-
sponding personal pronoun; ja was ez ie dîn site unde håst mir da
mite gemachet manege swære (HARTMANN VON AUE); alsobald
stunden seine schenkel und knöchel feste, sprang auf (LU.). Or
a da connected with an adverb may represent the demonstrative
pronoun: dâ mite so müezeget der muot und (das) ist dem lîbe ein
michel guot (GOTTFRID VON STRASSBURG). Finally, two different
elements may, as a single whole, form the subject of a following
verb; as in da vuorte si in bî der hant und sázen zuo einan-
der nider (HARTMANN VON AUE); dô nam daz Constantinis
wîb ir tochter, die was herlich, unde bâtin Dietherîche (ROTHER);
wie herzog Jason wardt verbrandt von Medea also genandt;
hetten doch vor viel zeit vertrieben (H. SACHS); so hertzlieb
•
ECONOMY OF EXPRESSION.
359
von hertzlieb musz scheiden vnd gentzlich kein hoffnung mehr
handt (ib.).
links.
545. We have seen in Chapter XVI. that two principal notions Absence of
may be connected by several links, which serve to define more
closely the nature of the connexion, whether this relation be at
once grammatical and psychological, or purely psychological and
at variance with the grammatical connexion. Now since expres-
sions frequently occur in which these links are dispensed with, we
are easily tempted to explain these as ellipses. This view is
however in many cases to be entirely rejected. If, for instance,
instead of Hectoris Andromache and Caecilia Metelli we can say
more exactly Andromache uxor Hectoris and Caecilia filia Metelli,
it by no means follows that in the shorter expressions the forms.
uxor or filia are to be supplied, for they may be independently
explained from the general function of the genitive; and whoever
assumes an ellipse here must, if he be logical, follow the gram-
marians of the sixteenth century in assuming ellipse for every
genitive. Beside these, however, occur expressions in regard to
which the term elliptical cannot be denied a certain justification,
in so far as they have originated in more complete expressions,
where, however, we are not on that account to assume the omission
of a definite word.
546. Indications of direction were doubtless originally devel-
oped only with verbs of motion. Direction is now often stated
with verbs of preparing and the like; as in MHG. sich bereite von
dem lande vil manic ritter starc (NIBELUNGENLIED); wir suln ouch
uns bereiten heim in miniu lant (ib.); dô soumte man ('accoutred
in') den degenen von dannen wafen und gewant (ib. C.); di sich
gegarwet hâten ze strîte ûf daz velt (ALPHART); do vazte sich der
herzoge in des kuniges hof (then the duke arranged himself to go
to the king's court') (Kaiserchronik), and often elsewhere in this
360
PRINCIPLES OF THE HISTORY OF LANGUAGE.
work; cf. Greek, φανερὸς ἦν οἴκαδε παρασκευαζόμενος (ΧΕΝ.);
similarly, ékéλevσav ẻπì тà Öπλa (ib.).¹ Similarly with MHG.
rûmen: heiz inz rûmen von dan (HARTMANN VON AUE); ich rume
dir daz rîche von hinnen vlühticliche (RUDOLF VON EMS). Cf.
also Greek, ἐκλείπειν τὴν πόλιν εἰς χωρίον. It is not to be
supposed that in such phrases there floated in the speaker's mind
the unexpressed infinite of a definite verb such as to go or ride.
The psychological process to which, e.g., the phrase таρаσкeváčeœÐαι
olkade owes its origin is rather the following. In the first place,
the two notions of preparing oneself, and of the local terminus
which the preparation is designed to attain, present themselves
to consciousness, and are directly connected as psychological
subject and predicate. Since, however, it was already usual in
such sentences as πορεύονται οἴκαδε οι παρασκευάζονται οἴκαδε
πоρeveσlaι to express the local terminus in a definite form,
was also applied in the present case. The process is thus the
work of two distinct factors: first, the capacity which existed in
language even before the rise of its formal elements, and which
never dies out of it, of expressing the mutual relation of two
notions in consciousness whether this relation be given im-
mediately, or only mediately through other notions-by mere
juxtaposition of the terms which denote these notions; secondly,
the analogy of the already developed forms of expression.
this
547. The same process occurs in numerous other cases. Many
of the expressions discussed in Chapter VI., such as scherz bei seite,
wer da? etc., belong to this head. When once the majority of
words had assumed formal elements, the capacity just described,
¹ When such combinations become usual, the meaning of the verb may be modified,
motion in a particular direction being treated first as an essential part of the meaning,
and finally as its principal constituent. Thus in NHG. schicken originally meant to
put right'; reise, a breaking up'; aufbrechen, the opposite of aufschlagen (i.e. the
camp).
ECONOMY OF EXPRESSION.
361
and more closely investigated in Chapter VI., could not come
into play without the meaning of these formal elements assert-
ing itself at the same time. We have now to consider a few
other constructions of this kind which are commonly treated as
elliptical.
548. Most nearly akin to those already discussed are terms
of direction following the verbs: können, mögen, sollen, wollen,
dürfen, müssen, lassen, e.g. ich mag nicht nach hause, ich lasse dich
nicht fort. These have become so far 'usual' that from the
standpoint of our present linguistic instinct they cannot possibly
be described as elliptical. Further, such phrases as er ist weg, er
ist nach Rom, which are to be understood exactly in the same
way as er ist in Rom, i.e. weg and nach Rom are to be taken as
predicates, ist as copula. Cf. Latin constructions such as quando
cogitas Romam? (CIC.); ipsest quem volui obviam (of whom I
wished that he should meet me,' TER.); puto utrumque ad aquas
(CIC.).
549. If we say ich möchte dich nicht anders, als du bist, this
will hardly be explained as an ellipse of haben. An ellipse of
sein would be more natural; but its insertion would produce
an un-German construction. And the same reason which forbids
us to supply a sein here, forbids us equally to supply esse in
the Latin Strato physicum se voluit (CIC.).
550. In Latin a subject-nominative is occasionally followed
by an accusative without a verb: sus Minervam, fortes fortuna,
manus manum, dii meliora; quae cum dixisset, Cotta finem (CIC.) ;
ego si litteras tuas (ib.); quid tu mihi testis? These constructions
are not explained by naming a verb to be supplied. We must
rather say: Two notions are here combined in the form of nomi-
native and accusative, because they stand in the same relation
as that which, in a more complete sentence, holds between subject
362
PRINCIPLES OF THE HISTORY OF LANGUAGE.
and predicate. Similarly to be conceived is the immediate con-
nexion of a subject-nominative with a prepositional determinant
or an adverb, as in itaque ad tempus ad Pisones omnes (CIC.), hæc
hactenus (where hæc could no doubt be equally conceived as accu-
sative), an tu id melius? ne quid temere, ne quid crudeliter; Taûτa
ταῦτα
μὲν οὖν δὴ οὕτως (PLA.). There are analogous instances even
in German; in lively narrative one can say: ich rasch hinaus, ich
hinterher, and the like; graf nun so eilig zum tore hinaus (GOE.);
cf. der sultan gleich dem tone nach (WIELAND).
551. In similar fashion a subordinate sentence may be directly
combined with a principal sentence which, if the thought were
more fully expressed, would have to be connected with it
mediately by another subordinate sentence, or by some sentence-
element. This mode of connexion may then become 'usual,' so
that no omission is any longer felt. Cf. wie Lavater sich hiebei
benommen, sei nur ein beispiel gegeben (GOE.), where for our present
linguistic instinct dafür is wanting before sei; und fragst du mich
nach diesen beiden schätzen: der lorbeer ist es und die gunst der
frauen (GOE.); dass ichs dir gestehe, da ergriff ihn mein gemüt (ib.) ;
besuche deine brüder, obs ihnen wohl gehe (LUT.). To this class
belong also such phrases as was das anbetrifft, was ich davon weiss,
and the like, which have analogies in the most various languages.
Similar is the case with infinitive phrases such as die wahrheit zu
sagen, es kurz zu sagen, um nur eins anzuführen, um von allem
übrigen zu schweigen; further, kurz, ('I do not know,') mit einem
worte, gerade heraus, beiläufig, à propos.
552. Words are supplied from the situation when, instead of
a substantive with a determining epithet, the latter alone is
expressed. We do not here refer to such usages as der gute, to
denote any good person whatever, or das gute to denote any good
thing whatever. There is here no ellipse at all. The notion of
ECONOMY OF EXPRESSION.
363
person in the last resort of male person, or of thing, is denoted by
the gender of the article. We have here to do only with cases in
which there is a reference to a more special notion, as in rechte,
linke (hand); calida, frigida (aqua); alter, neuer, süsser, Burgunder,
Champagner, etc., акρатоs (wine); agnina, caprina (caro): Appia
(via); aestiva, hiberna (castra); natalis (dies): quarta, nona (hora);
τῇ ὑστεραίᾳ, τῇ τρίτῃ ἡμέρᾳ); octingentesimo post Romam conditam
(anno); decima (pars); Ιόνιος (κόλπος); Μουσική, etc. (τέχνη) ;
OHG. frenkisga (zunga). If any one desires to assume ellipse here,
no great objection can be raised. Only we must not ignore the fact
that we supply from the situation, in very many cases, where we
do not think of assuming an ellipse. If we understand der alte to
mean 'old wine,' it is for the same reason that we understand the
same words to mean not any old man, but some particular one
whom we have before us, or who has just been the subject of
conversation. In the cases referred to, the special application of
the adjective has already become more or less 'usual.' The more
fixed the usage becomes, the less does it need to be enforced by
the situation. Thus the terms alter, neuer, are probably only
understood of wine in taverns, in the wine-trade, or where the
attention is already occupied with wine, and they are in general
used only in wine-growing districts; Champagner on the contrary
is, even in the absence of all such mental preparation, far more
readily understood of the wine than of the inhabitant of Cham-
pagne. And as soon as the word becomes intelligible without the
support of the situation, it is to be regarded no longer as an
adjective, but as a true substantive.
553. A perfectly analogous development meets us in the case of
genitive determinants. Cf. Latin ad Martis, ad Dianae (templum);
ex Apollodori (libro); de Gracchi apud censores (oratione); French
la saint Pierre (fête). In German the names of festivals Michaelis,
364 PRINCIPLES OF THE HISTORY OF LANGUAGE.
A psycho-
logical
subject or
predicate
may be
drawn from
Johannis, Martini, etc., and of places St. Gallen, St. Georgen, St.
Märgen, have become completely independent, and are no longer
felt to need supplementing, and therefore no longer felt as genitives.
554. In the cases discussed, an element of a sentence has its
meaning completed from the situation. But it is possible for such
an element, for a psychological subject or predicate, to be entirely
the situation. drawn from the situation. To this class belong the sentences of
apparently one member only, discussed above, p. 118, such as ‘fire,
'thieves,' etc. Even the form of such sentences may be influenced,
in the way described, by the analogy of more complete ones.
When one says, for instance, in a threatening tone keinen schritt
weiter, only the psychological predicate is expressed, the under-
stood subject being the person to whom the threat is addressed.
That the former, however, is in the accusative case, has the same
explanation as in sentences of the form Cotta finem. The same
holds of sentences like guten tag, schönen dank, herzlichen glück-
wunsch and the like. In such cases as glückliche reise, keine
umstände, viel glück, and many others, the form affords no evidence
whether the accusative is meant. In a sentence like manum de
tabula we may take manum as psychological subject, de tabula as
predicate; but the accusative manum shows that here also a
subject is to be taken from the situation, and that the relation of
manum to it is conceived on the analogy of object to subject. The
same is the case with ultro istum a me (PLAUT.), ex ungue leonem
¿§ оvúxwv Xéovтa, malam illi pestem (CIC.), etc. In German we
have such sentences as den kopf in die höhe, and hence also probably
such as gewehr auf, scherz bei seite, davon ein ander mal mehr,
even though the form affords no evidence of the accusative. Other
cases also, prepositional determinants and adverbs can be used in
this way, as the examples adduced show; cf. further, sed de hoc alio
loco pluribus (CIC.), de conjectura hactenus, nimis iracunde.
ECONOMY OF EXPRESSION.
365
555. Sometimes the psychological predicate is also to be taken
from the situation, in which case intelligibility may be promoted
by the tone, facial expression, and gestures. Thus e.g., in sup-
pressed threats: ich will (dich), cf., the well-known Vergilian quos
ego. Further, expressions of wonder or indignation, or of regret,
which only state the object upon which these emotions are directed.
The predicate is here chiefly indicated by the tone of feeling. Cf.
subject nominatives such as dieser kerl, diese fülle, der unglückliche,
ich armer. Also, infinitives such as so lange zu schlafen, so ein
schuft zu sein; Latin tantamne rem tam negligenter agere
(TER.), non puduisse verberare hominem senem (ib.); accusative with
infin. te nunc sic vexari, sic jacere, idque fieri mea culpa (CIC.); cf.
Draeg. § 154, 3.
sentences
dependent.
556. The same explanation applies also to isolated sentences, Isolated
which in form are dependent. They are originally either psycho- in form
logical subjects or predicates, the correlative element being under-
stood from the situation; they may, however, by 'usual' applica-
tion, generally attain the character of independent principal
sentences. Expressions introduced by dass are, like the above
adduced expressions of wonder and regret, originally subjects;
dass du gar nicht müde wirst! dass mir dass begegnen muss! dass
dir auch so wenig zu helfen ist! Further, conditional sentences as
threats: wenn er mir in den wurf kommt ertappe ich ihn nur— ;
Latin verbum si adderis (TER.). Conditional sentences to ex-
press wishes: wäre ich erst da! wenn er doch käme ! Conditional
sentences for which no completion can be found: wenn du noch
nicht überzeugt bist, wenn er aber nicht kommt; Latin si quidem
istuc impune habueris (TER.). Conditional sentences which serve
to reject assertions or proposals which are made in ignorance
of the real circumstances: wenn du in mein herz sehen könntest ;
wenn du wüsstest, wie leid es mir tut. Sentences of wishing and
>
366
PRINCIPLES OF THE HISTORY OF LANGUAGE.
proposing, with dass, are original predicate, or, according to the
grammatical form, objects: dass ich doch dabei sein könnte; MHG.
daz si schiere got gehæne; French que j'aille à son secours ou
que je meure; Italian che tu sia maledetto, and similarly in all the
Romance languages.
CHAPTER XIX.
RISE OF WORD-FORMATION AND INFLEXION.
WE
new creation
formation
its nature.
E have occupied our attention much with the question of Original
analogical new creation in the area of word-formation in word-
and inflexion. We must now direct our attention to original, non- secondary in
analogical creation in this department. This is not, like the
simplest syntactical combinations, primary in its nature, but
secondary only, and of slow development. There are, as far as I
can see, three ways, and three only, by means of which etymological
groups of words are formed out of simple single words standing in
no inner relation to each other. One of these is differentiation of
sound, which is itself followed by a differentiation of meaning. An
example to the point would be the division in IE. between the
imperfect and the aorist¹ (cf. supra, p. 287).
558. Similar divisions are no doubt conceivable enough even in
the primitive elements of language. Still, in most cases, our own
observation will give us proof that no groups are formed, seeing
that in this process the feeling of relationship is lost; and still less
are parallel groups formed, as in the case cited. A second method
¹ Of course the process is very different, even though it conduces to the same result,
when a secondary sound-difference, after the loss of the other differentiating tokens,
becomes the only token of the functional difference, as in the English foot - feet, tooth
-teeth, man men. Where such forms are to be found in our oldest traditions, it will
often prove impossible to decide whether they owe their origin to this process or to that
discussed in the text.
368
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Growth of
composition.
is the junction of converging development of signification with
converging sound-development (cf. suchen and sucht); of which we
have spoken, p. 233. It is obvious enough that any such process
can occur only as an isolated one. Thus the strictly normal
method of the growth of everything formal in language is, and
remains, the third: viz. composition.
559. We have plenty of opportunities of watching the growth
of composition. In the IE. languages there are two strata of com-
pounds which must be distinguished from each other: an older
one which has either come direct from the original language, or
has, at least, been formed on original models; and a later one,
which has been independently developed on the foundation of the
single-languages, and which, in modern languages, has gained a
great prevalence. This last method we can see to a great extent
actually growing before our eyes, and this consistently by means
of the conjunction of elements originally independent.
Con-
nexions of any kind serve this purpose. Thus, for instance, we
find that compounds take their rise from the junction of the geni-
tive with the governing substantive: cf. MHG. hungersnot, hasen-
fuss, freudenfest, kindergarten; French lundi (lunæ dies), Thionville
(Theodonis villa), connétable (comes stabuli), Montfaucon (mons fal-
conis), Bourg-la-Reine; Latin paterfamilias, legislator, plebiscitum,
caprifolium; [English Lady Day, Stagsfoot]; from the union of
the attributive adjective with the substantive: cf. NHG. edelmann
(MHG. still edel man, genitive edeles mannes), altmeister, hochmut,
Schönbrunn, oberhand, Liebermeister, Liebeskind, morgenrot; French
demi-cercle, double-feuille, faux-marché, haute-justice, grand-mère,
petite-fille, belles-lettres, cent-gardes, bonjour, prudhomme, prin-temps,
Belfort, Longueville, amour-propre, garde-nationale, ferblanc, vinaigre,
Villeneuve, Rochefort, Aigues-Mortes; Latin respublica, jusjuran-
dum; [English half circle, grandmother, upper hand, etc.]; further,
RISE OF WORD-FORMATION AND INFLEXION.
369
in NHG. einmal, jenseits (MHG. jensît), einigermassen, mittlerweile;
French encore (hanc horam), fièrement (fera mente), autrefois, autre-
past, toujours, longtemps; Latin hodie, magnopere, reipsa; from the
appositional connexion of two substantives, cf. NHG. Christkind,
gottmensch, fürstbischof, prinz-regent, herrgott, Basel-land; French
maître-tailleur, maitre-garçon, cardinal-ministre, Dampierre (dom-
inus Petrus), Dammarie (domina Maria), OFR. damedeus (dominus
deus); [English master-mariner, prince-regent, head-boy, etc.]; from
the co-ordination of two substances, in NHG. only, to denote the
connexion of two countries, as Schleswig-Holstein, Oestreich-
Ungarn; from the appositional or copulative connexion of two
adjectives, or that of an adverb with an adjective, a difference
which cannot always be distinctly marked, cf. NHG. rotgelb, bitter-
siss, altenglisch, niederdeutsch, hellgrün, hochfein, gutgesinnt, wolge-
sinnt; French bis-blanc, aigre-doux, sourd-muet, bienheureux, mal-
content; from the addition of two cardinal numbers, cf. NHG.
fünfzehn, English ‘fifteen,' Latin quindecim; from the connexion of
the adjective with a dependent case, cf. NHG. dusdrucksvoll, sor-
genfrei, rechtskräftig, Latin jurisconsultus, peritus, verisimilis
from the connexion of two pronouns, or it may be of the article
with a pronoun, cf. NHG. derselbe, derjene (at present only used in
the derivative derjenige), French quelque (quale quid), autant
(alterum tantum), lequel; from the connexion of an adverb, or of
a conjunction, with a pronoun, cf. NHG. jeder (from ie-weder), kein
(from nih-ein), French celle (ecce illam), ceci (ecce istum hic), Latin
quisque, quicunque, hic, nullus; from the connexion of several par-
ticles, cf. NHG. daher, darum, hintan, fortan, voraus, widerum,
entgegen, immer, French jamais, ainsi (aeque sic), avant (ab ante),
derrière (de retro), dont (de unde), ensemble (in simul), encontre ;
Latin desuper, perinde, sicut, unquam, etiam; from the connexion
of a preposition with the dependent case, cf. NHG. anstatt, zunichte,
;
2 A
370
PRINCIPLES OF THE HISTORY OF LANGUAGE.
zufrieden, vorhanden, inzwischen, entzwei, French contremont, par-
tout, endroit, alors (ad illam horam), sur-le-champ, environ, adieu,
affaire, sans-culotte [English uphill, downstairs], Latin invicem,
obviam, illico (= in loco), denuo (= de novo), idcirco, quamobrem ;
from the connexion of an adverb with a verb, cf. NHG. auffahren,
hinbringen, herstellen, heimsuchen, mislingen, vollführen, French
malmener, maltraîter, méconnaître, bistourner [English foretell, gain-
say, withstand], Latin benedicere, maledicere; from the connexion of
a dependent case with its verb, cf. NHG. achtgeben, wahrnehmen
(OHG. wara, with feminine stem) wahrsagen, lobsingen, handlangen,
hochachten, preisgeben; French maintenir, colporter, bouleverser ;
Latin animadvertere, venum dare-venundare-vendere, crucifigere,
usuvenire, manumittere, referre. It is also possible for more than
two component parts to unite into a compound, cf. NHG. einund-
zwanzig, einundderselbe, Latin decedocto (= decem et octo, cf. Corssen,
Aussprache des lat. ii. p. 886¹); French tour-à-tour, tête-à-tête, vis-à-
vis; French aide-de-camp, trait-d'-union, garde-du-corps, Languedoc,
belle-à-voir, pot-au-feu, Fierabras, arc-en-ciel, Châlons-sur-Marne;
Latin duodeviginti; NHG. brautinhaarn (a flower); Latin plusquam-
perfectum; NHG. nichtsdestoweniger [English notwithstanding, never-
theless]; Italian nondimeno. Compounds also rise from dependent
sentences, cf. NHG. newære, contracted from niur, etc. NHG. nur;
Italian, avvegna (adveniat), avvegnache, chicchessia; Latin, quilibet,
quamvis, quantumvis, quamlibet, ubivis. Similarly they arise from
sentences which are formally independent, but which are still
used in logical subordination, e.g. as intercalations; cf. NHG. weiss-
gott, MHG. neizwaz = AS. nat hwat Latin nescio quid, French
je ne sais quoi, MHG. deiswar (=das ist war), French peut-être,
piéça, naguère, Latin licet, ilicet, videlicet, scilicet, forsitan, Spanish
quiza (perhaps,' properly 'who knows '). Further, sentences may
=
1 See Darmesteter, La vie des mots, p. 23.
RISE OF WORD-FORMATION AND INFLEXION.
371
be turned into compounds by the aid of metaphors, and more
especially imperative sentences; cf. NHG. Fürchtegott, taugenichts,
störenfried, geratewol, vergissmeinnicht, gottseibeiuns; French,
baisemain, passe-partout, rendez-vous; Modern Latin, facsimile,
notabene, vademecum, nolimetangere; NHG. jelängerjelieber. It is
harder for a true sentence which has preserved its independence
to pass into a compound. For it is the very essence of a sentence
to denote the act of combining its several members, whilst it
would seem to be the essence of a compound to denote the process
of combination as a result already attained. In spite of this,
however, compound sentences do appear in the most different
languages, and particularly in the IE. and Semitic verbal forms.
from syntac-
position to
true com-
position.
560. The transition from syntactical juxtaposition to true Transition
composition is very gradual; no sharp line of demarcation can be tical juxta-
drawn between them. This is proved by the great uncertainty
which prevails in the orthography of modern languages with
regard to the treatment of various combinations as compounds
or as separate words. This uncertainty has led to an ortho-
graphical compromise the use of the hyphen. In English
the orthographical union of words is often neglected in cases
where, in other literary languages, it would appear indispensable.
In MHG. we often find compounds formed after the IE. model,
written separately.
561. The purely relative character of the difference between
a compound and a mere group of words can have only one
ground, viz., that the cause of such difference displays its activity
in very various degrees. We must not be seduced by the written
language to seek this cause in some supposed special intimacy
of union distinguishing the members of a compound from the
members of a group of words. Combinations such as article and
substantive, preposition and substantive, substantive and attributive
372
PRINCIPLES OF THE HISTORY OF LANGUAGE.
adjective, or dependent genitive, have exactly the same continuity
as a single word. Then, again, the cause has been sought in the
accent. It is doubtless indisputable that the unity of a word
consists in the graduated subordination of its other elements to
the one selected by the accent. But the case is precisely the
same with the unity of the sentence, and of every portion of the
sentence consisting of several words, of every group of closely-
connected words. The accent of an independent word may, in
many cases, be as far depressed as that of a subordinate member
of a compound. In a combination like durch liebe, durch receives
no stronger stress than in durchtrieben; zu in zu bett no stronger
than in zufrieden, herr in herr Schulze no stronger than in hausherr.
Indeed we cannot always consistently maintain this distinction,
that the position of the accent in a compound is fixed and invari-
able, while in a group of words it can change. I can say hérr
Schulze as contrasted with fráu Schulze, but I can say equally der
hausherr as contrasted with die hausfráu. Nor is any special posi-
tion of the chief accent necessary to the formation of a compound;
it may occur in any position at will. It is requisite indeed, in
order that the more recent method of composition may be brought
into line with the earlier, that the system of accentuation be
identical in the two cases. In order, for instance, that a formation
like rindsbraten or rinderbraten should be felt as substantially
identical with a formation like rindfleisch, it was indispensable that
the main accent should fall on the dependent genitive which
precedes. Where, however, the analogy of the older method
of composition does not come into consideration, in this case, even
in German, the stronger stress laid upon the second element in no
way prevents the formation of a nominal compound.
562. We have to seek the distinction between a compound and
a group of words united under one main accent, not in any physio-
RISE OF WORD-FORMATION AND INFLEXION.
373
logical cause, but in the nature of the psychological grouping.
The one essential point is that the whole as such be in some way
isolated from the elements of which it is composed. The degree
of isolation necessary in order to cause the fusion to pass into a
compound cannot be expressed in any universally applicable
definition.
563. In this question, all the different kinds of isolation which
we have previously discussed have to be taken into account. It
may either happen that the whole undergoes a development which
the single parts in their independent usage do not share, or that,
conversely, the single parts undergo a development in which the
whole does not participate, and this equally in respect of the mean-
ing and of the sound-form; or again it may happen that the single
parts in their independent usage disappear; or finally the particular
mode of combination may become obsolete and survive only in the
special formula.
564. The occurrence of any one of these processes may suffice
to convert a syntactical combination into a compound. But it
is by no means true that we commonly regard as a compound
every composite phrase in which such isolation has already made
its appearance. It is precisely such combinations as these which
call for our special attention, if we would observe the first
beginnings of the process of fusion.
565. The first step towards isolation commonly consists in
the syntactical combination as a whole acquiring a signification
which does not precisely tally with that which results from the
juxtaposition of the several elements. We have already become
acquainted with this process (cf. p. 82). The consequence is
that the single elements of the combination come no longer
clearly into consciousness. But by this very process the nature
of their combination is obscured, and this constitutes the first
374
PRINCIPLES OF THE HISTORY OF LANGUAGE.
step towards a syntactical isolation, which brings with it a formal
isolation also. The process of isolation, however, once started,
implies the possibility of further advance.
566. With regard to syntactical isolation we must distinguish
two cases. Such isolation may conceivably affect merely the
relation in which the members of the compound stand to each
other, as, e.g., in hungersnot, edelmann; but it is also possible.
for the combination as a whole to become isolated as regards
the other components of the sentence. The result is, then,
invariably an uninflected word; cf. keineswegs, gewissermassen,
jederzeit, alldieweil, zurecht, abhanden, überhaupt, vorweg, allzumal ;
Latin, magnopere, quare, quomodo, hodie, admodum, interea, idcirco,
quapropter, quamobrem ; French, toujours, toutefois, encore (=hanc
horam), malgré (=malum gratum), amont, environ, parmi, pourtant,
cependant, tout-à-coup. Such combinations as these become
capable of inflection by secondary development only, as, for
instance, zufrieden, débonnaire ( = de bonne air). Where the
capacity for inflection has not been destroyed by isolation, the
case may occur that the fusion of the members is arrested by
means of inflexion occurring in the interior of the group; e.g.
in such a combination as das rote meer, mare rubrum, in which
cases we are always reminded of the independence of the single
members by the inflexion des roten meeres, maris rubri, etc. A
further process must set in before complete fusion becomes pos-
sible, namely, the crystallisation of an inflexional form (as a rule
that of the nominative singular), in consequence of the obscura-
tion of its original function; a process which we have already
described (p. 194).
567. As we have seen on p. 253, the compound acquires the
same capacity of producing compounds from itself as the simple
word of the same class. Now we find that from a syntactical com-
RISE OF WORD-FORMATION AND INFLEXION.
375
bination which is not yet commonly regarded as a compound, a
derivative is composed on the model of the simple word, or that
this combination, like a simple word, is made into the member of a
compound after already existing models. We must draw the con-
clusion from this that the instinct of language has apprehended
these as a unity, and that thus, in any case, their development into
a compound is already, up to a certain point accomplished.
568. In the case of copulative combinations the process of
fusion begins to operate as soon as it is possible to bring the whole
combination under one single conception. This is the case, first,
when the combined elements are synonyms representing the same
thing from a different point of view; cf. art und weise, grund und
boden, wind und wetter, weg und steg, sack und pack, handel und
wandel, hangen und bangen, tun und treiben, leben und weben, wie
er leibt und lebt, frank und frei, weit und breit, hoch und teuer, angst
und bange, ganz und gar, drauf und dran, nie und nimmer [bag and
baggage, wind and weather, moil and toil, etc.]. In the second
place, when the elements combined are antithetic terms which
reciprocally supplement each other, e.g., stadt und land, himmel und
hölle, wol und wehe, alt und jung, gross und klein, arm und reich,
dick und dünn, lieb und leid, tun und lassen, dieser und jener, einer
und der andere, dies und das, ab und an, ab und zu, auf und ab, ein
und aus, für und wider, hin und her, hin und wider, drüber und
drunter, hüben und drüben, hie und da, dann und wann [hither and
thither, up and down, there and back]. There are besides many
other cases, such as haus und hof, weib und kind, kind und kegel,
mann und maus [house and home, kith and kin, etc.]. The two
members may also be formed by the same word, cf. durch und durch,
für und für, nach und nach, über und über, wider und wider, fort
und fort, der und der [again and again, out and out, through and
through]. In the case of der und der, however, the two members,
376
PRINCIPLES OF THE HISTORY OF LANGUAGE.
in spite of their formal identity, stand in contrast to each other.
In the case of some of these combinations a more thorough process
of isolation has set in. In the case of substantives we have a
criterion of the conception of a copulative combination as a unity,
in the agreement of an adjective attached to the combination with
its second member; cf. durch meinen trewen hilff vnd rat (H. SAchs);
mit allem mobilen hab' und gut (GOE.). Another test, which occurs
more frequently, is the incapacity for inflexion of the first member.
In the case of the above-cited combinations of words in themselves
capable of inflexion, the inflexion which would recall the indepen-
dent nature of the members is in most cases discarded: thus
cannot say mit sacke und packe or grundes und bodens. Cases also
occur, however, of inflexion in the second member alone; e.g. des
zu Abdera gehörigen grund und bodens (WIE.). Compare again von
tausend durchgeweinten tag- und nächten (GOE.); dem wenigen
glaube, liebe und hoffnung (GOE.); in Hans Sachs we even find
dem nimmer golt noch geldts gebrach. In German we constantly
find that the inflexion is omitted in the interior of the sentence
where two adjectives are conjoined; cf. die blank- und blossen wider-
sprüche (LE.) gegen inn- und äussern feind (GOE.), auf ein oder die
andere weise (LE.), mit mein und deinem wesen (LE.).¹
we
A
569. Even in modern German the internal inflexion is neces-
sarily dropped in a sentence like einer schwarz- und weissen fahne,
schwarz- und weisse fahnen, which differs in sense from schwarze
und weisse fahnen. The combinations, also written as such, einund-
zwanzig, einunddreissig (formerly inflected as eines und zwanzig),
are analogous to the expression schwarz- und weiss [cf. the English
¹ The omission, however, of the inflexion of the first member is not an absolute proof
that the two members have been conceived as a notional unity. It is common in early
NHG., and even in Goethe, where the combination consists of two adjectives. Hans
Sachs can even say weder mit böss noch guten dingen. It is less common in combina-
tions of two substantives, cf. von thier vnd menschen (H. SACHS), von merck vnd
steten (ib.)
RISE OF WORD-FORMATION AND INFLEXION.
377
one-and-twentieth]. Fixed combinations, no longer admitting of
any flexion in the middle are, further, all und jeder, ein und alles.
The word einundderselbe is written together; the ein being some-
times inflected and sometimes not. The Greek κаλокayalós was
probably formed from analogy under the influence of the old IE.
method of composition, otherwise the stem-form kaλo would be
hard to account for. Complete fusion would probably be of more
frequent occurrence if it were not checked by the copulative
particle. This check is removed when this particle, owing to
phonetic weakening, is no longer recognised as such; as in the
case of the Low German ritensplit, compounded of the imperatives
of riten and spliten ('tear and split,' Ger. reissen and spleissen). A
copulative combination without a particle fuses more easily. Thus
schwarzrotgolden and Oestreich-Ungarn, which logically are related
to each other as schwarz und weiss and Neapel und Sicilien, are
felt as real compounds. In the period of the IE. languages when
no inflexion and no copulative particle as yet existed, or when at
any rate both were not absolutely indispensable, the fusion into a
copulative compound (dvandva) must of course been very easy.
570. The combination of a substantive with an attributive,
genitive, or any other kind of determinant may pass through all
the phases of signification-change described in Chapter IV. without
the substantive, as such, being thereby affected. It happens very
commonly, in the first place, that the whole acquires a richer and
more specific meaning than that yielded by mere combination of
the parts. The determinant often serves, especially, to indicate
merely some one characteristic, while others which coexist with this
are ignored. Further modifications may then set in, in consequence
of which the epithet in its strict meaning is no longer appropriate.
Thus in the language of botany viola odorata does not mean a
sweet-smelling violet, but a special kind of violet, characterised by
378
PRINCIPLES OF THE HISTORY OF LANGUAGE.
other qualities than its sweet smell; and the name is applied even
to a dried violet which no longer emits any trace of a sweet smell,
and to the plant when not in flower. By the French moyen âge we
understand a strictly defined period of time, though no such defini-
tion is involved in the word moyen itself. Geheimer rat and wirk-
licher geheimer rat are titles which, as wholes, possess a definite
traditional value, which cannot be gathered from the words ge-
heim and wirklich in themselves; cf. further, der heilige geist, die
heilige schrift, die schönen künste, gebrannte mandeln, kaltes blut, der
blaue montag, der grüne donnerstag, der heilige abend, die hohe
schule; der stein der weisen; die weisen aus dem morgenlande. As
regards the substantival determinants, we have still to remark that
they are only able to fuse with the word defined into one single
conception if their signification is not, in 'occasional use,' indi-
vidualised; i.e. they must, except in the case of proper names,
and the denominations of such objects as are conceived of as
existing only once, be employed in an abstract sense. Many
compounds exist which are analogous to the instances cited of
syntactical combinations, including both cases in which the process
of fusion is historically traceable, as schwarzwild, weissbrot, dünn-
bier, rotdorn, sauerkraut, edelstein; haubenlerche, seidenraupe, blu-
menkohl, bundesrat; arc-en-ciel; and cases in which the mode of
composition belongs to a prehistorical epoch, as eisbär, holzwurm,
hirschkäfer, steineiche. It not seldom happens that the same con-
ception is expressed by a compound in one language, and in
another by a syntactical combination; cf. e.g. mittelalter with
moyen âge.
571. A subdivision of this great class is formed by generic
names of locality, which by the aid of some determinant, which
may itself be likewise unspecific, have become proper names. Such
are die goldene aue, das rote meer, der schwarze see, der breite weg
1
RISE OF WORD-FORMATION AND INFLEXION.
379
[the name of a street in Magdeburg and elsewhere], die hohe pforte
[the name of a gate in Magdeburg]; die inseln der seeligen, das cap
der guten hoffnung. With these may be compared such compounds
as Hochburg, Schönbrunn, Kaltbad, Lindenau, Königsfeld; Hirsch-
berg, Strassburg, Steinbach. Here too must be mentioned the case
in which an epithet added to a proper name as a distinguishing
mark passes into an integral portion of the proper name, being
learnt as attaching to a definite individual; cf. Karl der grosse—
der kahle—der kühne—der dicke, Ludwig der fromme―der heilige
-das kind, Wilhelm der eroberer; Davos platz - Davos dörfli;
Basel land-Basel stadt; Zell am see. With these we may
compare such compounds as Althans, Kleinpaul; Gross-Basel—
Klein-Basel, Oberfranken—Unterfranken, Eichen-Barleben; Kirch-
zarten.
572. The metaphorical application of a word is rendered recog-
nisable and intelligible not merely by the context (cf. p. 74), but
in a special degree by addition of a determinant; cf. der löwe des
tages, des haupt der verschworenen, die nacht des todes, der abend des
lebens, die seele des unternehmens [the heart of the enterprise, the life
of the undertaking]. The same effect is produced by a determi-
nant when it forms an element in a compound. Thus we ven-
ture with the aid of such a determinant to use metaphors such
as, if we considered merely the undetermined word, we should
refrain from using, because the determining element contains at
the same time a correction of the metaphor; cf. neusilber, katzen-
gold, ziegenlamm, bienenkönigin, bienenwolf, ameisenlöwe, äpfelwein,
namenswetter; hirschkuh, heupferd, seelöwe, buchweizen, erdapfel,
gallapfel, augapfel, zaunkönig, stiefelknecht, milchbruder.
573. We have to distinguish from these cases, those in which
the compound has a proper as well as a metaphorical meaning,
and only as a compound acquires its metaphorical use, as himmels-
380 PRINCIPLES OF THE HISTORY OF LANGUAGE.
schlüssel, hahnenfuss, löwenmaul, schwalbenschwanz, stiefmütterchen,
brummbär [swallow-tail, negrohead (of tobacco)].
574. The terms discussed above (p. 81), drawn from parts of
the body and mind or from clothes, are almost exclusively syn-
tactic combinations or compounds. The reason of this is that the
simple words, as in themselves lacking characterisation, would be
useless for any such application.
575. It is now our task to trace further how the fusion of the
determinant with the determinate is furthered by the syntactical
and formal isolation.
576. In the case of the coalescence of the genitive with its
governing substantive in German, we have, in the first place, to
observe that it only occurs where the genitive precedes. The
converse position is unsuitable for composition in the first place,
because where it occurs an inflexion occurs in the interior of the
combination whereby we are continually reminded of the inde-
pendence of the elements; and for this reason the fusion in
such a Latin word as pater-familias is less complete than in
plebiscitum. Further, where the genitive precedes, the accentua-
tion becomes analogous to that of genuine compounds
(OHG. táges stèrro = tágostèrro; on the other hand, stérro des
táges). The decisive impulse, however, to coalescence consists in
changes of the syntactical application of the article. As this
has in many cases sunk to the position of a mere case-sign, it has
especially become indispensable with the genitive of all appella-
tives not connected with an attributive adjective. Only the
clearly-marked genitive singular of the strong masculines and
neuters still occurs at times without an article, especially in pro-
verbs, such as biedermanns erbe, and titles such as schäfers klagelied,
geistes gruss, wandrers nachtlied. In OHG. the article is still
quite commonly absent. While this construction gradually died
RISE OF WORD-FORMATION AND INFLEXION.
381
out, certain combinations without article were traditionally handed
down, and thus the fusion became complete. It was still further
specially favoured by the usage, originally universal and then
likewise dying out, of placing the genitive as in Greek between
the article and its substantive. This construction long maintained
itself, especially in the language of the popular epic, only, it is
true, in the case of proper names and allied words; cf. in the Nibe-
lungenlied, daz Guntheres lant, das Nibelunges swert, diu Sivrides
hant, daz Etzelen wip, etc. Combinations such as der gotes haz,
segen, diu gotes hant, etc., are still habitual in the thirteenth century.
In the older period the genitive of every substantive might be thus
inserted without being itself connected with the article; cf. ther
mannes sun ('the son of man '), common in Tatian, then hîuuiskes
fater ('patremfamilias'), ib. 44, 16 (on the other hand, thes h. fater
72, 4; 147,8; fatere hîuuiskes 77, 5), ein ediles mann (‘a man of
noble descent'), Otfrid iv. 35, I. We find a similar insertion
between the ordinal number and the substantive in zwâ dûbono
gimachun ('two pairs of pigeons'), Otfrid i. 14, 24. As the
immediate juxtaposition of article and substantive gradually
became necessary, the combination was necessarily apprehended
by the instinct of language as a unity. In the course of time.
many formal cases of isolation have been added to these, the
older forms of the genitive having been preserved in composition
(lindenblatt, frauenkirche, hahnenfuss, schwanenhals, gänseleber,
Mägdesprung, nachtigall, etc.). These cases have arisen further
from the fact that, in the case of masculine and neuter mono-
syllables, the syncopated forms have become general in composi-
tion, and the non-syncopated in the simple word; cf. hundstag,
landsmann, schafskopf, windsbraut, as contrasted with hundes, etc.
(though we have also gotteshaus and liebeskummer). A final
source of such instances is that the genitive form in the compound
382
PRINCIPLES OF THE HISTORY OF LANGUAGE.
often corresponds with that of the nominative plural, and is
hence connected with the latter, there is a tendency by the
linguistic instinct where the signification agrees; cf. bienenschwarm,
rosenfarbe, bildersaal, äpfelwein, bürgermeister. In the last-
mentioned case the form also coincides with that of the nominative
singular in the case of Baierland, Pommerland (OHG. Beiero lant)
with this alone, while the plural of the simple word has changed
its inflexion.
577. The most ancient stratum of genitival compounds in
French proceeded from the old Latin genitives without the addi-
tion of the preposition de. In old French this method of con-
struction was, at least in the case of personal notions, still in
common and living use; e.g. la volonté le rei (='du roi'); it had
to give way gradually, because the form had come to coincide.
with that of the dative and accusative, and thus the relation
denoted was no longer clear. Certain traditional remains of the
old usage have maintained themselves down to the present day
without being orthographically denoted as compounds; rue St.
Jacques, etc., église Saint Pierre, musée Napoléon. In other cases
the fusion has become more complete, in part favoured by isola-
tion of another kind; cf. Hôtel-Dieu, Connétable (comes stabuli),
Château-Renard, Bourg-la-Reine, Montfaucon, Fontainebleau (f.
Blialdi). [In English, Lady-day.] In French, unlike German,
owing to the disappearance of all case-signs, the fusion is rendered
possible, even when the genitive is placed in the second place.
It must have been first completely effected in compounds where
the genitive preceded, since this order early became obsolete ;
thus we have Abbeville (abbatis v.), Thionville (Theodonis villa), etc.¹
578. The coalescence of the adjective with the substantive
1 Cf. such instances as Houlegate, v. Kluge, die Normannen, in Gröber's Grundriss
der romanischen Philologie, p. 394. See also Gröber, u.s. p. 423.
1
RISE OF WORD-FORMATION AND INFLEXION.
383
belonging to it originates, in German, principally in the so-called
uninflected form, which in its attributive use gradually becomes
obsolete; cf. above, p. 157. In MHG. (ein) junc geselle, (ein) edel
mann, (ein) niuwe jâr, are still quite ordinary constructions; in
NHG. junggeselle, edelmann, neujahr can only be apprehended as
compounds. A wider basis for the commencement of this process.
is afforded by the weak nominatives of polysyllabic adjectives
in r, l, n, which in MHG. reject their e, while this is in NHG. restored
after the analogy of the monosyllabic ones. In MHG. der ober roc,
diu ober hant, daz ober teil appear still as regular syntactical
combinations (whence we also find the accusative die obern hant
by the side of die oberhant); in NHG. der oberrock, die oberhand, das
oberteil can only be regarded as compounds, because they would
otherwise have to appear as der obere rock, etc. In spite of this
the simple plan of abiding by the old condition is insufficient for
the creation of actual composition, and many compounds of the
kind took their rise before the appearance of this syntactical
isolation. As early as in OHG. we find altfater, frîhals, guottât,
hôhstuol, and many others. The process is rather that the com-
bination becomes so stereotyped, the notion conveyed by it so
completely single, that internal inflexion is felt by the linguistic
instinct as an inconsistency, and it is then only natural that the
form adopted should be the really normal case, viz., the nominative
singular, which, owing to the disappearance of the inflexion-ending,
appears at the same time as the stem of the word. Since the
uninflected form had ceased to be employed attributively, the
fusion of the adjective with the substantive became correspondingly
more difficult. For the inflected forms of the nominative singular
(guter, gute, gutes) had, from the very outset, a much smaller
range, and were, from the very circumstance of their inflexional
termination, not so fitted to stand as representatives of the word
384
PRINCIPLES OF THE HISTORY OF LANGUAGE.
in itself. There was now, however, less need for such fusions,
since a quantity of compounds was already at hand with their
uninflected forms, which were also capable of producing analogical
new formations. Still, even in this period, certain cases of fusion,
complete or incipient, appear; the explanation being, in some
cases, that combination has been treated on the analogy of the
older combinations already fused (cf. geheimrat by the side of
geheime(r) rat); in other cases, that the inflected nominatival
form is generalised, as in krausemünze, jungemagd, in Gutersohn,
Liebeskind, and other proper names. In the case of certain words
the feeling for the unity of the conception betrays itself in the
fact that, instead of internal inflexion, the two elements are
written as one; cf. langeweile, hohepriester, hohelied. Lessing
even writes: ein Jüngstesgericht en mignature. Compare also
derselbe, derjenige.
579. Even where no complete fusion of the attributive ad-
jective with the substantive belonging to it has taken place,
derivations are still made from the combination; cf. hohepriesterlich,
langweilig, kurzatmig, hochgradig, vielzüngig, vielsprachig, rotbäckig,
einhändig, blauäugig, blondhaarig, kleinstädtisch, kleinstädter,
schwarzkünstler, tausendkünstler, einsilbler; which stand on pre-
cisely the same footing as grossmütig, edelmännisch, etc. We
are forbidden to treat these as noun-compounds by the circum-
stance that many of the simple words which would in this case
have to be prefixed, such as -weilig, -atmig, -gradig, do not
exist at all, and in fact have never existed.
580. Such combinations are also employed to form compounds
which, in spite of all the opposition of the grammarians, stubbornly
hold their ground. The ordinary objection made to such com-
binations as reitende artillerie-caserne ('mounted artillery barracks'),
that the 'barracks' are not 'mounted,' will not bear examination.
RISE OF WORD-FORMATION AND INFLEXION.
385
No one who uses the expression means to imply that they are,
and the true distribution is not reitende + artillerie-caserne, but
reitende artillerie + caserne. But a difficulty arises owing to the
inflexional and congruent nature of the adjective. The latter,
therefore, commonly accommodates itself to the second element,
not merely in cases where it could in any case be referred to
this as well as to the first, as in französischer sprachlehrer, freie
handzeichnung, but also in other cases, as in der sauern gurkenzeit.
In the case of many of these combinations it has become usual
to unite them continuously; cf. alteweibersommer, armesünder-
glöckchen, etc. In spite of this the adjective often agrees with the
latter element. Goethe writes: auf dem armesünderstühlchen ;
Heine, on the other hand, auf einem armesünderbänkchen; the
Kölnische Zeitung, nebst armsündertreppe. Klopstock even em-
ploys hohpriestergewand, and Luise Mühlbach den gutennachts-
gruss.¹ In English, where no inflexion produces confusion, such
combinations cause no difficulty.
difficult than in German,
Where only the distinc-
suggestion of composite-
581. In French coalescence is less
because the distinction of cases is lost.
tion of singular and plural remains, the
ness is necessarily far less frequent. Besides this there are many
combinations which from their very nature occur in the singular
only (e.g. sainte-écriture, terre-sainte), or again, in the plural only
(e.g. beaux-arts, belles lettres). Hence the sense of the essential
unity of such phrases commonly asserts itself by the use of the
hyphen. Another more important means of obtaining the evidence
of the linguistic instinct is afforded by the employment of the
partitive article. It is in French usual to say, for instance, il a des
belles lettres, just as it is to say il a des lettres, while it is correct
to say il a de belles maisons. Instances of isolation alike of form
1 Cf. Andresen, Sprachg. p. 152 and 64.
2 B
386
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Of a verb
with an
adverb.
and syntax may here come in to strengthen the union yet further.
In old French the adjectives which in Latin are inflected according
to the third declension have as yet taken no e in the feminine,
which was not added till later, according to the analogy of
adjectives of three terminations; e.g. grand = grandis; later
grande, after the analogy of bonne. In the case of compounds.
forms without e maintain themselves: grand'mere, grand messe,
Granville, Réalmont, Ville-réal, Rochefort. In the case of Vaucluse
(vallis clausa) the compound, apart from other phonetic peculiarities,
has not undergone the change of gender of the simple word (le val)
which has taken place in modern French. There follow assimila-
tions just as in German. In the case of adjectives which occur
more frequently in composition, the form of the masculine and
of the singular is generalised; thus in mi-, demi-, mal- (malfaçon,
malheure, maltôte), nu- (nu-tête, nu-pieds). By this means the
composition is plainly marked.
582. In cases where in NHG. the genitive has coalesced with
a governing adjective, it frequently appears that the construction
is either obsolete or grown unusual, and replaced by another;
e.g. ehrenreich reich an ehren, geistesarm-arm an geist, freuden-
leer leer von freuden.
583. In NHG. it is common in cases where adverbs, according
to the ordinary syntactical rules, precede the verb, to unite them in
writing with the verb: cf. aufheben, vordringen, zurückweichen, weg-
werfen, etc. We can gather that no composition, strictly so called,
has set in by the transposition (er treibt an, er steht auf, etc.) On
the other hand, the circumstance that the words are conjoined in
writing denotes that the whole has begun to be apprehended as a
unity.
584. In the case of most of these combinations we can distinctly
trace an isolation setting in against the elements. The old pre-
RISE OF WORD-FORMATION AND INFLEXION.
387
positional adverbs in general cannot any longer be employed freely
and independently, but are confined to a special group of combina-
tions. Instead of these, combinations with her and hin are chiefly
employed for free syntactical combination. Cf. hinaus gehen,
heran kommen, which differ essentially in meaning from ausgehen
and ankommen. There then sets in in most cases an independent
development of meaning of the combination as such: e.g., anstehn,
ausstehn, vorstehn, zustehn, auslegen, aufbringen, umbringen, zubrin-
gen, auskommen, umkommen, vorwerfen, vorgeben, etc. However,
the apprehension of these combinations as compounds is aided
by the parallel noun-compounds like ankunft, abnahme, zunahme,
vorwurf, ausspruch, zusage, anzeige, etc. These of course, influence
most easily the nominal forms of the verb in which the combina-
tion is already most firm, and becomes more so, the more they
approach the character of a pure noun (cf. the following chapter).
They possess the firmest character naturally when they only—and
not the finite verb-are usual in a definite meaning or remain so;
cf. aufsehen, nachsehen, abkommen; ausnehmend. In the case of
the participle the fusion may appear in the formation of the com-
parative or superlative, which are meaningless unless the whole is
apprehended as a unity, cf. die zwei entgegengesetzesten eigenschaften
(GOE.), der eingeborenste begriff (GOE.), unter nachsehendern gesetzen
(LE.); other examples may be found in Andr. Sprachg., p. 119.
From the union of the verb with the adverb further arise nominal
derivatives which are indisputably verbal unities, such as austreibung,
vorsehung, auferstehung, abschreiber, anstellig, ausgiebig, zulässig,
angeblich, absetzbar. For adverbial determinants the same rule
holds good as for the substantival words affixed to nouns, namely,
that 'occasional' individualisation stands in the way of fusion.
1 We might be tempted to regard these words rather as nominal compounds; but by
doing so we should set ourselves in opposition to the instinct of language, and in some
cases we should arrive at simple words which have no existence, such as stellig and geblich.
388
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Hence it is that the demonstrative adverbs of place, for instance,
maintain their independence; wer da ist, her kommt, and not daist,
herkommt. In the case of the nominal forms, it is no doubt true
that orthographical cases of union occur like sein hiersein, but we
do not in spite of these apprehend the whole so thoroughly as a
unit as we do such words as einkommen, zutrauen. The case is far
different with dasein in the sense of 'existence;' in this case precisely
da is not individualised. The conditions are much the same with
herkommen and darreichen, darbringen, etc., because dar has lost
its original function of demonstrative adverb equivalent to dahin.
There appear also to be symptoms of an inclination to turn the
finite verb as well into a true compound. In German newspapers-
and Germanists are at one with them on this point-it has become
the fashion to write er anerkennt [for erkennt an]. We thus see
plainly enough the path on which the old verbal compounds in
German (such as durchbréchen, betreiben), and in the other IE.
languages have taken their rise from syntactical combination.
585. An adverb derived from an adjective sometimes fuses with
the nominal forms of the verb. The first impulse to this fusion is
sometimes given by the metaphorical application of one of the two
component parts, cf. tieffühlend, weitgreifend, weittragend, hochflie-
gend. The combination becomes even closer when the first part
retains a function which, in general, it has lost. Under this head
we may class combinations with wol, such as wolleben, wolschmeck-
end, wolriechend, woltuend, etc., which have been handed down.
to us from the time when wol was still the common adverb of gut.
We may further compare the word erstgeboren, which dates from a
time when the word erst bore the sense of the zuerst of the present
day. No doubt the analogy of nominal compounds is here also at
work; cf. zartfühlend -zartgefühl, scharfblickend - scharfblick.
Here too the comparative and superlative forms may serve to test
RISE OF WORD-FORMATION AND INFLEXION.
389
the completion of the fusion; cf. bis zur schwerfälligsten, klein-
kauendsten weitschweifigkeit (SCHOP.); der tieffühlendste geist
(GOE.), die reingewölbteste stirn (ib.), die freigelegenste wohnung.
Superlatives like weitgreifendste, hochgeehrtester, hochverehrtester
are common. Still more remarkable are cases in which, from a
combination in which the adverb is already superlative, another
superlative is formed; cf. die zunächststehendsten (Frankfurter
Zeitung).¹
with an ob-
tive.
586. There are many combinations of a verb with an object- Of a verb
accusative which occupy a similar intermediate position between ject-accusa-
the compound and the syntactic group; cf. acht geben or achtgeben,
haushalten, standhalten, stattfinden, teilnehmen; further, combina-
tions of a verb with a predicate-adjective, like loskaufen, freigeben,
freisprechen, feilbieten, feilhalten, hochachten, wertschätzen, gut-
machen. The causes which in this case further the impulse to
composition are exactly the same as in the case of the combina-
tions which contain an adverb. There do occur, however, in the
process displacements of distribution, especially in the case of the
fusion of the predicative adjective, where this is normal, see p. 328.
Of course the transition to a compound is here also the easiest in
the case of nominal forms. There occur a great number of
participles which have coalesced with an objectival accusative;
cf. feuerspeiend, grundlegend, notleidend, leidtragend, wutschnaubend,
segenbringend, nichtssagend. Here too the comparative forms may
serve as a test whether real fusion has set in; cf. die nichtsbedeu-
tendsten kleinigkeiten (SCH.), das grundlegendste der maigesetze
(Kölner Zeitung), am gefährlichsten und feuerfangendsten (Deut-
scher Reichstag).2 But no sharp line can be drawn between
spontaneous fusion and formation by analogy after the model of
1 The examples are taken from Andr. Sprachg., p. 120 and 42-3, where still more
are adduced.
2 Cf. Andresen, u.s.
390
PRINCIPLES OF THE HISTORY OF LANGUAGE.
With a
prepositional
nominal compounds which certainly meets us in words like saft-
strotzend, kraftbegabt, mondbeglänzt, but which are almost wholly
confined to the high poetic style. We have transition into true
composition in the case of lobsingen, and wahrsagen (wahr is taken
substantivally and denotes truth), in which the influence of derivatives
from compounds, such as ratschlagen and weissagen, cf. p. 270, may
have contributed to the result. Derivatives are also formed from
combinations in which the fusion is not yet perfectly accom-
plished, cf. haushälter, teilnehmer, freigebig; even grundlegung,
preisverteilung, waffenträger, holzhauer, etc.; and further bekannt-
machung, kundgebung, lostrennung.¹
587. Like adverbs, substantives dependent on a preposition
determinant. coalesce up to a certain point with the verb. It is true that it is
not usual to write continuously combinations like su grunde legen
or in stand setzen, except in the case of the substantival use of the
infinitive, but we form the derivations zugrundelegung, instandset-
zung, ausserachtlassung, zuhülfenahme. Besides, we have such
superlative formations as an dem sichtbarsten, in die augen fallend-
sten orte (LE.)
Verbal Com-
have the
character of
588. I would call attention, further, to the numerous combina-
plexes which tions which, like the copulatives cited above, are not commonly
Compounds apprehended as compounds, but which, nevertheless, represent a
single notion; e.g. so wie so, vor wie nach, mann für mann, schritt
für schritt (cf. French vis-à-vis, dos-à-dos, tête-à-tête), von neuem,
von hause aus, sobald als möglich, so gut wie, was für ein, etc. In
many of these combinations the coalescence into an unity is
at the same time a displacement of the distribution in the sen-
1
¹ In this case also a doubt may arise whether the words in question are not to be
apprehended as nominal compounds, but the instinct of language again decides for the
view here taken. No doubt the analogy of nominal composition may to some extent
have cooperated, but formations like freisprechung, bekanntmachung could from their
signification not suit this analogy; otherwise they would necessarily be = freie sprechung,
bekannte machung.
RISE OF WORD-FORMATION AND INFLEXION.
391
If, for
tence, which announces itself in the mode of construction.
instance, Lessing says ein mehr als natürliches gift, the attributive
application of mehr als natürlich and the inflexion at the end has
only become possible by the fact that this combination is appre-
hended as a unity like übernatürlich, and that accordingly the
The
feeling for the method of combination has vanished.
same is the case in the following constructions: mit einer nichts
weniger als schönen bewegung (LE.), in so wenig als mögliche worte
(LE.), ausser der so lang als möglichen dauer (LE.). More striking
still, and differing in the fact that an inflexion occurs in the middle
of the combination, is the construction which occurs several times
in Lessing in der letzten ohn eine zeile. For so gut wie we may
compare phrases like er hat mirs so gut wie versprochen. The sub-
stantive belonging to was für (= qualis) depended originally on für.
Thus, for instance, the phrase was habt ihr für pferde means pro-
perly 'what have you in the place of horses.' If, however, we say
at the present day mit was für pferden, this shows that the words
was für are apprehended by the instinct of language as an inde-
clinable attribute to the substantive, which should properly depend
on für.
between the
compound
independent
589. The impossibility of drawing a sharp line between com- Coordination
pound and syntactical combination manifests itself also in this, elements of a
that frequently members of an otherwise undoubted compound are and an
placed upon the same footing with independent words. We do word.
not shrink from saying öffentliche und privatmittel, das ordinäre und
das feierkleid. Hans Sachs even combines gesotten, pachen vnd
pratfisch. Further, the first or determining member of a com-
pound may itself be followed by determinants, as though it were
an independent word; and not merely by such as might be also
applied to the whole, such as dankesworte für die gnade, but others
also, such as ein herausforderungslied zum zweikampf (LE.); ein
392
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Sound-
change with
isolating
effect.
böses erinnerungszeichen für ihn an die treulosen Griechen (HER.) ;
glaubensfreiheit an wunder und zeichen (GOE.); der vertragsent-
wurf mit Deutschland (Kölner Zeit.); hoffnungsvoll auf die zukunft
(GOE.); erwartungsvoll des ausgangs (WIE.); hopeless to circumvent
us join'd (MIL.); fearless to be overmatch'd (ib.). Lastly, pronouns
are referred to a member of a compound: cf. menschengebote, die
sich von der wahrheit abwenden (LU.); er hatte einen ameisenhaufen
zertreten, die seine herrschaft nicht anerkennen wollten (GOE.); es
gibt im menschenleben augenblicke, wo er dem weltgeist näher ist
als sonst (SCHI.).
590. Sound-changes which effect isolation are promoted by
various circumstances in the traditional groups.
We may pro-
bably with reason maintain, even though we be unable to follow
the development historically, that such changes, for the most part,
occur at first as a general result of a closer syntactical combina-
tion, but are afterwards displaced by levelling, and are maintained
only in cases where, as a consequence of the development of
meaning, the elements have already coalesced too closely for
change. The change most easily effected is the transference of a
final consonant to the following syllable; cf. NHG. hinein, hieran,
allein, einander; Latin etenim, etiam. Attraction of this kind has
no tendency to produce isolation in cases where, as in French, it
regularly occurs in all close syntactic combination. In cases, for
instance, like peut-être it cannot conduce to the formation of a closer
connexion, because it equally occurs in il peut avoir. Where, how-
ever, through the operation of the etymological principle it is con-
fined to the traditional forms, these are welded the closer thereby.
We further have to notice the contraction of a vowel in auslaut
with the vowel in anlaut of the following word, or elision of one of
them, as the case may be; cf. the Latin reapse, magnopere, alior-
sum, rursus (from *re-ursus); French aubépine (alba espina), Bon-
RISE OF WORD-FORMATION AND INFLEXION.
393
nétable (place in the department of the Sarthe), malaise, Gothic
sah ('this' from sa-uh), pammuh (to this,' from pamma-uh), MHG.
hinne (=hie inne), hûzen = NHG. haussen, NHG. binnen. The
elision in the French article (l'état) or in the preposition de, again,
does not produce composition, because it follows according to a
general rule, and is not confined to single formulæ. A third case
of frequent occurrence is the assimilation of a consonant in auslaut
to the anlaut of the following word: cf. NHG. hoffart, Homburg
(= Hohenburg), Bamberg (= Babenberg), empor (= entbor), sintemal
(=sint dem mal); Latin illico, affatim, possum. The most
thorough isolation, however, is effected by the influence of the
accent; cf. NHG. nachbar (= MHG. nâchgebûr), junker (=juncherre),
jungfer (=juncfrouwe), grummet (=gruonmåt), immer (ie mêr),
mannsen, weibsen (=mannes, wîbes name), neben (from in eban,
eneben); Latin denuo (=de novo), illico; French celle (ecce illa);
cf. the corresponding phenomena in the case of compounds formed
in the IE. method: NHG. adler (MHG. adel-ar), wimper (wint-brâ),
wildpret (wiltbrat or wiltbræte), schulze-schultess (schultheize),
schuster (schuochsûtære, 'schuhnäher' ['shoe-sewer,' 'souter']), glied
(gelit), bleiben (belîben); French conter (computare), coucher (collocare),
coudre (consuere); Latin¹ subigere (compared with agere), reddere
(with dare), surgere (from sub-regere), præbere (from præ-hibere),
contio (from conventio), cuncti (from cojuncti).
591. It is less common to find that phonetic changes occurring
in simple words occasion isolation. This happens, for instance, in
this way, that a consonant in auslaut holds its ground, through
having been attracted to the following word, while in other cases it
falls away: cf. NHG. da (OHG. dâr), wo (OHG. war), as against daran,
woran, etc., MHG. hieran, etc., as against hie, sârie as against sâ.
1 In order to understand the origin of the forms cited it is necessary to go back to
the prehistorical method of accentuation.
.394
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Limits with-
in which the
compound
appears as
such.
A different modification is avoided by this transition in vinaigre as
against vin. Just as the diminished strength of stress in the case
of the member of a compound may call changes into existence to
which the simple word is not liable, so, on the other hand, it may
act protectively where the simple word is changed under the
influence of the chief stress; cf. NHG. heran, herein as against her,
and the French cordieu, corbleu as against coeur. In NHG. the
vowel of a first member of a compound is preserved from the
lengthening to which the simple word is exposed by the following
double consonants, cf. herzog, Hermann, herberge, wollust.
592. The same phonetic changes which divide the compound
from the simple word divide also from each other the individual
compounds which contain the same member; and this process also
contributes to diminish the sense of the independence of the
members. It is, of course, also of special import for the coales-
cence of the elements if one of them is lost as a simple word: cf.
such instances as NHG. bräutigam (OHG. -gumo 'man'), nachtigal
(-gala 'songstress'), weichbild (wich- 'holy'), augenlid (-lid ‘a
covering'), einerlei (-leie 'kind'), wahrnehmen; French aubépine
(alb-), printemps (primum-), tiers-état (tertius-), minuit (media-), bon-
heur (-augurium), ormier (-mcrum).
593. We have hitherto only considered the contrast between
verbal groups and verbal unities, and endeavoured to apprehend
all the factors which serve to resolve the former more and more
decidedly into the latter. We have, however, to notice another
contrast. The development we have described must have pro-
ceeded up to a certain point before the combination produces the
effect of a compound; it must not, however, go beyond a certain
point if it is to continue to produce merely this effect, and not
rather that of a simple word. What from the point of view of the
linguistic instinct we may call a compound lies between these points.
rise of word-FORMATION AND INFLEXION.
395
594. Syntactical and formal isolation do not easily lead to an
advance beyond this second point; as a rule, this is occasioned by
the obsolescence of one element in independent usage or by
phonetic isolation, especially the shrinking together of the complex
of sounds under the influences of accent.
595. The liveliness of the feeling for composition manifests
itself especially in the capacity of a compound to serve as model
for analogical formations. If we have deduced composition from
syntax, we in no way intend to assert that each single compound
has taken its rise from a syntactical combination. Possibly, indeed,
the greater part of the so-called compounds in the different
languages are merely analogical formations modelled upon those
to which in its strict sense the name compound belongs. Thus,
for instance, every genuine noun-compound formed in the in-
flexional period of the IE. original language, and especially every
such compound formed within the time of the development of the
individual languages is to be regarded as a formation by analogy,
and not as the union of an obsolete crude form with an inflected
word. In the same way the NHG. genitival and adjectival com-
pounds were to a great extent in their origin not syntactical. This
may be best seen in cases where the s of the first member, derived
from the genitive termination, is transferred to words to which in
the genitive it does not properly belong (such as regierungsrat, etc.),
and again to words where the genitive is out of place; cf. wahr-
heitsliebend formed after wahrheitsliebe, etc.
596. When the limit is once overstepped, up to which the com-
pound appears as such to the linguistic instinct, the formation,
apart from the inflexional endings which may eventually belong
to it, makes the impression either of perfect simplicity, or that of
a derivative formed with a suffix or prefix. Thus, for instance,
words like NHG. amt (Gothic and-balti), öhmd (MHG. uo-mật),
396
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Origin of
derivative
and inflex-
ional suffixes.
=
schulze (MHG. schuldheize), echt (from MHG. chaht MHG. -haft),
heute (from *hiu tagu), heint (MHG. hî-naht), Seibt (OHG. Sigi-
boto), bange (from *bi-ango), gönnen (from *gi-unnan), fressen
(Gothic fraitan), nicht (from ni io wiht); Latin demere (from *de-
emere), promere (from *pro-emere), surgere (from *sub-regere), prorsus
(from *proversus), stand on precisely the same footing as (for
instance) stand, hase, bald, binden, pangere, versus; and words like
adler (OHG. adal-ar), schuster (MHG. schuochsiutare), wimper (OHG.
wint-brawa), drittel (= dritte teil), Meinert (= Meinhard) as
schneider, leiter, mittel, hundert. Also in words like nachbar, bräuti-
gam, nachtigal, the last syllable will be apprehended in the same
way as the full derivative syllables in trübsal, rechnung, etc.
597. We have here arrived at the origin of derivative suffixes
and prefixes. These always originate in a loss of touch between a
member of a compound and the originally identical simple word.
But much more must happen before a word-forming element can
arise. In the first place the other member must be etymologically
clear, must be associated, that is, with a related word, or a related
word-group; which is not the case in (e.g.) adler and wimper. In
the second place, the element must appear not merely in isolated
words (as in nachbar, bräutigam), but in a group of words, and in
all be used in the same sense. If these two conditions are fulfilled,
the group may become productive, and be augmented by means of
new formations modelled on the examples which have originated
by way of composition. It is, however, in the third place, necessary
that the signification of the member of the compound in question
should either, even in its usage as a simple word, possess a certain
abstract generality of meaning (such as being, quality, action),
or undergo a development within the compound from the more
individual and sensuous meaning of the simple word. This last
circumstance may even under certain circumstances be decisive,
RISE OF WORD-FORMATION AND INFLEXION.
397
even though the feeling of the connexion with the simple word be
not yet completely lost.
1
(
598. We have within the historical development, as far as we
can trace it, opportunity enough to observe how a suffix arises.
The best known of these among German suffixes are those in -heit,
-schaft, -tum, -bar, -lich, -sam, -haft. The type of such a word as
weiblich, for instance, reaches back to an old bahuvrîhi-compound,
the original German *wîbolikis ; properly woman's form,' then
by the agency of metaphor 'possessing woman's form.' Between
such a compound as this, and the simple word MHG. lich, NHG.
leiche, such a discrepancy, first of meaning, secondly of the sound-
forms, has developed itself, that all connexion ceases.
Above all,
however, from the merely material signification of the simple word
'form,' 'outward appearance,' the more abstract 'quality' has de-
veloped. In the case of such a word as schönheit, the development
of a compound and of a derivative from the compound took place
for the first time in the west-Germanic period. The original German
form was *skaunis haidus, 'fine quality'; hence was developed
according to the regular laws of sound-change OHG. scônheit. The
composition is completed by the transference of the uninflected
form to the oblique cases, exactly as in the case of hôchzît, etc.,
cf. p. 381. In virtue of its abstract meaning, the second member
then becomes a suffix, especially when it has become lost in its
independent meaning.
599. Even yet, at a later epoch, many second members of a
compound approach the character of a suffix. Thus, for instance,
the words schmerzvoll and schmerzensreich are in their meaning
identical with the Latin dolorosus, French douloureux; and the dif-
ference between anmutsvoll and anmutig, reizvoll and reizend, is
1 My object here and in what follows is merely to explain and illustrate the manner
of composition, and I will not maintain that precisely the word chosen as the example
must have belonged to the primitive formation.
398
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Formation
of suffixes
very small. The -tel (= teil) in drittel, viertel, etc., is to the instinct
of language a mere suffix. Again, in allerhand, allerlei, gewisser-
massen, seltsamerweise, a first step has been taken to the formation
of a suffix. As to the termination -weise, we can easily imagine
that had the generalisation gone further, it might have developed
into a regular adverbial suffix like -mente in the popular language
of the Romance nations.
600. The boundary line between the members of a compound
and a suffix can be defined by the instinct of language only. We
have objective proofs to aid our judgment ready to hand as soon
as modes of composition are formed which are inconceivable as
compounds. Thus, for instance, the French fièrement might
naturally be conceived of as fera mente; but such a word as récem-
ment could not rationally be referred to recente mente. The funda-
mental meaning of the German -bar (= MHG. -bære) is 'bearing,'
'bringing.' Words like ehrbar, furchtbar, wunderbar would to
some extent suit this meaning; but the MHG. magetbære (virginal)
meienbære (belonging to May), scheffenbære (admissible to the func-
tion of schöffen), would no longer do so. The suffix-character is
decisively present if the analogy leads to a transference into quite
different domains, as in vereinbar, begreiflich, duldsam, etc., which
can only be apprehended as derivatives from vereinen, begreifen,
dulden (cf. on this point ante, p. 265); or when fusions of suffixes
set in (see on this point ante, p. 265) as in MHG. miltecheit, miltekeit
from miltec-heit; from which next rise formations by analogy;
as frömmigkeit, gerechtigkeit, on the one hand, and, on the other
eitelkeit, heiterkeit, dankbarkeit, abscheulichkeit, folgsamkeit.
601. From these observations, to which we might append a
constant and number of similar ones based on other languages, we must
conclude that the formation of suffixes is not the work of a
definite prehistorical period, limited by a definite point of time;
unceasing.
RISE OF WORD-FORMATION AND INFLEXION.
399
but rather a process continually repeating itself as long as
language is living and developing. We may especially conjec-
ture that even the general IE. suffixes were not all in existence
before the rise of inflexions, as analytical grammarians are
too prone to assume, but that even the prehistoric inflexional
period cannot have been entirely barren in this respect. We must
judge of the prehistoric origin of suffixes entirely according to
the standard afforded us by historical experience; breaking with
all theories not based upon this experience, which at the same
time shows us the only way of making the process psychologically
intelligible.
suffixes.
602. There is one other important point to be insisted upon. Rise of new
The rise of new suffixes perpetually alternates with the disappear-
ance of old ones. We may say that the suffix, as such, has
disappeared as soon as it has ceased to be employed for new
formations. The manner in which sound-change especially tends
to destroy this capacity has been explained above (p. 201). Thus
from time to time the need recurs of replacing a suffix which has
become too much weakened, and is split up into a number of
phonetic forms by one fuller and more uniform.
This purpose
is often served by suffix-groups which have become fused to-
gether. We may consider, for instance, how in OHG. the more
ancient and simpler methods of formation have been expelled by
the nomina agentis in -âri, the nomina actionis in unga and the
abstracts in -nissa. In other cases, however, the compounds of
the kind described afford the substitute desired. These, as a
rule, appear in the first instance parallel with the ancient forma-
tions, but then speedily, owing to their greater plainness and their
closer connexion with the fundamental word, obtain a decided
preponderance over these, and succeed in ousting all traces of it,
save a certain number of traditional survivals. Thus schönheit
400
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Rise of
inflexional
suffixes.
Criticism of
the analysis
of Indo-
European
grund-
formen.'
succeeds in ousting the now obsolete schöne, and finsterkeit the
diu vinster, which was yet living in MHG.
603. Inflexional suffixes take their rise in the same way as
derivative suffixes. In fact no sharp line of demarcation between
the two exists at all. We have here too, in the processes his-
torically observable, a standard for prehistorical ones. The
coalescence of the pronoun into the tense-stem, for instance,
admits of illustration by means of processes drawn from the
Bavarian dialects of the present day which have been already
described (p. 348 seq.). The formation of the tense-stem
shows itself most conveniently in the Romance future: j'aimerai
= amare habeo.
But it seems superfluous, when the mass of
materials is so well known and so easy of access, to collect
further examples.
604. Whoever draws from our survey the conclusions as to
method which it involves must confess that the proceeding hitherto
commonly adopted in the theoretic construction of the IE. language
is very perverse. I shall indicate a few principles, obvious enough
after what has been said, in accordance with which the theories
in vogue require either qualification or absolute rejection.
605. If we dissect the IE. original form of a word (even assum-
ing it to be rightly constructed) in the usual way, into stem and
inflexional suffix, and then dissect the stem again into root and
derivative suffix or suffixes, we must not suppose that we have now
gained the elements of which the word is actually composed. For
instance, it should not be supposed that the second person singular
optative present *bherois (represented earlier as *bharais) really
and truly took its origin from bher+o+i+s. In the first place we
must bear in mind that while no doubt the first foundations of
word-formation and inflexion came into being by the coalescence
of elements originally independent, these foundations, once in
RISE OF WORD-FORMATION AND INFLEXION.
401
existence, were necessarily used immediately as models for ana-
logical formations. In the case of no single IE. form can we
absolutely decide whether it has arisen from a syntactic word-
group, or whether it is merely an analogical creation from a form
already in existence. We must not however, on the other hand,
assume off-hand that the type of a form must have originated in
the former way. We must rather assume the play of the same
factor in the earlier period which in the more modern periods has
played so great a part, namely the displacement of the formative
principle by analogical formation. Small grounds as exist for
referring such types as besuch, unbestreitbar, unveränderlich, ver-
waltungsrat to a syntactical combination, there are no greater
grounds for so treating many IE. formations. In the second place,
we must notice that even in the case of those forms which are
really of syntactical origin, the elements do not necessarily lie
before us in the phonetic shape which they had before their
coalescence. We have as little right to derive an IE. genitive
akmenos from ak+men+os, as schusters from schu+ster+s. Many
changes which may have occurred only after the formation of
the complex have long been recognised: others have lately been
demonstrated. And it is not only possible, but even likely, that
the whole sum of such changes is far from exhausted by what we
at present know.
606. Still less is it to be supposed that the elements discovered
by analysis are the fundamental elements of speech in general.
Our incapacity to analyse an element proves nothing whatever as
to its original unity.
607. Entirely to be rejected is the distinction, so habitual in the
history of IE. flexion, between a period of construction and a
period of decay. That which is termed construction occurs, as we
have seen, only as a consequence of decay, and that which is
2 C
402
PRINCIPLES OF THE HISTORY OF LANGUAGE.
termed decay is but the continuation of this process. Construc-
tion is impossible without the aid of syntax. Such a process of
construction may take place at any period, and new constructed
material never fails to offer its compensation whenever decay has
set in beyond certain proportions.
CHAPTER XX.
THE DIVISION OF THE PARTS OF SPEECH.
TH
IE. The division
of
of the parts
of speech due
to considera-
any logical
HE customary division of the parts of speech in the
languages, as handed down from the grammarians
antiquity, rests upon no systematic application of logical principles; tion not of
it is rather due to the consideration of very diverse sets of facts. principle,
It bears accordingly the mark of arbitrariness. Its defects are
easily pointed out. Yet it would be hard to replace it by anything
essentially better so long as we aim at assigning every word to a
particular class. The attempt to erect a system of strictly logical
division is in all cases impracticable.
points.
609. The usual division has been effected by the consideration but of three
of three points: the meaning of a word, taken by itself, its function
in the sentence, and its behaviour in regard to inflexion and word-
formation.
the usual
610. As regards the first point, the grammatical categories— Criticism of
substantive, adjective, verb-correspond to the logical categories division.
of substance, quality, and activity or, more accurately, occurrence.
But though it may be the proper function of a substantive to
denote substance,—a function of which the adjective and the verb
are incapable, yet there are also substantive terms for quality and
occurrence. There are also verbs which denote continuous states
and qualities. Regard to the meaning of words has further con-
tributed to the erection of pronouns and numerals into separate
404
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Finite verb.
Connecting-
words,
and preposi.
classes. To co-ordinate these with the classes of substantives and
adjectives involves a grave logical error. The contrast of sub-
stantive and adjective also pervades the pronouns and numerals.
And, on the other hand, if pronouns and numerals are separated,
as distinct species, from the noun-class, the same separation must
be made in the case of the adverb-class, since bene - huc- bis are
related to one another just as bonus-hic-duo.
611. If we have regard to function in the sentence, we might
perhaps, in the first place, divide words into those which can of
themselves form a sentence, those which can serve as members of
a sentence, and those which can only serve to connect such
members.
612. In the first class we might place the interjections, which in
isolation are to be considered imperfect sentences.
But they
occur also as members of a sentence, sometimes immediately,
sometimes with a preposition, as in woe to the land! o über die
toren, MHG. ach mînes lîbes.
613. More perfect as a sentence, with a suggestion of subject
and predicate, was originally the finite verb. We find it, however,
also occurring, even in the earliest stage of which we have a record,
as a mere predicate attached to a subject separately denoted; and
in our present language it occurs, apart from the imperative, only
thus. It is therefore after all not possible to define the verb as
that which can constitute a sentence. And the so-called auxiliary
verbs are still further degraded to connecting-words.
614. The connecting-words are, as we saw above, p. 326,
conjunctions derived by a 'displacement of the distribution' from independent
tions. words. This process is continually being repeated. Were it only
on this account therefore, they cannot be sharply distinguished. And
there is this further ground, that a word may be an independent
member of the particular sentence to which it belongs, and yet at
THE DIVISION OF THE PARTS OF SPEECH.
405
the same time serve to connect this with another. If I say, for
instance, a man who believes that is a fool, the who is at once an
independent member of the relative sentence and a connecting word
between the principal and subordinate sentence. This holds uni-
versally of the relative pronoun and adverb. It is true also of the
demonstrative, in so far as it refers to the preceding or following
sentence, but not where it applies to the idea actually in question.
615. If we then attempt a more complete division, we entangle
ourselves in fresh difficulties. The substantive has, in contrast to
the adjective and verb, above all, the function of serving as subject,
and hence, in the widest sense, as object. The formation of names
of qualities and of occurrences in addition to names of substances
probably originated in an imaginative mode of intuition which set
upon qualities and processes the mark of things and persons. But
further, the capacity of the substantive to serve at will as subject or
object is precisely the chief occasion for its formation. Neverthe-
less, it can also be used attributively, and predicatively like an ad-
jective, and on the other hand, other words may serve as subject; I
do not mean merely as psychological subject in the widest sense, but
also as grammatical subject in the usual limited sense. Cf. sentences
like frisch gewagt ist halb gewonnen, aufgeschoben ist nicht aufgehoben,
hin ist hin, verloren ist verloren, grün ist die farbe der hoffnung;
ehrlich währt am längsten, doppelt genäht hält gut, jung gefreit hat
niemand gereut, allzu scharf macht schartig, gleich wider ist die beste
bezahlung, geradezu gibt gute renner, [slow and steady wins the race].
An adjective may also occasionally act as object, as in er hält gut
für böse, or, depending on a preposition, schwarz auf weiss, aus arg
ärger machen.
616. If we now turn to the connecting-words, serious difficulty
is raised by the class of conjunctions as commonly described. In
the first place, their separation from the demonstrative and relative
406
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Intermediate
stages be-
adverbs, the position of which has already been described (p. 405),
is rather arbitrary,—where, for instance, being called an adverb, as,
while, conjunctions. In the single sentence we find prepositions
and conjunctions distinguished by the presence or absence of case-
government, ie. in general, according as the construction is
hypotactic or paratactic. It is true that these two distinctions do
not absolutely coincide. On the other hand, all connecting-words
which join sentences together are called conjunctions, although the
same distinction between hypotactic and paratactic construction
ought to be made here also. Thus, e.g., before, since, while, when
occurring in simple sentences, are called prepositions; when they
connect sentences, conjunctions; although their function is
analogous in both cases.¹
617. The division most capable of being systematically carried
out is that which starts from the mode of flexion. Such a division
is actually employed, and consists of three classes: nouns, verbs,
and uninflected words (indeclinabilia, particles). Here again, how-
ever, the nominal forms of the verb and the indeclinable substances
resist the division. And for further distinction the mere considera-
tion of inflexion affords no ground. The indeclinable particles
have accordingly to be treated as an indivisible class. The
pronouns deviate in their inflexion to some extent from the other
nouns, but only to some extent, and they differ among themselves.
The distinction between substantive and adjective inflexion is not
absolute. Even the formation of degrees of comparison is not a
decisive mark of the adjective, since the very meaning of some
adjectives renders them incapable of comparison.
618. If then the current division of the parts of speech involves
tween classes such diverse and incompatible points of view, it is natural enough
that it should not admit of being practically carried out. The
of words.
¹ On the use of prepositions in English to introduce subordinate sentences, cf. p. 168.
THE DIVISION OF THE PARTS OF SPEECH.
407
facts we have to deal with are too complex and too variously
combined to be comprehended under eight or nine rubrics. There
are a host of intermediate stages which render possible a gradual
transition from one class into another. Such a transition results
from the general laws of change of meaning and of analogical
formation discussed in the preceding chapters. If we follow out
these transitions we detect at the same time the grounds which
originally prompted the division of the parts of speech.
and
adjective.
619. Let us take first the division of substantive and adjective. Substantive
This formal distinction rests, in the IE. languages, on the
power of
the latter to admit inflexions of gender and degrees of comparison.
In individual languages still further distinctions have arisen. Thus,
especially, the Teutonic adjective has become susceptible of a
double, we may even say a triple, mode of inflexion (cf. gut-guter
-der gute) in which forms occur absolutely without analogy in the
substantives.
620. On the ground of such tests we have no difficulty, for
instance, in declaring hund to be a substantive, jung an adjective.
But in spite of all differentiation of form the adjective may at
any time receive, at first as an 'occasionally,' then 'usually,' the
function of a substantive. The content of meaning is at the same
time enriched, either the general notion of a thing or a person
being added, or else more special notions arising from the situa-
tion (cf. p. 363). This operation can take place occasionally
with any adjective whatever, the latter being recognised as a
substantive in modern German orthography by the use of a
capital. From this substantival adjective a pure substantive may
then be derived by traditional use, especially if it becomes in any
way isolated as against the other forms of the adjective. Its
approximately complete conversion into a substantive is chiefly
indicated, in regard to the construction, by its connexion with
408
PRINCIPLES OF THE HISTORY OF LANGUAGE.
an attributive adjective, which replaces an adverb, or with a
genitive, which ultimately takes the place of a dative governed
by the adjective; cf. Latin bonum publicum, malum publicum,
amicus fidelis; and even where the substantival quality is less
traditionally fixed: nonnulli nostri iniqui, nonnullis invidis meis
(cf. Draeger, § 16); cf. also English my like, equal, better, younger,
etc. (Mätzner, iii. p. 232); his worthier (MILTON); MHG. mîn
gelîche (whence NHG. meines gleichen). We also find mixture of
adjectival and substantival construction, as in Latin multorum bene
factorum (CIC.). The thought shows a different kind of confusion
when, in spite of the conversion into a substantive, a superlative
is formed: mei familiarissimi, pessimo publico (cf. Draeger, § 16).
In Latin the complete transition is effected without difficulty, since
the inflexions coincide. In German, on the contrary, even when
it is very far advanced, the adjectival inflexion continues to
recall the original nature of the word. Der bekannte, verwandte,
gesandte, vertraute, geliebte, verlobte, beamte, bediente, liebste, are
now felt as substantives, and constructed accordingly (der bekannte
des mannes, mein bekannter), but they still betray their ad-
jectival nature by the regular change of strong and weak inflexion
(der bekannte ein (mein) bekannter), and the corresponding
feminines in addition by the weak inflexion of the singular
which in the regular substantive has died out (die bekannte
[Gen. der bekannten], and die zunge [Gen. der zunge for -en]).
The following, however, are completely transformed: der junge
(ein junge), der greis (MHG. grise, from adjective gris), der
jünger (both changed from the weak to the strong declen-
sion), oberst. Of older growth are feind, freund, heiland, MHG.
wîgant ('warrior'), vâlant ('devil'), all old present participles,
also fürst (old superlative), herr (old comparative from hehr),
mensch (adjective mennisch from man), and the neuters gut,
THE DIVISION OF THE PARTS OF SPEECH.
409
übel, recht, leid, wild. This transformation of adjectives into
substantives is familiar to every one, and has left marks in all
languages.
621. Less familiar and far more interesting is the converse
process, the transformation of a substantive into an adjective.
This comes about through the elimination of some part of its
meaning - including at least the notion of substance- so that
only the qualities attaching to the substance remain. This trans-
formation virtually occurs as an 'occasional' use, whenever a
substantive is employed as predicate or attribute. For what
happens in such cases is merely the attribution of certain qualities.
to the substance of the subject, or of the determinate word, not
the positing of a fresh substance. Apposition approaches the
nature of an adjective especially where it is used to specify a class,
and particularly if the combination contains some further bold
irregularity. Cf. Greek ἀνήρ πολίτης, ῥήτωρ, ὁπλίτης, etc., γυνὴ
Séσπоia, πaρlévos xeíp; Latin exercitus victor (LIV.); tirones
milites (CIC.); bellator equus (VIRG. OV.); bos arator (SUE.);
French un dieu sauveur (VOLT.); [a boy warrior;] flatteur, and
other words in -eur must be regarded absolutely as adjectives.
The adjectival nature may disclose itself in the addition of an
adverb only suitable to an adjective, as in weg du traum, so
gold du bist (GOE.); diesen widerspruch, so widerspruch als er ist
(LE.); so kriegerinn als sie war (ib. and often); so ist er fuchs
genug (LE.); nemo tam puer est (SEN.).
622. Some substantives are in NHG. felt to be completely
adjectives when used predicatively, but are distinguished from
the adjectives in not being attributive and in not adopting the
adjectival inflexion. To this class we may probably refer the
German herr or meister sein (werden). Goethe says: als wenn sie
(Narciss and Landrinette) herr und meister der ganzen truppe
410
PRINCIPLES OF THE HISTORY OF LANGUAGE.
wären.
Here the two words still show a substantival nature, in
so far as a genitive is made dependent on them, but at the same.
time they are treated like predicative adjectives, since they could
not otherwise be attached without inflexion to a plural subject,
and, besides, would not agree with a feminine person. Still more
decisive is the case of einem feind sein, on account of the dative.
Also schuld sein, where the isolation in respect of the substantive
schuld is shown by the orthography; less clear are es ist not, zeit,
where es is at bottom a genitive. The isolation is carried still
further in es ist schade, the substantive being now usually schaden
[cf. 'tis pity beside 'tis a pity; ''tis true 'tis pity, and pity 'tis 'tis
true']. In MHG. the development had gone still further. Here
schade is also used as predicate to personal subjects, and a com-
parative and superlative are formed from it; e.g. in KONRAD VON
WÜRZBURG'S Trojan War: der was den Kriechen scheder dan
iemen anders bî der zît;¹ further, an adverb is formed to it as if
it were an adjective: swie schade er lebe (MHG. Wb. ii. 63¹). In
OHG. fruma (‘advantage') is used in the same way, e.g. Otfried, iii.
10, 33, nisť quad er thô, fruma thaz' (' that is no advantage'). Even
in MHG. this gives rise to a real adjective frum, NHG. fromm.
They said ein frumer man, etc. How far the border-line is obliter-
ated appears from a passage in the Flore, 1289, daz wirt in nütze
unde frume (:kume), where, if we regard the combination with
nütze, we must treat the word as an adjective; if the final e, as a
1 Comparatives are also formed from other substantives in MHG., even where the
syntax forbids their being regarded as adjectives. Thus from zorn, as in do enkunde
Gîselhêre nimmer zorner gesîn; from nôt, as in dîner helfe mir nie nater wart; from
durft, as in wand im nie orses dürfter wart. From angst a comparative occurs in
archaic NHG., as in also viel engster sol dir werden (LUTHER, Wb. i. 359a). In these
cases the result is due to the analogy rather of the adverb than of the adjective. This is
shown even by the frequent combination angst und bange (bange is originally only an
adverb). In Gottfried's Tristan, 17845, we read in was dô zuo einander vil anger und
vil ander; ange is adverb to enge, ande substantive ('schmerz'). We still use the adverb
in this way in mir ist wol, weh. Latin superlatives from substantives occur in Plautus:
oculissime homo, patrue mi patruissime, but probably with a designedly comic effect.
THE DIVISION OF THE PARTS OF SPEECH.
411
substantive. The adjective ernst also, which first appears in
Luther, has arisen in the same way as fromm from the substantive.
The substantive geck has become an adjective in Low and Middle
German dialects. Entwicht, from MHG. ein wiht, enwiht (properly
'an insignificant being' = 'nothing at all,' nichtig), is in the six-
teenth century a pure adjective, as in entwicht vnd ark (H. SACHS);
du bist vil entwichter (ib.); die bös entwichten (AYRER).
623. The same process had been already carried out in a far
earlier period of language. All the so-called bahuvrîhi-compounds
were originally substantives. For ῥοδοδάκτυλος, βαρύθυμος, βαθύ
Opi§, eveλπis, magnanimus, ignipes, misericors, are of course pro-
perly 'rosy finger,' 'melancholy,' 'deep hair,' 'good hope,' 'great
spirit,' 'fire foot,' 'pitiful heart.' The substantival origin is still
evidenced in part by a defective expression of the adjectival
function. The masculine pododáктvλos has also to serve for the
feminine.
624. Somewhat different is the course of development in barfuss
from bar vuoz ('bare foot'). It was first used as an absolute nomi-
native or accusative in the phrase barvuoz gân. It is now felt as
an adjective. It has true adjectival inflexion, e.g. in H. Sachs:
mit barfussen füssen.¹
625. If we disregard whether the noun is conceived as a thing
or not, there is no doubt a distinction of yet another kind between
substantive and adjective. The adjective denotes a simple quality
or one regarded as such, the substantive embraces a group of
qualities. If we treat this distinction as the fundamental one, we
may certainly understand orator in such phrases as Cicero orator or
Cicero est orator as a pure substantive. But this distinction itself
cannot be carried through. It traverses the other distinctions :
failing, for instance, to meet the case of, on the one hand, adjectives
1 Yet another variety of transition is discussed on p. 260.
412
PRINCIPLES OF THE HISTORY OF LANGUAGE.
like königlich, kriegerisch, etc., on the other, of adjectives used as
substantives, like der gute. Between these also there is an inter-
mediate stage which leads imperceptibly from one to the other.
The transition from the denotation of a simple quality to that of
a group of qualities is effected by the use κaт' ¿§oxýν of a sub-
stantival adjective, this usage then became fixed. Whoever first
employs the word in this way himself supplies the notions not
yet explicit in its hitherto current meaning. For a subsequent
speaker, however, employing the usage at second-hand, the notions
supplied by the former may attach themselves from the first just
as closely to the sound-group as the fundamental notion, and the
latter need no longer precede the others in entering his conscious-
ness. When this last is the case, the transition to a substantive is
complete as regards the meaning, and by means of further isola-
tions its complete severance from the adjective is then effected.
Cf. the above examples.
626. The inverse process, by which all the qualities of a group
retire in favour of one, may be observed in the case of adjectival
derivatives from substantives, which develop into denotations of
perfectly simple qualities. Especially instructive in this respect
are the names of colours, cf. Gr. #oppúpeos from Topдúpa ('purple-
snail), φοινίκειος from φοίνιξ, αέρινος (air colour), μήλινος
('quince-yellow'), Lat. coccinus from coccum ('scarlet-berry'),
crocinus from crocus, luteus from lutum (woad'), miniaceus from
minium ('cinnabar'), niveus, roseus, violaceus. In all these words
there is no essential limitation of the reference to the colour of the
object denoted, and they are in part used also without such limita-
tion, e.g., unguentum crocinum, vinculum roseum ('rosary'), etc.
Substantives may also be turned directly into names of colours;
e.g. πорþúрa, coccum, crocus, lutum, and the modern lilac, rose,
which are also used adjectively ('a rose ribbon ').
THE DIVISION OF THE PARTS OF SPEECH.
413
627. This process gives us the clue to the origin of terms for
simple qualities. That these arose later than terms for groups of
qualities is self-evident, if we start from the assumption that
complete perceptions are the primary stage of all. Here as else-
where it can be, at the outset, only the instantaneous conception in
the speaker's mind which makes it possible for one notion in a group
to thrust the others into the background. It is at bottom the same
process as in the figurative usage of a word. When we say, e.g.
the man is an ass, an ox, a sheep, a fox, etc., we are always regarding
a particular characteristic of the animal in question, abstracted
from the rest. This is only possible when a word is used as
predicate or attribute. For as soon as we connect the notion of
an independent thing with it, we connect also that of an entire
group of qualities. When this usage became fixed in a number
of words especially adapted to it, the first step was taken to the
formation of a separate word-class.
628. Again, in spite of their greater formal differentiation, the
distinction between noun and verb is far from absolute. There are
certain points which serve to characterise the verb and distinguish
it from a noun; such are personal terminations, distinctions be-
tween active and middle and passive, and forms for the denotation
of mood and tense. Hence springs the possibility of the existence
of forms which possess a portion only of these characteristics;
and the scope for variation is yet further increased by the fact that
forms may or may not possess the positive characteristics of the
noun, denotation of case and difference of gender. And lastly, the
differentiation of the construction attaching to the verb and noun
respectively gives occasion to manifold transitions and confusions.
629. Commonly speaking its personal endings are regarded as
the strictly formal characteristic of the verb. By this criterion
participles and infinitives would be excluded from the verbal
414
PRINCIPLES OF THE HISTORY OF LANGUAGE.
forms; and so would, in strictness, many forms of the second
person singular imperative: for a form like Bánλe or Báλe is
nothing else but the mere stem of the present or the aorist. The
personal terminations are accordingly, if we disregard the second
person singular imperative, a necessary original condition for the
function of the verb as a normal sentence, and further for its
function as predicate or copula in the normal sentence. They are,
however, not indispensable as a condition for the formation of
sentences, and other peculiarities of the verb are absolutely inde-
pendent of them.
630. The contrast of meaning which we commonly assign to the
verb as compared with the adjective (or substantive used predi-
catively or attributively), has nothing to do with the verbal termina-
tions as such. It may exist without these, and may be lacking in
spite of their presence. Such an expression as the Greek eуkoTeîs or
βασιλεύεις may have precisely the same signification as ἔγκοτος εἶ
and Baσiλeùs el. The contrast remains sharply marked only just so
long as the adjective or substantive expresses a lasting quality, and
the verb a process defined and limited by time. We must, however,
consider that the adjective can be employed not merely for the
description of a quality attaching to the essence of a thing, but like-
wise for the description of a transitory quality: and in this it
approaches the character of a verb. Conversely the verb may also
be employed to describe states, even permanent states. How
nearly the two significations of being in and passing into a state
touch each other, we have seen on p. 305.
631. When a signification denoting a process temporally limited
is connected with the form and function of an adjective, there
arises the participle, of which the special value is, that it gives us
the power of using the expression for an event attributively. We
are able in many cases to trace historically the transition from the
THE DIVISION OF THE PARTS OF SPEECH.
415
adjective strictly so called into the participle. This holds good,
among other cases, of the so-called German perfect or preterite
participle (such as gegeben or gelegt), which took its origin thus:
the adjectives in -no- and -to- which have survived from IE., have in
signification become approximated to the verbs formed from the
same root, and more particularly to the perfect (preterite) of these.
This has continued to operate and has entailed many further
formal approximations. The case is the same with the participle
perfect in Latin and in Slavonic. We must also assume a corre-
sponding origin for the older participles already existing in the IE.
language. We must decidedly decline to derive, as has been done
by many, the category of the adjectives from that of participles :
on the contrary, the former must have been completely developed
before the latter could arise at all. Participial formations must
have arisen from forms which could be as well apprehended as
derivatives from the stem of the aorist, or present as derivatives
from the root, on the model of which adjectival forms related to
other verbal stems were then formed.
632. The characteristic difference between the participle and
the so-called verbal adjective is the participation in the denotation
of tense. A further consequence of the approximation to the
forms of the verb is the assumption of its method of construction.
The participle is constructed as a noun merely as regards the
substantive to which it stands as an attribute. It may, however,
take an even further departure from the character of a noun by
striking out a special path in the further formation of the method
of construction. In German, the fact that in such phrases as er ist
gegangen, er wird gefangen, er ist gefangen worden, case and gender
are no longer indicated, weakens also the sense of the nominal
character of the participle, even although the construction in the
two former combinations is that of the ordinary adjective, and in
1
416
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the latter differs therefrom merely by the presence of worden,
instead of geworden. In the phrases er hat ihn gefangen, er hat
geruht, etc., we must admit a complete departure from the way of
constructing an adjective. No doubt it may be proved historically
that the former of these two phrases signified originally er hat ihn
als einen gefangenen; but this is a matter of indifference for our
modern linguistic instinct. At an earlier period they said habet
inan gefanganan, and this left no doubt as to the nominal character.
The circumstances in the corresponding combinations in the
Romance languages are peculiar and noteworthy. In these we can
watch the transition from the general adjectival into the special
participial construction. In French the idiom is to say j'ai vu les
dames, but je les ai vus, and les dames que j'ai vues. In Italian it is!
still possible to say ho veduta la donna, and ho vedute le donne as
well as ho veduto. In Spanish all inflexion in the case of peri-
phrases formed with haber is abolished; it is correct to say la carta
que he escrito, just as it is to say he escrito una carta. On the
other hand, in the case of the periphrasis with tener-a later intro-
duction into the language—the inflexion is always retained: tengo
escrita una carta is as correct as las cartas que tengo escritas.
633. Conversely, however, it is possible for the participle to be
gradually brought back to a purely nominal nature. This process
of reduction may be considered strictly speaking as accomplished
as soon as the present participle comes to be used for the lasting
or recurring activity, and the perfect participle for the result of the
activity, just as each form of the present or perfect can be
employed. A method of usage Kar' èçoxýv, or an usage in a
metaphorical sense or any other kind of isolation may have the
result of completing the transformation; cf. such examples as
schlagend, treffend, reizend, zwingend, bedeutend, getrieben, gelungen,
berufen, verstorben, verzogen, verschieden, bekannt, unumwunden,
THE DIVISION OF THE PARTS OF SPEECH.
417
verlegen, gewogen, verwegen, erhaben, bescheiden, trunken, volkommen,
etc. [Cf. agéd, learned, crabbéd, charming etc.]. Even the com-
bination with another word after the laws of verbal construction
does not impede this process; only then the entire composition
must be in a position to approximate to the analogy of nominal
composition, cf. ansprechend, auffallend, ausnehmend, anwesend,
abwesend, zuvorkommend, hochfliegend, hellsehend, wolwollend,
fleischfressend, teilnehmend; abgezogen, ausgenommen, hochgespannt,
neugeboren, wolgezogen, etc. [Well educated, ill behaved, etc., cf.
also the peculiar English compounds on the analogy of good-
natured, which seem to be a survival of an IE. method of
composition.]
634. The formation of a comparative and superlative may be
regarded as a test for the transformation of the participle into an
adjective pure and simple. Sometimes, however, this appears side
by side with a verbal construction; cf. dazu erschien mir nichts
wünschenswerteres, den character der nation ehrenderes (GOE.); die
Oestreich kräftigendsten elemente (Kölner Zeitung).¹ Another test
is the method of construction, e.g. in Latin the combination of the
participle with a genitive; cf. amans tuorum ac tui (CIC.), religionum
colentes (ib.), solitudinis fugiens-societatis appetens (ib.).²
635. The participle, like all adjectives, may become a substan-
tive, and the substantival may, like the participle adjectival, denote
a momentary activity or a state. The same resemblance is seen.
in its power of casting off its verbal nature; cf. der liebende, vorsit-
zende, geliebte, gesandte, abgeordnete, beamte (=beamtete), MHG. der
varnde, gernde (both='player'); of older date heiland, freund,
feind, etc., zahn-Latin dens = Greek ỏdoús (part. of essen, edere)
[l'amant, l'aimé].
636. The nomen agentis, just like the participle, is able to
2 See Draeger, § 207.
1 Cf. Andr. 119 sqq.
2 D
418
PRINCIPLES OF THE HISTORY OF LANGUAGE.
denote either a momentary or a lasting, or it may be a recurring,
activity. In the first-mentioned application it always remains
closely connected with the verb: and it might easily be conceived
that it might as naturally as the participle adopt the verbal man-
ner of construction; in fact, that it might be possible to say in
German, der erzieher den knaben, just as in the compound knabener-
zieher at least it is felt that the former part of the compound is an
accusative, and is constructed after the analogy of knaben erziehen.
Already in combinations like der sieger in der schlacht, der befreier
aus der not¹ the verbal character of the substantive is evident;
and, still more in such as the Greek ὑπηρέτης τοῖς νόμοις, or as we
actually find it in Latin dator divitias, justa orator. Conversely
the nomen agentis, as denoting lasting or repeated activity, may
isolate itself more and more as regards the verb, and thus finally
lose the character of a nomen agentis altogether [especially when
the meaning of the verb grows faint, as in tailor]; cf. schneider,
beisitzer, ritter, herzog (heerführer) vater, etc.
637. There is still another path from verb to noun. Beside the
nomina agentis stand the nomina actionis. These, like the sub-
stantival denotations of qualities, must owe their origin to a meta-
phor only, the activity being apprehended under the category of
the thing. These too may denote a momentary or a lasting and
recurring activity. These too may approximate to the verbal
construction; cf. die befreiung aus der not, ǹ roîs vóμois vπnρeσía,
knabenerziehung. And it is again the denotation of the lasting
and recurring activity which conduces to the loss of the character
of a nomen actionis. Thence develops the denotation of a lasting
condition; cf. besinnung, bewegung, aufregung, verfassung, stellung,
stimmung.
638. From this point a further development to the denotation
1 Cf. even more striking combinations with prepositions in Andr. p. 209.
THE DIVISION OF THE PARTS OF SPEECH.
419
of things is possible, as has already been shown (p. 86). In the
course of this development the correspondence of the meaning
with that of the verb may be broken; cf. haltung, regung, gleichung,
rechnung, festung, etc. And as the isolation proceeds, every trace
of the verbal origin may then be destroyed.
639. Thus far the nomen actionis bears an analogous relation
to the nomen agentis. It however approximates far more to the
verbal character than the latter, more even than the adjective or
participle: mainly from the fact that the infinitive and supine take
their origin from it. The infinitive bears in many ways a strong
analogy to the participle. While, however, the latter generally
speaking maintains the adjectival form and maintains the adjec-
tival method of construction by the side of the verbal, and only
occasionally develops for itself a special method of construction,
laying aside the formal characteristics of the adjective,-isolation
from the form and method of construction of the noun is a condi-
tion of the origin of the infinitive. The infinitive is, as formal
analysis proves, a case of nomen actionis, and must originally have
been employed after the analogy of those methods of construction.
which hold good for the combination of the noun with the verb.
But it must no longer be felt as a case, its mode of construction
must no longer be after the analogy of the original models, or else it
is no infinitive. The isolated form and the isolated mode of con-
struction become then the basis for further development. The
form and method of construction of the infinitive is on the one
hand verbal, like those of the participle: on the other hand, how-
ever, it is not nominal, but specifically infinitival.
640. The infinitive, too, admits of a gradual return to the
nature of a noun; but it meets with more impediments in the pro-
cess than the participle on account of its want of inflexion. The
approximation to the nominal character manifests itself therefore,
420
PRINCIPLES OF THE HISTORY OF LANGUAGE.
thero
as long as no special methods of distinction are resorted to,
especially in cases in which the characterisation by means of a
terminal inflexion is least indispensable; ie. in the employment as
subject or object. In the forms of such sentences as wagen gewinnt,
Latin habere eripitur, habuisse nunquam (SEN.), and certainly in
such as hic vereri (=verecundiam) perdidit (PLAUT.), we may
doubtless assume that the infinitive is constructed after the analogy
of a noun.
This is less certain in such cases as ich lasse schreiben,
ich lerne reiten. At any rate, if the infinitive is really and truly
employed in this case after the analogy of an accusative of the
object, this analogy no longer exists for the instinct of language
of the present day. The combination with prepositions finds its
way into language with even more difficulty. In MHG. durch is
especially common with the infinitive; in the Roman popular
language the combination of the infinitive with prepositions occurs
instead of the gerund (ad legere in place of ad legendum, etc.):
similarly sometimes in the poets and later prose writers, praeter
plorare, 'except lamentations' (HOR.), multum interest inter dare et
accipere (SEN.). A further approximation of the infinitive to the
noun requires special favouring circumstances. Generally speaking
those languages alone attain to this which possess in their article
a means of creating substantives and denoting cases. Hence it is
that Greek has advanced further in this direction than Latin
though in the latter it is true that the demonstrative pronouns too
may have a similar effect; cf. totum hoc philosophari (CIC.), inhibere
illud tuum (ib.). NHG., however, and the Romance languages
have gone even further than the Greek, for in them the infinitive is
employed as an equivalent to the noun pure and simple, even in
respect of inflexion. This equation is rendered possible in the
Romance languages by the general abolition of case-difference.
Old French and Provençal go so far as to give the infinitive the -s
THE DIVISION OF THE PARTS OF SPEECH.
421
of the nominative; li plorers ne t' i vaut rien; Meliers chanza es
donars que penres. In the case of NHG. the circumstance comes
into consideration on the one hand that the case-differences in the
substantives in -en are all cancelled excepting the genitive; on the
other side the approximation of the gerund (MHG. gebennes, ze
gebenne) to the infinitive, with which it originally had nothing
to do.
641. In the course of this development different steps are
again possible with respect to the maintenance of the verbal con-
struction. It commonly occurs without any process of prefixing
an article or a pronoun; cf. e.g. MHG. durch behalten den lîp, durch
aventiure suochen. In Greek, too, the article is no impediment
to this: we there find such expressions as τὸ σκοπεῖν τὰ πράγ
ματα, τὸ ἑαυτοὺς ἐξετάζειν, ἐπὶ τῷ βελτίω καταστῆσαι τὴν αὑτῶν
Stávolav. In NHG. the verbal construction is confined, in accord-
ance with the assumption of the nominal inflexion, to the same.
limits as in the case of the nomen actionis. In MHG., on the other
hand, a genuine verbal construction sometimes occurs: in fact a
relative referring to an infinitive may have a verbal construction;
cf. Hartman Greg. 2667, des scheltens des in der man tete; Tristan
1067, diz sehen das ich in hân getân. In the Romance languages
as well we find the verbal construction of the infinitive provided
with an article or a pronoun side by side with the nominal con-
struction; cf. the Italian al passar questa valle (but also il trapassar
del rio); Spanish el huir la occasion (but also al entrar de la ciu-
dad); Old French au prendre le congié; even in Montaigne, il se
penoient du tenir le chasteau; further, in Italian, il conoscer chiara-
mente; Spanish el bien morir; Old French son sagement parler.
642. As soon as the difference between infinitive and noun
resulting from inflexion is set aside, there remains nothing further
to hinder the transformation of the former into a noun pure and
422
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Adverbs.
simple, and this is accordingly of very common occurrence in
NHG., and of not unfrequent occurrence in the Romance languages;
cf. NHG., leben, ableben, leiden, scheiden, schreiben, tun und treiben,
wesen, vermögen, betragen, belieben, einkommen, abkommen, auskom-
men, ansehen, aufsehen, andenken, vorhaben, wolwollen, wolergehen,
gutdünken, etc.; French être, plaisir, pouvoir, savoir, savoir-faire,
savoir-vivre, etc. In this process it is possible for the same
differences of signification to occur as in the case of the nomina
actionis, and the same isolation as regards the verb.
643. The adverbs, as far as we can trace their origin, are almost
exclusively the outcome of crystallised cases of nouns, and to some
extent of the combination of a preposition with its case.
We are
thus led to suspect that the oldest stratum of the adverbs also
proceeded in like manner from nouns; only with the difference
that this process occurs in point of time before the development
of inflexion, and that hence no case as yet, but merely the stem-
form pure and simple, has come to be employed. The adverb
stands in the closest relationship to the adjective. It has a
relation in the first place to the verb, and then to the adjective as
well, analogous to that of an attributive adjective to a sub-
stantive. This proportional relation shews itself also in this, that,
generally speaking, an adverb may be formed out of any adjective
at will.
644. The formal difference between the adjective and the
adverb depends on the capacity of the former for inflexion, and
the consequent possibility of agreement with the substantive.
When this formal test is absent, the division between the two
parts of speech cannot be strictly maintained by the instinct of
language. In NHG. it is actually to some extent broken through
when the adjective, used predicatively, has become unchangeable,
and when the difference which, generally speaking, still exists in
THE DIVISION OF THE PARTS OF SPEECH.
423
MHG. between the flexionless form of the adjective and the ad-
verb (starc-starke, schane-schône, guot-wol, bezzer-baz) is cancelled.
There is, strictly speaking, no reason to distinguish the gut in
sentences like er ist gut gekleidet, er spricht gut, from that in er ist
gut, man hält ihn für gut, and to call the former an adverbial, the
latter an adjectival usage. The instinct of language knows nothing
of such distinctions. We can best gather this from the fact that
the adverbial form of the superlative has intruded into the place
which commonly falls to the flexionless form of the adjective.
The Germans say es ist am besten, and even du bist am schönsten,
wenn, etc.
645. On the other hand, many adverbs in different languages
take, when joined with an adverb, an adjectival inflexion. Thus,
in French, it is correct to say toute pure, toutes pures. Corre-
spondingly we find in Italian tutta livida, in Spanish todos desnudos,
etc.: just so, in Italian, mezza morte; Spanish, medios desnudos.
In many German dialects, too, we have such phrases as ein ganzèr
guter mann, eine ganze gute frau; solche schlechte ware; eine rechte
gute frau (LE.).
646. The function of the adjective corresponds especially with
that of the adverb used in connexion with nomina actionis and
agentis; cf. eine gute erzählung, ein guter erzähler ('a good story-
teller'). In this case the adjective denotes the manner of a
process in the same way as the adverb in other cases. The latter
combination is, however, capable of two interpretations, as we
might reasonably apply the word good, generally speaking, to the
person of the story-teller. This ambiguity would be removed if
the adverb were employed for one case after the analogy of the
verbal construction; and so it is in fact in English, as an early
riser. The Germans unite the conceptions into a single word;
cf. frühaufsteher, langeschläfer, schönschreiber, feinschmecker, etc.;
424
PRINCIPLES OF THE HISTORY OF LANGUAGE.
mann.
derivations from früh aufstehen, etc. The uncertainty attaching
to the meaning is, besides, not confined to the nomina agentis; cf.
ein guter kutscher, ein arger narr, ein grosser esel, ein junger ehe-
The adjective can either be simply referred to the person
or to the quality attributed to it by the substantive. In the latter
case it bears the same relation to the substantive as an adverb to
the adjective which it qualifies. The adjective holds a corre-
sponding position with respect to substantival denominations of
quality; cf. die hohe vortrefflichkeit, grosse güte.
647. The adjective and the adverb thus corresponding with
each other, the need is at once felt to have the two together for
each single case. There are, however, a great number of adverbs
which are not derived from any adjective, and which thus have
none parallel to them. In this case necessity impels to transfer
the function of the adjective to the adverb as well. The adverb
is most readily employed predicatively in this connexion; as
related to the adjective, the verb has sunk into a mere link-word.
In sentences like er ist da, er ist auf, die tür ist zu, alles ist vorbei,
er wird mir zuwider, the construction is apprehended by the
instinct of language as identical with that found in the phrases
die tür ist offen, er wird unangenehm. The adverb, however, added
as it is to a substantive as its definition, becomes undistinguishable
from the adjectival attribute. When the Germans say der berg
dort, die fahrt hierher, der baum drüben, the adverb marks its
difference from the adjective by its position in the sentence. It
is different with the Latin constructions (which are not, however,
very common), as nunc hominum mores vides? PLAUT., ignari
sumus ante malorum (VIRG.); discessu tum meo (CIC.)¹ The
adverb, however, approximates most nearly to the functions of
an adjective when it is, as in Greek, inserted between the article
1 Cf. Draeger, § 79.
THE DIVISION OF THE PARTS OF SPEECH.
425
and its substantive ; as in τὴν ἐκεῖ παίδευσιν, τὴν πλησίον τύχην,
tậ vûv yévei, ý Xíav τpuþý; in English, on the hither side, the above
discourse;¹ in Spanish, la sempre señora mia. In NHG. any such
employment of the adverb is an impossibility. The Germans
have created inflected words to meet the exigency. These words
are sometimes secondary derivations, employed only attributively,
not predicatively; cf. alleinig, hiesig, dortig, obig, jetzig, vorig, nach-
herig, sofortig, alsbaldig, vormalig, diesseitig; of rarer occurrence
are such as are employed predicatively as well, such as niedrig,
übrig (to which may be added alleinig in South German dialects).
On the other hand, many adverbs have simply adopted inflexional
terminations, a process which is favoured by the fact that in its
predicatival employment the adjective was not formally dis-
tinguished from the adverb, because its flexionless form was
employed; cf. nahe, fern, selten, zufrieden, vorhanden, behende (from
OHG. bi henti), täglich (from OHG. tagolich), ungefähr, teilweise,
anderweit. In dialects such expressions are used as ein zues
fenster, ein weher finger, ein zuwiderer mensch. The adjective
einzeln, recently formed from the adverb (strictly speaking a dative
plural), has driven the adjective einzel, which is at the root of
the word, from the field. The German word oft is affected by
adjectival augmentatives; cf. such expressions as the Latin propior,
proximus, which depend on prope, and the Greek forms ¿yyÚTEPOS,
ἐγγύτατος, which depend on εγγύς.
used as a
attribute.
648. The adjective, used as a predicatival attribute, closely Adjective
approaches the adverb. This portion of the sentence stands in predicatival
close relation to the subject, with which it is connected by concord,
but has become independent of it, and is hence enabled to enter into
direct relationship with the predicate. The adverb, on the other
hand, is connected with the predicate, but may in a similar manner
1 Cf. Mätzner, pp. 148-9.
426
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Origin of
prepositions
tions.
become independent of the latter, and so approximate more closely
to the subject. There are also cases, however, in which a qualifica-
tion is equally suitable to the subject and to the predicate. It is
thus intelligible that in many languages the adjective and the
adverb may be equally admissible, and that in one language one
of these parts of speech, in another the other, is customary. In
NHG. it often happens that the adverb stands as the equivalent of
an adjective used in other languages; cf. allein as against the
Latin solus, the French seul, etc.; zuerst and zuletzt and Latin
primus and postremus, etc.; gern and Greek ékáv, aσμevos, Latin
libens and libenter; ungern and Latin invitus, or the less common
invite. We find constructions like the following striking-which
indeed are not general in foreign languages—εἶδον παννύχιοι
(Hon.); κρήνη ἄφθονος ῥέουσα (ΧΕΝ.); Ασωπὸς ποταμὸς ἐῤῥύη
µéyas (THUC.); Latin beatissimi viveremus, propior hostem collo-
catus, proximi Rhenum incolunt, nocturnusque vocat clamore Cithae-
ron (VERG.); Aeneas se matutimus agebat (VERG.); frequens te
audivi (CIC.); in agmine atque ad virgilias multus (=frequenter)
adesse (SALL.); est enim multus in laudanda magnificentia (CIC.);
is nullus (=non) venit (PLAUT.); tametsi nullus moneas (TER.);
Italian che più lontana se ne vada (ARIOST.).
649. Prepositions and conjunctions as connective elements
and conjunc⚫ took their origin, in every case, from independent words through
a displacement of the distribution. This displacement must be
definite. In 'occasional' use the most diverse kinds of component
parts of sentences may, of course, be degraded to mere connecting
members. Not till a word is employed with some regularity as a
connecting word can it be regarded as possibly a preposition or
conjunction. Another indispensable condition, however, is that
the method of its construction should have become isolated from
that which it possessed as an independent word. But even then it
THE DIVISION OF THE PARTS OF SPEECH.
427
·
may at the same time fulfil its own function as an independent
word, so that it is thus impossible to bring it simply under the
head of a definite class of words. This strict classification is not
possible until the word in its independent application has disap-
peared, or until a phonetic differentiation has attached itself to the
two methods of application, or until some other kind of isolation
has crept in.
tion.
650. We are thus able to propose for the preposition the fol- The preposi-
lowing definition :-the preposition is a connecting word, with
which a case of any substantive may be connected without any
connective analogy to nominal or verbal methods of construction.
If we adopt this definition we shall not explain entsprechend in
such a sentence as er hat ihn seinen verdiensten entsprechend belohnt
as a preposition, because its construction is that of the verb ent-
sprechen. When we come to anstatt it is different. In anstatt des
mannes the genitive was originally the regular sign of the inde-
pendence of the noun. But whether the genitive is still felt as
such depends on the question whether we still feel anstatt to be a
combination of the preposition an with the substantive statt. If it
is not so felt, the construction with the genitive leaves the place it
had hitherto occupied in the group, and the preposition is created.
It is possible in this case for the instinct of language to be in a
high degree vacillating; nay, even different in the case of different
individuals. For there can be no doubt that statt is no longer a
substantive of general application, but is confined to certain
isolated combinations. If, however, we say an meiner statt we are
more emphatically reminded of the substantival nature of the word
statt. In other cases, the isolation has become absolute. The
German nach is originally an adverb, and identical in meaning
with nahe. But between seinem ende nahe and nach seinem ende all
relation is lost, though both go back to the same method of con-
428
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Conjunc-
tions.
struction. In this case the isolation of the method of construction
is due to the obscuration of the etymological relationship, owing
to divergency in the development of signification. In other cases
it is due to the disappearance of this method of construction from
living usage. In IE., as in Latin, the ablative was used after the
comparative. This construction was still maintained in OG.,
excepting that there the ablative, as is generally the case, became
confused with the instrumental and dative. As it disappeared
from general use, it maintained itself, among other instances, in
the case of two adverbial comparatives, which, owing to this isola-
tion, became prepositions, MHG. é (NHG. still found in ehedem) and
sît (NHG. seit)= Gothic seips in panaseips, which is, phonetically
speaking, the regular comparative of seipus. In the oldest preposi-
tions of the IE. the case must have been, in the first instance,
referred to the verb. For it denoted by itself the direction in
which or whence, or the condition of being in one place.
651. The particle was added merely with the view of a nearer
definition of the relation of space; it was in fact still an adverb.
As the case, apart from its combination with the preposition, lost
its old signification, a special method of construction was created
simply from this combination.
652. The genesis of conjunctions, like that of prepositions,
admits, to some extent, of being historically traced. Those which
serve to connect sentences are to a large extent developed from
the conjunctional adverbs, or from isolated forms of the conjunc-
tional pronouns, which, it may be, are connected with other words
(cf. daher, darum, deshalb, deswegen, weshalb, indem). These
words, therefore, actually serve to connect sentences before they
become conjunctions pure and simple. The question as to
whether they will be accepted as such must be one, to a great
extent, of subjective feeling: no sharp and definite line can be
THE DIVISION OF THE PARTS OF SPEECH.
429
drawn. A great deal must depend upon the degree in which the
origin of the word has been obscured. The existence of such
obscurity is a necessity, if we are to apprehend the word as a mere
link for sentences.
653. A special method of the genesis of conjunctions has been
touched on above (p. 333). In this instance, too, in most cases we
shall find a conjunctional—and that a demonstratival-pronoun at
the root, whether standing independently, or in connexion with
another word. There are, however, cases without any demonstra-
tive, such as those of the NHG. weil, falls, and the English because,
in case.
But, in this case as well, the indication of what is to follow
has attached itself to and affected the substantives which lie at the
root of the conjunctions.
654. A quantity of conjunctions take their rise from words
expressive of a comparison; cf. ingleichen, ebenfalls, gleichfalls,
gleichwol, andernfalls, übrigens; Greek öµws, åλλá; Latin ceterum ;
further the comparatives ferner, weiter, vielmehr; Latin potius,
nihilominus; French mais, plutôt, néanmoins. A relation is, from
the very outset, expressed by these words; on the other hand, an
expression is lacking for the object to which the relation tends:
this must be guessed from the connexion in which the word occurs.
655. The case is different, on the other hand, where words of
assuring have passed into conjunctions connecting sentences; cf.
allerdings, freilich, nämlich, wol, zwar (MHG. ze wâre, fürwahr);
Gothic raihtis ('but' or 'because'); Latin certe, verum, vero, scilicet,
videlicet, etc. These words are in themselves expressive of no
relation to another sentence. The logical relation in which the
sentence in which they are contained stands to another, is originally
thought as their adjunct, without finding any expression in thought.
As it is, however, precisely this relation which forces upon the
speaker the necessity of adding an express assurance, it comes to
430
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Difference
between
and conjunc-
tion in the
simple
pass that this relation appears to be expressed by the assurance.
Just as little does the Latin licet originally express any reference
to the governing sentence; in this case too, a reference-originally
existing in thought alone-has in a secondary way attached to this
verbal form, which has thereby passed into a conjunction.
656. One method of denoting the relations of two sentences or
portions of sentences is afforded by the anaphorical use of two
adverbs, not intrinsically conjunctional: bald-bald, jetzt - jetzt,
einmal einmal; modo modo, nunc
nunc, tum tum, etc. Of
course a similar employment of words which have already become
conjunctions has nothing to do with this.
657. The parallelism in the relation of members of a sentence,
and that of entire sentences to each other, appears in the fact that
the connecting words created for the one relation are analogically
transferred to the other. Thus from the oldest times the same
copulative and disjunctive particles have been employed for both
relations. The transference from the member of a sentence to a
sentence can be distinctly traced in words like weder, entweder,
MHG. beide, cf. p. 329. In the same way a parallelism exists in the
employment of the demonstrative and relative particles of com-
parison. In this case we shall have to assume the converse trans-
ference from sentence to member of sentence. For the other cases
of the employment of conjunctions which introduce a sentence
to introduce members of a sentence, cf. p. 166; for the method of
employing prepositions before sentences, cf. p. 168.
658. The difference between the preposition and conjunction in
preposition the simple sentence is sharply defined by the fact that the former
governs a case while the latter does not. Still confusions present
themselves even in this obvious difference. As far as the sense
goes, it matters little whether we say ich mit(sammt) allen übrigen
or ich und alle übrigen: and it thus happens that the predicate or
sentence.
THE DIVISION OF THE PARTS OF SPEECH.
431
the apposition to a connection introduced by mit is often placed in
the plural in cases where a regard for the strict grammatical
relation would demand the singular; cf. scherz mit huld in anmuts-
vollem bunde entquollen dem beseelten munde (SCHI.); Greek Anμo-
σθένης μετὰ τῶν συστρατηγῶν σπεύδονται (THUC.); Latin ipse dux
cum aliquot principibus capiuntur (LIV.); filiam cum filio accitos
(ib.); English old sir John with half a dozen more are at the door
(SH.); French Vertumne avec Pomone ont embelli ces lieux (ST.
LAMBERT); further examples cited from the Romance languages
are found in Diez, iii. 301, and from the Slavonic languages in
Miklosich, iv. 77, 78. In these cases we must regard the connecting
word, if we consider the case attached to it, as a preposition; if,
however, we regard the form taken by the predicate, as a conjunc-
tion. Examples of an absolute transference from a prepositional
to a conjunctival function are afforded by the NHG. words ausser
and ohne; cf. e.g. niemand kommt mir entgegen ausser ein unver-
schämter (LE.), dass ich nicht nachdenken kann ohne mit der feder in
der hand (LE.), kein gott ist ohne ich (LU.). On the other hand, the
conjunction wan in MHG. passes into a preposition constructed
with the genitive, cf. daz treip er mit der reinen wan eht des alters
einen (KONR. V. WÜRZB.). We can thus understand that in a
state of language previous to the formation of cases a line of
demarcation between prepositions and conjunctions can hardly
exist.
659. The transference from subordination into co-ordination is
rendered more easy when there is no case government from the
outset, and when consequently the connecting word is already a
conjunction or a conjunctional adverb. This is most easily seen in
such correlations as that of sowol with als auch, etc.; cf. die zurück-
weisung, welche sowol Fichte als auch Hegel.. erfahren haben
(VARNHAGEN v. ENSE); English your sister as well as myself are
432
PRINCIPLES OF THE HISTORY OF LANGUAGE.
1
greatly obliged to you (FIELDING); Latin ut proprium jus tam res
publica quam privata haberent (FRONT.); French la santé comme la
fortune retirent leurs faveurs à ceux qui en abusent (SAINT-EVRE-
MONT); Bacchus ainsi qu'Hercule étaient reconnus pour demi-dieux
(VOLT.).
CHAPTER XXI.
W
LANGUAGE AND WRITING.
and short-
comings
of writing
with spoken
HATEVER information we possess as to the divergences Advantages
of the linguistic conditions of the past from those of
to-day is due to the use of writing. No philologist should ever as compared
disregard the fact that what is written is not language itself; language.
that speech rendered into writing always needs to be rendered
back into speech before it can be dealt with. Such rendering
back is only partially possible - another fact which should be
constantly kept in mind; but as far as it is possible at all it is
an art which needs to be mastered; and in this the unprejudiced
observation of the relation of writing and pronunciation prevailing
at present among different peoples renders great service.
661. Writing is, however, an object of philological investiga-
tion not merely on account of the mediating part which it thus
plays it is also an important factor in the development of
language itself, which we have hitherto purposely avoided dwelling
on. It remains to us to define the extent and the limits of its
activity.
662. The advantages possessed by written over spoken matter
with regard to effective operation are sufficiently obvious. By
its means the narrow circle to which the influence of the individual
is otherwise confined may spread till it embraces the entire
linguistic community; by its means that narrow circle may extend
itself beyond the generation then living, and exert an immediate
2 E
434
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Alphabets
influence on all that follow. No wonder that these evident
advantages are commonly much over-rated-over-rated even in
the science of language, since it requires a somewhat deeper
reflexion to bring out those points in which writing lags behind
living language.
663. A distinction is commonly made between languages
which are, and those which are not, pronounced as they are
written. The former phrase must be understood in a very relative
sense if we would not involve ourselves in an error of serious
moment. Not merely is writing not language, but it is in no
way an equivalent for it. To rightly appreciate the relations
between them, we have to consider not this or that single dis-
crepancy, but a fundamental difference. We have seen above
(p. 48 sqq.) the importance of the continuity as well in the series
of possible speech-sounds as in the series of sounds consecutively
spoken, for a proper view of the phonetic side of language. But
an alphabet, however perfect it may be, lacks continuity in both
{ these respects. Language and writing bear the same relation to
each other as line and number. We may employ as many signs
as we like, and may define as closely as we like the corresponding
articulations of the organs of language: each will still remain the
sign, not for a single articulate sound, but for a series of infinitely
numerous ones. And even though the transition from one articu-
lation thus denoted to the other is in some respects a necessary
one, yet a free course remains open for many variations. And
then there remain quantity and accent.
664. The alphabets actually in use are, moreover, far behind
as useful as what might be attained. To make the sounds of one language
in use not
they might
be made.
distinguishable from those of another-nay, even those of one
dialect from those of another-can never be the purpose of an
alphabet which seeks to serve, not scientific phonology, but only
LANGUAGE AND WRITING.
435
Jotter
ordinary practical needs; and this need necessarily occur only
so far as the differences in question have a functional value.
And thus it is that most of the alphabets in question go no
further than this. It is unnecessary to denote the differences
conditioned by the position in the syllable, in the word, in the
sentence, by means of quantity and accent, provided only that
the conditioning factors in the dialect in question have always
the same result. If, for instance, the hard s sound in lust, brust,
etc., is represented by the same sound which elsewhere denotes
the soft s sound, while in words like reiszen and flieszen it is
represented by sz (ss), this no doubt depends upon a historical
tradition (MHG. lust-rîsen); but it is still a great question whether
the method of writing sz would have maintained itself, were it
not that in the beginning of the syllable a need had made itself
felt to draw a distinction between the hard and the soft sound
(cf. reiszen―reisen, flieszen-fliesen), while in the combination st
the st is always hard even in forms derived from words which
have otherwise soft s (er reist being in pronunciation undistinguish-
able from er reiszt). That the fact of its origin from MHG. z has
not been the only determining factor is confirmed by the form
it takes when written in auslaut. Here, again, there is no differ-
ence in pronunciation between the s which proceeded from NHG.
s and that which proceeded from MHG. z; the s in hasz, heisz, is
pronounced like the s in glas, eis. Sz is at the present day
written in auslaut (for MHG. z) only in cases where etymologically
allied forms with hard s in inlaut exist along with it; thus we
find heisz, heiszer, etc.; but, on the other hand, das,¹ es, alles,
aus; also blos as an adverb, and bischen ('a little'). Again, it
is not usual to write kreisz, kreises
=
MHG. kreiz, kreizes.
1 The exception in the conjunction dass is to be explained by the necessity felt by the
grammarians for differentiating.
436
PRINCIPLES OF THE HISTORY OF LANGUAGE.
665. From all this it is clear that the separation in the ortho-
graphy started only from cases in which more than a single
pronunciation was possible in the same dialect. Thus, too, in
the process of orthographically settling most languages, it was
not felt necessary to employ any special sign for the guttural
and palatal nasal; but the same sign is employed as for the
dental, the labial receiving its own peculiar sign. The reason
of this was that the guttural and palatal nasal always occurred
only before other gutturals and palatals; ie. in the combinations
nk, ng, etc.; and in this position occurred without exception,
while the labial and the dental were also common in auslaut, and
before vowels in anlaut and inlaut; and thus had to be differ-
entiated from each other. In French, again, where the guttural
also appears in the auslaut of words, and in that of syllables
before labials and dentals, no such pressing necessity presents.
itself for any special notation; and indeed such would have been
with difficulty introduced, even if in other respects a closer
equivalence of spelling and pronunciation had been carried out;
for a guttural nasal is a universal rule for the auslaut of a
syllable. It is again unnecessary in NHG. to denote the distinc-
tion between the guttural and palatal ch. For the pronunciation
is definitely fixed by the preceding vowel, and changes accordingly
within the same stem: fach-fächer, loch-löcher, buch-bücher,
sprach, gesprochen-sprechen, spricht. If, on the other hand, there
were such a thing as a palatal ch also after a, o, u, a guttural
ch´after e, i, ä, ö, i, then the need for differentiation would certainly
exist, and possibly would also be satisfied. Still less is it necessary
to denote such differences as are necessarily conditioned by the
position in the auslaut or inlaut of syllables; e.g. in the mutes
to denote whether the formation or the dissolution of the termina-
tion is audible. Everywhere we find kk, tt, pp written, and yet
LANGUAGE AND WRITING.
437
the same movement is in no case twice carried out; but the
latter is the inversion of the former. In no case, further, have
the numerous economies of movement in the transition from one
sound to another found phonetic expression; cf. on this, Sievers,
Grundzüge der lautphysiologie, p. 84 sqq.
666. No doubt there are also certain alphabets, such as the
Sanscrit, which go beyond the measure of what is demanded by
immediate practical needs, and which satisfy somewhat exacting
claims on the part of phonetic physiology, since they keep separate
similar, though not identical, sounds, even in cases where the
differentiation is self-intelligible for any one who knows the lan-
guage, even without any regard to sense or context. Far more
common, however, are alphabets which fail even in such modest
demands as we have described. The main reason of such short-
comings is that almost every nation, instead of creating an alphabet
independently to suit the demands of its own language, has adapted
the alphabet of a foreign language, as best it might, to its own.
Besides this we have to reckon with the fact that, as language
develops, new differences may well arise which could not be fore-
seen or takeri account of at the time of the introduction of the
alphabet. The same causes may, however, further produce an
unnecessary superfluity. Superfluity and defect often stand side
by side. NHG. may serve as an example. We have instances of
several signs serving to denote the same sound in c-k—ch—q,
c—z, f—v, v—w, s-sz, ä-e, ai-ei, äu-eu, i-y. Again there are
symbols capable of denoting different sounds, the nature of which
is not fixed by their mere position. Such a symbol is e, which
may represent both the French é and the French è. Thus in the
relations between ä and e we see superfluity and defect united.
It is the same with v (only in foreign words, it is true) in its
relation to ƒ and w. Ch also may in foreign words have different
438
PRINCIPLES OF THE HISTORY OF LANGUAGE.
1
values, as in the case of chor-charmant. To denote the length of
vowels several methods are in use, such as doubling the consonant,
h and e (after i); and yet there are many cases in which it remains
unnoted. These defects are in great part as old as the drawing up
of the records of the German spoken language; and at an earlier
period they made their influence felt in a still more provoking way.
Others which were before in existence have gradually disappeared.
Thus it was likewise an instance of luxury and poverty when
u and v, i and, were used each to denote not merely the vowel,
but the fricative as well, and were considered, in accordance with
orthographic traditions, to be interchangeable, In MHG. MSS. o—ö,
u (û)—ü (iu)—uo-üe are not distinguished. And there would be
no difficulty in citing numerous other instances of the imperfec-
tions from which German orthography has had to suffer in the
different periods of its development.
667. If we further consider that the accentuation commonly
remains either wholly or mostly undenoted, it must be clear that
even those among the conventional orthographical symbols in
which the phonetic principle has not been intruded on by regard
for etymology and the phonetic conditions of an earlier period
present a very imperfect picture of living speech. Writing bears
about the same relation to language as that of a rough sketch to a
picture worked out with the utmost care in colour. The sketch
suffices beyond doubt to satisfy a person who has a clear
remembrance of the picture that it is intended to represent this,
and also to enable him to identify the single figures in both. On
the other hand, one who has merely a confused idea of the picture
would, if he trusted to the sketch, be able to correct and fill out
its details at most in regard to certain main points. And one who
has never seen the picture is of course quite unable to rightly
imagine the details of the drawing, colouring, and shading. Should
LANGUAGE AND WRITING.
439
several painters simultaneously attempt to execute a perfect picture
from the sketch, their productions would show great divergences.
Suppose, for instance, that in the original picture there occurred
beasts, plants, utensils, etc., which these painters have never in
their life seen in nature or in exact copies of nature, but which
possess a certain likeness to other objects known to them, would
they not in their own picture substitute these familiar objects for
those of the sketch? It must necessarily be the same with all who
learn a foreign language or a foreign dialect from written specimens
only, and seek to reproduce it from these alone. What can they
do but introduce for each letter and combination of letters the
sound and the combination of sounds which they are accustomed
to connect with them in their own dialect; while quantity and
accent will further be regulated according to its principles, except
in so far as divergences are expressly indicated to them by symbols
which they understand. It is commonly agreed that in the process
of learning foreign languages, even when the alphabet they employ
is identical with our own, at least a detailed description of the
value of each sound is requisite, and that even this can in no way
supersede the necessity of listening to the spoken utterance,
especially when it is not presented to us as based upon phonetics.
It is obvious, however, that the same need is present when we wish
to attain a correct apprehension of the sounds of a dialect belonging
to the same larger group as our own. It is important for us not
to overlook the consequences which flow from this fact.
differences
writing.
668. In every linguistic area, divided into many different dia- Dialectical
lects, there exists, as a rule, a large number of different gradations concealed by
of sound; certainly far more, even when we regard only what can
be clearly distinguished, and neglect all barely perceptible nuances,
than the letters contained in the common alphabet used by all the
dialects alike. In each single dialect, however, there exists only a
440
PRINCIPLES OF THE HISTORY OF LANGUAGE.
definite fraction of these gradations, the most nearly related in
many cases excluding each other; so that if we count those, which
for practical purposes do not require to be distinguished, as one
only, their number will fairly tally with the number of letters to be
disposed of. If under such circumstances writing is applied in
different places to express the dialect spoken at each, the only
conceivable way in which this can be done is by employing each
letter for exactly the particular species of a wider family of sounds
which occurs in the dialect in question; in other words, the same
letter is used here for one sound, there for another. In this pro-
cess it also occurs that if two nearly related species occur in a
single dialect side by side, one symbol has to serve for both;
while, conversely, of two signs, indispensable for the other dialects,
one may be dispensed with for one or the other. We need only
examine some of the most important of such cases as they occur
in the German linguistic area, including not merely the dialect
strictly speaking, but also the language spoken by the greater
part of the educated classes. The distinction between hard and
soft fricatives exists in Upper as well as in Lower Germany. But
while in the former it rests on the greater or less energy of exspira-
tion, in the latter¹ a further characteristic comes in-the absence
or presence of the vocal tone. The Upper-Saxon and Thuringian
dialects, however, have no distinction either by vocal tone or by
the energy of exspiration. Accordingly b, for instance, denotes for
the Upper German a different sound (unvoiced lenis), from that
which it denotes for the Nether-German (voiced lenis), and again
another still for the Upper-Saxon (unvoiced fortis). Besides, k, t,
p in certain places denote for the Upper-Saxon and Thuringian
a different sound (unbreathed fortis) from that which they denote
There is no necessity to define the limits more closely; and I am unable to do so.
The fact was first established by Winteler, Grammatik der Kerenzer mundart, p. 20 sq I
LANGUAGE AND WRITING.
441
for the mass of other Germans (aspirate).¹ The Nether-German
pronounces the w as a labio-dental; the Central-German as a
labio-labial fricative; the Aleman as a consonantal vowel. The s
in the beginning of words, before t and p, is in a great part of
Nether-Germany pronounced as hard s, in the rest of Germany it
is spoken like sch. R is in one part a lingual, in an other a uvular
sound; and numerous other variations occur. G is in one
part of Nether and Central Germany, as well as in certain Upper
German districts, pronounced as a guttural or palatal fricative
sound, either invariably or only in inlaut. In the Germanic dia-
lects g has always denoted mute as well as the fricative. The
Alemannic does not distinguish the pronunciation of the ch
according to the nature of the preceding vowel. On the other
hand, it makes a distinction between f= NG. p and f= NG. f, which
is wanting elsewhere.
669. Where identity of symbol coincides with etymological
identity, but the pronunciation varies, the written form conceals a
dialectical difference. As this is of very common occurrence,
especially when we take into account the numerous variations
which, though insignificant in single cases, are, on the whole, per-
ceptible enough, as also, in most cases, the quantity and, above all,
the modulations of the pitch and of the energy of exspiration
remain undenoted, we must admit that a considerable portion of
the dialectic differences is taken no account of in writing. It is
precisely this fact which renders writing of special convenience as
a channel of communication in general intercourse. But it like-
wise renders it unsuited to influence pronunciation, and it is a per-
fectly mistaken view that we can operate as effectually with the
written word at a distance as with the spoken word on what is
near.
1 Cf. Kräuter, Zeitschrift für vergleichende sprachforschung, 21, 30 sqq.
442
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Incapacity of
writing to
alterations of
tion.
670. How, for instance, can any one, on seeing the written
symbol g, know which of the easily distinguishable pronunciations
current in Germany-they are at least seven in number, and in
part very divergent-was intended by the employer of the
symbol? How can we even know, merely from the writing, that
so many pronunciations exist at all? What can he do in such a
case but assume, as corresponding to it, the pronunciation current
in his own home?
671. Only the most extreme departures from one's own dialect
communicate can be detected in written speech; and even then without any
pronuncia- sure discovery of the special characteristics of the divergent
sounds. As far as we recognise departures we are, of course, able
to imitate these. This must, however, occur with full conscious-
ness and with full intention, the imitation of the foreign dialect
being undertaken as something distinct from the use of one's own.
It is a process distinguishable only in degree and not in kind from
the acquisition of a foreign language; and, on the other hand,
totally different from the unconscious process of being influenced
by the speech of our associates, already described in p. 43. The
fundamental condition of this is, as we saw, the small compass
within which the differences between individuals lie, and the
infinite capacity of the spoken sounds for gradation. Within the
sphere in which this kind of influence finds its scope, writing shows
no differences as yet, and therefore is incapable of influence.
672. And as with influence in the distance, so with influence
on the future. It is pure fancy to suppose that in writing we have
a control over sound-changes. Just as, in different places, sounds
markedly different can be denoted by the same letters, so, and even
more easily, in the same place at different times. No letter, of
course, is connected really with any particular sound by a real
bond capable of maintaining itself independently: the connexion
LANGUAGE AND WRITING.
443
depends exclusively upon association of the ideas. We connect
with each letter the association of just such sound as is in use at a
particular time. The process in the case of natural sound-change
is as we have seen (p. 44),-that in the place of this idea, another
slightly different is substituted, which, transmitted to the following
generation, is received from the outset as bound up with the letters.
The sound-image associated with the letter is thus unable to
exercise any check upon the sound-change, because it is itself
modified by this. And naturally, the actual phonetic value of
a letter is always assumed for it in the records of the past. There
is no possible way to compare the earlier phonetic value with
the present one. It is of course possible, by the aid of
scientific researches, to conjecture the variations; this does not
affect the present question. Commonly speaking, the changed
pronunciation may long go hand in hand with unchanged ortho-
graphy without any insupportable inconveniences arising from the
discrepancy. At all events, these do not declare themselves as
such until the change has become very violent. But then, a change
of language to suit orthography, if possible at all, can only be
brought about by conscious intention, and any change of the
kind would again be directly contrary to natural development.
As long as this follows its path undisturbed, it remains but to
put up with the inconveniences as well as we can, or to alter the
orthography to suit the language.
673. The defects of written-speech, so far enumerated, are far
from indicating the degree of divergence which may obtain between
writing and speech. We have hitherto regarded, strictly speaking,
only the condition incident to the period when language first
begins to be fixed by writing; when every one who writes is still
taking part independently in the creation of orthography; the
sign for each single sound being indeed settled, but not the spell-
444
PRINCIPLES OF THE HISTORY OF LANGUAGE.
ing of the word as a whole; so that the writer has always to break
the word up into its elements as best he can, and compound the
letters corresponding to these elements. There is no doubt, how-
ever, that mature practice in reading and writing renders the
process continually shorter. Originally the connexion between
the phonetic signs and the signification was always brought about
by the image of the sound and by the motory sensation. As soon,
however, as both of these are brought near together by this com-
bining element they form a direct connexion, and the combining
factor can be dispensed with. The possibility of fluent reading
and writing depends of course on this direct association. This
can be easily demonstrated by an inverse proof. If the written
symbols of a particular dialect are laid before some one who is
perfectly conversant with it, but who has hitherto known it only
aurally, he will always at first find some difficulty in adapting the
written symbols to the sounds, especially when the former do not
precisely conform to the system of the literary language with all its
inconveniences. And he may be more puzzled still if he be required
himself to write in such a dialect, even though it be that which he
has spoken from childhood. He will always evade a real solution
of the task, by submitting unduly to the influence of the familiar
orthography of the literary language. All modern dialectical
poets give proof of this. We must cast aside this background of
literary orthography which now always serves as an analogy, if we
would clearly realise the difference between the position which we
now-a-days occupy in committing our common language to writing
and that of, say the Old High German scribes, in finding written
symbols for their own dialect. We shall then be more ready to look
with compassionate toleration on our forefathers' want of skill. We
shall rather discover, especially if we do not confuse different writers,
but examine the orthographic usage of each separately, that they
LANGUAGE AND WRITING.
445
observed the sounds more correctly than is commonly done at the
present day; and this for a reason which, regarded from another
side, seems a defect in comparison with our present conditions,
viz., that they were contorted by no fixed standard of orthography,
and thus their unprejudiced observation of the sound was not
confused by the perpetual contemplation of such orthography. In
other words, they could not yet dispense with the phonetic image
as mediating link between the orthographic image and the
signification.
fixity of
674. The two stand in the closest mutual relation. If the direct Effect of
connexion between the orthographic image and the signification orthography.
is very close in the mind of all tolerably educated persons, this is
in great part due to the fixity of our orthography. We see this
particularly in such words as are in pronunciation identical, but
different in writing. Every divergence in spelling, even though
from a phonetic point of view it may be a decided improvement,
increases the difficulty of understanding. If this is a striking proof
of the direct combination of writing and speech, on the other side
the negative conclusion must be drawn from it, that the less fixed
the spelling, the less is any direct connexion possible between it
and the meaning. The want of fixity may arise from the un-
suitability of the material at hand, or from the awkwardness
of the writers, as when several signs are used interchange-
ably in the same acceptation, or, conversely, when a single sign
appears now in one acceptation, now in another; or it may be due
to the want of regulative authorities who might render possible a
concentration and union of the different orthographical efforts. It
may, however, be also due to the very perfection and consistency
of the phonetic-physiological phenomena. If, e.g., the spelling of
the stem in the different forms changes with the sound (MHG. tac
tages, neigen — neicte, etc.), or if, as in Sanskrit, even the spelling
446
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Natural
tendency in
towards
greater
fixity.
of the same form changes with its position in the sentence,—the
single meaning is confronted with a number of variations in spell-
ing, and it is impossible for an absolutely distinct orthographical
image to connect itself with the first. As long as fixity in spell-
ing is wanting, it is impossible, however great may be the skill
attained in writing and reading, to make the direct connexion
perfect. At the same time, however, the very practice tends to
produce a greater fixity. Every advance made in the former is to
the advantage of the latter, and every advance in the latter
facilitates the former.
675. We thus find that the natural progress of the development
orthography of the orthography of any language is towards greater fixity, and
this even at the cost of phonetic-physiological accuracy. It is no
doubt true that consistent progress in this direction is not the rule.
Violent phonetic changes are especially apt to produce variations
and retrogressive movements. There are three means by the aid
of which orthography strives to gain fixity: (1) the abolition of
the variation between several different manners of writing; (2) by
paying attention to etymology; (3) by keeping close to tradition
and disregarding changes of sound. The first-named method is,
viewed even from a phonetic stand-point, often an advance, or at
least not a retrogression, though it not unfrequently occurs that the
phonetic principle is discarded in the operation; the two others,
however, are direct contraventions of this principle. Naturally,
however, the tendency to bring language and writing into closer
connexion with each other still remains active: and this tendency
works partly in the direction of cancelling original deficiencies,
and partly in the reaction against the new inconveniences per-
petually arising owing to sound-change. As this tendency in most
cases comes into conflict with the effort to attain fixity, the history
of orthography exhibits the spectacle of a perpetual struggle
LANGUAGE AND WRITING.
447
between these two tendencies, the state of which at any given
moment gives a criterion of the relative strength of the combatants
at that moment.
between the
of writing
language.
676. If we follow the movement in detail, we find striking Analogies
analogies with the development of language side by side with development
notable differences. The latter depend chiefly on the following and that of
points. In the first place, changes in orthography are effected with
much more consciousness and purpose than changes in language;
though we must beware of exaggerating the degree of purpose.
In the second place, in the struggle over orthography, it is
not, as in the struggle over language, the whole linguistic com-
munity which is interested, but at most the writing (or printing
and publishing) portion of that community, and the individual
members of this in very different degrees and with very different
intensity; the preponderance of special individuals makes its in-
fluence felt to a much stronger degree than in language. In the
third place, since the capacity of influence is not limited to prox-
imity in space, it is possible in the department of orthography for
quite other ramifications of the reciprocal influences to develop
than in the department of language. In the fourth place, ortho-
graphic changes stand in decided contrast to sound-change in this
respect, that they are unable to advance by means of fine grada-
tions, but can only do so by leaps and bounds.
variations in
of identical
677. Let us examine, in the first place, the means by which Abolition of
fructuation between sound-signs of identical value is overcome. sound-signs
Such fluctuation may arise in several ways. The signs may have value.
been already employed indiscriminately in the language from
which the alphabet is borrowed: thus it stands in OHG. with the
doublets i-j, u—v, k—c, c—z. Or it may happen that two signs
had a different value in the original language, but that the language
which borrows them possesses no corresponding distinction, so
448
PRINCIPLES OF THE HISTORY OF LANGUAGE.
that now both signs fall upon the same sound. It especially
happens that both easily pass into use when the single sound of
the borrowing language lies between the two sounds of the other.
Thus in Upper German, at the time of the introduction of
the Latin alphabet there was in the guttural and labial series
no difference fully answering to the Latin distinction between
voiced media and tenuis; in the anlaut of syllables there was
nothing even approximately answering to it; but merely a sound
which differed from the Latin media in lacking the voice-tone;
from the tenuis by a weaker exspiration. Hence arose a vacillation
between g and k, b and p. The vacillation too between ƒ and v (u)
and in MG. that between v and b arose in the same manner.
Further, double symbols may arise only in the course of further
development where two originally different sounds fall together,
their two signs being then exchanged. Thus, for instance, in later
MHG. hard s and ≈ fall together, whence, sas written for saz, and
conversely huz for hus, etc.; though the latter case is certainly
rarer from the very outset. Finally, however, it is possible for a
division to appear owing to the different development of the same
written symbol; compare Latin i-j, u-v, and in German text
('fracturschrift') { and s. It is possible for these differences to
multiply, especially if at a later period we try to utilise an older
grade of development, as we see, for example, in the use of capital
letters side by side with small letters.
678. The superfluities thus arising are disposed of in a similar
fashion to that which removes the superfluities in words and forms.
The simplest way is by the gradual obsolescence of one of the two
signs. The other way is by the differentiation of the signs
originally used indifferently. This may be exclusively phonetic,
the removal of the superfluity at the same time serving to supply
a closely related want; e.g. when in NHG. i, u, and j, v were
LANGUAGE AND WRITING.
449
gradually differentiated as vowel and consonant. It happens not
unfrequently that the position of the sound within the word gives
the key to the differentiation without any phonetic difference being
present, or at least without any such being remarked by the
writers; as when we find j and v for a long continuous period
principally used in the anlaut of the word (even in the place of the
vowel); c in MHG. (irrespective of the combinations ch and sch) in
a preponderating majority of cases confined to the auslaut of a
syllable (sac, tac, neicte, sackes) and in NHG. only in the gemination
ck, it being in other cases ousted by etymological orthography;
in MHG., finally, f much more frequently before r, l, and u
and kindred vowels than before a, e, o. A third method lastly
consists in this, that without any phonetic or graphic motive,
according to individual caprice and fancy, one orthography becomes
usual in one word and another in that. This accounts for the
relation in NHG. between ƒ—v (fall—vater, etc.), t—th (tuch—thun,
gut—muth, etc.), r—rh, ai—ei, and further between the cases in
which length is denoted and those in which it is not denoted, and
between the different methods of such denotation (nehmen-geben,
aal-wahl, viel-ihr, etc.). An essential factor in this process, and
one of the main reasons which have prevented the consummation
of a uniform orthography, in fact which in modern times is ever
anew opposing itself to a consistent reform of orthography, is the
effort to distinguish words which have the same sounds but
different meanings. Compare, among others, ferse—verse, fiel—
viel, tau-thau, ton — thon, rein - Rhein, rede-rhede, laib - leib,
Main-mein, rain-rein, los-loos, mal-mahl, malen-mahlen,
war-wahr, sole-sohle, stil-stiel, aale-ahle, heer-hehr, meer-
mehr, moor—mohr. Even different meanings of words originally
identical are thus distinguished; cf. das-dasz, wider-wieder, etc.
Here, too, should be noticed, the gradual process whereby it
2 F
450
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Effect of
etymology.
became usual to employ capital letters for the initial letters of
substantives which were employed at an earlier period merely to
emphasise the word. Here again we see the tendency to utilise
writing for the expression of distinctions unknown to speech.
This manner of differentiation is one of the most characteristic
signs for enabling the written to become independent over-against
the spoken language. It also does not occur till such time as a
true written language has freed itself from the dialects, and it is
the product of grammatical reflexion. It is, however, worthy of
remark that even this reflexion does not in the first place create
differences in the way of writing to express its distinctions, but
merely utilises for its own purposes the variations which have by
accident arisen. In cases where no such variations exist, it is
possible for the impulse to differentiation to fail to make itself felt;
cf. the homonyms cited above, p. 228. Moreover, it does not
exhibit itself as operative in all those cases where we might have
expected that it would have done so.
679. Like the unphonetic differentiation, the operation of
etymology makes itself felt most strongly and most consistently
in written language; though it is in many cases unmistakably
present in dialectic records. We may compare the expulsion of
a more archaic phonetic orthography by an etymological one, with
the process of analogical creation whereby unmeaning sound-
differences are levelled; indeed we may actually designate it as a
kind of analogical creation peculiar to written language, for which
the laws hold good which we have brought before our readers'
notice. In this case, too, of course the etymology is not in itself
decisive, but rather the manner of grouping prevalent in the
language at the time. Isolation is a protection against assimila-
tion, and conversely, secondary rapprochement between sound and
meaning causes an attraction over into analogy.
LANGUAGE AND WRITING.
45I
=
680. Let us now consider from this point of view the most
important cases in which the NHG. has deserted the phonetic
spelling of MHG., and has allowed assimilation to set in. In MHG.
the media in auslaut, and before hard consonants in writing,¹ as
in pronunciation, becomes a tenuis; in NHG. this is the case in
pronunciation only, not in writing: MHG. tac, leit, gap, neicte
NHG. tag, leid, gab, neigte. We see the MHG. rule observed in
haupt (= houbet, houpt), behaupten, because there are no longer
any related forms with unsyncopated vowels remaining by its
side; in the proper name Schmitt, Schmidt; and in schultheiss,
where the composition with schuld is no longer felt. In MHG.
consonantal doubling is not expressed in writing in auslaut and
before another consonant: mann—mannes, brante—brennen. NHG.
represents the doubling in cases where forms etymologically
closely connected give the model for such; cf. mann, brannte,
männlich, männchen (though we find it already discarded in brand,
brunst, etc.); still in the pronoun man, and further in brantewein,
brantwein (which is no longer apprehended as gebrannte wein);
on the other hand, with a later attraction to herr we have herrlich,
herrschaft, herrschen, which appear in MHG. as hêrlich, hêrschaft,
hêrsen from hêr = = NHG. hehr. In MHG. it is true that the umlaut
of the long a is in most cases separated from ê as a; while that
of the short a is denoted by e. In NHG. is likewise employed
for the umlaut of the sound which was originally short, but is
now in many cases drawn out when we are still clearly conscious
of the relation to a form from the same root which has not under-
gone umlaut; thus, vater-väter, väterchen, väterlich, kraft—
kräfte, kräftig, glas-gläser, gläsern, kalt — kälter, kälte, land—
gelände, arg—ärger, ärgern, fahıre-fährst; in the same way in
the diphthongs, baum—bäume, haut-häute, häuten, bärenhäuter
¹ Less regularly, it is true, in the MSS. than in the critical editions.
452
PRINCIPLES OF THE HISTORY OF LANGUAGE.
(MHG. hût-hiute); on the other hand, erbe, ente (MHG. ant, gen.
ente) enge, engel, besser, regen (verb), although spoken with an
open e as well; leute, etc., since here kindred forms unaffected
by umlaut are wanting. The difference between ligen and legen,
winden and wenden, hangen and hängen, and fallen and fällen, is
worth noticing; in the first-named pairs we doubtless find also a
in the preterite (lag, wand), but it is merely the present which
is brought into relation with the present. Where the bond that
unites the group has been severed or loosened e remains; cf. vetter
with vater, gerben with gar, scherge with schar, hegen, gehege, and
hecke with hag, heu with hauen, fertig with fart (on the other
hand hoffärtig), eltern with älteren, behende with hände, ausmerzen
with märz (the a owing to the Latin a), strecke with stracks. The
assimilation likewise fails in cases where the form affected by
umlaut appears as the primary one; cf. brennen-brannte, nennen
-nannte, etc. The observation may be also made that the
accession of a further phonetic difference operates as a restraint;
thus we find hahn-henne, nass-netzen, henken, henker, as com-
pared with hängen. On the other hand, the e is in certain cases—
even where it has not arisen through umlaut, but is the equivalent
of the original German e (ë),-nevertheless apprehended as a case
of umlaut, if a word with a stands by it, from which that in e
can appear to be derived; cf. rächen (MHG. rëchen) referred to
rache (MHG. râche), schämen (MHG. schëmen) to scham, wägen,
erwägen (a form which arose from a confusion of the MHG. wëgen
with wegen) to wage (on the other hand we have bewegen).¹
681. Etymology plays an important part also in the re-
gulation of sound-vacillations spoken of above, p. 448. We
1 The correctness of the inductions given above is by no means impugned by the fact
that the a is found instead of e and e in some other cases as well where it has not been
motived by any reference to an a. To a certain extent the tendency to differentiation
also comes into account; cf. e.g. währen, gewähren, gewähr—wehren, gewehr.
LANGUAGE AND WRITING.
453
naturally write fahren-fahrt-gefährte-furt, etc., with an ƒ in
each case. Where his used as a symbol of lengthening, it is
as a rule consistently carried through in all related forms with
change of vocalisation; cf. nehmen—nahm—genehm—übernahme,
befehle—befiehlt-befahl-befohlen-befehl, etc. As examples of
isolation we may cite zwar (= MHG. zewâre), as against wahr;
drittel, viertel, etc., against theil; vertheidigen (from tagedingen) as
against tag.
682. This assimilation, however, is as a rule confined within
definite limits, as it only occurs in cases where it cannot render
the pronunciation doubtful. It is possible in NHG. to write lebte
with a b without any inconvenience, because the language does
not in any case distinguish between 6 and p in the auslaut of
a syllable. But we can consistently carry a long-symbol, for
instance, only so far through the related forms as the vowel
maintains its length (thus genommen as related to nehmen, furt to
fahren), and the doubling can only be consistently carried out as
far as the preceding vowel is short (hence kam as related to
kommen, fiel to fallen).
683. Analogy, too, operates protectively against changes of
the older spelling; and here, again, we have a divergence from
the conditions of spoken language. This is specially to be
observed in the case of the orthography of the French language.
If the consonants which in auslaut have become mute are
maintained in writing, the reason is that related forms in most
cases maintain themselves as well, in which these consonants
are still pronounced, and that they are also pronounced in the
same form when a word beginning with a vowel connects.
itself with them. If the French were to write, for instance,
fai, lai, gri, il avai, tu a, a startling contrast would appear be-
tween these forms and faite, laide, grise, avait-il, tu as été;
454
PRINCIPLES OF THE HISTORY OF LANGUAGE.
as has actually not been avoided in il a-a-t-il. Thus con-
sistency in writing would also be destroyed if we should endeavour
to introduce a special sign for the guttural nasal; in such cases
un would be differently written in un père and un ami. Further,
if it were desired to draw a distinction between the nasalised
and the non-nasalised vowel, it would be necessary to employ
different signs in the case of cousin and cousine, un and une, ingrat
and inégal. We may gather that the analogy of the related
forms has been decisive for the result from a number of isolated
forms like plutôt, toujours, hormis, faufiler, plafond (on the other
hand plat-bord), verglas (referred to vert), morbleu, morfil, Granville,
Gérarcourt, Aubervilliers, fainéant, vaurien, Omont (referred to haut).
We may also compare such cases of isolation as Clermont—clair.
684. If writing cannot keep up with the phonetic development
of language, it is easy to see that the reason of this is to be
sought in nothing else but in its want of continuity. In the phonetic
conditions it is unquestionable that continuity alone renders
possible the union of continuous advance with fixity of usage.
A similar fixity of usage in writing implies an unchange-
ableness of the latter; and this again implies a perpetual increase
of the discrepancy between writing and pronunciation. On the
other hand, the more vacillating that orthography is, the more
capable is it of development; or, conversely, the more it seeks
to follow the development of language, the more vacillating does
it become.
685. We must, however, in addition emphasize certain points
of view according to which a faithful adherence to the old
spelling, when the pronunciation is changed, becomes still more
intelligible. In judging of the relation of writing and sound in a
language it often happens that the standard of another language
intrudes improperly, while the orthography of every language
LANGUAGE AND WRITING.
455
claims to be judged according to its own conditions. As long
as ever a definite sound corresponds to a definite symbol, there
can be no question of any discrepancy between writing and pro-
nunciation. In one language it may be one sound, in another
a different one; this does not affect the question at all. Hence
if a sound consistently changes in all its positions, and in the
course of these changes does not coincide with another sound
already existing, no change in the orthography need ensue, and
the correspondence between writing and pronunciation is main-
tained. But even when the change is not consistent, a division
occurs, if only none of the different sounds coincides with
one already existing, no other alternative as a rule remains but
to maintain the old orthography; for, in order to distinguish
the sound, we should need at least one symbol more than we
have; and this cannot be created at will. The only resource is
where a superfluity existed before, which it is now possible to
employ to some purpose. In order to maintain the phonetic
principle intact certain violent innovations would from time to
time be indispensable, and it is hard to reconcile these with the
maintenance of unity in the orthography.
686. We must add to this that the operation of analogy
described counts for much in the preservation of forms. And
finally, we should remark that by the introduction of phonetic
writing many distinctions would be completely cancelled which
now are maintained in the written language. Thus in the French
language the plural would in most cases cease to be distinguished
from the singular; and in many cases the feminine would be undis-
tinguished from the masculine (clair-claire, etc.). In those cases,
however, in which differences still remained, the consistency of
the method of formation which now mainly prevails in writing
would be abolished.
Mixture in
the wider and
the narrower
sense.
Mixture of
distinct lan-
IF
CHAPTER XXII.
ON MIXTURE IN LANGUAGE.¹
F we start by assuming that individual languages are the only
ones which have any real existence, we are justified in
asserting that as soon as any two individuals converse, a mixture
in language is the result. The speaker in speaking influences the
idea-groups of the hearer which have reference to language. If we
accept the expression 'mixture in language' in this wide sense, we
must admit that Schuchardt is right in his assertion that of all the
questions with which the science of language at the present day
has to deal, none is of greater importance than mixture of lan-
guage. In this sense we have necessarily been discussing 'mix-
ture in language' throughout all the previous chapters, since it is
inseparable from the life of language. Here, however, we take the
word in a narrower sense, to denote a process not absolutely indis-
pensable to the life of language, though it be wholly wanting in
hardly any domain of speech.
688. Mixture of language in this more restricted sense is, in
guages, dia- the first instance, the influence exercised by one language upon
another, when the two languages are either wholly unconnected,
lects, and
stages of
language.
1 Cf. for this chapter Whitney, On Mixture in Language (Transactions of the Ameri-
can Philological Association, 1881); and particularly Schuchardt, Slavodeutsches und
slavoitalienisches, Graz, 1885 [also Sayce, Principles of Comparative Philology, Chapter
V., 1885].
ON MIXTURE IN LANGUAGE.
457
or, though originally connected, are so far differentiated that one
has to be specially learnt by those who are acquainted with the
other it also signifies the influence to which one dialect is subjected
by another belonging to the same unbroken and continuous lin-
guistic area, even when the two dialects do not yet differ so
strongly as to prevent a reciprocal understanding holding good
between the constituents of the two separate areas.
There is yet
another kind of mixture in language which consists in the re-
adoption into the language of material which existed at an earlier
period in that language, but has since died out.
originates in
689. We will consider, in the first instance, the mixture of dif- Mixture
ferent languages clearly and sharply divided from each other. In individuals.
order to understand the process in the case of mixture, we must, of
course, closely observe what passes in the case of single indi-
viduals. The strongest predisposition to mixture is given when
there are individuals who speak more than one language; possibly
several languages indifferently, but, at least, one other besides
their mother-tongue. Under any circumstances a certain mini-
mum understanding of a foreign language is indispensable. At
"least what is adopted from the foreign language must be under-
stood, even if only imperfectly.
ism.
690. The most ordinary cause of bilingualism, or a more or less Bilingual-
perfect apprehension of a foreign tongue is, of course, the situation
of a community upon the confines of two linguistic areas, and it
occurs in a greater or less degree according to the activity of the
intercourse between them. Again, it may be due to journeys and
the temporary residence of individuals in foreign countries; it
becomes more marked when individuals permanently migrate from
one country to another; and still more, when large masses of
population are permanently transplanted by conquests and by
colonisation.
458
PRINCIPLES OF THE HISTORY OF LANGUAGE.
691. Finally, the knowledge of the language of a foreign people
may be imparted by writing, without any direct contact between
the learners and teachers. In the last-mentioned case the know-
ledge thus gained commonly remains confined to certain excep-
tionally well-educated strata of society. It must thus be remarked
that the method of communication afforded by writing facilitates
not merely the process of borrowing from a foreign language, but
further, of borrowing from a primitive and distant epoch of its
development.
If one
692. In cases where two nations have been brought into con-
tact, and have been mixed on a large scale, bilingualism, and, con-
sequently, instances of reciprocal influence exercised on and
received by one language from another are common.
nation has a decided preponderance over the other, whether
this be due to its size or to pre-eminence-political or industrial
or intellectual-it will be found that the employment of its
language will tend to spread at the expense of the other: there
will be a gradual change from bilingualism again to unilingualism.
This process will be more or less speedy in its fulfilment, accord-
ing to the capacity for resistance of the inferior language, and this
language will leave traces more or less strongly marked in the
victorious one.
འ
693. The mixture again will not easily affect a single indi-
vidual, so as to cause his diction to embrace component parts
taken from one language in approximately similar proportions to
those taken from the other. Assuming that he is equally master
of both languages, it may be that he will find no difficulty in
passing from one to the other: but it will still be found that
within a single clause of a sentence one language will form the
strict basis; the other, though it may exercise a more or less modi-
fying influence, will still play but a secondary part. This, of
ON MIXTURE IN LANGUAGE.
459
course, applies even more to one who has as yet gained no real
command over the foreign language, but merely more or less
power of understanding it. The person who speaks two languages
will, of course, be liable to have each of these influenced by the
other his mother-tongue by the foreign language, and the foreign
language by his mother-tongue. The influence of the latter will,
as a rule, be stronger than that of the former. This is inevitable
as long as the foreign language is not completely mastered. Still
the influence of the foreign idiom may also be very strong upon
the mother-tongue when the speaker applies himself earnestly to
its study: and this is commonly the case when a foreign language
and foreign culture is more highly prized than the native language
of the speaker.
694. If, however, the impulse necessary to cause one language
to feel the influence of another must proceed from persons who
are masters of both languages, in however imperfect a degree; this
sensitiveness to influence may still spread within the same
linguistic community by means of the ordinary assimilating effect
of intercourse; and may thus reach persons who have not the
remotest direct contact with the foreign idiom. The latter are in
the process influenced not merely by the members of their own.
community, but under certain circumstances by members of a
foreign people as well, who have adopted their languagé. They
will, of course, in every case assimilate the foreign elements, but
slowly and in small quantities.
modes of
A. Adoption
695. There are two main ways in which a foreign idiom may Two chief
exert an influence. In the first place, foreign material may be influence.
adopted into the mother tongue. In the second place, though only of foreign
native material may be employed, still the methods of linking this B. Usage
into sentences, and its application to the conceptions it represents material after
may follow a foreign model: in this case the influencing force model.
material.
of native
a foreign
460
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Causes of the
adoption of
words.
extends merely to what Humboldt and Steinthal have described
as the inner language form (‘innere Sprachform').
696. Of course the first and foremost of the causes for the
adoption of foreign words into the mother tongue is the need felt
for them. Words are accordingly adopted for ideas which have as
yet no words to express them. As a rule, both the idea and the
word to express it are taken together from the same source. The
names of places and persons are the most obvious and common
among such adopted words; to which may be added, the names of
foreign products. If these are natural products their names may
pass, together with the objects themselves, from the least cultivated
people to the most highly cultivated; while on the other hand, the
introduction of artificial products with their names presupposes a
certain predominance of foreign culture, which may no doubt
extend to but a very limited area. Such predominance must be
presupposed more decidedly still when technical, scientific, re-
ligious, or political terms are introduced. A strong civilising
influence entails almost invariably a large importation of foreign
words. One other want must be mentioned as capable of causing
the adoption of words from a lower sphere of culture, namely, that
of describing the circumstances of a foreign country, whether this
description be designed as information, and so attempt to give
accurate delineation or narration, or whether it be applied for
poetical purposes. The process of borrowing transcends its actual
need when the foreign language and culture is prized higher than
the native, and when accordingly the mixture of words and phrases
taken from the foreign language passes as specially elegant or
tasteful.
697. There are practically no limits to the number of words
transferred from the mother tongue into the foreign one when a
speaker is compelled to employ that language without full mastery
ON MIXTURE IN LANGUAGE.
461
of it. Thus words of the most varied description may be intro-
duced into a language by persons who speak that language as a
foreign idiom.
process of
698. Borrowed words are subject to much the same laws as Stages in the
newly-coined words. The first employer of such words has no adoption.
intention commonly speaking of making them usual. He merely
employs them to satisfy the momentary need felt of making
himself understood. Such employment leaves no lasting effects
until it is repeated; as a rule, indeed, leaves none, except when it
proceeds spontaneously from different individuals. A loan-word
does not become usual at once. There are different grades of
usual employment. In the first place, there is a narrow circle
within a given community whose existence depends either upon
the ties of neighbourhood or similarity of culture, within which a
word becomes usual. It may be that there are several such circles.
Many words remain in this restricted application, while others
spread over the whole population. If these words have quite
generally become usual, and if they present no remarkable
phonetic irregularities, the instinct of language regards them much
as it regards native linguistic material. In fact, to the instinct of
language they are no longer foreign words at all.
of foreign
material.
699. The treatment of foreign phonetic material, in cases where Treatment
foreign words are borrowed, merits particular attention. We have phonetic
seen that in no two languages does the stock of sounds precisely
tally. To learn to speak a foreign language with exactness, new
motory sensations need to be acquired and put in practice. This
process is necessary unless the speaker would continue to work on
with the same motory sensations as those with which he utters his
mother tongue. Consequently he will, as a rule, replace the foreign
sounds by those nearest related to them in his own mother tongue,
and in cases where he attempts to utter sounds not occurring in
462
PRINCIPLES OF THE HISTORY OF LANGUAGE.
his own language, he will, in the first instance at least, fail in his
attempt. No doubt a long course of hearing and practising a
foreign language may impart a more correct pronunciation, but
still it is notoriously seldom that any one masters a foreign
language so perfectly as not to be taken for a foreigner. Thus
where a language spreads its domain over a nation that originally
spoke a different language, it must almost certainly result that the
earlier language of the people must leave some traces in the pro-
duction of sounds, and that other departures, more or less violent,
from the normal usage, should set in, because the motory sensations
connected with the two languages are not harmonious. Where the
acquirement of the foreign language is the result of writing alone,
of course there is no attempt made to imitate the foreign sounds,
but the sounds of the speaker's native language are assumed as
those of that thus acquired.
700. In cases where one people merely comes into contact with
another in the course of travel, or by literary intercourse, it will be
found that a minority only will understand the language of the
foreign people; a smaller minority will be able to speak it, and a
yet smaller to speak it accurately. It will thus happen that in the
course of borrowing a word from a foreign language, the persons
who first introduced the word will insert sounds belonging to their
own language among the foreign ones. Assuming, however, that
the foreign word be adopted with an absolutely correct pronuncia-
tion, such pronunciation certainly will not be able to maintain it-
self when it spreads to persons who are only imperfectly acquainted
with the foreign language, or, it may be, absolutely unacquainted
with it. The lack of a corresponding motory sensation makes in
this case the intrusive element, the sound-substitution (to speak
with Gröber) an absolute necessity. If a foreign word has once
made its way into a language, it is generally composed of the
ON MIXTURE IN LANGUAGE.
463
materials of the language into which it has entered. Even those
who, owing to their exact knowledge of the foreign language, are
aware of the departure it has taken, have to give in to the majority,
at the risk of appearing pedantic or affected. Only in rare circum-
stances does a foreign sound gain complete right of citizenship in
a language; those sounds, of course, succeeding most easily which
occur most frequently, and which contrast most sharply with those
native to the language. Thus, for instance, in spite of all the
numerous loan-words in the NHG. written language, one new sound
and one only has been introduced, viz., that of the French j or gin
jalousie, genie, genieren, etc. And even this sound is replaced in
popular dialects, and even in the ordinary language of intercourse
of towns, by the sound of the German sch, English sh.
701. It happens, not unfrequently, that several different foreign
sounds are replaced by the same native sound. Thus in OHG. the
Latin sounds of ƒ and v are both represented by ƒ (which is some-
times written as v or u), cf. fenstar, fiebar, fîra, etc.—fers, fogat
(vocatus), evangelio, etc. The reason why v as well as ƒ was
reproduced by ƒ is the lack of a sound exactly corresponding to
the Latin, a u, still consonantal, being spoken in the place of the
modern German w. Besides, in OHG. the Latin strong p, just like
the sonant lenis b, is reproduced by the surd lenis which lies
between the two, now written as b, now as p; cf. beh (peh)=pix,
bira=pirum, bredigôn=praedicare, etc.-becchi (pecchi) = baccinum,
buliz boletum, etc. The reason is that in Southern German there
was no voiced b after the sound-shifting, because the previously
existing one had ceased to be voiced, and no strong p, because that
which had previously existed was now shifted to ph. Conversely,
it may be possible to reproduce the foreign sound now by one,
1 Cf. Franz: die lateinisch-romanischen elemente im althochdeutschen. Strassburg,
1884, pp. 20, 22.
464
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Assimilation
of words
already
adopted.
now by another approximately similar native sound. It will,
however, commonly be found in cases where in the loan-words of
a language the same foreign sound is reproduced now by one now
by another sound, that the adoption of these words has taken
place at different periods. Thus it is that the Latin v is repro-
duced by w in the oldest German loan-words (cf. wîn, wiccha, pfâwo,
etc.), most probably because like the NHG. v it was still the
equivalent of the consonantal u, or in any case was still bilabial.¹
In the later OHG. loan-words it appears as ƒ (cf. above); in those
of more recent times again it appears as w.
702. In cases where a word passing from one language into
another depends merely upon the powers of hearing of those who
adopt it, and where they are but imperfectly acquainted with the
foreign idiom, it is quite easy for wider-spreading corruptions of
the word to make their appearance, depending on the one hand on
the imperfect powers of hearing of those who adopt the word, and
on its imperfect retention by the memory. One result, in special,
of this, is that unwonted combinations of sounds are replaced by
more ordinary ones; abbreviations are adopted; and very fre-
quently popular etymology begins to operate.
703. The changes affecting foreign words upon their adoption
must not be confused with those that do not affect them till they
have been naturalised in the language of their adoption. As,
however, we have only received many words long after their
adoption, it is not always so easy to prevent such confusion. The
foreign words which have become naturalised of course are subject,
no less than the native words, to the laws of sound-change. The
fact of the participation or non-participation of a word in the
sound-change may, in default of tradition, reveal to us the relative
epoch of each process of borrowing. Thus, if in OHG. the Latin t
1 Cf. Franz, u.s.
ON MIXTURE IN LANGUAGE.
465
appears in some words as t and in others as z (cf. tempal, turri,
abbât, altari—ziagil, strâza scuzzila), and the Latin p appears in
some words as por b and in others as ph or f (cf. pîna, priestar-
phil, phlanza, phifa, pfeffar), there is no doubt that the words with
z and ph or ƒ represent an older stage of borrowing than those
with t and p. For the changes in question could never have set in
if the words had not been adopted already before the sound-
shifting, so as to share the fate of the genuine German words.¹
704. Not only so, but foreign words are exposed in the course
of their extension to the same assimilating tendencies as at their
first adoption. A word may in the first place be exactly or
approximately so adopted with its foreign sounds correctly ren-
dered by persons who are perfect masters of the foreign language.
In the next place, however, as it is transferred to persons un-
acquainted with the foreign language, it may be corrupted by the
substitution of another motory sensation, through imperfect audi-
tion and through popular etymology. In cases where any such
corruption comes into common use among the great mass of
people, it may also have a reflex action upon those who are well
acquainted with the correct sounds of the original word. They
must, in spite of their superior knowledge, adapt themselves to the
pronunciation which has become current, if they would not remain
unintelligible or appear affected. In other cases, on the other
hand, a combination of sounds approximating to the original
model maintains itself in the mouth of the educated, while side by
side with this one or more popular variants develop, e.g. corporal
-kaporal, sergeant-scharsant, gensd'armes-schandarre (it is thus
pronounced in Lower Germany), kastanie—kristanje, chirurgus—
gregorius, renovieren—rennefîren, etc.
705. A special kind of assimilation consists in the transference.
[ Cf. Kluge: Etymologisches Wörterbuch, s.v. 'Pfingsten,']
2 G
466
PRINCIPLES OF THE HISTORY OF LANGUAGE.
of the native system of accentuation to foreign words. Such
transference probably does not set in as a rule when the words are
in the first instance transferred,—not till the word has become
quite naturalised. In English it is easy to trace how the French
words adopted originally with a French accent have only gradually
passed over to the German method of accentuation [Chaucer has
lángage and langáge, fórtune and fortune, báttaile and battaille:
láboure and labour, see Meiklejohn's edition of Chaucer's Prologue
(1880); so Pope, gallant]. In German the same remark may be
made as to foreign proper names. In OHG. (and this is partly true
of MHG. as well) the accent fell thus, Adâm, Abêl, David, etc.
Appellatives, on the other hand, appear already in the earliest
OHG. records with the accent thrown back, and with the effects of
this process ; cf. e.g. fogat (vocatus), mettina (matutina), fenstar. It
is probable, however, that in these cases too, the throwing back of
the accent did not follow closely upon the adoption of the word.
706. The phonetic modifications described tend to estrange a
word more and more from its origin, until it may at last become
unrecognisable, even by a person well acquainted with the lan-
guage from which it comes. But such estrangement may also be
enhanced by changes in the language from which the word
is borrowed. Thus, for instance, the German pronunciation of
the words borrowed by that language from French depends to
a great extent upon a pronunciation no longer current in France;
cf. Paris, concert, officier, etc. Words which have passed into the
Romance languages from the German have become even more
violently transformed in their passage; cf., e.g., the French words
tape, tapon=zapfen; Italian toppo=zopf; French touaille South
German zwehle, Middle German quehle; Italian drudo = traut. In
the same way it is possible for the signification with which the
word is borrowed to change in the parent speech as well as in the
ON MIXTURE IN LANGUAGE.
467
language into which it has passed, and finally it may entirely
disappear in the parent language.
adoption of
word.
707. The same word may be borrowed several times at different Repeated
periods. It appears then in different phonetic combinations, of the same
which the more recent connects itself more nearly with the parent
language, while the older has experienced changes more or less
violent. Sometimes the meaning with which a word is adopted on
the second occasion of its borrowing differs from that prevailing
on the first occasion, and the result is that no connexion between
the two forms is perceived; cf. such forms as the German ordnen—
ordinieren, dihten - dictieren, predigen —prädicieren, OHG. zabal
(draft-board) by the side of tavala, both from tabula; the two
words prüfen and probieren may also be mentioned as not precisely
tallying in meaning. [There are many so-called doublets in French
and English; cf. for French, Bréal, u.s., and for English, Skeat,
Principles of English Etymology, ch. xx.] In cases where the
meaning is perfectly identical the older form easily disappears, as
in the case of altar (in MHG. already alter), or it may be that the
older form is confined to the ordinary and popular diction; cf.
ade—adieu, melodei (regularly developed from MHG. melodîe)—
melodie, newly adopted from the French; phantasei-phantasie,
känel (kännel, kändel, kener)—kanal, kämi—kamin, kappel-kapelle,
keste-kastanie. Plurality of forms is very frequently found as a
result of various borrowing processes in the names of persons: it
frequently happens in this case that their origin from the same
stock is no longer recognised, the older forms merely appearing
still as family names; cf. Andres-Andreas, Bartel-Bartholomäus,
Michel Michael, Velten - Valentin, Metz - Mattis - Matthias,
Marx-Markus, Zacher-Zacharias, Merten-Martin, etc.
708. Sometimes it happens that instead of an absolutely new
loan being made, the loan-word already long-since naturalised
468
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Competition
among
languages
in the con-
loan-words.
and phonetically modified receives a merely partial assimilation to
the word in the foreign language which is at its base; cf. MHG.
trache=NHG. drache (draco), MHG. tihten=NHG. dichten (dictare),
MHG. Krieche=NHG. Grieche (Graecus). The word Jude also seems
to depend upon a secondary attraction to the word Judaeus, as
Jüde would be the only form correctly developed according to
the laws of sound.
709. Where two nearly related languages co-operate to influence
a third, it may easily happen that the words corresponding to each
tribution of other are adopted from both; these will then agree in their
signification; and in their sound-form differ but little from each
other. We find this markedly the case in loan-words borrowed
from Latin and from French-thus the German has the two
parallel forms ideal and ideell, real and reell, at the present day
differentiated in their meaning, but at an earlier epoch identical;
Schiller actually uses material in the sense of materiell. Goethe
writes religios instead of religiös. In the mouth of a south
German referendär answers to a north German form referendar.
Instead of the words trinität, majestät, etc., we find in MHG. such
forms as trinitât, majestât; in the sixteenth and seventeenth
centuries, both forms may be proved to have existed; the a
can only proceed from the French.¹
710. In these cases, even the form which has arisen from
the French would be naturally referred by a Latin scholar to
the Latin directly. In other cases words are not, generally
speaking, taken directly from the parent language, but merely from
another in which they occur as loan-words. Thus, for instance,
Greek words came to us in the first instance from Latin; hence
they had the Latin accentuation and the termination -us instead of
-os. In the same way, Latin words which on their part again may
1 Cf. J. Grimm: Kl. Schr. i. 337, who, however, appreciates the matter differently.
ON MIXTURE IN LANGUAGE.
469
be borrowed from the Greek, have come into English and German,
thanks to the French language; cf. such words as musik, protestant,
religion, etc., and also such proper names as Horace, Ovid, and
Livy. In this case too a direct relationship presents itself to a
person conversant with the original language, and the result is that
even when he takes words directly from the original language, he
gives these a phonetic combination analogous to those which have
entered by another channel; so that he, for instance, alters the
Greek into the Latin accent; that he lets the Latin terminations
-us -um and others drop; he changes the terminations of the Latin
words in -żo into -ion. It should here too be mentioned that words
taken directly from the Latin have received the termination -ieren
which comes from the French: cf. negieren, spazieren, poculieren,
praedicieren, annectieren, regulieren, prästieren, präparieren, etc.
From an older form personifieren, found in Lessing, a word personi-
ficieren has been coined, with direct reference to the Latin.
combination
with a foreign
711. We have seen above (p. 161) that a derivative formed with Pleonastic
an unusual suffix may easily receive in addition the normal suffix of a native
for its special function. A special instance of this process is when suffix.
a native synonymous suffix is added to the foreign one; as for
instance, Sicilianer, Mantuaner, Kantianer; Italiener; Athenienser,
Waldenser; Genueser, Bologneser; sicilianisch, italienisch, genuesisch,
idealisch, kolossalisch (both of these words were common in the last
century), kollegialisch, musikalisch, physikalisch, etc.; princessin,
äbtissin (MHG. ebbetisse). The words in -ieren took their rise from
the fact that the German verbal terminations attached themselves
to the OF. infinitive form in -ier already in existence.
derivative
712. Words are always borrowed in their entirety; never Borrowing of
derivative and inflexional suffixes. If, however, a large number suffixes.
of words containing the same suffix is borrowed, these range
themselves into a group just as easily as native words with
470
PRINCIPLES OF THE HISTORY OF LANGUAGE.
the same suffix: and such group may become productive in
its turn. The suffix thus adopted may be attached, by means
of analogical new-creation, to a native root. This case is far
from rare in the case of derivative syllables. German has
formed after the analogy of abtei, etc. such words as bäckerei,
gerberei, druckerei, etc.; while popular spoken German has crea-
tions like takelage, kledage, bommelage, etc., after that of bagage
(cf. Andr. Volkset. 98); again, after the analogy of corrigieren,
we find hofieren, buchstabieren, sich erlustieren, MHG. wandelieren;
and in Hans Sachs gelidmasieret. We find again Romance
creations like the Italian falsardo with a Teutonic suffix: and
English words like oddity, murderous, eatable with a French
suffix.¹ In German several suffixes are found of foreign origin,
current, however, in the language of the educated classes alone;
these proceed to enter into combination, not merely with elements.
drawn from the same language, but also with elements drawn from
another foreign language, and at times with native linguistic
material; cf. the German suffix -ist in jurist, purist, romanist,
tourist, manierist, hornist, hoboist, Carlist, etc.; -ismus in atavismus,
purismus, fanatismus, sonambulismus, etc.; -ianer, in Hegelianer,
Kantianer, etc. These formations are found too to some extent in
French, and are indeed partly borrowed from that language. If
creations like purist and purismus be objected to as hybrids between
Latin and Greek, the answer is that they are in truth neither Latin
nor Greek, but German or, it may be, French creations.
713. It is rarer for inflexional terminations to be thus adopted. 2
Such a process implies a very intimate contact between two
languages. The French plural formation in s is fairly common in
North Germany, as in kerls, mädchens, fräuleins, ladens: it is
Examples of Slavonic suffixes in German dialects may
2 Cf. Schuchardt, p. 8.
1 Cf. Whitney, u.s., p. 17.
be found in Schuchardt, p. 86.
ON MIXTURE IN LANGUAGE.
471
pleonastic in jungens. It has also found its way into the written
language in the case of words originally indeclinable; as for
instance, a's, o's, neins, abers, vergissmeinnichts, stelldicheins; and
again in the case of foreign words which end in a full vowel, and
hence do not naturally range themselves under any other declen-
sion: such are papas, sophas, mottos, kolibris; albums as the plural
of album is not so common nor so admittedly correct. The French
plural formation has spread further in Dutch, cf. mans, zons, vaders,
broeders, waters, euvels, lakens, vroukens, vogeltjes; and so generally
with neuters in -er, -el, en, and the diminutives; the s in jongens,
gebentes (from gebente), bladers (a by-form of bladen and bladeren),
benders (a by-form of benderen from ben), etc. The English
genitival termination has made its way into the Indo-Portuguese;
we find in that dialect, for instance, hombre's casa. The adoption
of inflexional terminations on the largest scale has taken place in
the Romany or language of the gipsies. There are thus Spanish,
English [and Hungarian] gipsy dialects.
the inner
form of
714. A language suffers influence in its inner linguistic form Influence on
principally in the mouths of those who speak it as a foreign linguistic
tongue. By no means exclusively, however. As regards the language.
literary language, the influence of translations in this respect is
noteworthy.
715. In cases where a word taken from a foreign language only
partially tallies in meaning with a word of the native language,
the temptation is great to invest the former with the full significance
of the latter. This is of course one of the most usual faults in
exercises of translation. Mistakes like this may very easily be-
come common in bilingual areas.¹ Thus a South Slavonic author
writes habt ihr keine scheu und schande, because the word sramota
may mean schande (disgrace') as well as scham (shame'). The
1 Cf. Schuchardt, u.s., 95 sqq.
472
PRINCIPLES OF THE HISTORY OF LANGUAGE.
word schnur is used among the German Ruthenians in the sense
of bride, because nevesta in Slovenian signifies at once daughter-in-
law and bride. We often find the word damals used [like olim in
Latin] in Slavo-German, of the future; in the same way wo is used
as the equivalent for wohin, because in Slavonic the same word is
used for both.
716. It is an essentially different process, when in order to
express an idea which has hitherto lacked a name, a word is
coined on the model of a foreign language, or when a transference
of meaning is effected in the case of one already existing on the
same model. This process is specially common in scientific and
technical language where foreign material is directly adopted. We
may compare, for instance, the attempts to reproduce the Latin
grammatical terms by means of German ones. The Latin ones
are in their turn mere imitations of the Greek.
717. Groups of words, again, which as such have developed a
special meaning, are translated word for word. Thus the Austrians
say es steht nicht dafür for 'it is not worth the cost or the trouble,'
after the model of the Czechish nestojé za to. [Cf. Schuchardt,
Romanisches und Keltisches, p. 280, who cites the following idioms
as occurring in Alsace : Est-ce cela vous goûte? Il a frappé dix
heures. Il brûle chez M. Meyer. Ce qui est léger, vous l'apprendrez
facilement. Cher ami, ne prends pas pour mauvais. Pas si beau-
coup. Attendez, j'apporterai une citadine.]] In South-west
Germany it is common to hear, after the French model, such a
phrase as es macht gut wetter.
718. Finally, we have to consider the influence undergone
by syntax.2 As the Slavs can employ one form for all genders
and numbers of the relative, we find in Slavo-German the
word was correspondingly employed; cf. ein mann, was hat
1 See Schuchardt, 2.s. p. 96.
2 Cf. Ibid., u.s. p. 99 sqq.
ON MIXTURE IN LANGUAGE.
473
geheissen Jacob; der knecht, was ich mit ihm gefahren bin; also
ich bin nicht in der stadt gewesen, was (='so long as ') er weg ist.
In the last century it was almost the universal custom to write,
after the French model, ich lasse ihm das nicht fühlen, etc. In
Lithuanian the German construction was für ein mann is literally
copied.
dialects.
719. Mixture in dialects, when occurring in one continuous linguistic Mixture of
area, distinctly emancipates itself from the normal assimilating effect
of intercourse when it occurs as between dialects whose territories
are not geographically coterminous. On the other hand, no
regular border can be drawn when the territories are geographically
neighbours, and are in constant communication with each other.
In drawing a distinction we can only look to this: whether a
sharply-marked contrast exists between the dialects in question,
or whether the differences are small, and are already bridged over
by transitional gradations.
720. Generally speaking, the same holds good here as in the
case of the mixture of different languages. Here, too, word-
borrowing is the process which most easily and frequently makes
its appearance. On the other hand, phonetic material is not
easily changed. Here, again, we meet with the substitution of
the nearest native sounds for the foreign ones. Hence a word
adopted from a kindred dialect appears commonly in the same
form as it would have obtained had it maintained itself from the
time of the ancient unity of language. This will, as a rule, be
the condition of things, though there will be unimportant differ-
ences in the phonetic developments. The case is of course
different when two dialects have, in the course of their develop-
ment, parted wider asunder, so that what corresponds etymo-
logically has lost all close resemblance on the phonetic side. Thus
e.g. the ch in the words sacht and nichte has not, in the adoption
474
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Borrowing
from an older
period.
of these words into HG., been replaced by the ft which, etymo-
logically speaking, corresponds to it.
721. It very commonly happens that a mixture occurs in a
literary terrain before the settlement of a general language in
this way that a record is transmuted from the dialect in which
it was originally drawn up into another. This is possible in the
case of written as of oral tradition. The transmutation remains,
commonly speaking, an imperfect one, especially when the rhythm
is opposed to it. This kind of mixture is wholly distinct from
that which takes place in the organism of liuguistic ideas in single
individuals.
722. The process of borrowing from an older linguistic grade
of language can of course result only by the agency of writing.
The phonetic material can accordingly never be influenced there-
by. This kind of borrowing takes place as a rule consciously
in the course of literary production. A distinction must be
drawn here. Either in the process certain actual or imaginary
advantages of the older language are intended to be stirred into
new life; or the old-fashioned characteristics of the language
are intended to serve as a characteristic of the time into which
the description transports us. In the latter case the tendency
is to go much further than in the former.
<
CHAPTER XXIII.
IN
THE COMMON-LANGUAGE.
with The common.
gen- has no real
language
existence,
task but is only
an ideal
N all modern civilised countries we find side by side
manifold dialectic ramifications a widely-diffused and
erally recognised common-language. The conditions of our
obliged us to defer the investigation of the nature and essence norm
of this language till the end.
first instance, the given circumstances which present themselves
to our immediate observation.
We must again observe, in the
724. We have up to this point always started with the view
of grasping the actual processes of the life of language. From
the very beginning we have insisted that in considering these
processes we must not regard what descriptive grammar calls a
language, the collection of the current usages, as an abstraction
which has no real existence. The common-language is naturally
an abstraction, and one of the first order. It is not a complex
of real facts, real forces, but merely an ideal norm prescribing rules
for speech. It bears the same kind of relationship to actual
linguistic activity as that of a particular code of laws to the
entire legal activity in the district to which the laws apply, or
of a confession of faith, or a dogmatic text-book, to the entirety
of religious views and feelings.
725. Regarded as such a norm, the common-language, like a
code or a dogma, is in itself unchangeable. Changeability would run
476
PRINCIPLES OF THE HISTORY OF LANGUAGE.
directly counter to its very essence. Wherever a change is pro-
posed it can only be imposed by a force outside of the norm, by
means of which a part is purposely removed from it, and replaced'
by something else. The occasions of such changes are analogous
in the different civilisations. A codex, however carefully executed,
will still always leave a certain freedom of movement; and it will
always happen that in practice a series of unforeseen cases present
themselves. The codex may, however, also contain difficulties,
and in places permit of several interpretations. Besides this, we
must take into account possible misunderstandings and lack of
knowledge on the part of those who should base their actions
upon it. Finally, it may be that the codex contains much that
is inapplicable, in some cases from its very inception, in some
cases owing to the changes in moral or economical conditions
which have set in since it was fixed. This inapplicability may
cause the legal instinct of the entire community, or of its most
influential circles, to set itself against the strict execution of the
letter of the law. The combined action of such circumstances
as these finally leads to a legislative alteration of the code. This
is precisely the condition of the common-language. It is nothing
but a rigid rule which would bring linguistic movement to a
standstill, if it were always strictly carried out, and is only com-
patible with change in so far as its authority is disregarded.
726. With all this, however, there is the difference that the
common-language is not actually codified. Generally speaking,
it remains true that usage determines the norm. This, however,
cannot be the usage of the entire community. For this very
community is very far from being homogeneous. We find even
in those areas in which the common-language has taken the
strongest hold, that the single individuals differ considerably
from one another, and this even when we regard them only so
THE COMMON-LANGUAGE.
477
far as they are expressly bent on speaking the written language.
And even if these divergencies ceased, it would be certain that
under the common conditions of the development of language
new ones would always arise. Both for the creation of a unity,
and for the maintenance of one already created, something is
needed which shall be independent of the linguistic activity of
the whole community, and stand in objective isolation from it.
This something is in every case to be found in the usage of a depending
definite narrow circle.
on the usage
of a narrow
circle.
written
727. We find, however, as far as our observation goes, that the
norm is settled in two different ways, viz.—(1) by spoken lan- Spoken and
guage; (2) by written authorities. If a norm, which can in any language.
degree be called settled, is to result from the first, the individuals
who are regarded as authorities must stand with respect to each
other in oral communication with each other, either continuous
or, at least, interrupted only by short pauses, the mutual inter-
course of these individuals being of the utmost frequency, and in
the most various combinations. As a rule we find the language of
some special district or town looked on as the model. But con-
sidering that in every case in which a real common-language has
developed, in however narrow an area, appreciable differences.
exist between the different classes of the population, the capacity
to serve as model must be restricted to the language of the edu-
cated classes of the district in question. But the model language
may even declare its independence of this; as is, for instance, the
case in Germany. It is pure prejudice when any particular dis-
trict is picked out in which it is alleged that the 'purest German'
is spoken. The language which serves as the German model German
is rather that employed in the theatre in serious drama, with the stage.
which the prevailing pronunciation of educated persons in no place
exactly agrees. The representatives of this 'stage-language' form
language of
478
PRINCIPLES OF THE HISTORY OF LANGUAGE.
a comparatively small, but geographically diffused, circle. This
geographical dispersion, however, contravenes only seemingly our
assertion that direct oral communication is an absolute necessity.
for the existence of a model language. In fact the unity, the con-
sistency, which characterises the language of the stage could not
be attained or maintained, were it not for the perpetual process of
shifting of the individuals attached to the several stages of Ger-
many, even the furthest removed from each other, and for the
existence, present or past, of certain central points which serve
again as examples to the others. There is, further, the fact that in
this case a shorter period of direct contact may produce the same
effect as a longer one in other cases; since, in this case, a regular
process of schooling is employed, a process which is actually based
on a careful attention to the physiology of sound. The reasons
why the stage-language, of all others, necessarily maintained itself
as a special unity, departing from all the local dialects, are not far
to seek. Nowhere else has there existed a community of persons
from the most different districts so closely bound together, and
compelled to co-operate in language. Nowhere else has a lin-
guistic community had such inducements to pay diligent heed to
its own pronunciation and to that of others, and to take deliberate
pains with it. Attention had to be paid, on the one hand, to
secure that the generality of a large audience should understand
the language employed, while, on the other hand, æsthetical calcu-
lations could not be left out of sight. Both of these considerations
alike forbid the admissibility of dialectic variations from the
standard language, even within the narrow limits in which they
still continued to distinguish the different local circles of the edu-
cated. It is obvious that a pronunciation consistently maintained,
to which the public gradually grows accustomed, is a main factor
in promoting intelligibility. And more, any inconsistency in this
THE COMMON-LANGUAGE.
479
respect does violence to the sense of beauty as well, unless indeed it
intended to aid characterisation. Thus, where no characterisa-
tion is intended, dialect must be avoided, precisely because its use
suggests characterisation. Now when different dialectic shades of
difference were struggling for the predominance, before any agree-
ment was settled on, it might well happen that, even though per-
haps on the whole, one predominated, still in this or that point it
yielded to another. The endeavour to attain the greatest possible
clearness would be the decisive factor in this decision. And this
endeavour must in every case have necessarily conduced to a
departure from the colloquial language. Such sounds and com-
binations of sounds as are only employed in this, when special
clearness is aimed at, were raised in the stage-language to the
regular and normal. Especially was it the case that the forms
which arose under the influence of the sentence, or even under that
of word-composition, and were affected by assimilation or by
weakness, as a result of the light stress they received, were, as far
as possible again rejected, and replaced by the sound-forms usual to
isolated words. There was in many ways a tendency to fall back
upon the orthography in cases where the pronunciation had already
deviated from the ordinary form. These peculiarities, caused by
the need felt to make things plain to a large audience, prevent, and
must prevent, the language of the stage from passing as an absolute
model for the colloquial language. In ordinary language such
laborious endeavour after clearness would appear mere affectation.
character of
728. The language of the stage then creates a more rigid norm Unyielding
for the sound-conditions than that created by the colloquial lan- its norm.
guage of any definite district. But its guiding influence is limited
to the phonetic side. Moreover its language is doled out to it by
the poets, and it cannot exert so much influence in other directions
as the colloquial language.
480
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Effect of
reduction of
language
729. The degree of agreement existing in the language of the
circle which passes for an authority can, of course, never be abso-
lute. It is difficult for it, in the case of any colloquial language, to
exceed the degree which exists in the idiom of a narrow district
which has grown up by natural processes. In an artificial stage-
language it is, no doubt, possible to effect a little more than this.
And just as the normal language is liable to variations, so is the
artificial language liable to gradual change like any other idiom.
In fact the conditions of its life are identical with those of the
other. Even though the norm may occupy an independent posi-
tion as regards a wider circle, it cannot do so to the same extent
in relation to the narrower circle which sets the fashion; but must
rather, from natural causes, be gradually modified by the speech of
this circle. This would be the case, even if this narrower circle
were able to maintain itself quite independent of the influences of
the wider one. But it is impossible to suppose that in the course
of an uninterrupted linguistic communication it should always
have merely been the contributor, and never the receiver. And in
this way the common-language also is determined by the gener-
ality of those who speak it; only the part which individual
speakers take in the determination is very different.
730. The second norm of the common-language, which is
the common- created by its reduction to writing, presents many considerable
to writing. advantages. Not till it has been fixed by writing does the norm
become independent of the individuals who speak it, and capable
of being transmitted unchanged to the following generations. It
may further be diffused without any direct intercourse. Finally,
in so far as it has merely to exert its influence again upon written
language, it has a much easier task, since, in order to accommo-
date ourselves to it, we do not need to habituate our motory sen-
sations anew as is the case when we have to master a foreign
THE COMMON-LANGUAGE.
481
pronunciation. On the other hand, it has the disadvantage of
leaving a very wide scope for variations in the pronunciation, as is
clear from our instances given in the previous chapter; and there-
fore it is not a good example to take as a model for this purpose.
731. With a view to the settlement of written language in
the strict sense of the word, it is at all events possible to propose
the usage of definite writers, definite grammars, and dictionaries
as sole authorities, and to adhere consistently to their usage.
This happens, for instance, when modern Latinists aim at repro-
ducing the style of Cicero. But even from this example we can
judge how all but impossible it is, even when we have a perfectly
definite model before our eyes, to produce its absolute equivalent.
It is of the essence of the process that we keep ourselves in un-
interrupted familiarity with our model, and scrupulously strive to
protect ourselves against any extraneous influences. The most
successful in this pursuit attains his success only by a self-
imposed limitation in the communication of his thoughts, by a
renunciation of all individuality, and at the same time of ex-
actness and clearness of expression. Rich as the writer's store of
thought may be, still even a cotemporary belonging to the same
stage of civilisation will fail to find in him the corresponding
power of representing all that he himself wishes to say, and this
will be much truer as applied to one of a later epoch still, when
the character of the civilisation has changed.
732. A written language to serve any practical purpose must
change with the times, just like a living dialect. Even where
it is based, in the first instance, on the usage of a single author,
or of a particular school of authors, still it must not attach itself
unconditionally for all time to its model, and, above all, must
not tend to exclude improvements in cases where the model is
inadequate. The individual must cease to keep the model before
2 H
482
PRINCIPLES OF THE HISTORY OF LANGUAGE.
his eyes in whatever he writes; he must, as in dialect, employ
his linguistic materials unconsciously with a firm confidence in
his own instinct; nay, this instinct should itself endue him with
a certain creative energy employed in the formation of language,
and enable him to influence others by his very creations. The
linguistic usage of the present must be his aim and model, if not
exclusively, at least in conjunction with ancient models. This
is the case with the Latin of the Middle Ages. The Humanists,
by cutting short the living development of the Latin language,
and re-establishing the ancient models as solely authoritative,
gave in fact the death-blow-quite against their intention-to
Latin as a world-literature, and rendered it henceforth completely
unfit to satisfy the general requirements of scientific or mercantile
intercourse.
732. As a literary language emancipates itself from its original
models, it is doubtless inevitable that it should lose in regularity,
and that numerous variations should occur in the usage of indi-
viduals. But it does not follow that it need split up into different
dialects geographically separated, as must inevitably be the case
under similar circumstances with spoken language. One, and
indeed the most important, source of dialectic differentiation is
completely wanting to the written language, viz., sound-change.
Inflexion, word-formation, word-signification, and syntax un-
doubtedly remain exposed to change, and hence to differentia-
tion; though even this is less the case than in the spoken dialect.
A principal disposing cause to change in this domain is, as we
have seen, the divergent character of the grouping which de-
pends upon the phonetic form of words from that which depends
upon their signification. It need hardly be said that even the
written language, as originally fixed, is not free from this diver-
gence; but it differs from the colloquial idiom, in which fresh
THE COMMON-LANGUAGE.
483
discrepancies are constantly called into play by sound-change;
nor are the different linguistic areas rendered, through dissimilar
development of their sounds, unequally adapted to analogical
creation. Thus it is that much less opportunity is given for
changes in the laws for the formation of inflexions and word-
composition. Nor is it simply and solely that fewer changes set
in; but those which do set in may easily spread over the whole
area, as long as the literary connexion is not broken. In cases
where they have not sufficient strength, they will, as a rule,
have to yield to overwhelming influences even in the narrow
area in which they may have established themselves. Least of
all will the unity of language be imperilled if the ancient models
continue to maintain a certain authority side by side with those
of more recent date, if they are much perused, and if rules are
abstracted from them which are generally recognised as authorita-
tive. The maintenance of agreement is most easily reconciled with
adaptation to changed conditions of civilisation, by keeping as
nearly as possible to the ancient models in syntax, and more
still in accidence, while preserving, on the other hand, a certain.
freedom in the creation of new words, and in the application of
new significations to old ones. This is, generally speaking, the
case with the more cultivated Middle-Latin authors.
733. It is in middle and in modern Latin that we are best
able to study the essence of a common-language which is a
written language only.¹ The national general languages, on the
other hand, are at once literary and colloquial languages. Hence
in these we find a norm proper to the written language, and
1 It is true that Middle-Latin was not a language living and developing exclusively
in writing. It was also employed in oral intercourse. This can, however, have had but
little influence upon its development, as it was always learned on the basis of written
examples. On the other hand, another factor totally unconnected with the influence of
writing was certainly of great importance, especially for the arrangement of the syntax,
viz., the mother-tongue of the writers.
484
PRINCIPLES OF THE HISTORY OF LANGUAGE.
another proper to oral intercourse existing side by side. It
seems obvious that both must be brought into agreement, and
continuously maintained in it. But, as we have seen in the
previous chapter, no such agreement is, strictly speaking, possible
on the phonetic side; and the emancipation of the written from
the colloquial language may go so far that their reciprocal
influence nearly ceases. And precisely the introduction of a
fixed norm favours this emancipation. It appears from this
how necessary it is that there should be a special norm for the
spoken language, seeing that on the basis of the mere written
norm it would hardly be possible to attain even an approximate
agreement in the sounds, though it would doubtless be easier
to attain this with an orthography like that of German than with
one like that of English.
734. We must further take the fact into account that between
written and colloquial language there exists a permanent dis-
tinction of style, which no one seeks to set aside. In conse-
quence of this, methods of construction, words, and combinations
of words, maintain themselves in the former, which in the latter
have become wholly obsolete; and on the other hand, much
that is new forces its way into the latter, which is rejected by the
former.
735. There is thus no absolute agreement of both depart-
ments in what is regarded in them as normal. They are, how-
ever, even apart from the two points mentioned, always liable
to the risk of separating in different directions. The authoritative
persons are in the two cases only partially identical, and the
degree of influence exerted by the individual is in the case of
the one not the same as in the other. To this must be added
in the written language the ever-renewed influence of the older
writers, while in the language of intercourse none but the living
THE COMMON-LANGUAGE.
485
generation operates directly. Thus, to avoid a wide gulf between
the two, a kind of compromise must always be arranged between
them, each yielding something to the other.
736. We have seen above (p. 33) that we have to seek that
which strictly characterises a dialect and differentiates it from
others in its sound-system. The same holds true of the common-
language in contrast to the individual dialects. We must not,
therefore, co-ordinate a technical language, or a poetical or arti-
ficial style, with a common-language any more than with a dialect.
in the
individuals
normal.
737. In every domain in which a normal common-language Gradations
exists, the languages of individuals are found to present very languages of
manifold gradations. Between those which approach the norm speaking a
most closely and those which represent the different dialects as
least affected by the norm, there are numerous connecting stages.
In these, most individuals employ two, sometimes even more,
languages, one of which approaches more nearly the norm, the
other the dialect. The latter is the cradle language which comes
naturally to the individual; the former is only acquired in later
years by artificial efforts. No doubt it happens now and then
that an individual learns at the very outset two side by side,
and special circumstances may even cause many, even in later
life, to learn and employ a language more widely differing
from the norm. The contrast between the two languages may
be of very different magnitudes. It may be so small that
they are distinguished in ordinary life merely as a somewhat
more careful and a somewhat more careless pronunciation; in
this case, again, it is easy for gradations again to set in between
the two. A wide gulf may, however, exist between the two.
The magnitude of the contrast depends of course not less upon
the degree in which the natural language differs from the norm
than upon that in which the artificial language approaches it. In
486
PRINCIPLES OF THE HISTORY OF LANGUAGE.
both respects there are great variations. If in ordinary life we
are contented to denote the artificial language simply as a literary
language, we leave out of calculation a quantity of not un-
important local and individual differences: if we denote the
natural language simply by the name 'dialect,' we overlook im-
portant variations occurring within the same narrow area. There
are of course cases of individuals who employ only one language;
on the one hand, such as in their natural language approach
or think that they approach the norm so nearly that it seems
to them no longer necessary to approach it any nearer by any
artificial exertions; on the other hand, such as are still unaffected
by the necessities which have conduced to the creation and em-
ployment of the common-language.
738. The further the natural language of an individual departs
from the norm, the more will he tend to feel the artificial language
by its side as something foreign; on the other hand, commonly
speaking, the greater will be the diligence which he expends
on mastering the artificial language, and the more complete his
final approach to the norm, especially in all those points which
can be fixed by writing. In Lower Germany a more correct
literary German is spoken than in Middle or Upper Germany.
In the same way the so-called gut deutsch of Switzerland is much
more correct than that, for instance, of the neighbouring territory
of Baden or of Würtemberg, because in these districts the town
dialects have approached the norm much more closely than in
Switzerland.
739. If on the same area many gradations exist side by side,
these must of course exercise a perpetual influence upon each
other. This must hold especially true of both the gradations
which are found together side by side in the same individual.
All gradations on the same territory must have certain peculi-
THE COMMON-LANGUAGE.
487
arities in common with each other. The gradations in the dif-
ferent territories which stand nearest to the norm must still always
stand in a relation to one another somewhat analogous to that
between those most remote from the norm.
740. The norm of the literary language is in all cases more
definite and freer from variations than that of colloquial lan-
guage; and the written language actually excels in this re-
spect in actual practice even those forms of colloquial language
which approach nearest to the norm. This is a statement
whose general truth may be found confirmed by experience
wherever we may look, and which, moreover, results necessarily
from the nature of the case. For in the first place, as we have
seen, all the finer differences of pronunciation must of themselves
disappear in writing; and in the second place, it is easier for an
individual to adopt a particular method of writing than a pro-
nunciation differing from his usual one. Thus a very little un-
prejudiced consideration will enable us to detect the fallacy of
certain hypotheses which assume for an older period a greater
degree of unity in the spoken than in the written language.
efforts pro-
approxima-
norm.
741. In the relation of the single individual languages to the Artificial
norm there occurs a continuous succession of displacements. While ducing an
these, on the one hand, are unable to emancipate themselves tion to the
from the general conditions of the natural development of language
and are hence impelled towards an ever-increasing differentiation,
and therefore to an ever-increasing distance from the norm; on
the other hand, artificial efforts succeed in producing an ever-
increasing approximation to the norm. It is of importance to
master the fact that both tendencies are in operation side by
side, and that it is not the case that where the latter begins to
operate the effectiveness of the former is checked. We are in
some degree able to directly observe the graduated approximation
488
PRINCIPLES OF THE HISTORY OF LANGUAGE.
to the norm. But besides this, we find all the stages of develop-
ment through which the single individuals gradually pass con-
temporaneously side by side. Let us now attempt to get a clear
appreciation of the single processes, by means of which such
approximation is effected.
742. In the first place, an individual learns, in addition to
the natural language alone hitherto employed, an artificial one
standing nearer to the norm. This is effected, in modern educated
countries, in the first instance, mainly by school-instruction; and
the written language in its strict sense is thus learnt contempo-
raneously with a colloquial language approximating to the written
language. It is possible, however, to learn an artificial language
by joining a fresh community which employs a language more
nearly resembling the norm than that of the community in which
one has hitherto lived, or at least by coming into nearer contact
with such a community than was the case when one first learnt
to speak. In the latter case it is not even necessary to learn
reading and writing at all. The relation of the individual to the
new language is of course for a long time a passive one before
it becomes an active one-that is, he learns, in the first place,
to understand the language and accustom himself to it before
speaking it. The individual has often a similar more or less
intimate passive relation to a great many dialects and graduations
of the colloquial language without ever passing from this point
into any active relation with them. To effect this a special
impulse, an unusually energetic influence, is necessary. The acqui-
sition of the artificial language is at first always incomplete; a
gradual advance may be made to an ever-increasing complete-
ness; but many never succeed in employing it confidently and
correctly. Under all circumstances a man's natural language,
earlier acquired, determines the specific character of his artificial
THE COMMON-LANGUAGE.
489
language. Even in cases where the latter takes the widest
possible departure from the former, it is still not learnt as an
absolutely foreign language, but always as having some relation
with the natural language, and this relation co-operates and aids.
in its application. We guide ourselves, in the first instance, as
indeed generally in the employment of any foreign language or
dialect, as much as possible by the motory sensations in which
we are versed. The finer phonetic divergences of the model-
language, which we are striving to imitate, remain disregarded.
Thus it may happen that even when the model language in
question stands as near as possible to the norm of the common-
language, still in the course of the endeavour to form language
on its model, small touches, reminding us of the original dialect,
make themselves apparent. We have, moreover, to consider that
the individual, as a rule, learns his artificial language from the
companions of his home, whose language is already built up
on the foundation of the same dialect. Further, as far as the
artificial language is learnt by reading, it is matter of course
that a process of inserting kindred sounds from one's own dialect
must take place.
743. But the stock of words and their meaning, the inflexions
and the syntax of the artificial language are not composed merely
after the models, but after the contents of the native natural
language. The stock of words especially drawn from the model
language is increased, if it is inadequate or does not come with
sufficient readiness, from the natural language: words are employed
which have never been heard in the model language or which the
person employing them, even assuming him to have heard them,
would be unable to reproduce if they did not likewise occur in this
model language. We proceed in this case, commonly speaking,
with a certain straightforward confidence, because indeed a great
490
PRINCIPLES OF THE HISTORY OF LANGUAGE.
portion, or indeed the greater portion of the words occurring in
the natural language, occurs in the model language as well, and
because we in many cases supply, and rightly supply, in this way,
the points wherein our knowledge of the latter falls short. It must,
however, in this process naturally occur that words are taken over
into the artificial language which are unknown to the model
language, or are known to it only with some discrepancy of
meaning. Where the same word occurs in the model language
and in the natural language, phonetic differences of form often
appear. If these differences appear regularly in any large series
of words, it follows that parallel series will form themselves in the
mind of the person who is master of both languages at once (e.g.
NG. water—HG. wasser; eten—essen; laten—lassen, etc.). There
grows up in his case a feeling, however dim and undefined, of the
regular relationship borne by the sounds of one language to those of
the other. This feeling may actually so operate as to transfer
correctly words with which it is familiar in its natural language
alone into the corresponding sounds of the artificial language.
Psychologically speaking, the process is similar to what we have
described as creation by analogy. In its carrying out, it is possible
for mistakes to arise owing to an incorrect generalisation of the
value attaching to a proportion: as a German boy brought up to
speak a Low German dialect has been heard in speaking High
German to say seller for teller. Such faults as these, however, are
generally speaking merely individual and of transitory effect, as a
counteracting control soon sets in. On the other hand, however,
the parallel series do not always show themselves operative, and
words are transferred into the artificial language in their dialectic
form, differing from the phonetic arrangements of the model lan-
guage. For the rest, it is in other respects as with the phonetics:
as a rule the colloquial language which in the first place serves
THE COMMON-LANGUAGE.
491
as each individual's model takes its shape through a co-operation
of the strict normal language with the home dialect.
744. In the second place, the artificial language affects the
natural language by borrowing from it words, and occasionally
inflexional forms and methods of construction. The words are,
of course, such as have reference to groups of ideas to express
which the artificial language would naturally be preferred. In the
case of any converse process of borrowing, these words are either
transformed into the phonetic shape proper to the natural language,
or are maintained in the phonetic form of the artificial language.
There is no single German dialect which has kept itself com-
pletely free from such a contagion as that described, although the
degree in which it has maintained its purity differs widely.
745. In the third place, in the case of persons who speak an
artificial and a natural language side by side, the use of the former
spreads at the expense of the latter. Originally the artificial
language is employed in cases only where a real need occurs for
its employment, i.e. in the case of intercourse with foreigners who
belong to an essentially different dialectic area. This comes about
more by means of written than by oral agencies: an artificial
written language is more a necessity for the purpose than an
artificial language of intercourse. In the case of communication
between persons from the same home, the artificial language does
not come to be employed until a foreign language at the same
time claims attention. As soon as it has taken firm root in its
application to literature and to official documents, it has a tendency
to spread generally to all written records, even those of a private
character, which are not intended for a foreign dialectic area.
This is the natural consequence of learning to read and write from
literary records,—a result of which is that it is easier to accom-
modate ourselves to the orthography prevalent therein, than to
492
PRINCIPLES OF THE HISTORY OF LANGUAGE.
master a system of writing applicable to one's own dialect as well,
or to invent one ourselves.
746. Further, the artificial language is customary for public
declamations delivered from written documents, for preaching,
instruction, etc. Not till it has found a wider application in all
the forms of intercourse mentioned, does it come so naturally to
any portion of the people that they begin to use it for private
intercourse in their homes as well, in fact that it becomes the
common-language of the cultured classes. The class who first
habitually employ it is naturally composed of those who first come
under the influence of literature, schools, etc. It is of course not
till this stage of development is attained that the employment of
a dialect in ordinary intercourse is able to pass as a sign of
inferior education: not till then does the dialect yield in pres-
tige to the artificial language. In Switzerland things have not
universally reached this stage. In the most highly educated circles
of Bâle, Bern, and Zürich, persons converse, unless indeed any
regard has to be paid to the needs of foreigners, in the natural
dialect of the speaker's youth, and the use of Swiss German, even
in political assemblies, gives offence to no one. The same or
nearly the same state of things was seen in Holstein, Hamburg,
Mecklenburg, and other Low German districts some twenty or
thirty years ago.
In the entire district of South and Middle
Germany a noteworthy variation from the strict normal language
is still tolerated, at least in the language of intercourse. The mere
consideration of the circumstances at present existing may serve
to teach how preposterous is the view that when two languages,
the artificial and the natural, co-exist, the latter need necessarily
sink in estimation as compared with the former, and how mis-
taken it is to set down as the first motive for the acquisition and
creation of an artificial language not the need felt for it, but the
THE COMMON-LANGUAGE.
493
desire to contrast with the great mass of the people by more
refinement of culture. To assume anything of the sort is to listen
to the prejudices of an unscientific pedantry, heedless of the facts
of historical development. The employment of the language in
daily intercourse may spread in very various and very finely-
marked gradations. In the first place they are employed with the
natural language interchangeably. In this stage a difference is
made according to the degree of acquaintance with the natural
language possessed by the person who is addressed, and the de-
gree in which he employs it. Finally, it may be that persons
arrive at a stage when they no longer employ the natural lan-
guage at all. There occur at the present day cases enough in
which it is possible to follow this entire course of development
stage by stage in a single individual. Persons never arrive at
the exclusive use of the artificial language without the previous
existence of a period of longer or briefer duration, in which
bilingualism was the rule.
747. If a number of persons has succeeded in employing the
artificial language, either exclusively or to a prevailing extent,
then the portion of the younger generation which chiefly falls
under their influence learns from the very beginning as its natural
language what was to them still an artificial language. The fact
that the elder generation arrived at this language by artificial
means is afterwards a matter of indifference as far as concerns its
existence, and its survival in the younger generation. This latter
stands towards that language much as the elder generation or
other strata of the people stand to their dialect uninfluenced by
the norm of the general language. We must beware of simply
confusing the contrast between artificial and natural language with
that between common-language and dialect. We must always
have a clear idea of whether we wish to judge the different indi-
494
PRINCIPLES OF THE HISTORY OF LANGUAGE.
vidual languages, after the objective shape they have taken, with
respect to their greater or less distance from the norm, or accord-
ing to the subjective relation of the speaker to them. Of two
languages which we hear from two different persons, A may stand
nearer to the norm than B, and yet A may be a natural, B an
artificial, language.
748. If in a single linguistic area one portion keeps to the
original dialect, while another employs a language artificially
introduced, even for the purposes of its daily life, there are, of
course, a number of persons who, from their earliest youth, are to
some extent equally influenced by both groups, and thus different
mixtures cannot fail to rise. Every mixture, however, favours the
rise of new mixtures. And thus a large quantity of manifold
gradations necessarily arises in the natural language as well.
In High and Middle Germany we can pass almost everywhere
from the dialectic conformations nearest approximating to the
norm to those which depart from it most widely; and this quite
gradually, and without a broad division occurring in any place.
On the other hand, in Switzerland, where the artificial language
has not yet penetrated into daily intercourse, a graduation is seen,
no doubt, between the dialects according as they are more or
less strongly affected by the written language; but between the
written language and the dialect, which is most strongly influenced
thereby, a contrast bridged over by no gradations exists.
749. If any one from his earliest years have learnt a language
standing nearer to the norm, he naturally has no such pressing
necessity to learn an artificial language in addition, as if he had
learnt the pure dialect of his home. It thus frequently happens
that a single language suffices for his oral intercourse. Circum-
stances may, however, compel him to aim at attaining a greater
proximity still to the norm, and then he becomes again bilingual,
THE COMMON-LANGUAGE.
495
and once again his artificial language may pass into the natural
one of a following generation, and this process may frequently
repeat itself.
norm for
language.
750. We have hitherto attempted to realise how circumstances Rise of a
settle themselves, on the assumption that a generally recog- the common-
nised norm for the common-language is in existence. We have
yet to consider the way in which such a norm is able to come
into existence at all. It must be pretty commonly admitted that
no such norm can have originally been found in the areas where it
exists at present; that there must have been a preceding period in
which merely pure dialects, with equal privileges, stood side by
side. But it seems difficult to many people to picture to them-
selves a language employed for literary purposes without a norm,
and the inclination seems very prevalent to date its origin as far
back as possible. I can merely see in this the after effects of
ancient prejudices, which regard the written language as the only
one with a right to exist, and the dialect as a perverted form of
the same.
The reason why any doubt at all is possible on the
subject is that olden times have bequeathed us only written.
records, and not spoken language. In consequence, there is much
room for conjecture as to the character of the latter.
The ex-
periences hitherto gathered as to the conditions of the life of
language are the sole possible test of the justness, or other-
wise, of these conjectures. All that fails under this test must
be content to submit and be convinced.
towards the
a common-
751. Of the impulses which operate towards the creation of a Impulses
common-language, the necessity felt for such a language must creation of
first come into consideration; as may be seen from our previous language.
investigations. Such necessity does not arise until the dialectic
differentiation has gone so far that all the members of a linguistic
community cannot conveniently understand each other: and even
496
PRINCIPLES OF THE HISTORY OF LANGUAGE.
Intercourse
between
areas lying
separate
from each
other.
in this case the necessity will only be felt for the reciprocal com-
munication of those whose native places lie far from each other, as
no overwhelmingly strong contrasts develop among near neigh-
bours. Nothing of more questionable veracity can be advanced
than the assumption that a common-language has formed, in the
first instance, within a narrow area which has as yet only trifling
dialectic differences to show, and that this has spread thence over
the further lying areas. It is far more natural to suppose, and
experience confirms our supposition, that a language becomes a
common-language and is taken as a model in areas whose dialect
is pretty far removed from itself, while smaller variations pass
entirely unnoticed. Indeed the whole character of a common-lan-
guage may be strengthened by the fact of its transference to an
unquestionably foreign linguistic area, as we may observe in the
case of the Greek kowý, and in that of the Latin language.
752. Assuming accordingly the existence of a pressing need,
the intercourse between the areas far separated from each other
must be already developed to a certain degree of intensity; active
relations, commercial, political, or literary, must already exist be-
tween them. To some extent too, the circumstances attending
the intensity of the wider communication must regulate the size of
the area over which the common-language spreads its domination.
The boundaries of the area do not by any means necessarily corre-
spond with those which we should draw as most useful if we merely
wished to regard the relation of the dialects to each other. If the
dialectic differences are approximately equal even in two different
linguistic areas, it may still happen that on one merely a single
common-language develops, on another two, three, or, it may be,
more. For instance, there is no doubt that greater differences
exist between Upper and Lower German dialects than between
Polish and Czechisch, or Servian and Bulgarian, nay, even than
THE COMMON-LANGUAGE.
497
between Polish and Servian. It is possible for two areas with very
nearly related dialects to take widely different departures as
regards the common-languages which establish themselves in them,
while two others with dialects widely contrasting may assume the
same common-language.
common-
be created
for it must
753. How much depends on the necessity, the following con- For a
sideration may serve to show. It is difficult, if not impossible, language to
after a common-language has to a certain extent established itself the necessity
in an area of any large size, to create a special common-language be felt.
for a portion of such area. It is too late to hope to create a Low
German or a Provençal common-language. Even the efforts which
have been made to create a special Norwegian common-language
have proved useless before the existing predominance of the
Danish. Conversely, it is not easy to enforce the predominance of
a common-language over any large area, if the single parts of this
area have already their special common-language, which itself
sufficiently responds to the calls of the greatest necessity. We can
see this in the failure of the efforts of the Panslavists. An entirely
foreign language too, if it has once become the accredited language
of literary and official communication, tends to hinder the for-
mation of a national common-language. Thus the efforts to found
a Flemish literary language have been crowned with but little
success; French has taken too deep root in Flemish-speaking
Belgium. Latin, from its claim to be considered a universal
language, exerted this impeding influence in a very marked way.
754. Direct communication alone feels the necessity in its
greatest extent. No necessity in many cases exists for indirect
communication at all, even if the persons between whom the com-
munication takes place are dialectically far removed from each
other. If the communication be made by the agency of other
persons whose dialects lie between the two extremes, it may, owing
2 I
498
PRINCIPLES OF THE HISTORY OF LANGUAGE.
to numerous cases of transference, receive such a shape as to be
easily intelligible even to those to whom it would have appeared
unintelligible in the original dialect. Such a transference obviously
occurs if the productions of poetry pass orally from one place to
another. But even records committed to writing and spread
further by copyists yield to it also. In any case, the transference
remains always imperfect, so that mixed dialects arise. Very
numerous examples of this proceeding are afforded by the different
national literatures of the middle ages. In this way a literary link
becomes possible without the agency of a common-language be-
tween territories, which, dialectically speaking, stand far removed
from each other. Indeed this obvious process absolutely prevents
a dialect in which any literary records of special prominence are
drawn up from gaining a prevailing influence in its own territory,
because it is not spread with the records in question; at least not
in its simple form. Matters are very different when once printing
has begun to do the work of extension. By its aid it is possible to
spread a written record in every place without restriction, in its
genuine form, in the shape given it by the author or the printer.
And, for the advantages of printing to have full play, one type of
letters must, if possible, suffice for the whole linguistic area, and it
is of course also indispensable that the language expressed by this
type shall be everywhere understood. Thus the introduction of
printing introduces on the one hand the necessity of a common-
language, while on the other hand greater conveniences for the
satisfaction of this need are provided. For the rest, not until
printing was discovered was any diffusion of a knowledge of read-
ing and writing possible in large circles. Before the application of
printing there must always have existed but a narrow circle
capable of being touched by the operation of the norm of a written
language.
THE COMMON-LANGUAGE.
499
language
based upon
an existing
dialect.
755. The want is not of course in itself sufficiently strong to A common-
create the norm of a common-language. It is also incapable of must be
giving the impulse to create such a language voluntarily. Neither
does purpose prevail to this extent in this department, however
much greater may be its power than in the case of the natural
development of language. In every case the norm is represented
not, in the first instance, by any new creation, but by one of the
existing dialects. Nor is one of these chosen as the result of
deliberate consultation. It is rather true to say that the dialect
destined to become the norm must already possess some natural
preponderance, it may be in commerce, politics, religion, literature,
or in respect of several of these simultaneously. The intentional
creation of a common-language is a later matter, and only arises
after the first steps have been taken towards it. At all events
it seems to be of quite modern occurrence that the plan has been
deliberately adopted of creating a common-language without some
previously existing basis for its existence; and even in cases
where such attempts have been made [as in the case of Volapük],
they have rarely proved successful. The creators of any such lan-
guage have taken as their guide the circumstances of other lin-
guistic areas already possessing a common-language. When the
common-languages of the great civilised countries of Europe were
founded, there were as yet no guides to follow. The discovery
had to be reached that the existence of such a language was
possible before the attempt could be made to create one.
stances
necessarily
756. Any impulse to the formation of a common-language Circum-
implies the existence of a number of persons compelled by circum- which must
stances to make themselves acquainted with one or with several precede the
foreign dialects, so that they can easily understand them, and to of a common-
some extent apply them as well. This may be the result of their language.
having passed into another territory and settled there; or of their
formation
500
PRINCIPLES OF THE HISTORY OF LANGUAGE.
having made a long stay there in passing through; or of constant
association with persons who have arrived from a foreign land; or of
close familiarity with written records which took their origin there.
The relationships established in this way may be very manifold.
One who belongs to dialect A may learn dialect B, another may
learn C, another D; and again, conversely, one who belongs to
dialect B, C, or D may learn dialect A. As long as the reciprocal
influences of the different dialects are fairly balanced, no progress
is possible. If, however, one dialect presents in a marked degree
more occasion to master it than all the rest, and if this occasion
appeal to those who belong to all the dialects, this is then pre-
destined to be the common-language. Its predominance manifests
itself in the first place in the communication between the persons
who belong to it and those who belong to the other dialects, as it
is in this process mastered more generally and with more facility
by the latter than their dialect is mastered by the former; while
the other dialects remain more in a position of equality with each
other. The first really decisive step, however, is not made until
the prevailing dialect is actually employed for communication
between those who belong to other different dialects. A natural
result of this is that a larger quantity of persons become acquainted
with it. For it is then more convenient to employ it, when once
the native dialect has ceased to suffice, than to learn a third or
fourth dialect in addition. It presents itself most naturally when
the speaker desires to appeal not less to those who by birth belong
to it than to the rest of the nation, as is the case in literary
intercourse, and political intercourse as well, presuming always the
existence of political unity. Immediately that the consciousness
that such dialect is suitable for further intercourse has arisen,
the propagation of the development by voluntary means begins
as well.
THE COMMON-LANGUAGE.
501
757. The suitability, however, of a special dialect to serve as
model, is, as a rule, only a transitional step in the development of
the norm of the common-language. The imitations of the model
remain, as we have seen, in every case, more or less imperfect.
Mixtures arise between the model and the different native dialects
of the single individuals. It can scarcely fail to occur that these
mixed dialects as well obtain a certain authority, especially when
eminent writers employ them. On the other hand, the original
model dialect is exposed to incessant change as such,-while the
normal language must be more conservative in its tendencies, and
indeed, cannot maintain itself but by remaining true to the models
of past times. Thus it is that the dialect must gradually lose its
character of absolute model, and must strive for the mastery with
others of varying shades of difference.
necessary
artificial
758. Accordingly, the artificial language of a large area has a Development
tendency to become dialectically differentiated in a certain stage of even in an
development, in much the same degree as the natural language language.
within a particular territory. Centralisation of a more perfect
kind is commonly attained only by laying down actual rules, by
directions given orally, by grammars, lexicons, academies, etc.
But, however consciously, and with however set purpose, the norm
of a common-language may be created, the involuntary develop-
ment described in the former chapters can never be stayed; such
development is inseparable from all linguistic activity.
abendmal, 88.
abends, 194.
A
abgewinnem (einem), 324.
abhinc, 169.
abominable, 233.
abseit, 235.
INDEX.
absitzen (infin. used in demands), 124.
abspannen, 235.
Abstract meaning, 66.
abstreifen, 235.
Accusative, 151.
acetum, 55.
acht (different meanings of), 68.
achsel, 244.
(ahte), 209.
ἄγγελος (ὁ--ἡ), 290.
ἄγε, 254.
ἀδικῶ, 305.
Adjective used as a predicatival attribute,
425.
Adoption, stages in process of, 461.
of words, causes of, 460.
Adverbs, 422.
Adverbial determinance, psychological
treatment of, 317.
achtere, 85.
aeque
after, 168.
• •
•
ac, 329.
ac, 329.
Analogy in sound-substitution, 107.
ander, 225.
anders, 195.
anima, 85.
anklagen (einem), 324.
anliegen (einem), 324.
anreden (einem), 324.
anstatt, 427.
ansts, 200.
antequam, 333.
antwort, 214.
ἀνυσθῆναι, 57.
ardere, 256.
ἀρειότερος, 161.
arm, 94.
armbrust, 235.
armen, 247.
Assimilation of words already adopted, 464.
a-t-il, 206.
-atmig, 384.
atque, 168.
auf, 154, 195.
aufdrängen (einem), 324.
aufsitzen (infin. used in demands), 121.
auge, 245.
aurichalcum, 235.
ausdrucksvoll, 369.
ausschliesslich (emphatic word), 315.
ausserhalb, 225.
alius
•
allem (emphatic word), 315.
als, 172, 282, 283.
wie, 168.
ἀμφορεύς, 56.
amo, 94.
ἀμφότερος, 298.
ana, 205.
Analogy in inflexion and word-formation,
TOO.
balg, 248.
B
band (double gender of), 278.
bank, 244.
-bar, 265, 283.
barfuss, 411.
barn (Danish), 218.
baz, 169.
504
INDEX.
be- (function of making intrans. verb.
trans.), 153.
bechern, 84.
befallen, 153.
befahren, 230.
befel-befehlen, 262.
beide, 254, 298.
beizen (different meanings of), 68.
belabour, 153.
belt, 240.
berg, 214.
bescheiden, 277.
beschreiben, 153.
beschwichtigen, 234.
besonders (emphatic word), 315.
bewegt, 282.
bewegungsgefühl, 36.
bewogen, 282.
Bilingualism, 457.
bind (double meaning of), 74.
bittschriften, 118.
blindlings, 252.
bock (different meanings of), 68.
bogen, 240.
brachte (brâhte), 209.
brennen, 237.
brinne, 201.
brother (double plural of), 280.
buch, 279.
burg, 245.
burgher, 84.
buttern, 84.
cantitare, 161.
C
carina (fig. of speech, pars pro toto), 77.
castra, 297.
Categories psychological and grammatical,
288.
cattivo (sound of), 46.
*
cominciare, 160.
Composition, growth of, 368.
Compound sentences, 330.
Compounds, displacement of relations in,
268.
Concord, 339.
Concrete meaning, 66.
Conjunctions, 428.
Connecting words, origin of, 326.
Contamination, 160.
Convenience, 47.
Conversion of formal difference into differ-
ence of meaning, 218.
Creation, analogical in syntax, 98.
crêpe (double gender of), 279.
crocodilus, 55.
Crompare, comprare, 55.
Crystallisation, 253.
Culture of speaker, conditions, meaning,
91.
d&, 358.
dach, 82.
D
dagas (Old Friesic), 225.
dass, 165, 166.
(illogical), 170.
de (French), 324.
débonnaire, 374.
deinde, 169.
-dem (Conjunction suffix), 334.
Dependence united with independence,
140.
Dependence to independence, transition
from, 331.
derer, deren, 281.
derselbe, 369.
Derivative and inflexional suffixes, origin
of, 396.
Determinant and determinate, changes of
functions between, 319.
celestial (Spanish adjective), 161.
ch— palatal and guttural, 436.
Development necessary even in an artificial
language, 501.
Change of usage results from the ordinary Deviations from prescriptions of motory
exercise of speech, 12.
of meaning, 65.
linguistic, 24.
larus, 85.
cloth (double plural of), 280.
sensation inevitable, 43.
Dialectical differences concealed by writing,
439.
Dialects, mixture of, 473.
διδάσκαλος (ὁπή), 290.
INDEX.
505
dienstag, 235.
Differentiations mutually independent, 27.
Differentiation, real and apparent, 276.
ding (double plural of), 279.
-dings (in Adverbs), 252.
dismal, 233.
Displacement of motory sensation, 44.
diu, 249.
Divergence of analogical formation from
usage, 102.
divinal (Spanish adjective), 161.
Division of parts of speech, 403.
doppelt, 161.
Doublets, 275.
dröchst, 209.
dröcht, 209.
drücke, 279.
δρύφακτος, 55.
E
ehrenhold, 235.
eindruck, 257.
erziehen, 94.
es (a genitive), 321.
etsi, 167, 332.
Etymology, effect of, 450.
examen (different meanings of), 68.
ezzih, 55.
F
Factors in production of sound, 36.
falconer, 84.
fallen, 309.
fällen, 309.
fallieren, 234.
falls, 195, 252.
fang, fangen, 262.
feder, 85.
feld (sound of d in), 39.
felleisen, 235.
fiah, 213.
finfi, 55.
Finite verb, 404.
fio, 309.
first, frist, 55.
einfluss, 257.
einöde, 235.
einwirken, 256.
einzel, 225.
fisch, 94.
fisher, 84.
fisk, 248.
ἔκπαγλος, 56.
elira-eller, 55.
-eln (verbal suffix), 267.
Emphasis as a means of distinguishing
subject from predicate, 115.
empfangen, entfangen, 46.
en (negation), 171, 195.
endekrist, 232.
-endus (participial ending), 308.
enkel (double meaning of), 229.
-ens (Adverbs in), 252.
fix, fisc, 55.
flauen, 262.
flog, 221.
flüchser (flugs), 252.
fluobra, 55.
flur (double gender of), 278.
fluss (noun of action derived from a verb),
87.
fold, 214.
former (double comparative), 161.
fox, 76.
entweder
oder, 329.
enweste, 165.
-er, 263.
erbeizen (different meanings of), 68.
erde, 215.
ἐρεισθῆναι, 57.
erfahren, 308.
frau, 226, 244.
freitag, 231.
freitags, 162.
friedhof, 232.
friesen, 210.
Frisian, 33.
erila, erle, 55.
erkenntniss (double gender of), 279.
ermen, 247.
-ern (verbal suffix), 267.
erwähnen, 231.
fruit, 82.
frumentum, 82.
fuchs (different meanings of), 68.
führen, 94.
fünfzehn, 369.
funke, 245.
506
INDEX
fuss, 299.
futter (different meanings of), 68.
G
gar, 215.
garbe, 229.
garde (le-la), 291.
garten, 240.
gas, 12, 174.
gast, 222, 238.
geben (root-syllable in), 220.
(change from e to i in), 209.
gedanke, 240.
gedeihen-gediegen (change between and
haaren, 84.
-haft, 265, 283.
H
(double gender of), 278.
halba, 249.
halber, 254.
hand (expressions formed with), 89.
hangen, hängen, 309.
hantieren, 234.
hanun, 246.
hass-hassen, 262.
hausen, 84.
haut, 244.
-head, 87.
heide, 240.
heiss, 200.
-heit (-keit) 87, 245.
gin), 108.
gefi, 215.
gegessen (gezzen), 161.
geiz, 262.
gèle (il) (impersonal), 120.
gelernt, 308.
Genealogical tree, image of inadequate,
28.
Gender, 289.
Genitive, 150.
and adjective, 321.
genus, 227.
substantive, 322.
geschehen, geschichte, 108.
geschick (different meanings of), 68,
geschickt (different meanings of), 68.
geschwister (die), 296.
geschworen, 308.
gesicht (double plural of), 279.
gewiss, 318.
gewohnt, 161.
sewon, 161.
ginesan, 210.
glaube, 240.
glücklicherweise, 318.
gold, 214.
got, 214.
gouch, 180.
-gradig, 384.
Grammar and logic, 18.
grolioso-glorioso, 55.
Groups, destruction of, 192.
material and formal, 94.
quot, OHG., 286.
ἡμέδιμνον, 56.
ἥμισυς, 163.
her (uses of), 387.
herde, 214.
herr (as a prefix), 234.
herz, 226.
herzen, 84.
hier, 205.
hin (uses of), 387.
hinderniss, 244.
hinterseite, 225.
hoch (sound of ch in) 214.
hoher hoch, 108, 110.
hold, 214.
holz (hölzern), 214.
horn, 88.
hortus, 94.
huld, 244.
humanal (Spanish adjective), 161.
hunger-hungern, 262.
hungers (sterben), 194.
Hypotaxis, 145.
'Iáλvσos, 292.
Ιατρός (ό—ή), 290.
I
Idea-groups, organisms of the basis of all
linguistic action, 3.
idem
if, 167.
•
ac, 329.
-ig (marking derivatives from nouns), 263,
283.
INDEX
507
-igen (verbal suffix), 266.
ihrer, 161.
in (usages of), 99, 196.
inan, 161.
klack, 180.
klacks, 180.
klaff, 180.
klagen, 118.
Incapacity of primitive man to produce klanges (sound of ng in), 95.
speech, 187.
Individual spontaneity and social influence,
26.
klatsch, 180.
klecks, 180.
koch, 214.
Infinitive and an element dependent on it, | kokodrille, 55.
325.
infra, 230.
Inflexional suffixes, rise of, 400.
innerhalb, 225.
Inner linguistic form of language, influence
on, 471.
können, 230.
köpfen, 84.
korn, 82.
kost, 244.
krach-krachen, 262.
kragen, 240.
krist (e), 240.
int-, 231.
inter, 230.
κύανος, 291.
Intercourse, varying states of, 24.
kukuk, 180.
Interjections, 181.
L
ιπνός, 292.
ἵππος (ὁ—ἡ), 290.
laden, 230.
ipse, 253.
irgend, 67.
irnan, rinnan, 55.
island, 232.
Isolated sentences in form dependent, 365.
Isolation, 190.
J
jactitare, 161.
jamdudum, 164.
je, 67.
jemine! 181.
jubeljahr, 235.
juris-consultus, 369.
K
haladruis, karadruis, 55.
καλοκαγαθός, 377.
kalp (declension of), 227.
kaum, 146, 162, 173, 318.
κελαινεφής, 56.
kennen, 230.
κέραμος, 292.
Kéxʊka (dissimilation of aspirate), 63.
kiel (double meaning of), 229.
kiesen, 210.
κισσός, 292.
lagh (Danish), 218, 219.
land (sound of n in), 39.
lange (g not heard in), 39.
Language, the common, 475.
a common, must be based upon an
existing dialect, 499.
lanzknecht, 232.
laute, 233.
leben (en pronounced m), 39.
(lengthening of vowel in), 209.
legal—loyal (doublets), 277.
legen (last e not heard), 39.
leiden, 213, 246.
leider! 181.
λείπω, 219.
-ler, 266.
lesan, 210.
lesen (change from e to i in) 209.
letz, 161.
leutemund, 233.
-lich, 265, 283.
-liche (adverbial suffix), 267.
licht (double plural of), 279.
lichten, 230..
lieb (sound of ʊ in), 39.
liebstöckel, 235.
liegen, 305.
-ling, 266.
508
INDEX.
links, 195.
litterae, 297.
los (double meaning of), 229.
löschen, 250.
-μa, 292.
magis, 169.
M
mäher, 264.
mähder, 264.
mahlen, 230.
mähre (double meaning of), 229.
mais, 169.
maiza, 161.
mal (different meanings of), 68.
μᾶλλον, 169.
manen, 237.
mann (plur. in er), 227.
manniz, 248.
μάραθος, 292.
mark, 299.
marmul, marmor, 55.
mauern, 84.
martel, martyrium, 55.
maulwurf, 235.
Meaning, change of, in department of
syntax, 149.
meduz (u-declension in original German),
222.
mein, 162.
mensa, 94.
mensch (double gender of), 278.
mêriro, 161.
messe (different meanings of), 68.
metu, 223.
mîde-meit, 48.
minimissimus (double superlative), 161.
Mixture in language, 456.
modo, 147.
mord—morden, 262.
morgens, 194.
mutter, 244.
N
nach, 215.
nachts, 194.
nadel, 69.
nahe, 215.
näher, 264.
nähter, 264.
name, 226.
-ndus (Latin participial ending), 304.
nehmen (change from e to i in), 209.
neigen, 237.
-ner, 266.
nescio quid, 370.
New creation still operative, 174.
New suffixes, rise of, 399.
newt (=an eft), 110.
nichts (a genitive form), 321.
-niss, 282.
niuwen, 246.
Number, 295.
nuptiae, 297.
Nursery language, 183.
nur (emphatic word), 315.
(derivation of), 254.
O
-0. (OHG. adverbs in), 252.
oberarm, 225.
oberhalb, 225.
Object, double, 131.
Occasional meaning, 65.
octo (sound of), 46.
oder, 167.
öffnen, 284.
öftrer, 161.
ohm (double meaning of), 229.
ohne, 172.
οἶδα, 220.
olt, 235.
Onomatopoeia, 179.
Oratio obliqua, 142.
Organic nature, analogies from, 20.
orison—oration (doublets), 277.
ort (different meanings of), 68.
(double plural of), 279.
Orthography, effect of fixity of, 445.
-ws (Greek adverbs in), 252.
otto (sound of), 46.
οὐκοῦν, 258.
palliolum, 55.
par. ac, 329.
Parataxis, 143.
πάρος, 303.
Р
INDEX.
509
παρθένος (ὁ), 290.
Participial constructions, 158.
paterfamilias, 368, 380.
πεισθήναι, 57.
pelegrinus-peregrinus, 55.
penitence-penance (doublets), 277.
penny (double plural of), 280.
pensioner, 84.
people, 296.
répevya (dissimilation of aspirate), 63.
Periphrasis, the discrepancy avoided by,
315.
perire, 256.
peristroma, 292.
pfund, 299.
phellel, 55.
φεύγω, 305.
plebiscitum, 368, 380.
πλέκω, 219.
Pleonasms arising from contamination, 168.
Pleonastic combination of a native with a
foreign suffix, 469.
pluit (impersonal), 120.
Plurality of meaning, 67.
Popular etymology, 231.
pluriores, 161.
posthumous, 233.
quelque, 369.
quin, 257.
quindecim, 369.
quinque, 55.
quisque, 254.
rabe, 240.
R
rad (plural in er), 227.
rappe (different meanings of), 68.
rat (noun of action derived from a verb), 87.
rauch-rauchen, 262.
rauh, 215.
rauscht (es), (impersonal), 120.
recht, 214.
rechts, 195, 198, 252.
rechtskräftig, 369.
rěden, 28.
redete (middle e weak), 60.
regal-royal (doublets), 277.
reiter, 264, 278.
Relation of ai and ei, 449.
""
f
,, V, 449.
r
""
rh, 449.
t
""
th, 449.
postquam, 333.
poulterer, 84.
""
Repeated adoption of same word, 467.
Reproduction, 189.
rîbe-reip, 48.
rîchen, 246.
Power of distinguishing sounds, limits to, rings, 252.
42.
ritensplit (Low German), 377.
Predicate, psychological and grammatical, ritter, 264, 278.
II2.
Prepositions, 154.
Preposition, the, 427.
Prepositions and conjunctions, origin of,
426.
Preposition and conjunction in the simple
sentence, difference between, 430.
prior, 55.
priusquam, 333.
rock (double meaning of), 230.
rosary, 88.
rouse (double meaning of), 230.
row (
), 230.
""
ruch (sound of ch in), 213.
sal, 233.
s, soft and hard, 435.
puff-puff, 183.
punch (
), 230.
-sam, 265, 283.
purl
""
""
), 230.
sammeln, 284.
pump (double meaning of), 230.
sallow, 230.
S
quamquam, 332.
quamvis, 167.
Q
sangen, 226.
sänger, 264.
sapere, 156.
sau (double plural of), 279.
510
INDEX.
1
schaden, 240.
schaffen, 281, 282.
-schaft, 86, 245.
schall-schallen, 262.
schanz, 230.
schenken (ein kind), 153.
scheuchen, 278.
scheuen, 278.
schild (double gender of), 278.
schirm, 81.
-ship, 87.
ship (a general word), 91.
sich, 255.
setzen, 305.
sichern, 284.
sicut, 333.
sihe (sound of h in), 213.
simulac, 333.
schritt fahren (infin. used in demands),
124.
schlag-schlagen, 262.
schlingen, 230.
schmälen, 234.
schmecken, 156.
schmelzen, 247.
schmerz, 240.
schneide, 200.
schneiden, 213.
schneider, 264.
schnitter, 264.
schnur (double meaning of), 229.
scholar, 83.
schöpfen, 281.
schule, 225.
schüssel, 225.
schwammen, 221.
schweigen, 305.
schwein (schweinefleisch), 299.
schwerlich (with negative following), 173.
schwulst (double gender of), 279.
see (double gender of), 279.
sehen (used for aussehen), 156.
sehr, 199.
-seits (adverbs in), 252.
selber, 253.
semestris, 56.
Sentence, definition of, III.
Sentences, extension of, 129.
negative, 122.
interrogative, 125.
of assertion and demand, 122.
sentir, 156.
Separation of languages, 31.
sessel, 246.
shed (double meaning of), 229.
sheer (double meaning of), 229.
sheet (double meaning of), 74.
similis.
singer, 264.
sît, 159.
situ, 223.
ac, 329.
Situation, the, forms logical subject, 118.
sitze, 200.
sitzen, 305.
sitzende, 157.
sluog, 213.
so, 283.
bald dass, 333.
dass, 333.
sorgenfrei, 369.
Sound picture, 48.
Sounds, essential mark of a dialect, 33.
sovereign, 232.
so wie, 333.
Specialisation of meaning, 74.
sprang, 221.
springen, 201.
spruch (sound of ch in), 95.
st, soft and hard, 435.
starb, 221.
States of language, requisites for descrip-
tion of, 8.
stehen, 305.
stein (different meanings of), 68.
Stem, elements of, converted into inflex-
ions, 226.
steuern (different meanings of), 68.
stinken, 156.
στίξω, 200.
stracks, 252.
streit, streiten, 262.
studiert, 308.
Subject and copula, psychological, 313.
double, 129.
of science of languages, 1.
psychological and grammatical, 112.
Substantive and adjective, 407.
suchen, 233.
INDEX.
511
!
}
=
sucht, 233.
Suffixes constant and unceasing, formation
fusion of two, 265.
Union in formal groups, 246.
unserseits, 225.
unterarm, 225.
unterhalb, 225.
unterlegen (einem), 324.
unterwegs, 252.
of, 398.
συνέδριον, 235.
sunu (u- declension of OHG.), 222.
utrum
•
·
an, 329.
swaeren, 247.
uuesan, 210.
Superfluity in language, 272.
swâren, 247.
Syntactical juxtaposition to true composi-
tion, transition from, 317.
T
tag (sound of g in), 39.
tags, 194.
Technical and poetic language, 34.
teils, 195.
tenebrae, 297.
Tense, 300.
till, 168.
todes (verblichen), 194.
tor (double meaning of), 229.
törichterweise, 318.
trabs (fig. of speech, pars pro toto), 77.
tragen (lengthening of vowel in), 209.
Treatment of foreign phonetic material,
461.
treten (change from e to ¿ in), 209.
trotz, 154.
tuch (double plural of), 279.
-tum, 282.
tun, 316.
τύραννος (ὁ), 290.
turtur, 55.
über, 154.
U
überwerfen (einem), 324
veɩ (impersonal), 120.
ullus, unquam, 67.
um, 154.
umringen, 232.
unbillich (labial nasal in), 39.
-undus (participial ending), .308.
-ung (designates activity), 198.
ungnade (guttural nasal in), 39.
Unification, 202.
Uniformity of laws of sound-change, 57.
vater, 223.
ventitare, 161.
V
Verb and adverb, 324.
Verbal complexes which have the charac-
ter of compounds, 390.
Verbs (impersonal), 119.
verderbt, 282.
verdienst (double gender of), 279.
verdient, 308.
verdorben, 282.
verliesen, 210.
versammeln, 284.
versichern, 284.
verstummen, 305.
vielleicht, 318.
vix, 146.
Voice, 306.
voir (infinitive used in demands), 121, 124.
vol (use of), 156.
volitare, 161.
vollends, 252.
vor allem (emphatic words), 315.
vorderseite, 225.
vormund, 232.
vorstellen (einem), 324.
vortrab (noun of action derived from a
verb), 87.
vorweisen, 232.
vroüde (declension of), 241.
wahnwitz, 232.
W
während (use of), 159.
wahrnehmen, 323.
wagen, 225, 246.
-wald (in proper names), 232.
wald, 238.
512
INDEX.
wan, 231, 358.
wandel, 265.
ward, 287.
-ware, 296.
was für, 391.
wau-wau, 183.
•
weder
weg, 115.
noch, 329.
weges (gerades), 194.
weide (different meanings of), 68.
weil, 159.
-weilig, 384.
weleras, 55.
wîso, 249.
without, 168, 172.
wo (for dass), 166.
Word, series of sounds, 39.
wort (double plural of), 279.
wurde, 287.
wenen, 237.
werden (variation in sing. and plur. in),
221.
309.
whence, 169.
while, 167.
willens, 194.
wini, 238.
winter, 246.
wis (from wesan), 160.
zer-, 230.
Ꮓ
ziehen, 213.
(change between h and g in), 108.
ziemen, 215.
zil-zilen, 262.
zoll,* 299.
zu, 154, 195.
zucken, 282.
zücken, 282.
zukunft (noun of action derived from a
verb), 87.
zunge (declension of), 241.
zuo, 204, 231.
zusprechen (einem), 324.
INDEX OF AUTHORS REFERRED TO.
Andresen, 137, 138, 141, 157, 158, 172,
232, 356.
Bechtel, 55.
Braune, 31.
Bréal, 69, 467.
Brentano, 121.
Brugmann, 14, 55, 161, 300.
Darmesteter, 283, 370.
Darwin, 13.
Delbrück, 125.
Firmenich, 3I.
Gabelentz, Von der, 113, 115.
Marty, 119, 121.
Max Müller, 41.
Miklosich, 119, 120, 121.
Nörrenberg, 31.
Osthoff, 264, 265.
Sayce, 289.
Schuchardt, 456, 470.
Skeat, 467.
Steinthal, 3, 4, 36, 67, 68, 113, 114, 133,
135, 185.
Wegener, 69, 115.
Wenker, G., 31.
Printed by T. and A. CONSTABLE, Printers to Her Majesty,
at the Edinburgh University Press.




3
}
'00

UNIVERSITY OF MICHIGAN
BOUND
MAR 29 1948
UNIV. OF MICHI
LIBRARY
3 9015 01031 3131
"2

י
43
어
​10 »
น.ส
$5.
R
11
J
प