/ ^ N 
 
 
 & 
 
 / rkef X 
 
 ^ 
 # 
 
 & 
 
 \x 
 
 <<fc 
 
 rh 
 
 % 
 
 '<*, 
 
 **<? 
 
 % 
 
 %, 4 
 «> ^ 
 
"*&ZM> J? \ '*&$m 
 
 v* 
 
 ./ 
 
 
 \ 
 
 V '*73^ 
 
 
 t <T 
 
 (Mi 
 
 
 • \ 
 
 %. 
 
 *) X (PP 
 
 
 \ 
 
 \, 
 
 s\ 
 
 *b. 
 
 <*,<? 
 
 </*\ 
 & ^ 
 
 UV 
 
 ***# I PI \ %^ 
 
 • % 
 
 **o 
 
 
 
 °*. 
 
 
 
 ^ 
 
 %, 
 
 € * 
 
 ,0^ °A 
 
 f 
 
 % 
 
 X, 
 
 o*. 
 
 <f 
 
 "■*>. 
 
 tf 
 
 \ • 
 
 
 rh 
 
 %& 
 
 / % 
 
 ■%. 
 
 % 
 
A SHORT 
 
 COMPABATIVE GBAMMAR 
 
 OF 
 
 ENGLISH AND GERMAN. 
 
■THE ACADEMY," IN REVIEWING THE FRENCH 
 EDITION (6th January, 1894), write* :— 
 
 " The object of this very able and interesting book is to enable 
 students who know something of modern German and modern English, 
 to understand the relation in phonology and grammar between those 
 languages, and their common relation to Greek and Latin. The attempt 
 is novel and somewhat daring, but M. Henry has been in a surprising 
 degree successful. The book, in fact, forms an excellent introduction 
 to the comparative philology of the Indo-European languages, its 
 peculiar value being due to the fact that the principles of the science 
 are throughout presented in their application to the explanation of 
 phenomena with which the learner is already familiar. The index of 
 English words illustrated contains about nine hundred entries, and the 
 index of German words nearly as many." — The Academy. 
 
A SHORT 
 
 COMPARATIVE GRAMMAR 
 
 OF 
 
 ENGLISH AND GERMAN 
 
 AS TRACED BACK TO THEIR COMMON ORIGIN AND CONTRASTED 
 WITH THE CLASSICAL LANGUAGES 
 
 BY 
 VICTOR HENRY 
 
 Deputy-Professor of Comparative Philology in the University of Paris, Doctor of Letter: 
 and Doctor of Laws 
 
 AUTHOR OF "A COMPAEATIVE GRAMMAR OF GREEK AND LATIN " 
 
 TRANSLATED BY THE AUTHOR 
 
 Itonbon 
 
 SWAN SONNENSCHEIN & CO 
 
 NEW YORK: MACMILLAN & CO 
 
 1894 
 
 S Y or the r \ 
 
 f university; 
 

 syyrs' 
 
 BOTLKB & TaNNEB, 
 
 Thk Sklwood Pbikting "Works, 
 Fbohs, and London. 
 

 In compliance with current copyright law, 
 
 U. C. Library Bindery produced this replacement volume on 
 
 paper that meets the ANSI Standard Z39.48-1984 to replace 
 
 the irreparably deteriorated original. 
 
 2000 
 
PREFACE. 
 
 The French edition of this work, published in November, 1893 
 (Paris, Hachette), was reviewed shortly afterwards in the 
 Academy (no. 1131), in so friendly and sympathetic a spirit 
 that 1 feel bound to express my gratitude to the anonymous 
 critic, both for the valuable suggestions he made, and for the 
 praise he gave to my book. Of the former I have, as he will 
 perceive, availed myself as far as possible, and have deferred to 
 his opinion in almost every case. There is, however, one point 
 on which he appears to have misjudged me ; he evidently 
 missed the note on page 23 (p. 21 of this translation), and was 
 therefore led to suppose that I had stated as a fact that an 
 r-vowel actually existed in English. The confusion is only 
 apparent ; for the sake of brevity I thought it advisable not to 
 separate the English final r from the other English and from the 
 German final liquids and nasals, but at the same time I reminded 
 my French readers that the r in this position had become an 
 untrilled vowel, and I referred them for further details to a 
 subsequent section. And I am still inclined to believe this 
 course the best for my purpose, especially when we take into 
 consideration the fact, of which there is hardly any doubt, that 
 .6 recently as two centuries ago the r was no less trilled in 
 mother than it is at the present day in raven. 
 
 A few words are necessary in regard to the terminology 
 - adopted in this work. Since the words " phonemes " for 
 vi " sounds " and " apophonie " for " vowel -gradation " were not 
 3 v b 
 
 ■3 
 
VI PREFACE. 
 
 adopted by the skilful translator of my Comparative Grammar 
 of Greek and Latin — of which the present book is a symmetrical 
 counterpart — I did not think it right to introduce them here : 
 convenient as they are, I have tried to do without them. But 
 as the term "metaphony" for G. "Umlaut" has no exact 
 equivalent in scientific language, I have ventured to introduce 
 it to the notice of English scholars and teachers in the belief 
 that, when once naturalised, they will find it a serviceable and 
 almost indispensable acquisition. " Pregermanic," though a 
 somewhat loose translation of G. " Urgermanisch," is at any 
 rate brief and unmistakable, considering that the notion equiva- 
 lent to " Vorgermanisch," — which in the present state of our 
 knowledge is scarcely distinguishable from "Indo-European" — 
 never occurs in my work, and, when occurring elsewhere, 
 may easily be expressed by some such unobjectionable term as 
 " Ante-Germanic." Lastly, I have retained the term " deflected 
 grade," which Mr. Elliott adopted in his translation, for the 
 o-shade of an Indo-Eui*opean syllable containing an e, seeing that 
 the only reason which can be urged against it is that it means 
 nothing. And this indeed is precisely the reason why I adopted 
 it many years ago and still adhere to it, since, as it means 
 nothing, it suffices to represent the fact we desire to express, 
 without introducing any misleading connotation. Wo know 
 what a " normal " syllable is, — namely, one that is accented, — 
 and what a " reduced " syllable is, — namely, one that is nnac- 
 cented, — but we are quite unable to account for the fact that, 
 in certain derivations, the e of the root or the suffix is regularly 
 changed to o. Such terms as " nebentonige Hochstufe," besides 
 being rather cumbrous, are merely fitted to conceal our own 
 ignorance and to prejudge the solution of a question which had 
 as yet better remain unsettled, and which indeed may never 
 receive a complete and satisfactory answer. 1 
 
 1 The theory that the deflected grade of the root has any connection with 
 the primitive accentuation becomes more and more improbable or, at any 
 rate, impossible to demonstrate. See Henry, Mutton, III. (1884), p. 502, 
 and now Streitberg, die EnUtehung der lie inttttfe (1894), p. 62. 
 
PREFACE. . 
 
 The present work, as my friendly critic points bnt in the 
 review already referred to, is intended to introduce the com- 
 parative method to students who have already some knowledge 
 of both the English and German languages. An English 
 reader will find it accessible if he has mastered the general 
 outlines of the grammatical structure of German. So much 
 indeed is necessary — for the comparative method does not pre- 
 tend to teach the rudiments — but at the same time it is amply 
 sufficient ; and I venture to emphasize this point at the outset, 
 lest the student should fear to be uneqnal to the task and 
 should decline to undertake it for lack of knowledge, which, far 
 from being required beforehand, is presented to him as it 
 becomes necessary in each chapter of this Grammar. He must, 
 too, bear in mind that the comparative study of the Germanic 
 languages forms an independent whole in itself and does not 
 necessarily presuppose a knowledge either of Sanskrit, or> 
 Greek, or even Latin, and that all the instances quoted in the 
 following pages from foreign languages are intended to aid 
 those who are already acquainted with them, not to disconcert 
 those to whom they are unfamiliar. It must further be re- 
 membered that, while a teacher may sometimes be compelled to 
 enter into minute details in order to provide for the wants of 
 more mature students, the beginner may very well pass these 
 over, or at any rate need not remember them. He must use 
 his own tact and judgment in making a choice between the 
 facts and ideas suggested to him, in dwelling upon those which 
 will aid his memory and in overlooking such as might encumber 
 it, and, in a word, in limiting his researches to the objects he 
 has in view, — according as he intends merely to compare 
 German and English, or to obtain a general idea of the complete 
 history of the Germanic languages, or even to glance at the 
 primitive features of the Indo-European mother-tongue. 
 
 The student should also beware of too readily accepting con- 
 clusions without understanding the evidence on which they are 
 based, nor should he wait to examine one subject thoroughly 
 before passing to another. Tin's fragmentary method of learn- 
 
Vlll PREFACE. 
 
 ing, although of great practical utility, must uecessarily seem 
 rather discouraging to the linguistic student, since he will know 
 nothing of the goal to which his steps are leading him, and 
 when he at length reaches it he may have forgotten the road 
 he has traversed. Hence I should advise him, in the first 
 place, to read the book as a whole, omitting nothing, of course, 
 but without waiting to consider specially any passages which 
 at first sight may appear difficult or obscure. It must, for 
 example, be confessed that pure phonetics have little charm for 
 the uninitiated reader; it almost requires a special sense to 
 admire a beautiful correspondence of vowels or consonants, no 
 less than a neat formula in mathematics. Word-formation 
 itself, especially when confined within the narrow limits which 
 a primer imposes, seldom appears sufficiently clear and cogent 
 to compensate for the inevitable dryness of its main statements. 
 It is not until he comes to grammar proper, that is, to the 
 study of declension and conjugation, that he will find himself 
 quite at home. Here, at any rate, nothing is unfamiliar to 
 him, and he may even verify for himself facts of which he has 
 hitherto remained ignorant. He will now begin to understand 
 the reasons for the phonetic and etymological processes which 
 may have at first proved bewildering. Were he now to re-open 
 the book after this first . perusal, he would find a new and 
 unexpected light cast upon most of the points which he had 
 wisely reserved for further examination. In this second reading 
 he should avail himself, to a large extent, of the numerous 
 references given in the notes at the foot of the page; he should, 
 for instance, associate the theory of final s with the study" of 
 plurals and genitives, or that of medial s with the formation 
 of German plurals in er, or vowel -gradation with the classifica- 
 tion of the so-called Germanic strong verbs; in short, he should 
 take a broad survey of the vast tract of knowledge into which 
 a sound method has led him step by step. 
 
 In the next place, I would recommend my readers to make 
 frequent use of the indexes of words which are given at the 
 end of the book. The examples have been purposely multi- 
 
PREFACE. IX 
 
 plied, and, though of course the most typical among them must 
 occur several times in different places, constant care has been 
 taken to vary them as much as possible, in such a way as to 
 include nearly all the important words of both languages, and 
 thus to supply the student with a double etymological vocabu- 
 lary, reduced to its simplest elements. He will find it advan- 
 tageous to take cognate words, and to compare them with each 
 other, vowel for vowel and consonant for consonant, and so 
 obtain a clear idea of the correspondences or phonetic discords 
 they disclose. He may then turn for verification to the com- 
 plete and detailed dictionaries in which Prof. Skeat and Prof. 
 Kluge have displayed so much erudition and industry. 
 
 Finally, although the immediate object of this work is the 
 comparative study of Modern En glish and Jjerman, I may 
 perhaps indulge in the hope that for some students it may also 
 serve as a primer of the earlier languages, and enable them to 
 translate, without further help than that of an appropriate 
 glossary, such easy texts as may be found in an elementary 
 chrestomathy of Middle English or Middle High German, of 
 Old English or Old High German, or even of Low German or 
 Gothic. This, for a tolerably experienced student, is by far the 
 most valuable of all trainings, and he would hardly be likely to 
 meet with any serious difficulties, apart from the variable and 
 arbitrary character of many old transliterations ; but if he has 
 once been informed, that Otfrid, for instance, still writes every- 
 where th for initial d, or that Notker, in certain well-defined 
 positions, replaces b, d and g, respectively, by p, t and k, or that 
 the Old English texts spell as ed, eo, etc., the diphthongs which 
 were actually sounded ea, eo, it will only require a little care 
 and attention on his part to triumph over these obstacles. 
 Whatever languages he chooses for study, I should advise him 
 in any case to begin with the Gospels, which are usually trans- 
 lated word for word, so that the difficulties of syntax are 
 avoided ; here, too, it is easy to verify and compare. Other 
 texts, for more advanced students, are mentioned below in a 
 bibliographical note. 
 
X PREFACE. 
 
 It is perhaps almost superfluous to add that this is by no 
 means a servile translation of the French. Besides silently 
 correcting errors which were pointed out by my critics or 
 discovered by myself, I have introduced into it some slight 
 modifications which seemed convenient in a book intended for 
 the use of English students and teachers, and, as a matter of 
 fact, if the two texts are compared, few pages will be found 
 the same. For any further suggestions from English critics — 
 supposing they find the work in its new shape worthy of their 
 attention — I shall be truly thankful. I must also express my 
 indebtedness to Mr. D. B. Kitchin, M.A., late of Trinity 
 College, Cambridge, who has been good enough to revise my 
 manuscript and to polish my rough continental English ; — to 
 Messrs. Swan Sonnenschein & Co., who undertook the publica- 
 tion of this translation and entrusted me with its completion 
 without any other guarantee than that of my name ; — and. 
 finally, to the printers, whose skill, judgment and care hav 
 overcome the difficulties of their task and enabled the work to 
 attain at least the material accuracy which the reader expects 
 and the author desires. 
 
 V. HENRY 
 
 Paris, May 1st, 1894. 
 
TABLE OF CONTENTS. 
 
 PAGE 
 
 Preface . . V 
 
 Table of Contents xi 
 
 Transliteration xvii 
 
 Conventional Signs xxiii 
 
 Bibliographical Note xxv 
 
 (i) Intboduction 1 
 
 (8) FIRST PART. 
 
 SOUNDS . . ... . .15 
 
 (g) Chapter I.— The Elements of Physiological Phonetics . 18 
 
 (io) Section I. — The Production of Sounds 19 
 
 (n) Section II. — Classification of Sounds 23 
 
 (ia) § 1. — Vowels, Semi-vowels and Diphthongs . . .23 
 
 (13) § 2. — Consonant- Vowels 25 
 
 (14) § 8. — Consonants 26 
 
 (15) Chapter II. — Vowels and Diphthongs 29 
 
 (16) Section I. — Vocalic Laws in English and German ... 29 
 
 (17) § 1. — Recent Vowel Change . . . . . .80 
 
 (20) § 2. — Shortening and Lengthening 89 
 
 (a 1) § 8.— Old English Vowel-Breaking 42 
 
 (22) § 4. — Metaphony (Vowel Mutation or Modification) . 43 
 
 (24) § 5. — Pregermanic Compensatory Lengthening . . 49 
 
 (25) Section II. — Primitive Vowels and Diphthongs and their 
 
 Evolution 51 
 
 (26) § 1.— Short and long e 52 
 
 (27) § 2. — Short and long i ....... 54 
 
 (28) § 3.— Short and long u 55 
 
 (29) § 4 — Diphthongs of short e 57 
 
 (3°) § 5.— Short a and o, and their Diphthongs ... 59 
 
 (33) § 6. — Long a and o 62 
 
 (34) Section ILL— Vowels in Final Syllables 63 
 
 § 1.— Final Vowels 64 
 
 § 2. — Non-final Vowels 65 
 
 xi 
 
Xii TABLE OF CON'TF.XTS. 
 
 TACE 
 
 (35) Chapter III. — Sbmi-Vowei.s axd Coxsosant- Vowels . . 66 
 
 Section I. — Semi-vowels 66 
 
 (36) § 1- — Semi-vowel y 6G 
 
 (37) § 2- — Semi-vowel to 6S 
 
 (38) Section II. — Consonant- Voueis 69 
 
 (39) § 1.— Nasals 69 
 
 (41) § 2. — Liquids 72 
 
 (43) Section III.— ludo-Enrojtcan Voicel-G 'radalion (Aj/ophony) . 74 
 
 (44) § J. — The Principle of Vowel-Gradation .... 75 
 
 (45) § 2. — Germanic Vowel -Gradation 77 
 
 (46) CnAi-rKit IV. — Explosive Coxso.na.xts axd their Substitutes 82 
 
 (47) Section I. — The Second Consonantal Shifting .... S3 
 
 (48) § I.— Labials 84 
 
 (49) § 2. ~ Dentals 8S 
 
 (50) § 3.— Gutturals 91 
 
 (51) Section II. — The First Consonantal Shifting . . . . 9S 
 
 (52) § 1. — Grimm's and Vomer's Laws 100 
 
 (S4) § 2. — Primitive Voiceless Explosives .... 101 
 
 (56) § 3. —Primitive Voiced Aspirates 10S 
 
 (57) § *• — Primitive Voiced Explosives 109 
 
 (58) Chapter V. — Sibilant Coxsoxaxts Ill 
 
 (59) Section I. — Initial Sibilant Ill 
 
 (60) Section II. — Medial Sibilants 113 
 
 (61) § 1. — Voiceless Sibilant .114 
 
 (6a) § 2.— Voiced Sibilant . . .'-' 116 
 
 (63) Section III— Final Sibilant 117 
 
 (64) Chapter VI. — Accent 118 
 
 (65) Section I.— Word- Accent 119 
 
 (66) Section II. — Sentence-Accent 123 
 
 (6 7 ) SECOND PART. 
 
 WORDS. 127 
 
 (70) Chapter I. — Primitive Deiuvatiox 133 
 
 (71) Section I. — Primary Suffixes . . . • , . . 133 
 
 (72) § 1.— Nominal Stems 134 
 
 (81) § 2.— Verbal Stems 145 
 
 (86) Section II. — Secondary Suffixex 150 
 
 (87) § 1. — Nominal Stems 150 
 
 (92) § 2.— Verbal Stems 156 
 
 (g4) Chapter II. — English and Geiimax Dkkivation . . . 160 
 
 (95) Section L — Prefixes 160 
 
 (96) § 1.— Nominal Prefixes 161 
 
 (97) § 2.— Verbal Prefixes 164 
 
TABLE OF CONTEXTS. XI 11 
 
 PAGE 
 
 (101) Section II. — Suffixes properly so-called .... 170 
 
 § 1.— Nominal Suffixes 170 
 
 (102) A. — Nouns 171 
 
 (105) B.— Adjectives 176 
 
 (106) § 2.— Verbal Suffixes 177 
 
 (108) Section ML— Old Words changed to Suffixes . . .182 
 
 (109) § 1.— Nouns 182 
 
 (no) § 2.— Adjectives 184 
 
 (hi) § 3.— Adverbs 188 
 
 (112) Chapter III. — Composition 190 
 
 (113) Section I. — Classification of Compounds .... 190 
 
 (114) § 1. — Grammatical Classification 191 
 
 ( IX 5) §2. — Functional Classification 193 
 
 (116) Section IL — Formation of ComjMunds 195 
 
 (117) • § 1.— Form of the First Term 196 
 
 (118) § 2.— Form of the Last Term 200 
 
 (119) Chapter IV. — The System of Numeration .... 203 
 
 (120) Section I. — Cardinal Numbers 203 
 
 (iai) § 1.— Units and Sums of Units 203 
 
 (122) § 2.— Tens 206 
 
 (123) § 3.— Hundreds and Thousands 208 
 
 (124) Section 1 1. — Derivatives from Cardinal Numbers . . 209 
 
 (125) THIRD PART. 
 
 DECLENSION. 213 
 
 (126) Chapter I. — Articles 215 
 
 (127) Section I. — Definite Article 216 
 
 (128) § 1. — Origin and Primitive Declension .... 216 
 
 (129) § 2.— Modern State 218 
 
 (134) Section II. — Indefinite Article 222 
 
 (135) Chapter II.— Nouns 223 
 
 (136) Section I. — Gender 223 
 
 (137) Section II. — Number 226 
 
 (138) § 1.— General Remarks 228 
 
 (139) § 2.— Plurals in -$ 230 
 
 (140) § 3.— Plurals in -en 234 
 
 (143) § 4. — Plurals in -e without Metaphony .... 240 
 
 (144) §5. — Metaphonical plurals with or without an -e . . 243 
 
 (145) A. — English metaphony 244 
 
 (146) JB.— German metaphony 245 
 
 (147) § 6. — Metaphonical plurals in -er 248 
 
 C 
 
XIV TABLE OF CONTENTS. 
 
 PAGE 
 
 (148) Section III.— Cases 249 
 
 (149) § 1.— Accusative 249 
 
 (150) § 2.- Genitive 252 
 
 (152) § 3.— Dative 255 
 
 (153) Chapter III. — Adjectives . . . . . . . . 259 
 
 (154) Section I. — Declined Adjective 260 
 
 (i55) § 1.— Strong Declension 260 
 
 (156) § 2.— Weak Declension 261 
 
 (157) Section II. — Invariable Adjective ...... 263 
 
 § 1.— In German 263 
 
 § 2.— In English 264 
 
 (158) Chapter IV.— Pronouns 266 
 
 (159) Section I. — Demonstratives 266 
 
 (160) § 1.— Demonstratives properly so called .... 268 
 
 (161) § 2. — Interrogative and Indefinite Pronouns . . . 268 
 
 (162) § 3.— Relative Pronouns 270 
 
 (163) Section II. — Personal Pronouns 271 
 
 (164) § 1.— First Person 271 
 
 (165) § 2.— Second Person 278 
 
 (x66) § 3.— Third Person 274 
 
 (167) § 4.— Reflexive Pronoun 276 
 
 (618) § 5.— Possessives .278 
 
 (170) FOURTH PART. 
 
 CONJUGATION. 282 
 
 (171) Chapter I. — Tenses 284 
 
 (173) Section I. — Perfect: General Survey 285 
 
 (175) Section II. — Strong Perfect and Participle .... 289 
 
 (176) § 1. — Perfects with vowel-gradation .... 290 
 
 (179) A. — Type drive— treiben 292 
 
 (180) B. — Type choose=kiesen 295 
 
 (181) C. — Type drink=trinken and swell=schwellen . . 297 
 
 (182) D.— Type steal=steJden 300 
 
 (183) E.— Type see=selien . . . ... . .303 
 
 (184) F. — Type slay=schlagen 804 
 
 (185) § 2.— Reduplicated Perfects 80S 
 
 <?.— The only Type being fall = fallen . ... 806 
 
 (186) Section III.— Weak Perfect and Participle . . . .810 
 
 (187) § 1.— Apparent Anomalies 810 
 
 (188) § 2.— Principle of Formation 815 
 
 (189) § 3.— Applications 819 
 
TABLE OF CONTENTS. 
 
 XV 
 
 (190) 
 
 (191) 
 (192) 
 
 ( x 93) 
 (196) 
 (198) 
 
 (*99) 
 (200) 
 (204) 
 (208) 
 
 (211) 
 (213) 
 (214) 
 
 (215) 
 (216) 
 (218) 
 (222) 
 
 (223) 
 (224) 
 (225) 
 (226) 
 (227) 
 (228) 
 (229) 
 (230) 
 
 (231) 
 (232) 
 C233) 
 
 (234) 
 (235) 
 (236) 
 (237) 
 (238) 
 
 (239) 
 (240) 
 
 PAGE 
 
 Chapter II. — Moods 321 
 
 Section I. — Subjunctive 321 
 
 § 1. — Present Subjunctive 321 
 
 § 2.- Perfect Subjunctive 323 
 
 Section II. — Imperative 327 
 
 Chapter III. — Person- Endings 330 
 
 Section I. — Ordinary Conjugation 330 
 
 § 1.— Present-Endings 330 
 
 § 2.— Present-Metaphony 335 
 
 § 3.— Perfect-Endings . . 341 
 
 § 4. — Subjunctive-Endings 343 
 
 § 5. — Imperative-Endings 345 
 
 § 6.— The English Verbal Plural 346 
 
 Section II. — Anomalous Conjugations 349 
 
 § 1.— The Verb " to be " 349 
 
 § 2.— Other Root- Verbs 352 
 
 § 3.— Preterito-Presents 355 
 
 Type 4 357 
 
 Type C. 358 
 
 TypeZ) 359 
 
 Type.E. 360 
 
 Type F. ... 360 
 
 Unclassed 361 
 
 Chapter IV. — Verbal Periphrases 362 
 
 Section I. — Periphrastic Tenses 365 
 
 § 1.— Present 365 
 
 § 2. -Past 365 
 
 § 3.— Future 367 
 
 Section II. — Periphrastic Moods 369 
 
 § 1.— Indicative 369 
 
 § 2.— Subjunctive 370 
 
 § 3.— Conditional 370 
 
 § 4. — Imperative 371 
 
 Section III. — Periphrastic Aspects 372 
 
 § 1. — Reflexive Aspect ' 372 
 
 § 2. — Passive Aspect 373 
 
 Conclusion 375 
 
 Index of "Words.— I. English 377 
 
 II. German 384 
 
 Index of Terminations.— I. English 392 
 
 II. German 393 
 
/ or the * \ 
 
 (UNIVEESITTl 
 
 TRANSLITERATION. 
 
 In the transliteration of words and forms quoted either from 
 the Indo-European or Pregermanic mother-tongue, or in gene- 
 ral from any language, which, having no literature, naturally 
 lacks a conventional orthography, — such as rural dialects, — 
 as also whenever it is necessary to emphasize the true pronun- 
 ciation of any form even of a literary language, the spelling 
 is strictly phonetic, so that every symbol must be given its 
 specific and precise value, as defined at the beginning of this 
 work (no. 10-14) : thus, the vowels, a, e, i, o, u, are the so-called 
 Latin, never the English vowels, and ought to be sounded 
 everywhere, for instance, a = E. a in father, i = E. ee in see, u = 
 E. oo in soon (German u), and so forth; 3, }>, tS and b are the 
 spirants which respectively represent the German g of ewige, E. 
 hard and soft ih, and E. v x ; s is E. sh, and z is French j (z in 
 glazier) ; w and y, semivowels, E. w and y in young, etc. The 
 latter sound, however, is denoted by a j in Pregermanic and 
 Gothic, as it was in Old High German and is still in all the 
 continental Germanic languages. 8 
 
 In regard to accent and prosody, the reader must always dis- 
 tinguish very accurately such symbols as : a or simply a, 
 unaccented short vowel ; a, accented short vowel ; a, unaccented 
 long vowel ; d, accented long vowel ; occasionally, a, short 
 vowel with a secondary accent ; and the like. 3 
 
 1 Though not quite exactly, the B being a bilabial sound, whereas E. v is 
 rather denti-labial. 
 
 2 The reason, here, for infringing the law of uniformity in transliteration, 
 is merely a practical one, namely, because all Germanists agree in using this 
 spelling: wherefore also the diphthongs ay and aw will be written ai and au 
 in Pregermanic. 
 
 3 In such languages as no longer exhibit any written evidence of accentu- 
 
XV111 TRANSLITERATION. 
 
 The theoretical transliteration adopted for the mother- 
 tongues, as a rule, holds good for their offspring, apart from 
 the additions and modifications which are forced upon it by 
 usage or conventional orthography, and may be briefly summed 
 up as follows. 
 
 Sanskrit. — The c, in any position, like E. ch (in church) ; 
 the j, like E. j. Non-italic t, d, and n, in italic words, are 
 cacuminal consonants ; but, of course, we may be allowed to 
 blend them with the corresponding dentals. The s is never 
 soft and is sounded in every position like initial E. s or 
 French c . 
 
 Greek. — The so-called Erasmian pronunciation x is not quite 
 correct ; but it may be deemed sufficient for our purpose, pro- 
 vided the long vowel be always carefully distinguished from 
 the short one. The F, a merely dialectic sound, is E. w. 
 
 Latin. — The rules for correct Latin pronunciation are both 
 plain and short, viz. : — each vowel has its phonetic value ; — 
 c, in any position, like k, and g, in any position, as in E. give ; — 
 j = German j (E. y), and u = E. w; — s hard (never like z), even 
 when medial ; — the accent, always on the penult, or on the 
 antepenult if the penult is a short syllable. 
 
 Gothic. — The ei is not a diphthong, but a simple I (E. ee). 
 The ai is a short open e (E. set), and the au a short open o (E. 
 not), whereas ai and du are true diphthongs, sounded as they 
 would be in Modern German (E. approximately I and how). 
 If the symbol exceptionally lacks the distinctive accentuation, 
 then it must be inferred that its real value is as yet unknown. 
 In the diphthong iu, the stress is on the i, and the glide on the u. 
 
 The sound of b, d and g is a double one, — a fact, however, 
 which it will be sufficient to bear in mind, without troubling 
 to reproduce it, — namely: medial between vowels, they are 
 changed to spirants and become respectively b, 6" and 3. The 
 
 ation (Gothic, Old English, Old High German, etc.), accent should be 
 neglected, and there remains but the distinction between short and long 
 vowels. The Latin accent, though well known, is never marked, because it 
 is of no importance for Indo-European comparison : see my Gramtn. of Gr. 
 and Lat., no. 80-82. 
 
 1 Each vowel with its phonetic value, and 6 = t+h (not E. tli): cf. my 
 Gramm. of Gr. and Lat., no. 23, 24 and 54. 
 
TRANSLITERATION. XIX 
 
 >, of course, is E. hard th. The s is always hard, and the z, as 
 in English, is the corresponding soft sibilant. 1 The v, which is 
 written w by many Germanists, is equivalent to the E. w, and 
 the q is to be pronounced like hw. Lastly, the g, before a A; or 
 a g, represents the guttural nasal A (briggan like E. bring and 
 G. bringen). 
 
 Old Norse. — The ce, the ce and the y, whether short or long, 
 respectively, like German a, o and u (French u). Spelling 
 keeps the voiced dental spirant (6"), apart from the voiceless 
 ()>) ; but the labial spirant, whether voiceless or voiced (jf or 
 v), is always written /. Other symbols present no difficulty. 
 The quotations, moreover, are very few. 
 
 Old English. — The exact pronunciation of the vowels and 
 diphthongs is lost altogether ; but there is little chance of error 
 in pronouncing the words as they are written, that is, each 
 letter with its phonetic value. 2 We should only observe that 
 the ce is a simple vowel, not a diphthong; for it arises from a 
 primitive a and has returned to E. a : it is the sound of a in 
 bag, cab, with but this difference, that it may be either short or 
 long. The y is French u (G. u), likewise either short or long. 
 The groups ea, ea, eo, eo, ie, ie do not form two syllables : they 
 are mere diphthongs, the stress being laid on the first com- 
 ponent, and the second uttered very swiftly as a kind of semi- 
 vowel. 
 
 Old English indifferently spells 6" (later J}) the dental spirant, 
 whether voiced or voiceless : hence we are unable to distinguish 
 them from each other with absolute certainty, and the best 
 course will be, always to pronounce the o" as a hard E. th. The 
 /, likewise, is sometimes an /, and sometimes a v. The c is 
 sounded k before any vowel, though it may have assumed, quite 
 early, in certain positions, the slightly palatal shade which later 
 on resulted in the present sound of E. ch 3 The g was a spirant 
 
 1 Beware of pronouncing it as Modern German z. 
 
 * Beware, above all, of sounding the Old English, and even the Middle 
 English vowels and diphthongs, with the strange value they have now ac- 
 quired in contemporary speech (cf. infra no. 17-20) : gos (a goose), not gus, 
 and ges (geese), not gls, and so forth ; of course, Old English u is German u. 
 
 3 Thus, in ic (I), pr. ich: some E. dialects have the form chill for I will, 
 where the apparent initial of the vej4rf^^,lly|(jtufyj^ie final of the pronoun. 
 
 y fv 5 - CFTHF '^> \ 
 
XX TBANSLITEBATION. 
 
 in all the numerous cases in which, it has now become E. y : 
 the symbol for it was j ; indeed, it is from the Old English 
 alphabet itself, that our contemporary phonetics have borrowed 
 this peculiar letter to denote the voiced palatal spirant. 1 But 
 ordinary pronunciation need not dwell upon these minute 
 details. The semi-vowel w is E. w, and the i, when represent, 
 ing the Germanic semi-vowel /, — as, for instance, in the ending 
 of verbs in -ian, — is also confined to the value of a simple glide 
 or semi- vowel. 
 
 Middle English. — Take each letter almost for its phonetic 
 value, obscuring however such unaccented sounds as later have 
 become mute. The th, either hard or soft, according as it is 
 pronounced in the corresponding Modern English words. 
 
 High German (Old and Middle). — The vowels and diph- 
 thongs are to be sounded as they are spelled, that is, very 
 nearly as in Modern German, but with a clear distinction be- 
 tween short and long vowels : thus, the student should beware 
 of lengthening the first syllable in geban ( = geben), or shorten- 
 ing the final in haben (=haben). It is superfluous to make any 
 difference between e and e, the former being written for primi- 
 tive Germanic £, and the latter for posterior e arising from the 
 metaphony of a. In the diphthongs ea, ia, eo, io, uo, the stress 
 is on the first component. 
 
 The consonants, in general, are those of Modern German, 
 except in the case of four : medial h before a consonant, and 
 final h, like G. ch (naht = nacht); 3 and 33, approximately a hard 
 s 2 ; s, hard in any position ; w, as in English. 3 
 
 Slavonic. — Immaterial, owing to the rareness of quotations. 
 Follow the spelling : u is an obscure vowel ; s = E. sh ; c = Sans- 
 krit c or E. ch. 
 
 Modern Languages. — The reader is supposed to be ac- 
 quainted with their pronunciation, or else must be referred to 
 
 1 In this Grammar, however, the O.E. palatal has been written g, simply 
 to avoid an unfamiliar symbol, and to emphasize the perfect correspondence 
 of both the English and the German sound. 
 
 8 But f or zz, just as at present (ts). 
 
 8 In Middle German, 10 verges on its present sound, namely, English v. 
 Old and Middle German v has the sound of /, with only a slight difference 
 which is now quite lost. 
 
TRANSLITERATION. XXI 
 
 pronouncing dictionaries. The traditional orthography of 
 Modern German has been retained in this work, because it is 
 still the best known and most current out of Germany, and 
 differs too little from the new one, to involve here any incon- 
 venience ; though, indeed, the recent reform ought to be 
 encouraged ; for, inconsistent as it may be, England and France 
 have a great deal to do for themselves, before they can be 
 allowed to find fault with the modest improvement in spelling 
 attained in Germany. 
 
-' ' rr ;,;; :,, y x 
 
 CONVENTIONAL SIGNS. 
 
 ace. 
 
 accusative. 
 
 nomin. 
 
 adv. 
 
 adverb. 
 
 nt. 
 
 advb. 
 
 adverbial. 
 
 O.E. 
 
 erf. 
 
 compare. 
 
 O.F. 
 
 daf. 
 
 dative. 
 
 O.H.G. 
 
 d«/?. 
 
 deflected. 
 
 O.N. 
 
 E. 
 
 English. 
 
 part. 
 
 •#• 
 
 for instance. 
 
 Pf- 
 
 F. 
 
 French. 
 
 pi. 
 
 /m. 
 
 feminine. 
 
 pi. 1, 2, 3 
 
 G. 
 
 German. 1 
 
 
 gen. 
 
 genitive. 
 
 pr. 
 
 Go. 
 
 Gothic. 
 
 pref. 
 
 Gr. 
 
 Greek. 
 
 Preg. 
 
 i.e. 
 
 that is. 
 
 pres. 
 
 I.E. 
 
 Indo-European. 
 
 red. 
 
 impf. 
 
 imperfect. 
 
 sg. 
 
 ind. 
 
 indicative. 
 
 tg. 1, 2, 3 
 
 inf. 
 
 infinitive. 
 
 
 L. 
 
 Latin. 
 
 Sk. 
 
 Lith. 
 
 Lithuanian. 
 
 SI. 
 
 M.E. 
 
 Middle English. 
 
 tq. 
 
 metaph 
 
 metaphonical. 
 
 subj. 
 
 M.F. 
 
 Middle French. 
 
 tuff. 
 
 M.H.G 
 
 Middle High German. 
 
 vb. 
 
 Mod. 
 
 Modern. 
 
 voc. 
 
 msc. 
 
 masculine. 
 
 All other abbreviations 
 
 Zd. 
 
 will be self- 
 
 nominative. 
 
 neuter. 
 
 Old English. 
 
 Old French. 
 
 Old High German. 
 
 Old Norse. 8 
 
 participle. 
 
 perfect. 
 
 plural. 
 }. 1st, 2nd, 3rd person of the 
 plural. 
 
 pronounce. 
 
 prefix. 
 
 Pregermanic. 
 
 present. 
 
 reduced. 
 
 singular. 
 J. 1st, 2nd, 3rd person of the 
 singular. 
 
 Sanskrit. 
 
 Old Slavonic. 
 
 and following. 
 
 subjunctive. 
 
 suffix. 
 
 verb. 
 
 vocative. 
 
 Zend (Avestic). 
 -explanatory. 
 
 The sign of equality ( = ) between two forms implies their complete iden- 
 
 1 That is, what is called " German " in English, namely HochdeuUch. 
 The word " Germanic " is never abbreviated. 
 
 2 That is, especially, Old Icelandic. 
 
 xziii 
 
XXIV CONVENTIONAL SIGNS. 
 
 tity, whether the one proceeds from the other, as in E. book = O.E. hoc, 1 or 
 both are to be traced back to a common ancestor, as in E. book = G. buch. 
 A formula a:b = c:d denotes a proportion, to be read as in arithmetic. 
 
 An asterisk before a form denotes that it does not rest on any historical 
 evidence and is merely restored by conjecture. 2 Of course, this is the case 
 with all the so-called Indo-European or Pregermanic forms. 
 
 A hyphen, placed before or after a form, denotes an element of language 
 which never appears by itself and cannot be used but by becoming united 
 with some other element, namely : — the form followed by the hyphen is 
 either a prefix detached from the compound it belongs to (E. be-, G. ver-), or 
 a bare stem curtailed of its grammatical endings (G. seh- " to see ") ; — and 
 the form preceded by the hyphen, is either a stem which does not occur 
 without a prefix (G. -kunft 8 ), or (more often) a derivative suffix or a gram- 
 matical ending (E. -y = Q. -ig as in holy = heilig, E. sg. 2 -st = G. sg. 2 -st, 
 etc.). 
 
 The work has been divided into 240 sections, numbered uninterruptedly 
 from the beginning to the end, each of which forms as homogeneous a whole 
 as possible. All the references introduced by the words supra and infra refer 
 to these divisions. 
 
 The Indexes will be found at the end of the volume. 4 
 
 1 The signs > and <, which respectively should mean "resulting in" and 
 " derived from," are not used here : 1. because the exposition could do with- 
 out them ; 2. because they are equivocal, scholars having not yet come to a 
 full agreement as to their value. 
 
 2 Quite exceptionally, as on p. 1, it precedes a form which neither is nor 
 ever was extant. On the contrary, it cannot be too strongly insisted upon, 
 that the restored forms are not at all fictitious, but based on a strictly scientific 
 induction and, therefore, as sure in most cases as if they were actually found 
 in some book. Thus no word *varm-t, which would mean "warm," is to be 
 read in any Gothic text ; but both E. and G. have warm, to which a Go. "varm-s 
 is the strictly phonetical correspondent ; and, on the other hand, Gothic has a 
 derivative vb. varm-jan " to warm," which necessarily presupposes a nominal 
 basis *varm- : hence, we may affirm with absolute certainty that Gothic 
 once possessed an adjective *varm-s, and the asterisk here is an almost 
 superfluous symbol. 
 
 8 In the compounds aus-kunft, zu-kunft, ein-kiinft-e, no word 'kunft being 
 now extant ; but the adjective to zu-kunft is kiinft-ig, which irresistibly points 
 to the nominal basis * kunft. 
 
 4 The reader will be pleased to remember that a given word may occur 
 more than once under the same number, and to peruse the notes as well as 
 the text. 
 
BIBLIOGRAPHICAL NOTE. ] 
 
 Behaghel (0.). Die Deutsche Sprache. Leipzig und Prag, 
 Freytag, 1886. Engl. Transl. by Emil Teichmann sub tit. 
 Short History of the German Language. London and New 
 York, Macmillan, 1891. 
 
 Beitrage zur Gesehichte der Deutschen Sprache und Litera- 
 tur, herausgegeben von H. Paul and W. Braune (E. Sievers). 
 I. -XVIII. Halle, Niemeyer, 1874-94 (in progress). 
 
 Beitrage zur Kunde der Indogermanischen Sprachen, heraus- 
 geben von Dr. Ad. Bezzenberger. I.-XX. Gottingen, Pepp- 
 miiller, 1877-94 (in progress) . 
 
 Braune (W.). Gotische Grammatik, mit einigen Lese- 
 stiicken und Wortverzeichniss [1880]. 3* auflage. Halle, Nie- 
 meyer, 1887. Eng. transl. by G. H. Balg, sub tit. Gothic 
 Grammar, with selections and glossary, New York, Westermann 
 (London, S. Low &■ Co.) 1883. 
 
 Braune (W.). Althochdeutsche Grammatik [1886]. 2 te 
 auflage. Halle, Niemeyer, 1891. 
 
 Braune (W.). Abriss der Althochdeutschen Grammatik. 
 Halle, Niemeyer, 1886. 
 
 Braune (W. ; ed.). Althochdeutscbes Lesebuch, zusammen- 
 gestellt und mit Glossar versehen [1875]. 3 te auflage. Halle, 
 Niemeyer, 1888. 
 
 Brugmann (K.). Grundriss der vergleichenden Grammatik 
 der Indogermanischen Sprachen. Strassburg, Triibner, 1886- 
 93. Eng. transl. sub tit. Elements of the Comparative Gram- 
 
 1 This list has been purposely confined to the narrowest limits. Readers 
 who desire to prosecute their studies further, will find in the books quoted 
 here new references to guide them. 
 
XXVI BIBLIOGRAPHICAL NOTE. 
 
 mar of the Indo-germanic Languages, vol I. [Introduction and 
 Phonology], by Dr. Joseph Wright. London, Triibner, 1888 ; 
 vols. II.— III. [Morphology] by R. Seymour Conway and W. H. 
 D. Rouse. London, Paul, 1891-2. 
 
 Champneys (A. C). History of English, a sketch of the 
 origin and development of the English language. London, 
 Percival, 1893. 
 
 Delbruck (B.). Yergleichende Syntax der Indogerman- 
 ischen Sprachen. I. [ = vol. III. of Brugmann's Grundriss, ut 
 supra]. Strassburg, Triibner, 1893. 
 
 Earle (J.). The Philology of the English Tongue. Oxford, 
 Clarendon Press (New York, Macmillan), 1892. 
 
 Henry (V.). A Short Comparative Grammar of Greek and 
 Latin. Authorized translation by R. T. Elliott [1890]. 2nd 
 edition. London, Swan Sonnenschein & Co. (New York, Mac- 
 millan & Co.) 1892. 
 
 Indogermanische Forschungen, Zeitschrift fur Indogerman- 
 ische Sprach- und Alterthumskunde, herausgegeben von K. 
 Brugmann und W. Streitberg. I.-III. Strassburg, Triibner, 
 1892-94 (in progress). 
 
 Kluge (Fr.). Nominale Stammbildungslehre der Altger- 
 manischen Dialecte. Halle, Niemeyer, 1886. 
 
 Kluge (Fr.). Etymologisches Worterbuch der Deutschen 
 Sprache [1882-83]. 5 te auflage. Strassburg, Triibner, 1894. 
 Engl, transl. by J. F. Davis. London, Bell, 1891 [made from 
 the edition of 1889]. 
 
 Mayhew (A. L.). Synopsis of Old English Phonology, being 
 a systematic account of Old English Vowels and Consonants 
 and their correspondences in the cognate languages. Oxford, 
 Clarendon Press (New York, Macmillan), 1891. 
 
 Noreen (A.). Altnordische Grammatik, I. Altislandische 
 und Altnorwegische Grammatik, unter Beriicksichtigung des 
 Urnordischen [1884]. 2* auflage. Halle, Niemeyer, 1892. 
 
 Oliphant (T. L. Kington). The Old and Middle English 
 [1873]. 2nd edition. London and New; York, Macmillan, 1891 
 [Based on his The Sources of Middle English.] 
 
 Passy (P.). Etude sur les changements phonetiques et leurs 
 caracteres generaux. Paris, Didot, 1890. 
 
BIBLIOGRAPHICAL NOTE. XXVI 1 
 
 Padl (H.). Principien der Sprachgeschichte. 2*® auflage. 
 Halle, Niemeyer, 1889. English translation by Prof. H. A. 
 Strong [1888]. 2nd edition. London and New York, Longman, 
 1891. 
 
 Paul (H.). Mittelhochdeutsche Grammatik [1881]. 3 te 
 auflage. Halle, Niemeyer, 1889. 
 
 Paul (H.). Grundriss der Germanischen Philologie. I. 
 Strassburg, Triibner, 1891. 
 
 Scherer (W.). Zur Geschichte der Deutschen Sprache 
 [1863]. 2 te ausgabe (Nener Abdruck). Berlin, Weidmann, 
 1890. 
 
 Schleicher (A.). Die Deutsche Sprache [I860]. 5* auflage, 
 hrsg. J. Schmidt. Stuttgart, Cotta, 1888. 
 
 Sievers (B.). Grundziige der Phonetik [1876]. 4** auflage. 
 Leipzig, Breitkopf und Hartel, 1893. 
 
 Sievers (E.). Angelsachsische Grammatik [1882]. 2 te 
 auflage. Halle, Niemeyer, 1886. Engl. tr. by A. S. Cook, sub 
 tit. Old English Grammar [1886]. 2nd edition. Boston, U.S., 
 Ginn (London, Arnold), 1888. 
 
 Skeat (W.). An Etymological Dictionary of the English 
 Language. Oxford, Clarendon Press (New York, Macmillan), 
 1884. Abridged Edition [1881] 1891. 
 
 Skeat (W.). Principles of English Etymology. Part I. 
 Oxford, Clarendon Press (New York, Macmillan [1887], 1892. 
 Part II. 1891. 
 
 Soames (L.) Introduction to Phonetics : English, French, 
 and German. London, Swan Sonnenschein & Co. (New York, 
 Macmillan), 1891. 
 
 Societe de Linguistique de Paris (Bulletin et Memoires de 
 la). I.-VII. Paris, Vieweg (Bouillon), 1869-92 (in progress). 
 
 Streitberg (W.). Zur Germanischen Sprachgeschichte. 
 Strassburg, Triibner, 1892. 
 
 Sweet (H.). A Primer of Phonetics. Oxford, Clarendon 
 Press (New York, Macmillan), 1890. [First pubd. sub tit. 
 Handbook of Phonetics, 1877]. 
 
 Sweet (H.). A History of English Sounds, from the earliest 
 period, with fall word-lists. Oxford, Clarendon Press (New 
 York, Macmillan), 1888. 
 
XXV111 BIBLIOGBAPHICAL NOTE. 
 
 Sweet (H.). An Anglo-Saxon Primer, with Grammar, 
 Notes and Glossary. 6th edition. Oxford, Clarendon Press 
 (New York, Macmillan), 1890. 
 
 Sweet (H.). An Anglo-Saxon Reader, in Prose and Verse, 
 with Grammatical Introduction, Notes and Glossary [1876]. 
 6th edition. Oxford, Clarendon Press (New York, Macmillan), 
 1892. 
 
 Sweet (H.). A New English Grammar, logical and histori- 
 cal. I. Introduction, Phonology and Accidence. Oxford, 
 Clarendon Press (New York, Macmillan), 1892. 
 
 Vietor (W.). Elemente der Phonetik des Deutschen, Eng- 
 lischen und Franzosischen [1884]. 3 U auflage. Leipzig, Reis- 
 land, 1894. 
 
 Wilmanns (W.). Deutsche Grammatik (Gotisch, Alt-, Mit- 
 tel- und Neuhochdeutsch). I. Lautlehre. Strassb u rg, Triibner, 
 1893. 
 
 Wright (J.). A Primer of the Gothic Language, with 
 Grammar, Notes and Glossary. Oxford, Clarendon Press (New 
 York, Macmillan), 1892. 1 
 
 Zeitschrift fiir vergleichende Sprachforschung, herausgege- 
 ben von (Th. Aufrecht), A. Kuhn (E. Kuhn und J. Schmidt). 
 I. -XXXIV. Berlin, Diimmler (Qutersloh, Bertelsmann), 
 1852-94 (in progress) . 2 
 
 1 From the same scholar we have a recent and detailed, most interesting, 
 dialectical monograph, namely : a Grammar of the Dialect of Windhill, in 
 the West Biding of Yorkshire (London, 1892). 
 
 2 Headers who are desirous of an easy and agreeable introduction to the 
 dialects of Southern Germany, — as those of Central Germany differ but 
 slightly from standard German, — are referred to Hebel's Alemannische Ge- 
 dichte and to Arnold's Pfingstmontag. 
 
U N 
 
 CAUF CRN!*' 
 
 A SHORT COMPARATIVE GRAMMAR 
 OF ENGLISH AND GERMAN. 
 
 INTRODUCTION. 
 
 (i) When parents or masters are teaching a child a spoken 
 or a literary language, they usually frame it on some dogmatic, 
 arbitrary, and often contradictory rules, the reasons for which 
 he is unable to make out for himself, the more so as his elders 
 would be quite at a loss to make them out for him. Why, for 
 instance, is the plural of ox, oxen, and not *boxen the plural of 
 box ? Why, again, the plural of ship, ships, but not *sheeps the 
 plural of sheep ? Why should it be a vulgar blunder to say he 
 *hnowed, though it is quite correct to say he bowed ? Because 
 we ought to speak so and so. No further answer is ever given. 
 Learnt in that way, a language must appear a mere chaos. 
 The pupil may master it perfectly well : that is merely a matter 
 of practice and memory ; but his knowledge of it lacks any 
 basis of reason, no trace of scientific spirit is ever allowed to 
 pervade it, and thus it is he obtains a grossly misleading view 
 of the whole of a large domain of human thought. While the 
 other works of man are stamped with the seal of his genius, 
 language, the first and noblest of all, and the chief character- 
 istic which differentiates him from other living beings, seems 
 some strange and monstrous building, swarming with traps and 
 blind-alleys, planned at random and founded on caprice. 
 
 The fact is that a language, taken by itself, is as unable to 
 account for its own existence as would be any other human 
 fact. Let us glance at the map of Europe, and inquire why the 
 countries in it, and even the provinces in these countries, are 
 cut up into such irregular patterns. Why should they not 
 have been so many squares carefully drawn across the ground ? 
 
I ENGLISH AND GERMAN GRAMMAR. 
 
 Well, their history answers the question. In the same way we 
 may trace back the history of a language, and the higher we 
 trace it back, the better we become aware of the causes of such 
 discords as offended us at the outset, seeing them gradually, as 
 it were, blended into some superior harmony. This is the task 
 of Historical Grammar, which does for languages what 
 History does for nations. 
 
 Yet however far it may be carried back, the investigation 
 must come to an end before most of the problems have received 
 a solution. If he lacks witnesses, the historian becomes silent ; 
 so, where written documents are wanting, language escapes the 
 attempts of the grammarian. Now a language, before it is ever 
 written down, has already lived for centuries as vulgar speech, 
 unknown to learned people, or despised by them. Even the 
 most cultivated tongues are descended from popular dialects. 
 How easily, too, might the meagre documents of a language 
 when first written be either disfigured or lost ! The earliest 
 French text, short as it is and corrupted by generations of 
 scribes, goes back to the ninth century. The spoken French 
 which superseded Latin in Gaul was certainly very much older. 
 
 What are we then to do, if the authorities fail ? A political 
 historian is at a loss ; not so he who deals with the history of 
 languages. When he lacks the language he is studying, he 
 may refer to that from which it has arisen. So, as French 
 down to the ninth century stands beyond his reach, but is well 
 known to be descended from Latin, he will require Latin itself 
 to account for French. This must often answer his purpose, 
 since with the aid of Latin he can go back to the third century 
 B.C., whence the field of his researches is more than doubled. 
 Even supposing that Latin also were lacking, that is to say, 
 the whole Latin literature, epigraphy and civilization had 
 perished without leaving any relics but its mere name, he 
 would nevertheless find a way to it. The Latin parent had 
 other children besides French: Italian, Spanish, Portuguese, 
 Roumanian, descend from the same stem. Let the linguist 
 therefore put together and compare with one another all these 
 languages, each of which is sure to have preserved more or less 
 pure some feature of the original type; let him unify what 
 
INTRODUCTION. 3 
 
 they have in common, eliminate or conciliate their discords, 
 and then he may be able to restore the deficient Latin, not of 
 conrse in its minutest details, much less in the delicacies of its 
 style, but sufficiently in the general outlines of its grammatical 
 structure to bring back into unity its diverging offspring and 
 thus mend up the tie accidentally loosened. This at any rate 
 he would be compelled to do, if he wished to go back beyond 
 Latin, since the unknown Prelatin whence it has come can 
 never be avouched by any written document. It is irretriev- 
 ably lost, and yet, through the comparison of Sanskrit and 
 Greek, the grammarian has succeeded in restoring it. 
 
 Hence, as sooner or later evidence is wanting, and literary 
 or monumental tradition vanishes in the mists of the past, 
 grammar cannot be truly and consistently historical with- 
 out being at the same time comparative. 1 
 
 (2) Such is in particular the case with English and German. 
 They are well known to be near relations ; and yet they are 
 not descended from one another, nor both from any language 
 historically known. English is not derived from German, as 
 might be inferred from the large number of erroneous state- 
 ments still current, which teachers and learners cannot too 
 carefully avoid. 2 The two languages sprang, before the fifth 
 century a.d., from a common stem, and thereafter flourished 
 apart, diverging more and more down to the present time. Let 
 us now inquire what this common stem was. Certainly not 
 
 1 One point should be carefully insisted upon, namely, that the comparison 
 of two linguistic types, as of two natural species, does not pretend to restore 
 an intermediate type, exhibiting the characteristics of both, but a type prior 
 to both, wherein the divergent characteristics find a common origin. Such 
 a mistake would not be made by any one conversant with scientific methods. 
 Thus Darwin teaches us (Origin of Species, London, 1872, p. 265) : " I have 
 found it difficult, when looking at any two species, to avoid picturing to 
 myself forms directly intermediate between them. But this is a wholly false 
 view; we should always look for forms intermediate between each species and 
 a common but unknown progenitor ; and the progenitor will generally have 
 differed in some respects from all its modified descendants. To give a simple 
 illustration : the f antail and pouter pigeons are both descended from the 
 rock-pigeon ; if we possessed all the intermediate varieties which have ever 
 existed, we should have an extremely close series between both and the rock- 
 pigeon ; but we should have no varieties directly intermediate between the 
 j antail and pouter; none, for instance, combining a tail somewhat expanded 
 with a crop somewhat enlarged, the characteristic features of these two 
 breeds." 
 
 2 Cf. Skeat, Principles, i. p. 73. 
 
4 ENGLISH AND GERMAN GRAMMAR. 
 
 Gothic, 1 as perhaps is still wrongly believed by some students. 
 Gothic is not the ancestor, but at most an elder brother. The 
 documents inherited from it are four centuries earlier than the 
 most ancient texts written in Old English or Old German, so 
 that it reflects a more archaic if not altogether purer form of 
 the primitive Germanic type ; for it will be seen that in many 
 respects English and German stand nearer to the original 
 standard. 2 
 
 Now, to proceed to an exact account of their relationship, 
 Continental German (Deutsch) admits of two main divisions : 
 High German or High Teutonic (Hochdeutsch), and Low 
 German (Netherlandish, 3 Flemish, Plattdeutsch along the Baltic 
 shore). What we properly call German, whether literary or 
 conversational, is genuine High German, with here and there 
 scanty borrowings from Low German. On the other hand, a 
 long time before the division of Continental German had taken 
 place, a dialect called Anglic or Saxonic or, in one word, 
 Anglo- Saxonic, had emigrated from its native land, and, for 
 ever insulated by the sea, though subject to many foreign 
 influences which did not affect Continental German, has now 
 become Modern English 4 (Englisch). The common language, 
 historically unknown, but capable of restoration through lin- 
 guistic comparison, from which both English and German are 
 descended, is now known by the conventional term of West 
 Germanic. 
 
 1 In the French edition of this work the word is spelt without an h ; but 
 we cannot venture to introduce this spelling into English, as there the th is 
 not only written, but even pronounced. For the sake of correctness, how- 
 ever, it would seem advisable to drop this h altogether — allowing Gothic 
 writing and architecture to preserve it if they choose — as it lacks any support 
 in Germanic orthography. 
 
 2 Thus the so-called West Germanic branch has preserved, and English 
 and German still sometimes preserve, the primitive Indo-European e, which 
 Gothic almost everywhere changed to {. 
 
 3 To these alone we now apply the name (Dutch) that properly belongs to 
 all Teutons. Curiously enough the sense has become restricted. The 
 French language, on the contrary, has extended to the whole Teutonic race 
 the name of a southern tribe (L. Alamanni) which was first known to the 
 Franks after their conquest of Gaul. 
 
 4 English has modified its vowels, while its consonants are still even purer 
 than they are in Low German ; High German, on the contrary, has shifted 
 all its explosives : thus, High German and Low German stand nearer to each 
 other in their vocalism, whilst in their consonantal system English and Low 
 German are more closely related. 
 
INTRODUCTION. O 
 
 West Germanic, again, had two sisters through which we 
 are able to trace back one degree more. Gothic (Gotisch) is 
 indeed nearly akin to English and German, though far less so 
 than these are to one another. The same is the case with Old 
 Norse (Altnordisch), which still survives in the so-called 
 Scandinavian languages : Icelandic, Norwegian, Danish, and 
 Swedish. 1 Now, Gothic and Old Norse show more traces of 
 relationship with one another than with West Germanic, 
 though a serious amount of specific differences hold them apart, 
 forming two groups called respectively East Germanic 
 (Ostgermanisch) and North Germanic (Nordgermanisch). A 
 step farther we come back to the great parent, lost of course 
 though also restorable. Its existence is necessarily presupposed 
 by the fact of Germanic having spread out in three branches, 
 Eastern, Northern, and Western. This is the primitive or 
 common Germanic language, in short the Pregermanic 
 (Urgermanisch), a language never written, but spoken, at the 
 time when Hellenic civilization began to flourish, among the 
 barbarous tribes that were waudering about through the woods 
 and moors of Central Europe. 
 
 The historical and prehistorical series thus described will, if 
 now traced downwards, assume the form of a pedigree, as given 
 below, in which the asterisk denotes the dead languages from 
 which no written document survives. 
 
 "Primitive German or Pregermanic. 
 
 'East Germanic. 'North Germanic. *West Germanic. 
 
 I I . 1 
 
 Gothic. 2 Old Norse. 3 
 
 (died out without 1 Old English. 'Common German. 
 
 leaving any off- | 
 
 spring). Scandinavian languages. English. | | 
 
 Low High 
 
 German. German. 
 
 1 Modern Icelandic has preserved the original features of Old Norse better 
 than any other modern language. 
 
 2 Gothic is the language of the civilized and Christianized Goths (L. Goti) 
 of Moesia, a Roman province between the Danube and Balkans. It is pre- 
 served in the text of the Four Gospels, St. Paul's Epistles, and a few 
 iragments of other books of the Bible, the remains of a complete translation 
 of the Old and New Testament, composed for the use of his diocese by 
 Bishop Ulfilas (Vulfila, that is, " little wolf ") in the fourth century. 
 
 8 The earliest remains of Old Norse are quite asoldas those of Gothic* 
 
 •<«Ssi IS BR .*V. \ 
 
6 ENGLISH AND GEEMAN GEAMMAK. 
 
 (3) Thus, as surely as the Romance languages carry us- 
 back to Latin, the Germanic languages point to a primitive 
 German, save that Latin was both spoken and written, whereas 
 Pregermanic never was more than spoken. But are Latin and 
 Pregermanic two isolated tongues, without a tie between them Y 
 Have we now come to the limit of our knowledge, and to the 
 end of prehistorical induction ? No, the way remains still open 
 A science already mature though scarcely a hundred years old,, 
 has led us by infallible methods to the power of uniting into 
 one family a considerable number of languages, still represented 
 at the present day by numerous offspring which cover more 
 than half the habitable world. In Asia, Sanskrit, Zend,. 
 Persian, and their modern posterity; in Europe, Greek, 
 Latin, and the Romance languages transplanted in our days 
 into America; Celtic with its now dying offshoots; Ger- 
 manic, spread in every direction mainly through the expan- 
 sion of English ; Lithuanian, spoken eastwards to the Baltic- 
 Sea ; and Slavonic, 1 occupying the whole of the East and part 
 of the Centre of Europe (Old Slavonic, Bulgarian, Croatian- 
 Serbian, Ruthenian, Russian, Polish, and Czech). Some of the- 
 hymns of the Rig-Veda, the earliest documents of Indian 
 literature, may be as old as the tenth century B.C., and the 
 primitive stock of the Homeric poems carries us back almost, 
 to the same date. Hence Sanskrit and Greek are justly deemed 
 the best witnesses of an original speech, 2 optionally termed 
 Aryan, Prearyan, Indo-Germanic (Indogermanische Ursprache), 
 or still better the common Indo-European language. 3 
 
 perhaps even earlier (third to fourth century) ; but they merely consist of 
 short Scandinavian inscriptions written in so-called Runic characters. Only 
 as late as the twelfth century appear the first literary texts of Old Icelandic,, 
 and a little later the rich literature of the Eddas. 
 
 1 The last two are akin to one another, and can be referred to an earlier 
 Letto- Slavonic or Balto-Slavonic. We can also restore Celto-Latin. Less. 
 certain, though still possible, would be the so-called Greco-Italic and Ger- 
 mano-Letto- Slavonic groups. 
 
 2 Sanskrit has better preserved its consonants ; Greek has kept the vowels, 
 purer. As far as vowels only are concerned, Greek is the standard Indo- 
 European speech, except that it has often changed their genuine accent ; in 
 this respect Sanskrit is far superior. 
 
 8 To this term I adhere, cumbrous as it is, for the sake of greater accuracy. 
 Terminology, of course, is of secondary importance. Whatever its name, it 
 
INTRODUCTION. 7 
 
 Bat all these tongues, even the latest, must be brought into 
 the discussion, as the scattered pieces of a broken mirror 
 should be framed together to reflect the likeness of yore, and 
 thus every one, or at least some one, among the Indo-European 
 languages must be consulted whenever we are attempting to 
 trace back the history of any given word or form of English 
 and German. 
 
 A further pedigree will show the above stated division of the 
 Indo-European family into six main groups : 
 
 *Parent-speech. 
 
 *Indo-Iranian. 
 
 I 
 
 Sanskrit. 
 
 Iranian 
 (Zend and 
 Persian). 
 
 •Hellenic. 
 
 Old Greek 
 
 and Mod. 
 
 Greek. 
 
 "Pregermanic. 
 
 •Italic. 
 I 
 
 I 
 
 Latin 
 
 and Romance 
 
 Languages. 
 
 » Celtic. 
 
 Gaulish 
 
 and Mod. 
 
 Celt. Lang. 
 
 "Letto- Slavonic. 
 
 "East. 'North. 'West. *Lettic. 'Slavonic. 
 
 Mod. Germ. 
 Languages. 
 
 Mod. Lett, 
 and SI. Lang. 
 
 Having thus stated the place occupied by Modern English 
 and German in the great family to which they belong, we must 
 return to these, the special objects of our study. 
 
 (4) English is the official language of the British Empire ; 
 the current tongue of both islands — excluding the few rural 
 
 must be understood from our previous statements that Indo-European is a 
 prehistoric language, restored but in no way fictitious. It is also clear that 
 this language is the limit of our present knowledge : beyond it lies an un- 
 limited past, but this past we are unable to reach. We can go back beyond 
 English and German by means of Gothic and Old Norse ; even beyond 
 Pregermanic, by means of Greek and Sanskrit ; but we cannot pass beyond 
 Indo-European since there is no other left besides it. The case would be 
 different if we could find out any relationship between it and another family 
 of languages, e.g. the Semitic group (Assyrian, Hebrew, Syro- Arabian). 
 This, however, can scarcely be expected. 
 
O ENGLISH AND GERMAN GRAMMAR. 
 
 districts in which Celtic dialects still survive (Welsh, Scotch, 
 and Irish Gaelic) — of the United States of North America, of 
 nearly the whole Dominion of Canada, Australia, and several 
 other colonial centres scattered through the world ; lastly, the 
 common speech of trade and intercourse in all the great havens 
 and marketplaces of Eastern Asia, inasmuch as they are thriv- 
 ing either under the actual protection or at least the prevailing 
 influence of the United Kingdom. To this marvellous growth 
 of a language that has now become universal, let us oppose its 
 humble and obscure beginnings. 
 
 In the course of the fifth century some tribes, originating 
 from the north of Germany (the isthmus of Sleswig and the 
 estuary of the Elbe), severed themselves from the nations then 
 called Angles, Saxons, and Jutes, whose idioms were nearly 
 related though not identical, and repaii'ed by successive 
 emigrations to the great island of Britain, then occupied by a 
 Celtic population. Here they founded seven kingdoms gener- 
 ally known by the name of the Heptarchy. The settlements 
 of the Jutes lay in the south-eastern corner of the land (Kent) 
 and in the Isle of Wight ; those of the Saxons, on the river 
 Thames and the Channel (Essex, Middlesex, Sussex, and 
 Wessex) ; the Angles conquered the eastern protuberance (East 
 Anglia), viz. Suffolk and Norfolk, the great central square 
 (Mercia) — the corners of which in the main coincide with the 
 mouths of the Severn, the Mersey, and the Humber, and the 
 present seat of the Metropolis, — lastly, Northumberland up to 
 the boundary of Scotland. It is chiefly from the dialect of the 
 Angles, or more accurately speaking from Mercian that our 
 Modern English is derived : hence it is quite appropriately 
 named. But it must be well understood that a great many 
 Saxonic elements have intruded into it, the more so because 
 the Metropolis lies on the boundary of the two domains. Un- 
 fortunately enough, moreover, no important document has come 
 down to us from Old Mercian ; for it entered rather late into 
 its literary period and was fairly overrun by the neighbouring 
 Saxonic, 1 the Wessex dialect. In this, the very numerous 
 
 1 Northumbrian had preceded even Saxonic ; but the brilliant Northum- 
 
INTRODUCTION. V 
 
 literary texts l go back to the ninth century, and the little 
 Latin-Saxon glossaries to the eighth at least. It is therefore 
 this language, usually termed Anglo-Saxon or Old English, 
 that represents to us the most ancient form of our present 
 tongue that we can reach. Yet we must never forget that, 
 when tracing back the latter to the former, we are contrasting 
 it not with a real ancestor, but, as it were, with a twin brother 
 of this ancestor. 2 
 
 The historical and still surviving dialects of English are 
 three in number : the Northern (Northumbrian and Lowland 
 Scotch), and the Central (Midland) dialect (this again divided 
 into Eastern and Western), are descended from Anglic; the 
 Southern is Saxonic. The Jutic tongue has left but insignifi- 
 cant traces. 3 
 
 Chronologically the language may be divided into three 
 periods : Old English, conventionally limited to the end of 
 the twelfth century ; Middle English, till about the begin- 
 ning of the sixteenth century, chiefly represented by Chaucer 
 (1328-1400), and Modern English, from the year 1500 down 
 to the present day. 
 
 (5) Since Old English is a fortuitous though unequal mixture 
 of Anglic, Saxonic, and Jutic, English, of course, cannot be ex- 
 pected to be a language altogether pure. Many other causes, 
 however, occurred in the course of these fourteen centuries to 
 alter its character. Firstly, a few words were borrowed from 
 the Celtic idioms still living beside and beneath it. In the 
 ninth century, the struggle against the Danes brought into the 
 island a considerable number of Scandinavian elements, 
 
 brian civilization perished under the Danish invaders, who did not reach the 
 southern part of Britain. 
 
 1 Works of Alfred the Great (died 901), viz. translation of Gregory the 
 Great's Gura pastoralis, of Orosius' Chronicles, etc. ; Saxon Chronicles ; a 
 Northumbrian version of the New Testament, called The Lindisfarne Gospels, 
 etc. The poetical works (Beowulf = " the bees' wolf," is the most often 
 quoted) appear only in the following centuries. 
 
 2 Thus the vocalic phenomena called " Anglo-Saxonic breakings " (infra 
 21) do not occur in Mercian, wherefore pure English is free from them. 
 
 8 Neither Irish nor American-English are true dialects, but mere varieties 
 of classical and official English, as introduced by relatively recent conquest 
 or immigration. 
 
 r OF THE r \ 
 
 university) 
 
10 ENGLISH AND GERMAN GRAMMAR. 
 
 strengthened afterwards by the political supremacy of the 
 invaders under the reign of Knut the Great (1006). Much 
 greater still were the results of the battle of Hastings (1066), 
 when the Norman yoke was laid upon England. French be- 
 came the official language, the more so because noblemen could 
 and would speak no other, and even at the time of the Great 
 War against France, the English court was nearly as French as 
 the French court itself. When the national tongue prevailed, 
 there yet remained a considerable stock of French words 
 pronounced with an English accent. 1 Lastly, from that time, 
 and uninterruptedly up to the present, but especially under the 
 influence of the literary revival in the sixteenth century, 
 English as well as French adopted a great many learned 
 terms, borrowed either from French itself, or from Latin, or 
 more lately from Greek. These new words became every day 
 more necessary, thanks to the daily increase of new ideas, so 
 that the present English vocabulary is, to the extent of more 
 than a half, of Romance origin. 2 
 
 In spite of this mixture, English remains a truly, purely T 
 and exclusively Germanic tongue. For the vocabulary of 
 any language whatever, and, up to a certain point, even its 
 syntax, are mere accidents, complacently yielding to every 
 process of admixture. Grammar alone is able to resist externa) 
 influence. Now English grammar, excluding a few derivative 
 types borrowed from Romance, 3 does not show one feature that 
 is not Germanic, inasmuch as all the foreign words we now 
 use as native ones have been forced into all the laws of Ger- 
 
 1 Hence, in English, so many doublets, hue and colour, kindred and rela- 
 tions, husband and spouse, etc., sometimes with a slightly varying sense, as 
 sheep and mutton, shape and form, fulness and plenty (=O.F. plente = Jj. 
 plenitdtem). The whole question must be left here, and the reader referred 
 to Skeat's Principles (II., pp. 3-248), and Behrens, Franzos. Elem. im Engl. t 
 in Paul's Grundriss, I., p. 799 sq. 
 
 1 Scientific language, in particular, admits many liomance and Latin 
 elements ; but literature, and even poetry, by no means exclude them, much 
 less so the common speech.. In a recent letter from London I counted IS 
 Germanic and 12 Latin words, that is, among the nouns, adjectives, and 
 verbs, since, of course, the prepositions, conjunctions, and pronouns are all 
 Germanic. Counted without this distinction, the words of the Lord's Prayer 
 are but four French to forty-four Germanic words. 
 
 3 As, for instance, the feminine suffix -ess, infra 69 and 87. 
 
INTRODUCTION. 11 
 
 manic accidence. 1 Whilst the latter has been kept unchanged 
 in its general outlines, the former elements are but moving and 
 fortuitous atoms, which should never be taken into account in a 
 scientific comparison of English and German. 2 
 
 (6) German is geographically far less extensive than English. 
 It comprises hardly more than the greater part of the German 
 Empire, apart from the French, Danish, and Polish districts,* 
 three-fourths of Switzerland, the German provinces of Austria, 
 and in part the Baltic provinces of Russia. In the New World,, 
 however, especially in the United States and Brazil, prosperous 
 centres of immigration have given rise to conspicuous groups of 
 German-speaking population. 
 
 In regard to dialects, Northern Germany being occupied by 
 those of Low German, High German proper, with which alone- 
 we are dealing, is divided into two groups : Central or Frank- 
 ish-Saxonic, and Southern or Alamanno-Bavarian. The- 
 Central group, again, comprises Saxonic, Thuringian, High and 
 Middle Prankish.* The Southern dialects are Alamanno- 
 Swabian (Wurttemberg, Baden, Alsace, Switzerland), and 
 Bavarian, including Austrian German. Among all these 
 Saxonic is well known to have exerted a prevalent influence 
 upon literary German, since the time of Luther and the 
 diffusion of his Bible. But the earliest documents bequeathed 
 to us from mediaeval High German belong for the most part to 
 the Alamannic group. 5 
 
 1 Thus we say the barber's shop like the father's house, he save-d like he 
 Uve-d, declining and conjugating the borrowed term as though it were a 
 thoroughly English one. 
 
 2 One instance will suffice. Considering the identity of the two words for 
 butter (G. butter), who would doubt them to be related through the parent 
 speech ? Yet it would be an egregious error ; for, if they were, they would 
 be less alike : E. butter requires a G. 'butzer or 'busser (cf. E. water = G. 
 wasser) ; to G. butter would correspond E. *bodder (cf. E. j "odder = G. j 'utter). 
 So, as a matter of fact, the two words have been formed and separately bor- 
 rowed, in each language, from the L. butyrum. 
 
 3 German, indeed, is the official tongue of all these countries, as it is like- 
 wise of the Northern provinces of Germany, in which the current speech is 
 Plattdeutsch. 
 
 4 Niederfrankisch belongs to Low German. 
 
 5 Of Schwdbisch we have but scanty remains, but a fair number of Alarn- 
 annisch, namely, several collections of glossaries (eighth century), the 
 Dominicans' Bule, the Hymns (ninth century), and the works of Notker, a 
 
 ^ 
 
12 ENGLISH AND GEKMAN GRAMMAR. 
 
 In its historical evolution, German, like English, is successively 
 assigned to three periods. Old High German (Althochdeutsch) 
 begins with the most ancient texts, dating from the eighth 
 •century, and ends with the end of the eleventh. These consist, 
 firstly, of glossaries of Latin words with brief comments, and 
 little charters, such as the famous oath taken at Strassburg by 
 the sons of Louis le Debonnaire (842) ; 1 later, of short frag- 
 ments, such as the mutilated epic called Hildebrandslied,* and, 
 lastly, of long formal works, viz. Tatian's Evangelical Harmony, 
 Otfrid von "Weissenburg's Evangelical Poem (High Frankish), 
 •considerable extracts from the Bible, and the Alamannic docu- 
 ments mentioned above. According as final syllables are 
 weakened, 3 and the process of metaphony 4 gradually gains 
 ground, Middle High German (Mittelhochdeutsch) appears 
 with features more and more distinct, and then lasts, like our 
 Middle English, down to the year 1500. In this language were 
 written the famous poem of the Nibelungen (twelfth century), 
 and the songs of such Minnesinger or troubadours as assembled 
 ;at the poetical and perhaps legendary meeting of Wartburg in 
 Thuringia (1206 ?) The Saxonic translation of the Bible opens 
 the era of Modern German, which, however, in forms or 
 constructions, differs very slightly from its immediate prede- 
 cessor. Of course, the words have become a little shorter ; 
 some final syllables, already half-mute, have been entirely 
 dropped ; and the grammar is simplified, while, on the other 
 hand, the syntax displays new means for the expression of new 
 ideas ; but, on the whole, the language is still one and the 
 same, so that, at least, in this primer, it will seldom be neces- 
 
 monk of St. Gall, who died in the year 1022. This dialect, through an 
 -erroneous extension of Grimm's second law (infra 47), was once taken for 
 High German proper (Strengalthochdeutsch), and termed by this name. This 
 view has now been abandoned. 
 
 1 But the manuscript of the historian Nithard, who has preserved it for 
 us, is not earlier than the eleventh century, or the tenth at most. The dialect 
 is the Frankish spoken along the Rhine (Rheinfrankisch). 
 
 2 A curious but corrupt text, because High and Low German have been 
 interwoven in its transliteration. The poem of the Saviour (Heliand), a short 
 "time later, is written in old Saxonic, that is, in pure Low German. 
 
 8 O.H.G. demu = M.H.G. deme dem=G. dem is a good illustration of this 
 gradual change. But see below, the examples in 19. 
 
 4 For German metaphony [Umlaut) the reader is referred to 22 infra. 
 
INTRODUCTION. ] 8 
 
 sary to quote the Middle German form as an intermediate link 
 in the clear and obvious genealogy traced from High to Modern 
 German. 
 
 German, as may be understood from what we have stated, is 
 also a mixed tongue, but not nearly so much so as English ; for 
 it has grown on its own soil, and thus it neither borrows words 
 from a conquered race, nor receives any from the foreign influ- 
 ence of a conqueror. Its vocabulary, it is true, contains some 
 Low German terms, always easily distinguishable ; Roman 
 civilization has poured a great many into it ; the Southern 
 dialects infused into the softness of the Central pronunciation 
 a shade of their own rudeness and energy, and contributed on 
 a large scale to enrich the general tongue ; x but the main stock 
 remains, nevertheless, Central German, or rather Saxonic, such 
 as is spoken, more or less purely, by all educated people, from 
 Gottingen to Konigsberg, from Berlin to Bern or to Vienna. 
 Further, in regard to literary or scientific words, drawn from 
 Greek and Latin, it could not help following the fashion of all 
 civilized nations ; and yet it has adopted fewer of them than Eng- 
 lish and French, preferring — whether rightly or wrongly — such 
 as its native genius could afford. 2 More recently, as the French 
 literature and fashions of the seventeenth and eighteenth cen- 
 turies gained in Germany an almost overwhelming vogue, the 
 language of the higher classes was turned into a macaronic cant 
 of French words supplied with Germanic endings, of which many 
 have survived to the present time. 3 But these have been lately 
 
 1 Thus, the diminutives in -chen (L.G. -ken=E. -kin), belong to Central 
 German; those in -lein (O.H.G. -lin, Swiss -li), are taken from the Southern 
 dialects. Literary German now admits of both. Cf. infra 103 (IV). 
 
 2 A German would say, for instance, eindruck, ausdruck, gleichung, wasser- 
 stoff, and even fernsprecher, whereas we use the words impression, expression, 
 equation, hydrogen, and telephone. Scientific terms thus manufactured are 
 clear enough to a German child, but perplex foreigners, and do nothing to 
 advance that desideratum of the future, an international scientific language ! 
 
 3 Hence have come such words as genie, mode, marsch, marschieren, par- 
 lieren, rasonnieren (the latter two somewhat contemptuous), altered some- 
 times by fanciful and popular etymology. Thus, abenteuer (M.H.G. aventiure, 
 " a tale," an old borrowing from F. aventure), sounds to the illiterate ear like 
 abend-theuer, " dear to eventide " (because evening companies are fond of 
 story-telling), and F. trottoir (footway), has become Berlin, tretoir or trittoir, 
 as though from the verb treten (to tread), er tritt (he treads), etc. 
 
14 ENGLISH AND GERMAN GRAMMAR. 
 
 proscribed, and are disappearing, under a perfectly justifiable 
 sense of national independence, wbicli, however, is not altogether 
 free from a certain ridiculous exaggeration. 
 
 (7) Our comparative examination of English and German 
 must comprise four parts ; the first devoted to Phonetics or 
 the Science of Sounds ; the other three, to Morphology 
 or the Comparison of Forms. In other words, after we 
 have compared the vowels and consonants of the two languages, 
 we have to contrast in both the Formation of words, then 
 the accidence of declinable words (nouns and pronouns) or 
 Declension, and lastly, the accidence of conjugable words 
 (verbs) or Conjugation. Syntax, as already hinted above, is 
 only an accessory part of Grammar properly so-called. 
 
FIRST PART. 
 SOUNDS. 
 
 (8) The sounds of every human language are in a 
 state of perpetual change. The son's speech is never quite 
 the same as the father's ; but the difference is so slight that 
 they can scarcely be aware of it, else the father would try to 
 correct the son. It grows greater between grandfather and 
 grandson; but these still understand each other pretty well, and 
 are conscious of uttering the same sounds. This, however, is 
 not at all the case between more widely-separated generations, 
 a,s sooner or later the gap must appear, and, after five or ten 
 generations, the ancestor and his offspring would no longer 
 understand one another. But, let the changes be ever so slow 
 or so fast, it is clear, to any one who has ever considered a 
 language at two distinct stages of its evolution, that they all 
 obey or follow some natural and consequently constant laws. 
 This is as much as to say that, if a given sound or group of 
 sounds, in a given position, has undergone such or such muta- 
 tion, it must needs have undergone the same change in every 
 word (of the same language) wherein it happened to occupy the 
 same position. 
 
 Common sense itself forces this conclusion upon us. Thus, 
 supposing an English group, e.g. initial fl, and an individual 
 physically incapable of pronouncing it correctly, and compelled 
 to replace it by something like fy, 1 he will, of course, not be able 
 to pronounce it in any one word more than in any other, and, 
 if he says a fyower, he ought to also say the fyoor. Briefly, this 
 man's tongue will never admit of an initial fl, at least before 
 
 1 Compare Italian fwre and ¥ . fieur =Jj. florem. 
 15 
 
16 ENGLTSH AND GERMAN GKAMMAE. 
 
 a vowel o. In his children's speech this peculiarity might be 
 neutralized by some external influence ; otherwise, an English 
 dialect will arise from him, in which every initial fl. is replaced 
 by fy. This necessary consequence we express by saying that, 
 in a language absolutely pure, the laws of sounds are 
 constant. In other words, a given sound, in one and the 
 same position, cannot give rise to two different sounds. 
 
 But, as we have been taught by the history of English and 
 German, no language in the world can claim absolute purity, 
 and none is free from some foreign admixture. Side by side with 
 the individual who pronounces fyower and fyoor, lives another 
 who says correctly flower and floor. If the latter's pronuncia- 
 tion also spreads, there will appear side by side two parallel 
 dialects, in one of which initial fl is changed to fy, whereas it 
 is kept in the other. Supposing them to be for ever severed, 
 they would in time become distinct dialects ; but as soon as 
 conquest, trade, or mere proximity brings them into contact, it 
 may easily happen either that the /jyower-dialect will borrow 
 the word floor from the other, or that the floor- dialect will 
 borrow the word fyower from the former ; and thus there may 
 strangely appear, in one and the same dialect, the two contra- 
 dictory forms fyower and floor. In either case, the phonetic 
 law, "fl becomes fy" or "fl is kept unchanged," will seem to 
 be at fault. Now, nothing is more common, in any language 
 whatever, than a similar process of borrowing. Each one of us 
 is subject to it daily, inasmuch as we talk with others, and our 
 language must needs become a kind of compromise with the 
 others that surround us. Hence we see that the constancy of 
 phonetic laws, though theoretically necessary, can never be 
 directly observed in any dead or living language, since a col- 
 lective language is but a fortuitous aggregate of a great many 
 individual speeches. 
 
 The principle of constancy, therefore, must be deemed, 
 above all, a methodic one. It tells us to beware of hasty 
 guessings, specious and arbitrary identifications or fanciful 
 analogies. Soundly interpreted it amounts to this ; we must 
 collect and classify all the phenomena which agree together, — 
 these, at any rate, will be seen to far outnumber the exceptions, 
 
SOUNDS. 17 
 
 — and, when this is done, we must try our best to explain the 
 apparent irregularities. 
 
 Bat we cannot understand how the sounds are transformed, 
 unless we have some knowledge of the means through which 
 they are formed. This is the province of physiological pho- 
 netics, a difficult and delicate science, though its elementary 
 and unquestioned principles may be summarized in a very short 
 chapter. 
 
 (university 
 
CHAPTER I. 
 THE ELEMENTS OF PHYSIOLOGICAL PHONETICS. 1 
 
 (9) Like every wind-instrument, the vocal apparatus may 
 be said to consist of a pair of bellows, emitting a current of 
 air; a sonorous tube, in which the current M air, more or 
 less impeded, causes vibration, and of a sounding-board, 
 by contact with which the volume of the sound is increased. 
 
 The bellows are the lungs. As they can only supply air 
 during the process of expiration, the moments of inspiration 
 are intervals of rest, such as are denoted by punctuation. 
 There are, at any rate in the European languages, scarcely 
 any inspiratory sounds, with the exception of a few instinctive 
 exclamations. 2 
 
 The air expired, escaping through the windpipe, reaches the 
 larynx, which forms a gristly protuberance that can be easily 
 seen at the upper end of the windpipe. The larynx, in its turn, 
 opens into the pharynx by a round aperture called the glottis. 
 The upper margins of the glottis, called vocal chords, are 
 hard and elastic, and, by contracting, are able to oppose an 
 obstacle to the current of air, and consequently vibrate while 
 it is passing through. 
 
 The sounding-board consists of the double cavity of the 
 mouth and nostrils. The shape and size of this cavity may 
 vary, in such a way as to modify the sound emitted through 
 the glottis, under the influence of three chief factors. 
 
 1. The elasticity of the inner and outer walls of the 
 
 1 The following pages are taken from my Grammar of Greek and Latin, 
 with, however, such slight alterations as were advisable in a Grammar of 
 English and German. 
 
 2 The most obvious example would be the sound of a kiss uttered as an 
 admiratory exclamation. 
 
 18 
 
THE ELEMENTS OF PHYSIOLOGICAL PHONETICS. 19 
 
 mouth, which can be made longer by being narrowed and 
 shorter by being widened. 
 
 2. The action of the soft palate (velum palati). In front, 
 that is, for two-thirds of their extent, the nose and mouth are 
 completely separated by the bony arch of the palate ; but from 
 the pharynx to the nasal cavities there is a communicating 
 passage, which can however be closed by means of a fleshy and 
 movable prolongation of the palate, called very appropriately 
 the " veil of the palate." When, the mouth being at rest, the 
 veil falls like a loose curtain, the two cavities are in free com- 
 munication with one another ; but when the veil rises and rests 
 on the back part of the pharynx, it isolates the nasal cavities 
 and so renders the whole of the upper half of the sounding- 
 board ineffective. The soft palate has a small continuation, of 
 the shape of a grape, called the uvula, which can be perceived 
 in a mirror by opening the mouth wide. The share it has in 
 modifying sounds will be seen below. 
 
 3. The extreme mobility of the tongue, which by resting 
 successively against the soft palate, the back, middle, or front 
 part of the palatal arch, the gums, the teeth, etc., is capable 
 of producing an infinite variety of modifications in the shape 
 of the mouth and its mode of opening. 
 
 Whilst the sounding-board reflects, increases and varies the 
 musical sounds emitted through the glottis, the movements of 
 the tongue and lips, at the same time, produce noises, which 
 may be either momentary and explosive, when the mouth opens 
 or shuts suddenly, or continuous and fricative, when the mouth 
 being almost closed allows the air to escape at any point 
 through a very narrow passage. The musical sounds are 
 the vowels ; the noises, whether accompanied or not by 
 glottal sound, are the consonants. 
 
 Section I. 
 
 THE PRODUCTION OF SOUNDS. 
 
 (io) 1. Before coming into action, the vocal apparatus may 
 be said to rest in an indifferent position, the mouth being 
 
20 ENGLISH AND GERMAN GRAMMAR. 
 
 very slightly open, the soft palate lowered, the tongue resting 
 flat on the bottom of the mouth, and the glottis permitting the 
 air to pass freely through it : in short, the position assumed 
 during deep thought or tranquil sleep. Neither sound nor 
 noise is then produced, although during expiration there passes 
 a gentle current of air, which contains in itself the potential 
 utterance of a vowel. 1 This is the inaudible sound which in 
 certain modes of writing is represented by a particular symbol, 
 e.g. the Greek soft breathing or the French and Spanish h. 
 If the air is expired with more energy and a certain amount 
 of effort, we have the English or German h, what is quite 
 improperly called the aspirated h. 
 
 2. The organs being in the first position, the soft palate is 
 raised and cuts off all communication with the nasal cavities ; 
 at the same time the vocal chords contract and vibrate. In 
 this way a pure or oral vowel is produced, a, i, u, etc. 
 
 3. If the vibration takes place without the soft palate being 
 raised, the vowel is sounded in both cavities at the same time, 2 
 and so we obtain a nasalized vowel, written in French an, 
 in, un, etc. 
 
 4. If the mouth, when in the third position, is closed by 
 means of the lips or the tongue at any point, then the air 
 expired being only able to escape by the nostrils, no oral sound 
 can be produced. The result is a nasal sound, m, n, etc. 
 
 5. 3 The open mouth lets the current of air pass through ; 
 but its passage is impeded by an elastic obstacle, which is 
 displaced, and afterwards returns to its original position with 
 a rapid alternate quivering or trilling sound, viz. the trilling r 
 {infra 13, 1 A). 
 
 1 That is, supposing the position to remain unchanged, a vowel is heard 
 as soon as the vocal chords vibrate. 
 
 2 This can easily be proved by experiment. A looking-glass placed in 
 front of the mouth and nostrils, and protected by a screen from the breath 
 of the mouth, remains clear ajter the pronunciation of o, but not after the 
 pronunciation of the nasalized vowel on. 
 
 3 In this and all the following positions, the soft palate is raised, and 
 consequently the nasal cavity plays no part in the production of sound, 
 except in the case of persons who, from some fault in the structure of the 
 organs, or from idleness in using them, are unable to raise the soft palate, 
 whence they " speak through the nose." 
 
THE ELEMENTS OF PHYSIOLOGICAL PHONETICS. 21 
 
 6. The mouth is open, but the tongue completely obstructs 
 the middle part of it, leaving only the two sides free ; the 
 current of air, being thus impeded, is obliged to split itself up 
 into two portions in order to find an outlet, and vibrates with 
 forcing a passage for itself in the narrow space between the 
 cheeks and teeth. This is the lateral trill I. 
 
 According as the nasal and trilled or liquid sounds are 
 accompanied or not by a slight vibration of the vocal chords, 
 they are said, like all other consonants (infra 7), to be sonorous 
 (voiced) or surd (voiceless). The former ease is by far more 
 frequent ; but a nasal or a liquid may become voiceless, when 
 it is either preceded or followed by a voiceless consonant, to 
 the character of which it becomes adapted. 
 
 It is now time to ask whether the various sounds correspond- 
 ing to positions 4, 5, and 6, are consonants or vowels. Of course, 
 a nasal or liquid is a consonant, when preceded or followed by 
 a vowel that combines with it to form a syllable, 1 as in E. note, 
 undo, mare, rare, lame, elbow, etc., and in G. nein, mein, baum, 
 rad, fern, lahm, salz, etc. But let us now consider such very 
 common English and German final syllables as E. even, buxom, 
 sister, middle, etc., and G. rasen, gutem, mutter, mittel, etc. ; 
 and, putting the spelling quite out of account, for in 
 phonetics spelling is but a conventional and deceptive element, 
 let us allow only our mouth and our ears to bear testimony 
 of what the end of these words contains. It is not a vowel 
 followed by a consonant; for we do not pronounce them as 
 tven, boksem, sist&r, mtdel, nor as rdzen, gutem, muter, mitel, 
 but simply as ivn, boksm, sistr, midl, and rdzn, gutm, mutr, 
 mill ; that is to say, the nasal or liquid itself tills the whole 
 syllable and supports the preceding consonant. Hence, in this 
 position, it is a true vowel, and we are led to the conclusion, 
 that the nasals and liquids are alternately consonants 
 and vowels : consonants when they are supported by a vowel ; 
 vowels generally whenever they support a consonant, and par- 
 ticularly when they occur between two consonants (E. fathei'less, 
 
 1 Apart from the well-known (13, 1A) and peculiar pronunciation of 
 E. r. 
 
22 ENGLISH AND GERMAN GRAMMAR. 
 
 G. vergehen); 1 for which reason they are termed here con- 
 sonant-vowels. 
 
 7. Further, if the mouth, when closed at any point, is opened 
 suddenly in order to let the current of air escape, or if, when 
 already opened in order to pronounce a vowel, it then is closed 
 completely at any point and suddenly arrests the current of 
 air, the result is not a sound, but a noise, a momentary con- 
 sonant, called explosive in the former case, and implosive 
 or occlusive in the latter. 2 If this noise is not accompanied 
 by voice in the glottis, the consonant is called surd or voice- 
 less, k, t, p ; if however, while the current of air is passing 
 through, there is a slight contraction of the glottis, together 
 with vibration of the vocal chords, then we have a sonorous or 
 voiced momentary consonant, g, d, b. 3 
 
 8. Lastly, if the . mouth, instead of being completely closed 
 and then opened wide, is obstructed at any point, in such a 
 
 1 It must be again understood, once for all, that the spelling should 
 always be neglected. Conventional orthography is one thing, and pronun- 
 ciation quite another. 
 
 8 Thus, in the group pa, the p is purely explosive, as the closed lips are 
 suddenly opened in order to utter it. In such a group as appa, if the two 
 p's are pronounced, the first is implosive (uttered by closing the mouth after 
 it has been opened for the emission of the a), and the second is explosive 
 (uttered by opening the mouth again in order to pronounce the following a). 
 In the similar group abba, the occlusion and explosion, though less energetic, 
 are likewise quite perceptible. Hence we may infer that, in a group apa or 
 aba, the single p or b is both implosive and explosive. In abma, the b is 
 implosive ; for the lips do not open until the m is sounded. And, lastly, 
 in amba, the b is explosive ; for the lips have been closed, immediately after 
 the a, in order to utter the m. All these distinctions ought to be taken into 
 account by those who wish to obtain a correct view of the reciprocal influ- 
 ences of contiguous sounds in any vocalic or consonantal group. 
 
 3 The reader may prove by experiment the existence of this unconscious 
 vibration of the glottis. First practise the pronunciation of p or b by mere 
 explosion, without letting any vowel follow. This result obtained, if you 
 pronounce p, and at the same time tightly close the ears, no sound is heard ; 
 whereas, if you pronounce b, you will be conscious of a deep rumbling 
 sound. This is the "glottal buzz " or vibration of the vocal chords, which 
 penetrates into the ear through the internal auditory meatus. It can often 
 be perceived even more simply by touching with the finger the protuberance 
 of the throat (Adam's apple). Certain ethnic groups however pronounce 
 the voiced consonants almost without voice : this is the case with South 
 German and Alsatian d and b, which to a French or English ear sound like 
 t and p ; and, as a matter of fact, apart from a certain weakness in the 
 utterance, there is scarcely any difference between a voiceless and a thus 
 unvoiced consonant. 
 
THE ELEMENTS OF PHYSIOLOGICAL PHONETICS. 23 
 
 way as to allow the expiratory current to escape only through 
 a narrow opening in the centre, the air passes through this 
 opening with a noise of friction, which constitutes a con- 
 tinuous, spirant or fricative consonant. According as it 
 is or is not accompanied by glottal vibration, this consonant 
 also is called voiced, e.g. z or v, or voiceless, e.g. s or/. 
 
 Section II. 
 
 CLASSIFICATION OF SOUNDS. 
 
 (n) A brief analysis of the action of the vocal apparatus, 
 reduced to eight main positions, has allowed us to divide all 
 human sounds into four distinct classes : mere expiration (case 
 1); vowels (cases 2, 3); consonant-vowels (cases 4-6), and 
 simple consonants (cases 7, 8). We must now go a little 
 further into particulars. 
 
 § 1. Vowels, Semi-vowels and Diphthongs. 
 
 (12) 1. Oral vowels. — The two opposite poles of vocalism 
 are i (E. ee), the high-toned vowel, and u (E. oo), the low-toned 
 vowel. In pronouncing i, the larynx rises and the corners of 
 the mouth are widened in such a way as to give to the sonorous 
 tube the least possible length ; whereas, in pronouncing u, the 
 larynx is lowered *■ and the lips are thrust forward, so that 
 the length becomes as great as possible. Between these two 
 lies the vowel of equilibrium, a (E. father), the sound which 
 is produced when, the organs being in the indifferent position 
 (supra 10, 1), the soft palate is raised and the glottis begins 
 to vibrate. 
 
 Between these three chief notes of the vocalic scale there 
 is naturally room for a large number of intermediate sounds. 
 Thus we ascend from a to i through open e (E. net, G. netz) 
 and close e (E. ere, G. ehre)) and again we descend from a to u 
 
 1 These movements may be verified by placing the finger on the protu- 
 berance of the throat whilst uttering alternately i and u with some energy. 
 Everybody moreover has experienced the extreme difficulty of singing an ee 
 syllable on a low tone, or inversely the word wood on a high one. 
 
 . 
 
24 ENGLISH AND GERMAN GEAMMAE. 
 
 through open o (E. hot, G. soil) and close o (E. home, G. lohn). 
 The o sounds and the e sounds, in their turn, are connected 
 together through the intermediate sounds of G. b, E. u in a close 
 syllable (but), and English, German or French so-called e mute. 
 Lastly, if the larynx takes the position required for i, while 
 the lips are placed in the position required for u, we shall hear 
 the mixed sound represented by G. u or F. u, which does not 
 appear in the English pronunciation. 1 
 
 2. Nasalized vowels. — From the description given above (10, 
 3) it will have been understood that to each oral vowel there 
 may correspond a nasalized vowel. Such, however, is rarely 
 the case in practice. French, for instance, has but four nasal- 
 ized vowels. As they do not occur at all in the correct pro- 
 nunciation either of English or German, we need not dwell 
 further upon them. 
 
 3. Diphthongs. — Let us consider such a group as G. ai, 
 and inquire of what it consists. Not of two vowels, certainly ; 
 for it is not sounded a-i in two distinct expirations, but simply 
 ai in one syllable. In other words, the a in it is a vowel, but 
 the i is not a vowel, since we miss in its utterance the expira- 
 tory effort described above (10, 1) as the necessary condition 
 for the emission of a true vowel. The same is to be observed 
 of the u in the G. group au. This attenuated i or u, which 
 forms only one syllable with the preceding or following vowel, 
 is what we call a semi-vowel. Semi-vowels, in a consistent 
 phonetic spelling, are written respectively y and w, thus ay, aw, 
 and ya, wa. The real combination of a vowel and a semi- vowel 
 into one syllable constitutes a diphthong. 2 English has a great 
 many Diphthongs ; German, perhaps even more. 
 
 4. Long and Short vowels. — Every vowel, whether oral, 
 
 1 Here in particular the student must not be deceived by appearances and 
 spelling : each language uses no more than five or six vocalic symbols, more 
 or less diversified by accessory signs ; but there is not a language in the 
 world •which has not, at the very least, ten vowels ; and English, German, 
 and French have many more, both short and long. 
 
 2 Real, because it may be only apparent and due to conventional spelling : 
 thus, E. ai and G. ie are not diphthongs, since they are merely sounded as 
 e and i ; -whereas the Bingle i is a true diphthong in our word fine, which is 
 pronounced fayn. 
 
THE ELEMENTS OJF PHYSIOLOGICAL PHONETICS. 25 
 
 nasalized, or in a diphthong, may either be uttered very quickly 
 or prolonged through the whole of a single expiration : hence 
 an indefinite number of degrees of quantity, which may be 
 easily observed in language. For the sake of simplicity, gram- 
 marians have reduced these varieties to two, short and long, 
 thus a and a, and have also agreed to regard the duration of a 
 long vowel as about twice that of a short one. 1 
 
 § 2. Consonant-Vowels. 
 
 (13) I. Liquids. — A. "We have seen the medial liquid r to 
 be caused by a vibrating obstacle interposed to the passage of 
 the expiratory current. This obstacle may be either the mar- 
 gins of the glottis, or the uvula, or the tip of the tongue. 
 Glottal r, however, is very rare in European languages. 2 But 
 uvular and lingual r occur very often in all of them. The 
 latter is the English r as it is sounded when correctly trilled at 
 the beginning of a word or between vowels (raven, caring) ; but, 
 when final (fair, sister), or before a consonant (earth, careless), 
 it is untrilled, and becomes a vague vocalic sound, easier to 
 reproduce than to describe. Uvular r is produced by the uvula 
 vibrating against the back of the tongue, as in the Northum- 
 brian burred r. German r generally appears, with many local 
 and individual exceptions, lingual in the Northern provinces 
 and uvular in the Southern. If, when final, it happens to be 
 untrilled, it may become an uncertain vocalic sound approach- 
 ing to a. 3 
 
 B. In-order to form the lateral trill 7, the tongue usually 
 
 1 French sometimes denotes the long vowel by a circumflex accent. Eng- 
 lish has no particular sign, but often doubles a medial consonant in order to 
 indicate that the preceding vowel is short. German occasionally uses the 
 doubled vowel (saat = E. seed), or writes an e after an i (viel, vieh), or intro- 
 duces an h (zahm, sohn). But none of these conventional and irregular spel- 
 lings have any etymological value. (O.H.G. sat, filu, fihu, zam, sunu). 
 
 2 But frequent in the Semitic languages, as in Modern Arabian. In some 
 dialects of German Switzerland the pronunciation of the deeply guttural ch 
 produces a rasping of the throat which much resembles glottal r. 
 
 8 Thus I have seen quoted the curious word fyamenna, heard at a railway 
 station as the train stopped. Though it could be^eftei ly mi staken for an 
 Italian word, it is merely the G. fur manne7^^pj&s&i4e&~ ifi "thalj affected 
 way. f ^, ?Z2*J: . ^ ,, 
 
 ^ 
 
26 ENGLISH AND GERMAN GRAMMAR. 
 
 rests against the sockets (alveoli) of the upper teeth. This is 
 the dental or, better still, alveolar I. Now, the tongue may 
 also rest against some inner part of the month. This is mainly 
 the case when the I is followed by another consonant, whence it 
 is modified to a deeper sound rather akin to that of a w : thus, in 
 the English words false=-\i. fdlsum, falcon=Jj. falconem, etc., 
 the I has caused the preceding vowel to change to an o sound. 
 So also in their French equivalents faux, faucon, whereas in G. 
 falsch and falke the a and the I are sounded separately and 
 distinctly heard. 
 
 2. Nasals. — The complete closure "which determines the 
 utterance of a nasal may take place at any point whatever of 
 the cavity of the mouth. If in front, and through the lips 
 joining together, we have the labial nasal, m ; if through the 
 tip of the tongue resting against the upper teeth or the upper 
 sockets, we get the dental or alveolar nasal, n ; if again, 
 further back, through the root of the tongue resting against 
 either the hard or soft palate, the nasal becomes either palatal 
 or velar, both of which are often included under the common 
 and inaccurate, but convenient name of guttural nasal, fi, 
 which is simply written n in English and German (E. pink, ink, 
 ringing, tongue ; G. schlank, henker, schwingung, zunge, etc.). 
 
 § 3. Consonants. 
 
 (14) 1. Explosives. 1 — The closing of the mouth, necessary 
 for the production of a voiceless or voiced explosive, may also 
 take place at various points. With the lips closed, and then 
 opened, we get the labial explosive, p, b ; with the tongue ex- 
 ploding against the teeth or sockets, the dental or alveolar, t, 
 d 2 ; against the hard palate, the palatal, k, g, sounded as in 
 king, gift ; against the soft palate, the velar, q, g, sounded as 
 
 1 According to the general custom of English grammarians, this term is 
 henceforth applied to all momentary consonants whatever, whether ex- 
 plosive or implosive. 
 
 2 The tongue may also be slightly rounded, in such a way as to touch the 
 dome of the palate : then the consonant becomes cacuminal. Thus, t and 
 d are rather cacuminal in the standard English pronunciation, whereas they 
 are strictly alveolar in German and French. 
 
THE ELEMENTS OF PHYSIOLOGICAL PHONETICS. 27 
 
 in the G. words huh, gunst ; the latter two classes being like- 
 wise included under the less precise appellation of gutturals 
 with a more or less deep utterance. 1 
 
 2. Spirants. — Imperfect closure of the mouth may naturally 
 vary in position as much as complete closure. If the lips are 
 half closed, we hear the labial spirant (either bi-labial, or 
 denti-labial), voiceless/, voiced v. Between the teeth is uttered 
 the interdental spirant, voiceless }>, voiced b (E. th, re- 
 spectively in thin and this). Against the sockets, the dental 
 or alveolar, voiceless s, voiced z. Against the upper part of the 
 palatal arch, the cacuminal spirant, voiceless s (E. sh = Gr. 
 sch), voiced z (E. z in glazier). In the back part of the hard 
 palate we have the palatal spirant, either voiceless (G. ch in 
 ich, blech), or voiced (G. g in wiegen, luge); and, against the 
 soft palate, the velar spirant, either voiceless (G. ch in doch = 
 E. though, nacht, hoch), or voiced (G. g in tage, gelogen). 2 We 
 may conveniently unite the two latter under the common term 
 of gutturals, and denote them by the same symbol, h for the 
 voiceless, 3 and g for the voiced, whether palatal or velar. 
 
 3. Modifications of the Consonants.— The two chief 
 possible modifications of the consonants are aspiration and 
 palatalization (French mouillement). 
 
 A. Aspiration affects scarcely any but the explosive con- 
 sonants. It consists in the explosion being more energetic, 
 and accompanied by the forcible expiration 4 which we have 
 designed by h. Hence the consonants thus modified are denoted 
 in phonetic spelling by ph, th, Teh, qh (voiceless), and bh, dh, gh, 
 gh (voiced). These aspirates can hardly be said to occur in 
 
 1 It has already been observed that the Southern Germans show a general 
 propensity to confuse, in all these classes, the voiced consonants with the 
 voiceless. 
 
 2 This difference of pronunciation in the ch and the g depends, as is well 
 known, on the nature of the preceding vowel. In standard German, how- 
 ever, the g is never sounded as a spirant, save in the final syllable of words 
 ending in -ig. 
 
 8 The reader must not confuse the expiratory h, which only occurs as 
 initial (O.H.G. hus = G. haus), with the medial h written for the guttural 
 spirant (O.H.G. naht = G. nacht). 
 
 4 Thus for these consonants also (see above, 10, 1) the term " aspirate " 
 is very inappropriate ; but this terminology being consecrated by usage will 
 be retained. 
 
28 ENGLISH AND GERMAN GRAMMAR. 
 
 Modern English. German has no voiced aspirates, but its 
 voiceless explosives almost always, and chiefly when initial, 
 receive some more or less energetic aspiration : thus pabst 
 (pope), nearly as phdpst. Sometimes it appears even in writ- 
 ing : thun = H. (to) do, that = ~E. deed, thal = ~E. dale, etc. Kh, 
 not a mere k, is heard in kind (child), and the initial of Jcuh = l&. 
 cow is a real qh. 
 
 When the explosion of the consonant gradually coalesces 
 with the expiratory breath which follows it, the two sounds 
 end by uniting into one, that is to say, into the corresponding 
 spirant. Thus the transition is easy from ph to pf and /, from 
 th, either to tj> and f> (E. th), or to t$ (G. z) or s (G. sz, ss) x ; and 
 the initial aspirate guttural of G. kind and kuh has become a 
 decided spirant in the Alamannic dialects of Switzerland. 
 
 B. Palatalization consists in a consonant (chiefly I and n) 
 coalescing into one sound with a following y. This phenomenon 
 is well known in the Romance languages ; but, apart from its 
 influence in causing met aphony, 2 it may be said to have very 
 little importance in the Germanic family, and none at all in 
 modern English or German. 
 
 1 This observation is of the utmost importance in the history of Germanic 
 languages, inasmuch as nearly the whole system of Grimm's laws (infra 47 
 sq.) is based upon it. 
 
 2 Or vowel-mutation, infra 22. 
 
CHAPTER II. 
 VOWELS AND DIPHTHONGS. 
 
 (15) A chapter devoted to the study of English and German 
 vowels and diphthongs must naturally also include the semi- 
 vowels so far as they form a diphthong with the preceding 
 vowel. As for the semi- vowels by themselves, it will be more 
 convenient to take them with the consonant- vowels, with which 
 they will be seen * to exhibit some remarkable analogies. 
 
 If the object of our inquiry has been made sufficiently clear 
 in the Introduction, the reader will understand that our pre- 
 sent task is to examine in detail the vocalic systems of English 
 and German, to contrast and compare them, in their historical 
 and prehistorical stages, either with one another or with the 
 vocalism of other Germanic languages, and so to trace them 
 back to that Pregermanic system from which they are de- 
 scended ; and then, this Germanic unity being restored, to carry 
 our inquiry, through it and the other Indo-European groups 
 (Sanskrit, Greek, Latin, Slavonic), up still further to the 
 original vocalism of the Indo-European family. Thus, we 
 begin with the latest phenomena. 
 
 Section I. 
 
 VOCALIC LAWS IN ENGLISH AND GEEMAN. 
 
 (16) Under this name we comprise all the causes which are 
 known to have altered the vowel systems of the two languages, 
 either recently and within the historical period, or earlier in 
 
 1 In the section of Vowel-Gradation, infra 43-45. 
 
 29 
 
30 ENGLISH AND GERMAN GRAMMAR. 
 
 West Germanic, and even up to the Pregermanic stage in 
 which they are brought to agreement with Gothic and Old 
 Norse. 
 
 § 1. Hecent Vowel Change. 
 
 (17) I. All who have any acquaintance with the two 
 languages must have observed at the first glance that their 
 near likeness, though obvious when they are written, vanishes 
 strangely enough as soon as they are spoken. If we put a 
 page of English, with a German translation, under the eye of a 
 student ignorant of both languages, he would immediately point 
 oat several pairs of similar words. But, if we were to read a 
 few English lines to a German hearer, he would not under- 
 stand a single word. We need only contrast together such 
 forms as are almost identically spelled : E, fare, G. fahren ; E. 
 bare, G. baar ; E. even, G. eben ; E. slide, G. schlitten ; E. dumb, 
 G. dumm; E. maid, G. maid; E. (he) was, G. (er) war, etc. 
 Thus, even when the written symbol betrays clearly enough 
 the original vocalic identity, the pronunciation actually dis- 
 guises it in such a way as scarcely to let it appear at all. 
 
 Continuing the examination we soon perceive that English 
 is especially responsible for this discord. For the German 
 vowels are still pronounced with nearly the same sound they 
 once represented in the Latin alphabet, from which they are 
 borrowed as in English and French. The a is a Latin a, 
 whether short or long; the e is a Latin e, either open or 
 close ; and the u, as in many Romance languages, has pre- 
 served the primitive value which corresponds to our double 
 symbol 00. Now, to the same letters English assigns quite 
 different values : our phonetic a is called e ; our e, I ; our u, 
 yu; and so forth. The conclusion lies near at hand : English 
 has shifted its vowels at a time when its spelling was 
 already fixed. 1 These alterations are so various, so delicate, 
 
 1 This would be plain enough from the contrast with German, even if 
 the history of both were quite unknown. But history throws a new light 
 on the results of comparison. We can ascertain from indubitable evidence, 
 and particularly from the study of rhyme and assonance, that Old English 
 and even Chaucer's English were very nearly pronounced as they were 
 spelled. Cf. Skeat, Principles, I., pp. 24 sq., 51 sq. 
 
VOWELS AND DIPHTHONGS. 31 
 
 and sometimes so capricious, through the infusion of so many 
 dialects into the literary language, 1 that we are compelled to 
 refer the reader to special grammars or, better still, to common 
 use, the only true guide in details. The general outlines, how- 
 ever, may be here briefly and conveniently sketched. 
 
 1. A. — Long a in a close syllable 2 usually remains a {far, 
 hard), whereas, in the same position, a whether primitively 
 short or lately shortened, 3 verges on the sound of open e {bag, 
 cab, bath, to have).* But long a in an open syllable is almost 
 everywhere — excluding father, rather, etc. — changed to a close 
 long e, so long and so close indeed that it is now followed by a 
 slight i sound, which forms with it a kind of diphthong : thus 
 cave is pronounced nearly as keyv 5 ; so also E. knave = G. knabe ; 
 E. to lade — G. laden, etc.; occasionally even in some close 
 syllables, as in E. haste = G. hast (both borrowed from M.F.). 
 
 Before or after ic, or before a consonantal group which begins 
 with Z, 6 a, whether short or long, is influenced of the conso- 
 nant, and, being shifted one degree lower in the vocalic scale, 
 assumes a duller sound almost identical with open o ; while the 
 following w or I becomes more or less blended with the vowel : 
 E. draw = G. tragen, E. {he) saw = G. {er) sah, E. wash = G. 
 waschen, E. water (long open o) = G. wasser (short and pure a), 
 E. all = G. all, E. fall = G. fallen, E. false = G. falsch (cf. L. fal- 
 sus), E. balk (the I is quite mute) = G. balken, etc. In a few 
 
 1 Cf. supra 4, 5, 8, and infra 21. 
 
 8 A syllable is said to be open when it ends in a vowel, and close when 
 ending in a consonant. In other words, a syllable is close when final and 
 ending in a consonant (fur), or when medial, if its vowel is followed by two 
 consonants [farther). The only exception is when the two consonants are 
 an explosive followed by a liquid, as in table, the I being then really a vowel. 
 
 3 Cf . infra 20. The reader is advised to note such references and even- 
 tually to multiply them himself. Phonetics suppose the knowledge of a 
 great many laws, which complete or contradict each other : they cannot be 
 well understood if learned by fragments. Here, for instance, the a of hard 
 (cf . G. hart) is long only because it is followed by rd ; and, on the contrary, 
 the a of bath (cf. G. bad, and E. to bathe) has been kept short only because 
 it stood in a close syllable. 
 
 4 The final e is quite mute ; hence the syllable is actually a close one. 
 
 5 Here the e is also mute. But it was sounded in Middle English, at 
 least in the plural canes (pr. kaves, now keyvz) : thus the syllable ca was 
 open. 
 
 6 Cf. supra 13, 1 B, and add F. cheuaux (horses) = O.F. chevals =li. cabal- 
 los. 
 
32 ENGLISH AND GERMAN GRAMMAR. 
 
 cases before I, the vowel was even shifted further to a pure o, 
 and so spelled: E. oZd = O.E. did eald = G. alt; E. cold = O.E. 
 cald ceald=G. halt; E. (to) hold = O.J&. haldan = G. halten. 1 
 
 2. E. — Accented e in an open syllable becomes i, which, when 
 long, is diphthongized almost to iy : E. mere = G. meer (sea); 
 compare the unshif ted pronunciation of the same e in the close 
 syllable of the compound word mer-maid. Still more so, of 
 course, with e, whether long or lengthened, usually spelled ee 
 or ea : E. see — G. sehen 2 ; E. to breed = G. bruten (cf . the corre- 
 sponding nouns brood and brut); E. to speah — O.lE. specan = G. 
 sprechen; E. queer = G. gwer (oblique, transverse, awry). Short 
 e, even when unaccented, assumes the t sound in monosyllables 
 (he, me, we), and a sound akin to i in the prefixes in an open 
 syllable (be-fore, be-cause, cf. G. 6e-). 3 In a close syllable short 
 E. and G. e (open) have been kept quite alike (net=netz, to set 
 = setzen), apart from the peculiar sound imparted to an e or ea, 
 as likewise to an i in English, by a following consonantal group 
 beginning with r (her, serve, person, heard, earth = G. erde, etc.). 
 
 3. I. — E. and G. i correspond pretty well to one another ; for 
 short i in a close syllable remains i in both languages (he will = er 
 will, bit = bisz, to swim = schwimmen, wind=G. wind); and long 
 i in an open syllable, which is now sounded ay in English, is 
 likewise pronounced as ay, though spelled ei, in German : wine 
 = wein (borrowed from L. vinum), by = bei, while =weile (a space 
 of time), mile = meile (borrowed from L. mllle "thousand"). 4 
 Yet not seldom the quantity of the vowel differs in a close 
 syllable before I and n : E. wild (pr. wayld) = G. wild (pr. vild) ; 
 E. to wind = G. winden; E. blind — G. blind, etc., because English 
 has lengthened the i. The gap is also large between the two cor- 
 
 1 Cf. Dutch oud (old), houden (to hold), stadhouder = G. statthalter ; and 
 for the Old English vocalic variations, see below, 21. 
 
 2 The two apparent e'e in the English form are not to be mistaken as 
 representing the two real e's in the German one ; for, in reality, the word 
 see contains but one e, namely the long e of the stem seh-, here transliter- 
 ated by ee ; whereas the final -en has been dropped in English, like all 
 other infinitive endings, infra 19, 2. 
 
 8 But here the i is primitive, infra 19, 1 and 4, and 66, II., 2. 
 
 4 The pronunciation sin, bi (pi), evil (a while), etc., still persists in South 
 Germany (Alamannic), but it is proscribed in classical and official (Central) 
 German. 
 
VOWELS AND DIPHTHONGS. 33 
 
 responding i's when followed by a consonantal group beginning 
 with r. Under the influence of this consonant, E. i has be- 
 come a duller sound, which cannot better be compared than 
 with the sound of E. u in a close syllable, or of G. open 5 : E. 
 birch = G. birke; E. birth = OJE. (ge-)byr&, cf. G. (ge-)burt; E. 
 first = O.E. fyrst = G. fiirst (prince), the E. vowel being the 
 same that is heard in burst = OM. berstan = G. bersten (to burst); 
 circle=-Jj. circulum ; virgin = Jj. virginem, etc. Lastly, final 
 unaccented i, usually spelled y, as in lusty, manly, is well known 
 to stand at an equal distance between % and e (cf. G. lustig, 
 mdnnlich). 1 
 
 4. 0. — Open o is the same in both languages : ox = ochse, horse 
 — rosz, bishop = bischof=h. episcopum. But, as ancient E. a, 
 now pr. e, becomes a diphthong with semi- vowel y, so also long 
 o is now followed by a sound of the semi-vowel nearer akin to 
 it, that is, w the semi-vowel of u, and thus such a word as bone 
 is sounded with a diphthong that could almost be spelled 
 bown. This accessory sound sometimes prevailed in such a 
 way as to change the original o to a decided u : who, move, and 
 even E. gold ( — G. gold) is vulgarly pronounced guld. The 
 same is almost always the case with o, long by nature, its 
 length being now denoted by the spelling oo : E. loose (bor- 
 rowed from Scandian), cf. G. Zos = O.H.G. Ids ; E. hoof—OM. 
 hof= G. huf ; 2 though the original sound is retained in E. floor 
 = G. flur, and E. door = G. thor. The contrast between o 
 (sounded u) and b (spelled u, but sounded o) can be nowhere 
 better seen than in room and the derived verb to rummage. 
 
 5. U. — O.E. u was undoubtedly pronounced like G. u, and 
 even E. u has still this value in several words, especially after 
 a labial consonant, after an r, and before an I: E. put, bull = G. 
 bulle; bush = G. busch, butcher = F. boucher; rule = O.F. reule = L. 
 regulam; full = G. voll, etc. Everywhere else, in an open 
 syllable, u is diphthongized to yu ; but this law concerns 
 
 1 In German also, the i in every position has a mixed sound, more akin 
 to close e than the sound of pure i ; we need but compare the pronunciation 
 of spielen and spinnen. 
 
 2 As for the latter type, both tongues agree in the sound, and vary in the 
 spelling : cf. infra 18, 2. In bolt = bolz, E. has close o. 
 
 D 
 
34 ENGLISH AND GERMAN GRAMMAR. 
 
 scarcely any other but borrowed words 1 (tune, music, suit) ; 2 
 and, in a close syllable, u becomes that indistinct sound (tub) 
 which resembles open 6: B. dung = G. dung; E. funk = G. 
 funke ; E. hut, borrowed from F. hutte, the latter itself bor- 
 rowed from Gr. hiitte; negative prefix E. un- = G. un-, thus un- 
 even = un-eben, etc. 
 
 Besides the u, O.E. had an u, short or long, then written y, 
 and now entirely lost in English, 3 where it is changed to i and 
 always pronounced i, even where it happens to be still written 
 u : thus, busy = OM. bysig, and dizzy = O.E. dysig, have the same 
 vowel in their first syllable, though differently spelled ; further, 
 E. to fill = O.E. fyllan = G. fiillen ; E. otw = O.E. synn = G. siinde; 
 E. pillow = O.E. pyle (pylwe), borrowed from L. pulvinum* etc. 
 Of course this i comes to be sounded either ay or o, under the 
 conditions which thus modify an original i : E. a fo'e = O.E. 
 lygen = G. luge; E. first, see above 3. 
 
 6. Many observations of this kind might be suggested here 
 about English diphthongs, whether true or false, now more or 
 less sounded as long and even short vowels, but still spelled as 
 diphthongs. Thus, the development of ea is quite parallel to 
 that of e, becoming I in an open syllable, and open e in a close 
 one: E. to read = O.E. rcedan, cf. G. reden (to speak), but he 
 ■read == O.E. roedde, cf. G. beredt (eloquent); E. clean = O.E. 
 
 1 Because primitive E. u had already become ow, infra 18, 1. 
 
 2 Hence the pronunciation of u has been confused with that of the group 
 ew (dew = G. thau, sounded like due = F. du, and screw = G. schraube, like ac- 
 crue— F. accru), and consequently some few words have been written with a 
 u, whereas the spelling ew would have been etymologically correct : E. hue = 
 O.E. hlw; E. Tuesday = O.E. Tiwesdccg (the day of the god Tiu, which is the 
 the same as Gr. Zetfs and L. Ju-piter ; G. Dienstag is corrupted by popular 
 etymology, as if it were dingestag, "the day for law business," instead of 
 the regular M.H.G. ziestac, which survives in the Southern dialects, e.g. in 
 High Alsatian tsiUik). On the contrary, the ew; has occasionally prevailed 
 in some cases when we should rather expect a u or oo : E. view = F. vue ; E. 
 he slew = G. er schlug. 
 
 8 Englishmen are well aware of the difficulty in pronouncing correctly a 
 F. u, especially when followed by i, as in pluie (rain). 
 
 * G. « also is much nearer to i than F. u, and in certain dialects it does 
 not differ at all from a pure i. Thus, sometimes even an etymological i is 
 spelled u: G. /itn/=O.H.G. fnf= Go. fimf=O.B. fif=Fu.Jive; here, how- 
 ever, the word is corrupted, cf. infra 121, 5. But compare M.H.G. wiste, 
 now become wiiszte, infra 223, 3, 4. 
 
VOWELS AND DIPHTHONGS. 35 
 
 clcene = G. klein (little), 1 but E. weather = O.E. weder 2 = Gc. 
 wetter, etc. Farther, oa is a false diphthong, since its sound 
 does not differ from close 5, and it therefore alternates with 
 simple o in the transliteration of the long vowel derived from 
 O.E. a: E. 6o»e = O.E. ban, but E. loaf= O.E. hlaf=Go. hldif-s 3 
 = G. laib; E. road = O.E. rod, and, with the same vowel, O.E. 
 rdd, now spelled (he) rode, pf. to the verb n*<2e = O.E. ridan = G. 
 reiten, etc. But it will be better to let the student multiply 
 these instances for himself, 4 and to conclude with the almost 
 superfluous statement that in the whole world there is no 
 language, including French, and excepting only Tibetan, 
 wherein symbol and sound have so much diverged from one 
 another as they have done in English. 
 
 (18) II. With the numerous cases in which English and 
 German show the- same spelling and varying pronunciation, we 
 ought to contrast the equally important cases in which they 
 are spelled differently and sounded alike. In such a case, and 
 provided that the two words compared may be traced back to 
 an earlier common form, both languages, separately undergoing 
 an evolution either parallel or divergent, have come to a similar 
 result, whereupon either or each of them has altered its spell- 
 ing according to its own conventional use of written symbols. 
 
 1. The evolution has run parallel in the two languages. 
 We have seen that every primitive I has been shifted to ay in 
 German as in English : German spells it ei, 5 whereas English 
 retains the symbol i. Primitive u has undergone a quite simi- 
 lar process, with this difference only, that the spelling has been 
 altered in both languages ; it is now a diphthong with semi- 
 
 1 The original meaning is "pretty" (cf. G. klein-od, "jewel"), and then 
 the transition is, in English, " pretty— neat — clean," and in German, 
 " pretty — fine — little." In the preceding example, the G. reden is only 
 quoted for the sake of the analogy in pronunciation ; for it does not corre- 
 spond to E. read, which is equivalent to rathen (to guess). 
 
 2 Here, of course, there is no reason for spelling the word with ea. 
 
 8 Go. final s is the ending of the nominative singular, which has been lost 
 in English and German. 
 
 * The reader may be here referred to the very complete and suggestive 
 statistics in Mayhew's Synopsis of Old English Phonology, and especially to 
 the Appendices, pp. 257-259. 
 
 5 Because, at a certain time, it was actually pronounced eg. 
 
36 ENGLISH AND GERMAN GRAMMAR. 
 
 vow«l w and vowel intermediate between o and a, the whole 
 spelled E. ow or ou, G. au: 1 E. brown = O.E. brun, and G. braun 
 = O.H.G. brun ; E. house, borrowed from Scandinavian hies, 
 and G. 7iatts = O.H.G. Ms, cf. the same u shortened in E. hus- 
 band = Scand. hus-bondi "he who dwells (G. bauend) in a house"; 
 E. town = O.E. tun, 2 and G. zemw (an enclosure) = M.H. G. zun, 
 etc. As, however, vowel lengthening and vowel shortening 
 did not obey precisely the same laws in the two languages, 3 it 
 is but natural that we should occasionally find a u answered by 
 a mere u, as in E. found = * 'fund, instead of G. (ge-)fvnden, and 
 E. pound = G. pfund=^h. jpondo; but inversely E. thumb = G. 
 daumen = O.H.G. dumo, etc. 
 
 2. The evolution, though it has been divergent, has led 
 to a similar result. — The vowel is nearly the same in foot 
 and fusz, brood and brut, and many others. But the two lan- 
 guages have not reached the same point by the same road. 
 The 6 in O.E. fot (Go. fbt-u-s, O.N. fot-r 4 ), now written oo, has 
 been merely shifted to u, as we have seen above. But in Ger- 
 man it was at first diphthongized to uo, M.H.G. vuo^, O.H.G. 
 fuo%, 6 whereupon the semi-vowel o, gradually uniting with the 
 u, finally lengthened it. If, on the other hand, this o has since 
 been shortened in English, we then get the new correspond- 
 ence : E. o = G. u (the latter itself eventually shortened to u) ; 
 E. brother — O.E. bro&or, for G. bruder = O.H.G. bruodar ; E. 
 mo^er = O.E. modor, for G. mutter (#) = O.H.G. muotar ; E. 
 goose, but shortened gosling. Inversely, the 6 has been kept 
 long, and consequently become u, in E. moon — O.E. mona, and 
 
 1 The same dialects which do not diphthongize the i (supra 17, 3) have 
 also kept the primitive u : Swiss pruun = braun, huus = haus ; High Alsatian 
 tr tume (with a long u) — der daumen. 
 
 ' 2 English exhibits in its own dialects exactly the same phenomena of 
 preservation as have just been stated for Southern Germany, namely : in 
 Western dialects (Cornwall) the verb shine = G. scheinen is Sin ; whereas, in 
 Northern dialects (thus John Browdie in Dickens' Nicholas Nickleby), town 
 is sounded as tun (spelled toon) ; and so also, respectively, Wild for child, 
 dun for down, even kit for cow. 
 
 8 Cf. supra 17, 3, and infra 20. 
 
 * O.N. final -r is the ending corresponding to Go. -s, supra 17, 6. 
 
 5 The process of diphthongization is still quite perceptible in Southern 
 Germany, the pronunciation being fues, pruet, puep = bube, muetr = mutter, 
 etc. 
 
VOWELS AND DIPHTHONGS. 37 
 
 the consonantal group has not made it short in G. mond = 
 M.H.G. mande = 0.1L.Gc. mdno; but compare the identical vowel 
 of Monday and Montag. 
 
 (19) III. All the above statements, with the one exception 
 of final E. y, concern only the vowels of more or less accented 
 syllables. The laws of unaccented vowels, though they some- 
 times may seem arbitrary owing to the numerous irregularities 
 in spelling, show a remarkable conformity in both languages, 
 and may be reduced to five main principles. 
 
 1. The unaccented vowel, whether in prefixes or in finals, 
 assumes a dull and vague sound, usually represented by an e : 
 pref. bi-, in Go. bi-gitan (to find), bi-satjan (to beset), bi-saihvan 
 (to look at), bi-hlahjan (to laugh at) ; E. beget, beset; G. besehen, 
 belachen, etc. ; pref. ga- in Go. ga-baurj?-s (birth), ga-juh (pair) ; 
 G. geburt, gejoch, etc. ; Go. haban (to have), haba (I have), E. 
 have, G. haben, habe; Go. fiskon (to fish), fisko (I fish), fiskoda (I 
 fished) ; E. fish, fished ; G. fischen, fische, fisch(e)te, etc. 
 
 2. In the unaccented final -en, chiefly of infinitives and par- 
 ticiples, the n was dropped early in M.E. (have, cf. O.E. habban 
 and G. haben) ; the £ then became absolutely mute, and was 
 either written or omitted, according to the conventional pecu- 
 liarities of English orthography (fish = G. fischen). Thus, com- 
 pare E. to find = OM.findan with G. finden, and E. found = O.Hj. 
 (ge-)funden with G. gefunden. 1 
 
 3. In German as well as in English, if this £ has come in 
 contact with a consonant- vowel, m, n, r, I, it has simply disap- 
 peared, and the consonant has become a vowel, m, n, r, ^, 2 so as 
 to support the syllable : E. oxen, G. ochsen, pr. oksn ; pref. 
 fra-, in Go. fra-liusan-s (lost), E. for-lorn, G. ver-loren (ipr.fr.-) ; 
 
 1 Final n, however, is often kept, as in E. heathen = G. heide, E. maiden 
 (but also maid) = G. magd, in the plural forms children, oxen — G. ochsen, 
 and in a great number of strong participles (bidden and bid, hidden and hid, 
 fallen, known, as opposed to the infinitives fall, know, etc., infra 179-185). 
 Without going into particulars, we may here observe : (1) that final n 
 was sounded when the following word began with a vowel (cf . in Mod. E. an 
 and a) ; (2) that certain dialects (Saxonic) lost the final n sooner than some 
 others (Anglic) ; (3) that, in consequence, the common tongue formed from 
 all these dialects kept, or even restored, the n, wherever it appeared to have 
 a decided grammatical value. 
 
 2 Cf. supra 10, 6. X» £ 
 
38 ENGLISH AND GERMAN GRAMMAR. 
 
 E. mother (pr. mdb"r) = O.E. modor, like G. mutter (pr. mMf) = 
 O.H.G. muotar, etc. 
 
 4. Before any other consonant, e has likewise been dropped 
 in most cases, so that the word has lost one of its syllables : 
 G. bleiben = O.H.Gr. biliban = Go. bi-leiban (to remain); G. glaub- 
 en=O.H.G. gilouben = Go. ga-laubjan (to believe) ; G. begleiten 
 (to accompany) = be-ge-leiten, cf. G. geleiten and Zet£e» = E. to 
 lead; E. gen. sons, G. sohns = sohnes; E. sZepi = O.E. *(gre)sZ*ped; 
 G. gre/wita x = O.H.G. gihabet, etc. And even when it is marked 
 in writing, it is well known to be almost always eliminated in 
 pronunciation, as in E. walked (pr. wokt, just like slept), chiefly 
 in rapid and vulgar speech, thus b , lieve = believe. In fact, it 
 cannot be said to persist necessarily, except when it stands 
 between two consonants of the same order, which could not be 
 sounded at all without some intermediate vowel : E. pi. sons, 
 rats, but kisses, houses ; E. slept, walked, but blotted, mended ; G. 
 geliebt, angeregt, but geleuchtet, geredet. Even in this position it 
 is liable to disappear : E. pi. oaths, paths, months, etc. ; G. gere- 
 det (spoken), but beredt (eloquent). 2 "When, the e being dropped, 
 two incompatible consonants become contiguous, an assimila- 
 tion takes place, as already seen, or the first consonant is 
 entirely dropped ; E. hast=*havest, and G. hast = *habest, etc. ; 
 E. had = *havde and *haved, like E. head = OM. heafod = Go. 
 hdubi]} = O.H.G. houbit = G. haupt, instead of *haubet. 
 
 5. Lastly, even where writing does not denote the neutral 
 character of the unaccented vowel by the use of the symbol e, 
 its dulness and vagueness are quite perceptible in actual pro- 
 nunciation. Thus, though we spell with an o the second 
 syllable of the word buxom = OM. buhsum (flexible, cf. G. bieg- 
 sam), we really pronounce it boksm, with a consonant- vowel, as 
 in G. allem actually sounded dim; and, in spite of orthography, 
 the same vowels are heard in E. thousand and G. tausend. 
 
 1 Actually pr. gehapt, the b being assimilated to the t. 
 8 Thus also geredet becomes kret, for instance, in Alsatian. Here the syn- 
 cope is very early : cf. infra 187. 
 
VOWELS AND DIPHTHONGS. 39 
 
 § 2. Shortening and Lengthening. 
 
 (20) We have just seen English and German vowels to suffer 
 various changes of value, according as they were short or long. 
 But which of them were short, and which long ? Were these 
 the same in both languages ? In other words, is the quantity 
 of each vowel in both now exactly such as it was in their Pre- 
 germanic unity ? No, indeed, since we have seen the same 
 vowel treated as long in English and short in German, and 
 vice versa. Thus, before the period of sound-shifting, 1 there 
 must have been a previous period, during which the Preger- 
 manic vowels were either shortened or lengthened, in English 
 and in German, apart from one another, and according to the 
 different laws which prevailed in each. To this period we are 
 now going back. 
 
 At the outset a great principle pervades the whole evolution 
 of vowel-quantity : whether in English or in German, an 
 accented vowel has a decided tendency to lengthen 
 in an open syllable, and shorten in a close one. Thus, 
 the reader may compare: E. to keep = O.J£. cepan, and E. he 
 kept = O.E. cepte ; sleep and slept, leave and left, read and read 
 (respectively O.E. ruedan and rcedde), lose and lost, etc ; G. ich 
 sage, er sagt ; ich lege, ich legte (I lay, I laid) ; tragen (to bear), 
 tracht (dress) ; moge (may) and macht (the might), etc. ; G. stube 
 (a room; P. etuve, "a warmed room ") = O.H.G. stuba = O.E. 
 stofa = 1 E. stove, etc., etc. 
 
 But, if the law be the same in the two languages, we should 
 
 1 The chronological succession of phonetic facts is always to be taken 
 into account, as being, at least, as important as the facts by themselves. 
 This is a consideration of which the student cannot be too earnestly re- 
 minded. In the study of language, as of geology, every fact bears its own 
 date, if properly observed. Supposing a geological stratum, and a sinking 
 of the ground in the same place, has the former taken place before the 
 latter, or the latter before the former? We know that if the stratum 
 is older than the sinking, it will be broken up like the strata below; if 
 not, it will have remained level. So also, has the E. u of *fund (cf. G. 
 gefunden) been lengthened to u before or after the u of hus had been shifted 
 to ou ? The answer is : before, since this u also has become ou, inasmuch 
 as we pronounce found like house. If the u of "fund had become long after 
 the shifting of u to ou, we should now have the dissimilar forms house and 
 'fund. 
 
40 
 
 ENGLISH AND GEEMAN GRAMMAR. 
 
 expect to find in both the same long 1 and the same short vowels. 
 Such, however, is far from the case, as may be inferred from 
 the last example quoted. Whence arises the difference ? The 
 principle, of course, is identical ; but its effects have been car- 
 ried on separately, in various dialects and various periods of 
 English and of German, and consequently have resulted in the 
 most violent and striking contrasts. 
 
 1. The fact is, that our principle, as given above, does not 
 indicate any precise and positive phenomenon, but a mere 
 general tendency, which was not equally observed by all dialects. 
 In German, for instance, the law of shortening in a close 
 syllable does not belong as such to the common language, but 
 mainly to the dialects of Low German, the peculiarities of their 
 pronunciation having sometimes intruded into the literary lan- 
 guage. Hence many words hesitate between the two quantities : 
 G. genug and geniig, compared with E. enough (the final always 
 short). German shortening appears nowhere earlier, and there- 
 fore more consistent, than before the medial group ht : G. (er) 
 brachte =M.H.G. fcra&te = O.E. brohte = 'Et. (he) brought; and yet 
 even here it is seen not to have taken place before the modern 
 period. English shortening began far earlier. 1 
 
 2. Such later changes in the pronunciation as took place in 
 the two languages may have caused a given vowel to stand in 
 a close syllable in the one, and in an open syllable in the 
 other. Thus, the £ in an unaccented syllable being only 
 slightly sounded or entirely dropped, the preceding syllable 
 varies accordingly. We have the short vowel in he read, he led ; 
 whereas German has the long one in the corresponding forms 
 er redete, er leitete, in which the 8 mute has been retained or 
 rather restored in contrast to O.E. rcedde, etc. 2 So also, in the 
 genitive, E. son's, but G. sohnes, hence sohns, etc. 
 
 3. In the course of declension or conjugation, the radical 
 vowel might occur alternately in a close or open syllable, and 
 
 1 Though in this case the corresponding vowel seems to have remained 
 long, it is only long because the sound of the h (gh) has coalesced with the 
 preceding vowel and lengthened it. 
 
 2 The dialects in which geredet is pronounced kret sound the e short (as in 
 E. viet). 
 
VOWELS AND DIPHTHONGS. 41 
 
 consequently it was, according to rule, in turn short and long. 
 Now grammatical analogy naturally levelled most of these 
 differences, 1 and thus either the long vowel or the short one 
 was carried throughout the whole flexion. German, for in- 
 stance, has everywhere the short one in gemacht, er macht, ich 
 machte, ich mache, machen, whereas English shows everywhere 
 the long one in made (O.E. macode), he maketh, I made, I make, 
 to make, etc. G. glas has the gen. glases, pr. gldses, whence the 
 long vowel also pervades the nomin., pr. glds = TZ. glass. Simi- 
 larly the gen. sohnes caused the nomin. to become sohn 2 instead 
 of O.H.Gr. sunu = Go. s«»u-s = O.E. sttm* = E. son. Thus, in 
 German, the long vowel of the pi. form waren = M.H.G. wdren 
 has been transported to the sg. war = M.H.G. was, whereas, in 
 English, the long vowel in the pi. were = O.E. wceron had no in- 
 fluence upon the short one in the sg. was, because the sound of 
 the latter was different. 3 More instances of the kind the reader 
 may easily discover by himself. 
 
 4. Lastly, some subsidiary laws, especially in English, have 
 modified the original quantity of the vowels. 
 
 A. Before a group no", ns, nf, in Old English, any vowel is 
 lengthened while the nasal disappears : 4 E. us, shortened in a 
 close and unaccented syllable, 5 from O.E. us = Go. uns = G. uns; 
 E. five (shortened in the close syllable of fifth and fifty) = O.T$. 
 fif= Go. fimf= G. funf. 
 
 B. From the earliest period an accented English vowel is 
 lengthened before a group consisting of a nasal or liquid and a 
 voiced explosive. Thus disappear the differences in quantity 
 which have been partly stated above, between E. find, mild, 
 gold, old (O.E. did), word, sound (in health), hound, and G. 
 finden, mild, gold, alt, wort, (ge)sund, hund, etc. Moreover, in 
 
 1 On the effects of analogy and the part it plays in language, see Henry, 
 Grammar of Gr. and Lat., 83 and 183, and below 22, 55, 177, etc. 
 
 2 Pr. zon. The h only denotes that the vowel is long, supra 12, 4. 
 
 3 Thus is explained the double discord in sound and quantity mentioned 
 on the second page of this chapter, For r=«, see infra 61, I. 2. 
 
 4 This is the Old English compensatory lengthening ; cf. infra 24. The 
 phenomenon is the same in Greek (toi>s tirirovs=Tbvs lirirovs) and in Latin 
 \equos = *equo-ns). See Henry, Gramm. of Gr. and Lat., 189, 2, and 206, 3. 
 
 5 Cf. supra 19, and infra, 65, 5, and 66, H. 4. 
 
42 ENGLISH AND GERMAN GRAMMAR. 
 
 Mod. English, an r before any consonant whatever lengthens 
 the preceding vowel by coalescing with it : 1 thus, for instance, 
 compare hard with hart, heart with herz, learn with lernen, hark 
 with imper. horche, sharp with scharf, etc. Even in German, 
 however, we find bdrt (beard), zdrt (tender), erde = earth, and a 
 few others. 
 
 § 3. Old English Vow el- Breaking. 
 
 (21) The English process of lengthening lastly mentioned is 
 bnt one effect of a more ancient and more general cause, which 
 also produced the curious process of the diphthongization of the 
 vowels, called Vowel-breaking (G-. brechung, E. fracture). 
 Though this phenomenon does not properly belong to English, 
 since it seems not to have taken place in Mercian, yet, as it 
 was immensely developed in the Southern dialects (Wessex), it 
 could not fail to find its way somehow into the common lan- 
 guage. 2 Omitting many details, we may briefly summarize the 
 effects of this law as follows : before a consonantal group 
 beginning with r, I or h (including the x = hs), the two vowels a 
 and e became respectively " broken " to ea and eo. Thus, Old 
 English opposes wearm, feallan (Merc, fallan), heord, seolfor 
 (Merc, sylfur), seox, cneoht, neaht, healf (Merc, half), etc., to E. 
 warm = G. warm, E. to fall=G. fallen, E. herd = G. herde, E. 
 silver =G. silber, E. six = G. sechs,~Ei. knight = G. knecht,^!. night 3 
 = G. nacht, E. half= G. halb, etc., etc. 
 
 In all these cases English appears quite free of any breaking. 
 The process is exclusively Saxonic, 4 and the only counterpart 
 of it in classical English is a mere lengthening. A lengthen- 
 ing it is also, most likely, when the a is shifted to in the 
 words quoted above, old, cold, hold, sold, etc., in which Saxonic 
 has the broken vowel (eald, ceald, healdan, seald) instead of 
 
 1 Cf. supra 13, 1 A. 
 
 2 A point already alluded to, supra 4. 
 
 8 In these three words, the vowel i is due to a kind of metaphony (infra 
 22), which is regularly produced by the following palatal consonant. 
 
 4 Hence it follows, as a matter of course, that the Southern dialects must 
 contain a very large number of broken forms. It is mainly for this reason 
 that they so widely differ from classical English. 
 
VOWELS AND DIPHTHONGS. 43 
 
 the pure Mercian vowel (did, cald, said). 1 But the change 
 appears already much greater in worth = O.E. weorb~=Q. werih 
 = Go. vair^-s, E. sword = G. schwert, E. work=&. werh, E. 
 world 2 = O.E. weorold, cf. G. welt = M.K.Gr. weralt 3 ; hence, 
 in these words and some other, we may perhaps recognise a 
 process akin to breaking, if not the breaking itself. The more 
 so with the E. a corresponding to Germanic e in such words as: 
 E. far = O.E. feor, as opposed to Go. fairra and G./er-?i = O.H.G. 
 verr-ana; E. star = O.E. steorre, compared with M.H.G. sterre, now 
 replaced by stern = M.H.G. s<erne = O.H.G. sterno = Go. stairno, 
 etc. Lastly, the breaking is still clearly visible in the spelling, 
 and partly preserved in the pronunciation of such words as E. 
 beard, earth, learn (O.E. leornian), heart (O.E. heorte), hearth, 
 as opposed to G. hart, erde, lernen, herz, herd, and some others. 
 
 § 4. Metaphony (Vowel Mutation or Modification).* 
 
 (22) Our search has now brought us back very near to the 
 Pregermanic period ; but we have not yet reached it. The 
 remarkable process we call Metaphony (G. umlaut) took place 
 in English and in German separately ; yet it is in both so early 
 and general, that it seems impossible not to attribute it to a 
 tendency inherited from West Germanic. Gothic alone is free 
 from it. It is found everywhere else, though at various dates : 
 in English metaphony appears fixed before the earliest written 
 documents of Old English; in German, we can still trace its slow 
 
 1 G. alt, halt, halten, and M.H.G. sal (delivery) = E. sale, still surviving in 
 sal-buch (a register of lands) ; for the vb. sell(en), now used in English in the 
 sense of L. " vendere," has reached it through the sense of " assigning " [a 
 piece of land] . 
 
 2 Here the effect of a preceding w, as described, supra 17, 1, may perhaps 
 have combined with a kind of breaking. 
 
 8 The word would be Go. *vair-ald-u-, and mean " a man's age." For the 
 first part, compare L. vir, and G. wer-geld (compensation for a man, that is 
 to say, for a man's slaughter), G. and E. wer-wolf, etc. The original mean- 
 ing of the word was L. "sseculum," whence it became "world," just as seecu- 
 lum has done in Ecclesiastical Latin, and F. seule in Eulalia's Song (line 24). 
 
 4 Since the usual term " Mutation " is altogether vague, designating any 
 kind of vowel -shifting whatever, I have ventured to introduce into the 
 English terminology the word "Metaphony" for German "Umlaut," as 
 already in my French edition. It seems indeed convenient to adopt a separate 
 term for such " mutations " as are described under this head. 
 
44 ENGLISH AND GERMAN GRAMMAR. 
 
 progress, step by step, from the Old High German, wherein it 
 begins, down to the year 1150, when it is entirely accomplished ; 
 and it may, therefore, be deemed a very convenient distinction 
 between Old and Middle High German. 1 Under this name are 
 comprised the mutations undergone by a vowel, a, e, o, u, when 
 immediately followed by a syllable which contains an i 
 (or its semi-vowel, written j in Germanic), under the influence of 
 which it acquired a slight i- sound, and altered accordingly. 2 
 
 Let us, for an instant, consider only German metaphony, as 
 it is more recent, and, in consequence, clearer ; and, further, let 
 us consider it only in such pairs of words as mann. mannlich, 
 erde irdisch, gott gottliclt, (zu)kunft kunftig, etc., where its cause 
 is quite obvious. Here we may very well understand how the 
 palatal vowel of the second syllable at first palatalized the pre- 
 ceding consonant (supra 14, 3 B), which in its turn modified the 
 sound of the preceding vowel, so as to shift it one degree nearer 
 to i (supra 12, 1). The effect, indeed, is wholly adequate to the 
 cause. 3 
 
 But, at the very outset, an objection must be faced : in con- 
 trast with the numerous cases which actually exhibit metaphony, 
 a great many words might be quoted, wherein a following i has 
 no such effect ; thus gelb (yellow) gelblich, gold goldig, ruhe (rest) 
 ruhig, etc. How can this be ? We must never forget that a 
 language constantly and daily creates new words, whereas a 
 particular phonetic law cannot outlast in any language a given 
 
 1 More accurately speaking, the mutation of S is certainly Teutonic (West 
 Germanic), and perhaps even partly Pregermanic (infra 26, I. 3) ; the muta- 
 tion of a begins in O.H.G. about 750 ; that of u is scarcely at all visible in 
 it ; that of o, as will be seen, is late and due to analogy. Germanists, how- 
 ever, generally agree in thinking that the mutation took place in the usual 
 pronunciation long before it was marked in writing : Wilmanns, Deutsche 
 Grammatik, I. p. 192. 
 
 8 We have also in English a metaphony caused by a following palatal 
 consonant (supra 20), and in German, another caused by a following u (infra 
 26, I. in fine) ; but they need only be mentioned here for the sake of com- 
 pleteness. On the other hand, there are some groups of consonants which 
 impede or seem to impede metaphony. For these particulars, the reader 
 must here be referred to more detailed works. 
 
 8 In some cases, the spelling itself may illustrate the process of meta- 
 phony : thus, metaphonized u is spelled ui or iu in O.H.G., and G. u is well 
 known to be called ui. 
 
VOWELS AND DIPHTHONGS. 45 
 
 period. If, for instance, ruhig has been derived from ruhe at a 
 late date, when the law of metaphony had long exhausted its 
 effect, and if, as is the case, a regularly metaphonical form 
 gulden, still mentioned in every German vocabulary as archaic 
 and poetical, has been currently replaced by a word golden on 
 the analogy of the primitive gold, it is clear that the cause and 
 effect of the phenomenon remain invariable and unquestionable, 
 in spite of exceptions due, in particular cases, to some accessory 
 and accidental causes which have nothing to do with the prin- 
 ciple. Nay, the contrary would appear far more wonderful ; 
 for supposing, for instance, that the language had ever created 
 such a regularly metaphonical compound as *hals-binde (a neck- 
 cloth); since both terms of it, namely, hals and binde, must 
 always remain present to the linguistic consciousness of any 
 speaking subject, he cannot fail, sooner or later, to alter the 
 compound accordingly, and restore the non-metaphonical form 
 instead of the seemingly corrupt one. 
 
 On the other hand, we meet with a great number of similar 
 cases of metaphony, although no i appears in the second 
 syllable : G. lamm (lamb), pi. Idmmer ; G. geben (to give), pres. 
 sg. 2 du gibst ; G. hoch (high), hoher (higher) ; G. buck (book), 
 pi. biicher, etc. ; and compare E. man men, brother brethren, goose 
 geese, foot feet. But if, instead of confining our attention to 
 the modern form of these metaphonical words, we proceed to 
 examine them in O.H.G., or, even beyond O.E., in that Gothic 
 speech, through which we are able to trace them back to Pre- 
 germanic, then we shall, almost invariably, find the second 
 syllable to have contained a primitive i or j, more lately dropped 
 or changed to e in consequence of the laws of final and unac- 
 cented syllables. 1 Thus we shall find, for instance, O.H.G. 
 lamb lembir, geban gibis, hoh hohir, etc. The vowel which has 
 vanished still betrays itself through the mutation it has 
 effected, and the mutation, which seems arbitrary at first sight, 
 is quite satisfactorily explained by the vowel as preserved in an 
 earlier state of the same language. 
 
 Lastly, if we find metaphony taking place in cases where we 
 
 1 See supra 19, and infra 34. 
 
46 ENQLISH AND GERMAN GRAMMAR. 
 
 should not expect it, namely before a syllable that does not and 
 never did contain either an i or a j, even then we need not fear 
 for our theory. For, as non-metaphonical forms may have 
 influenced metaphonical forms and abolished some regular 
 mutations, — thus, ruhig and goldig modelled on ruhe and gold, — 
 so also, of course, the latter may have altered the former, that 
 is to say, Germans might as well have come to say *giild for 
 gold, from gulden, as they have come to say goldig and golden, 
 from gold. Hence arise some false analogies in metaphony, as, 
 for instance: G. thrane (a tear)=M.H.G. trine (but O.H.G. 
 trahan trdn), altered under the influence of its pi. *trini = Old 
 Saxonic trahni ; E. friend for O.E. friond (G. freund), and E. 
 fiend for O.E. fiond (G. feind), both altered on account of the 
 datives O.E. friend = *friond-i, and fiend = *fiond-i. 1 The con- 
 clusion is thus forced upon us, that a metaphonical vowel in 
 either language may very well correspond to a non-meta- 
 phonical vowel in the other, and the case will be seen to occur 
 frequently. 
 
 All that has been said above about German applies, of 
 course, as well, and even better, to English. English also has 
 created new words since the end of its metaphonical period, 
 and in much greater number than German could do, because 
 in English this period ended far earlier ; English also has lost, 
 after metaphony had wrought its effects, the i's and fs that 
 caused it, and, still for the same reason, on a much larger 
 scale than German 2 ; lastly, English has also levelled, during 
 a longer period, the metaphonical and the non-metaphonical 
 forms, in such a way as to cause either the former or the latter 
 to extend throughout a whole system of derivation, declension, 
 or conjugation ; thus gold golden, man manly, god godly, con- 
 trasted with G. gulden, mannlich, gottlich. Besides these con- 
 siderations, English metaphony, less visible in spelling of older 
 date, and partly disguised by the effects of Old English break- 
 ing, cannot be reduced to the exact parallelism exhibited by 
 the almost algebraic formula of present German metaphony, 
 viz : 
 
 1 See declension, infra 139, 1, 5, and 152, 3. 
 
 2 Most of them are already dropped in Old English. 
 
VOWELS AND DIPHTHONGS. 47 
 
 Pure vowels : * a 
 
 Metaphonical vowels : e, a 2 
 
 o 
 (il) o 3 
 
 Whereas English has the following correspondences : 
 
 Pare : a a e 
 
 Metaph. : e ce \ i 
 
 o o 
 
 (broken or ea eo 
 otherwise) ie y 
 
 (23) Some instances of metaphony have been quoted above ; 
 a great many others will occur in the study of forms, so 
 tbat here it will prove sufficient to mention a few mutations 
 selected among the most interesting derivations of words con- 
 taining either an i or a j- suffix. 
 
 A. — G. ameise = O.H.G. ameiffa = O.E. cemette (E. ant); but 
 O.H.G. emiy^ig (busy), M.H.G. em^ic, G. emsig. — G. reltich and 
 O.E. rcetic, 5 borrowed from L. radicem. — G. pfanne and E. pan, 
 borrowed from L. patina ; but, with a derivative suffix, G. 
 pfenning pfennig, E. penny (on account of its being round like 
 a pan ?). — E. Angle (ethnic name), but English ; so also, E. 
 France, but French = *Franhish (cf. G. Frdnkisch), E. Wales, but 
 Welsh (cf. G. Walsch, opposed to Deutsch), etc. — E. a tale, but 
 to tell = 0.1&. tellan for Germanic *tal-jan ; quite as G. zahl 
 (number) and z'dhlen (to count) = O.H.G. zellen = *zal-jan. 6 
 
 1 With barely any distinction between short and long vowels, that is to 
 say, a is mutated to short a, and a to long a, and so forth. 
 
 8 The variation is merely one of writing. O.H.G. has been seen to write 
 e. Mod. German has been influenced by an etymological scruple : tbe vowel 
 is spelled a when another form survives wherein the pure vowel a is pre- 
 served ; but the spelling e is retained if such is not the case, or if the 
 language has forgotten the metaphonical origin of the vowel : thus, nuvnn 
 manner, kalb kalber *= 0, .H.G. chalbir, etc.; but ende = O.H.G. enti = Go. 
 andeis showing the i whence the metaphony arose (cf. E. e«d = O.E. ende, 
 from a Preg. *an$ja- which is akin to Sk. dnta-, "a limit "); thus also, hand 
 hdnde, but in composition behende^ O.H.G. *bi henti, literally "by the 
 hand " (hent-i is a dative case), whence the sense " convenient, easy, swift." 
 
 8 When is mutated to u, as in gulden, the stands for a primitive u, and 
 the mutation has taken place in the early period when the vowel was still 
 sounded as u. Cf. infra 28, I. 
 
 * Long or short, that is to say, u becomes y, and u becomes y ; and simi- 
 larly, ea and eo are mutated, the one to ie, and the other to y. What these 
 again may become in the shifting from Old to Modern English, has been 
 stated above. 
 
 5 E. radish is borrowed from F. radix. 
 
 6 For the sense, compare G. er-zahlen (to relate). 
 
48 ENGLISH AND GERMAN GRAMMAR. 
 
 E. — G. gern (willingly) = O.H.G. g'erno (Go. gairn-s, "cove- 
 tous"), vb. begehren (to wish for) = O.H.G. geron; but gier 
 (avidity) = O.H.G. girl, gierig (covetous), begierig, begierde, etc. — 
 G. feder=~E. feather, but fittich (wing). — G. pilgrim and E. 
 pilgrim, borrowed from L. peregrinum, cf. F. pelerin and Ital. 
 pellegrino. — E. silly = OM. syllic = *seollic (broken vowel) = Go. 
 silda-leik-s (strange, queer); cf. E. seld-om = G. selt-en, G. selt- 
 sam (strange), etc. 
 
 0. — G. gold, but gulden ; O.E. gold, but gylden, the latter now 
 altered to golden, though the metaphony survives in the vb. to 
 gild = 03$. gyldan = *guld-jan. — E. /ore = G. vor; but E. first = 
 O.E. fyrst = *fur-ist; cf. G. furst (prince). — E. fox, but vixen 1 ; 
 cf. G. fuchs and fiichsin. — O.E. dom (judgment) and E. doom ; 
 but E. to deem = O.E. deman = *dom-jan. So also, food and to 
 feed, brood = G. brut, and to breed = Gr. briiten = O.H.G. bruoten, 
 which latter would be Go. *brod-jan, etc. 
 
 . U.— O.E. and E. full, O.H.G. fol and G. voll= Go. full-s ; but 
 O.E. fyllan and E. to fill, G. fullen — O.H.G. fullen = Go. full-jan. 
 — E. dizzy = 0.E. dysig, contrasted with O.H.G. tusig, M.H.G. 
 dusel (disorder), G. dusel (dizziness) ; cf. also E. bustle and the 
 actual pronunciation of busy 2 = O.E. bysig. — O.E. ful and E. 
 foul, O.H.G. ful and G. faul (Go. ful-s, 6.N. full) ; but M.E. 
 file (to defile), E. to defile and filth ; cf . G. fdule fdulnisz. — E. 
 mouse = O.E. mils, pi. mice = O.E. mys; cf. G. maus, pi. mause. 
 
 It seems superfluous to dwell upon the English broken 
 vowels, whereof silly has afforded an instance. But we may 
 recall here, by way of general recapitulation, the chief gram- 
 matical classes in which German, remaining truer to its origins 
 than English from the main reason that the origins are more 
 recent, has preserved the regular process of metaphony, though 
 occasionally it may be obscured by false analogies. 
 
 1. Feminine nouns in -in : gott gottin, etc. 
 
 2. Neuter collectives, with prefix ge- : berg (mountain) and 
 gebirg = O.H.G. gi-birg-i; so also stern (star) and gestirn (con- 
 stellation), tosen (noise) and getose, etc. 
 
 1 The word belongs to a Southern dialect, which changes initial / to v. 
 Cf. E. vat=G.fasz (tub). 
 8 Svpra 17, 5, though often spelled O.E. bisig. 
 
VOWELS AND DIPHTHONGS. 49 
 
 3. Abstract feminine nonns in -e (formerly -i) : hoch (high) 
 and hohe (height) = O.H.G. hoh-i. 
 
 4. Diminutive nouns in -el, -chen, -lein, and -ling (O.H.G. 
 -ila, -kin or -chin, -lln and -line) : these types are common and 
 well known (infra 103). 
 
 5. Adjectives, denoting material, in -en (formerly -in) : gulden 
 (golden), irden (made of earth), etc. 
 
 6. Adjectival derivatives in -icht, -ig, -lich, etc. 
 
 7. Comparatives, hoch = Go. hduh-s, but hbh-er — hdnh-iza\ 
 and superlatives, h'6ch-(e)st =■ Go. hduh-ist-s, etc. 
 
 8. So-called metaphonical plurals, as fusze (feet), kdlber 
 (calves), vater, etc. 
 
 9. Causative and denominative verbs, because there is a 
 primitive suffix -jan hidden in their ending : to the previous 
 instances, add to set — G. setzen = Go. sat-jan (to cause to sit 
 down, to settle). 
 
 10. Formation of the imperfect subjunctive in the so-called 
 strong verbs : er nahm (he took), er ndhme (O.H.G. nam and 
 ndmi) ; trug, triige ; zog, zoge, etc. 
 
 11. Conjugation of strong verbs in the 2nd and 3rd person 
 sg. of present indicative, sometimes also in the 2nd person sg. 
 imperative, thus : sprechen (to speak), er spricht (he speaks), 
 sprich (speak) ; schlagen (to strike), er schldgt ; ich fahre (I am 
 travelling), du fahrst, er fdhrt, wir fahren = O.H.G. faru, ferist, 
 ferit, faren. 1 
 
 § 5. Pregermanic Compensatory Lengthening. 
 
 (24) "We are passing at length the boundary of the primitive 
 Germanic language ; and, on this very line there appears to 
 
 1 Let us end with a well-known but curious derivation with a twofold 
 metaphony, the G. compound elend = O.H.G. eli-lenti, "from another land" 
 (eli- = Go. alji-<=li. alio- appears likewise with metaphony in E. else), hence 
 " foreign, outcast, wretched," and " exile, wretchedness." — It must be borne 
 in mind that metaphony has been considered here only as a general phonetic 
 phenomenon, abstracted from the enormous differences, either in chrono- 
 logy or in application, which may characterize it in the historical life of 
 English and German. It will be met with in every province of both 
 languages, and seen to exhibit itself in very various conditions. Cf., for 
 instance, infra 26 (i., 1, 2, 3), 28 (i.), 80 (xvi.), 144-146, 147, 194, 197, 
 204-206, etc., etc. 
 
 E 
 
50 ENGLISH AND GERMAN GRAMMAR. 
 
 us a phenomenon, which, though it is common to all Germanic 
 dialects and must therefore be attributed to the parent-speech, 
 can nevertheless be verified "without any intervention of other 
 Indo-European idioms. In a single word, as exhibited by the 
 generality of Germanic vocabularies, we observe a regular 
 vocalic and consonantal variation, namely : here, a short vowel, 
 followed by a nasal and a guttural explosive ; and there, the 
 same vowel lengthened, while the nasal has disappeared and 
 the guttural has become a spirant. Hence the law may be 
 easily inferred : every Pregermanic short vowel, &, i, u,, 1 
 followed by a group of nasal and surd guttural spirant (i\h), is 
 lengthened, whilst the nasal is dropped. 
 
 1. Group a + fih. — Go. pagkjan (to think), O.E. d"cencean and 
 E. to think, O.H.G. denchen and G. denhen, 2 everywhere with 
 short vowel and nasal, from a Preg. *}>a'hkjanam. But, if the 
 k becomes an h, 3 as in Preg. *J>a'i\hta (he thought), the a be- 
 comes a, and the i\ is dropped : hence Preg. *}>ahta, and its 
 offspring, Go. }>dhta, O.E. (Sbhte and E. he thought, O.H.G. 
 ddhta and G. er dachte.* So also: Go. briggan = 'Preg. *hrefv&a- 
 nam, and pf. brdhta = Preg. *hrafthta; E. to bring, brought, 
 brought ; G. bringen, brachte, gebracht, etc. 
 
 2. Group i + fih. — O.N". ping (lawsuit, business), O.E. Sing 
 and E. thing, with a more general sense, as in O.H.G. dine and 
 G. ding, everywhere with nasal and short vowel, from a Preg. 
 *Jjh\%-. But, from a secondary verb *J>iiVian, changed to 
 *J?lhan, the Go. gadeihan, G. gedeihen (to thrive, to succeed), 
 and O.E. gefteon, which still shows the nasal and the short 
 vowel in its participle geSungen. 5 
 
 3. Group u + fih. — Go. frugkjan (to seem), the root being that 
 of Jjagkjan, only with reduced vocalism. 6 O.E. b"yncan and E. 
 
 1 The case cannot take place with enor o, because in this position they 
 respectively become or remain {oru: infra 26, i., 5, and 28, i. 
 
 2 The vowel is metaphonical owing to the suffix -jan. 
 8 On this change, see below 53 C, and note. 
 
 4 Shortened in Modern German, supra 20, 1. 
 
 6 G. gediehen is formed by analogy with gedeihen, and such verbs as con- 
 tain a primitive i (schreiben geschrieben, leihen geliehen). But geduvgen and 
 bedungen still exist as participles of the verbs dingen and bedingen. 
 
 6 See the Section on Vowel-Gradation, infra 45, 4. 
 
VOWELS AND DIPHTHONGS. 51 
 
 think, the latter being confounded "with earlier ftcencean, and 
 having thus assumed its meaning. O.H.G. dunchan, M.H.G. 
 dunlcen, G. diinhen (to seem). Thus, everywhere the nasal and 
 the short vowel suggesting a Preg. *pui\kjanam, pf. *Jmfi.hta, 
 whence *}>uhta : Go. ]>uhta (it seemed) ; O.H.G. duhta and G. 
 (mich) deuchte * or dduchte, with metaphony (borrowed from the 
 subjunctive) instead of *dauchte ; and, an infinitive being formed 
 on this new pattern, the analogous vb. dauchten. 
 
 Section II. 
 
 PRIMITIVE VOWELS AND DIPHTHONGS, AND THEIR EVOLUTION. 
 
 (25) We have examined the series of transformations under- 
 gone by the vowels of English and German during the his- 
 torical and prehistorical life of both languages. Now, if 
 Pregermanic were known to us through any direct tradition, 
 it would only remain to show how its sounds are reproduced 
 in each of them. But, since Pregermanic is lost, and can only 
 be restored by the comparison of its offspring, and, subsidiarily, 
 of the other Indo-European languages, we must now apply to 
 them in order to discover that necessary link of the chain 
 wJfich we have so far followed upwards, but which we are now 
 attempting to trace downwards. We shall thus, as it were, 
 prove the operation, provided that Sanskrit, Greek and Latin 
 agree with English and German in confirming what we have 
 already learnt about Pregermanic vocalism ; and then our 
 Germanic phonetics will rest on a true scientific basis. 
 
 According to the evidence afforded by the Indo-European 
 languages, the primitive vocalism of their parent consisted of : 
 five short vowels, a, e, %,o,u; the six corresponding diphthongs, 
 ay, ey, oy, and aw, ew, ow ; the five corresponding long vowels, 
 a, e, i, 0, u \ 2 and, lastly, an indeterminate vowel a, which, at 
 least in the European languages, has been entirely confounded 
 
 1 This is the archaic and proper form. The usual one mich diinkte is due 
 to the analogy of diinhen. 
 
 2 Occasionally also in a diphthong, thus ay, ew, etc. But in this short 
 grammar we are compelled to omit these rare sounds. 
 
52 ENGLISH AND GERMAN GRAMMAR. 
 
 with &, and therefore need not be distinguished at all from it 
 in this work (I.-E. *pdter-, Gr. narip-, L. pater, Go. fadar, E. 
 father, G. vater). 1 
 
 The Indo-European a and o, whether short or long, have 
 become blended together in the Germanic branch, a 
 and o being Preg. a, and a and o being Preg. o. Apart from 
 some partial mutations, the other vowels are quite dis- 
 tinctly kept. Thus, it will seem advisable to begin our study 
 with e, i and u. 
 
 § 1. Short and long e. 
 
 (26) The vowel e may be said to be the touchstone of the 
 European languages ; for the Asiatic (Indo-Iranian) group of 
 our family 2 has changed it to a in every position. The Ger- 
 manic group shows it still almost as clearly as any other- 
 European Language. 
 
 I. I.-E. e = Gr. e = L. e — Preg. 8: I.-E. *bhero (I carry, cf. 
 Sk. bhdrd-mi), Gr. <pep<o, L. fero, Preg. *t>ero, whence Go. 
 bair-an 3 (to carry), O.E. beran and E. to bear, O.H.G. beran, 
 M.H.G. b'ern and G. (ge)bdren (to bring forth a child), cf. the 
 original meaning preserved in bdhre (a hand-barrow) ; I.-E. 
 *pell-a (skin), Gr. ireXk-a, L. pell-i-s, Preg. *fell-a, whence E. 
 and G. fell ; I.-E. *ed-6 (I eat ; cf . Sk. dd-mi), Gr. tSw, L. edo, 
 Preg. *eto, whence Go. it-an (to eat), O.E. etan and E. eat, 
 O.H.G. eifian and G. essen, etc. 
 
 Preg. S has been generally kept. In five main cases, how- 
 ever, it became % at the very earliest period. 
 
 1. In the primitive diphthong ey, first changed to ty owing 
 to the law stated under 3 below, and this again then contracted 
 *to I : I.-E. *steygho (I go up), Gr. trra^a), but Preg. *stl%6, 
 whence Go. infinitive steigan, O.H.G. stlgan and G. steigen.* 
 
 1 The Asiatic languages here have an 1, thus, Sk. pitdr-. 
 
 2 Cf . supra 3. 
 
 8 As a matter of fact, Gothic, less pure in this respect than any other 
 Germanic tongue, changes every Preg. e to 1 ; the e, however, is restored, 
 though spelled ai, before an r or an h : hence arises the contrast between 
 6airan = E. bear, and tton = E. eat. 
 
 4 O.E. stigan and E. sty ; the latter being obsolete and replaced by F. 
 mount. On the later evolution of this 1 in both languages, cf. supra 17, 3, 
 and infra 27, II. 
 
VOWELS AND DIPHTHONGS. 
 
 53 
 
 2. In the same group before a vowel: the group first be- 
 comes %y, which then is reduced to a simple y. Such is the 
 case with all causative verbs. These verbs will be seen to 
 require in I.-E. the deflected root, 1 accompanied by a suffix 
 -eyo-, which suffix takes the person-endings : thus, Sk. bhdr-d-mi 
 (I carry) and bhdr-dyd-mi (I cause to carry), Gr. <f>op£w ; root 
 mH (to think), L. moneo (I cause to think, I warn) = *mon-eyd; 
 root nek (to die, cf. L. nex nec-is), L. noc-eo, etc. Consequently, 
 a root sed (to sit down, Sk. sdd-as, " seat," Gr. cSos, L. sed-eo 
 sed-es, etc.), when assuming the causative meaning " I make 
 to sit," will become I.-E. *sod-eyo. Now, the Go. sat-jan (to 
 place) refers us to a Preg. *sat-ja-, which is quite identical with 
 *sod-6yo, the only difference being that the ey is changed to y ; 
 hence, with the necessary metaphony, O.E. settan and E. to set, 
 O.H.G. sezzen and G. setzen, as opposed to sit, sitz and sitzen. 
 
 3. When the following syllable contains either a prehistorical 
 i or y 2 : I.-E. *medh-yo-s (placed in the middle, cf . Sk. mddhyas), 
 Gr. fJL€<ro<s = ^co-cros = */i€0-i/o-s, L. medius, but Preg. *mift-ja-z, 
 whence Go. midjis, O.E. midd and E. mid, O.H.G. mitti and 
 G. mitte ; I.-E. *es-ti (he is, cf. Sk. dsti), Gr. «tti, L. est, but E. 
 (he) is, G. (er) ist, etc. 
 
 4. As final, always, and sometimes in certain other unaccented 
 positions less easy to define : I.-E. final *-e in the 2nd person 
 sg. imperative, as *nem-e (take), Gr. ve'/t-c (divide), L. em-e 
 (buy), but Preg. *mm-i, whence O.H.G. nim 3 and G. nimm; 
 I.-E. final *-8s in the nomin. pi., as *pod-es (feet), Sk. pdd-as, 
 Gr. 7rd8-€s, but Preg. *fot-%z, whence the metaphony in O.E. jet 
 (sg. fot) and E. feet. Here the change of 8 to K seems attribut- 
 able to the following z. 
 
 5. Before any group beginning with a nasal : I.-E. root 
 
 1 See the Section on Vowel-Gradation, infra 44. 
 
 8 This is the first appearance of Germanic metaphony, already alluded to, 
 supra 22, note. The same influence which, in Pregermanic, had altered 
 the £ before an i, more and more extended to other vowels in E. and G^ 
 This is the reason why we find such scanty instances of the mutation of e 
 (supra 23 E) separately in the two languages, and especially in English : 
 S was mutated before their separation. 
 
 8 Final % has vanished, but shows its existence through the metaphony it 
 has caused. — Observe that L. emere preserves the meaning " to take" in the 
 compounds sumere, eximere, etc. 
 
54 ENGLISH AND GERMAN GRAMMAR. 
 
 bhendh (to bind), Sk. bandh (to bind), bdndh-u-s (a relation), 
 Gr. irev6epos = *<f>ev6-ep6-<i (father-in-law), Tr€io-T^p = *<f>ev$-T^p (a 
 cable), but Go. bind-an, O.E. bindan and E. to bind, O.H.G. 
 bindan and G. binden ; I.-E. root we (to blow), cf . G. wehen, and 
 L. ve-ntu-s, but E. and G. wind ; I.-E. *pefi,qe (five), whence a 
 probable derivation *pefi,q(e)-ro-s (fivefold), Go. figgr-s (finger), 
 O.N.fingr, G. and E. finger. 
 
 The e persists everywhere else. Only in O.H.G. it becomes t, 
 if the following syllable has a u : I.-E. *septm (seven), Sk. saptd, 
 Gr. «rra, L. septem, Preg. *sebtm, whence Go. sibun, O.E. seofon 
 and E. severe, but O.H.G. sibun, M.H.G. siben, G. sieben. 1 
 
 II. I.-E. e has become, in Pregermanic as in Greek (v), a 
 very open e : from this have arisen, in Gothic and Old English, 
 a very close e, and, on the other hand, in Old High German, 
 the still more open sound of a. Examples are : I.-E. root ed 
 (to eat), pf. *ed-, Sk. ad-imd (we ate), Gr. (lh-)rfi-a-p.€v, L. 
 ed-imus, Preg. *et-ume, whence Go. etum, O.E. ceton and E. (we) 
 eat (pf.), O.H.G. djum and G. (wir) aszen ; I.-E. root dhe (to 
 place, to do), Gr. #17-0-0) from TL-$v-p.i, L. je-c-l, Preg. *b~e-tSi-s 
 (action), whence Go. dejy-s, O.E. deed and E. deed, O.H.G. tat 
 and G. that ; I.-E. root se (to sow, to throw), Gr. iir)p.t. = *<ri-crr)-fii 
 (I throw), rj-p,a (a throw), L. se-men (seed), Preg. *se-b~i-s (seed), 
 whence Go. (mana-)se}>-s (mankind), E. seed, G. saat ; I.-E. root 
 ne (to spin, to sew), Gr. vrj-p.a (tissue), L. ne-re (to spin), thus 
 I.-E. *ne-tro- or *ne-tlo- (an engine for sp. or s.), Gr. vrjrpov 
 (distaff), Go. nefila (needle), E. needle, G. nadel. 
 
 § 2. Short and long i. 
 
 (27) I. The » is a very consistent vowel, as it remains un- 
 changed in all the Indo-European languages, Pregermanic 
 included : I.-E. root bhid (to split), Sk. bhid (id.) L. ftdimus (pf.), 
 Preg. *hit-ume (we bit), Go. bitum, O.H.G. bivfium and G. (wir) 
 bissen, E. (we) bit ; L. pisci-s, Go. fislc-s, O.N". fisk-r, E. fish, 
 G. fisch ; Sk. vidhava (a widow), L. vidua, Go. viduvo, O.E. 
 widewe and E. widow, O.H.G. wituwa and G. wittwe. 
 
 Before an r or an h, Preg. £, in Gothic only, becomes & (spelled 
 
 1 The vowel lengthened in an open syllable, supra 20. 
 
VOWELS AND DIPHTHONGS. 
 
 65 
 
 ai) : I.-E. root migh (to be moist, to make water), Sk. mih 
 (id.), Gr. o-/nx-€a) (id.), 6-^t'x-Xr/ (a cloud), L. ming-ere, but Go. 
 maih-stu-s (moisture, fog, mist) = E. mist=Q. mist (dung), 1 the 
 latter both preserving their x. 
 
 In the five cases wherein e becomes i, % of course remains 
 unaltered. But, excluding these, Preg. % is changed to e, when- 
 ever the following syllable contains an a or an o : I.-E. *wir-o-s 
 (man), L. vir, Preg. *wir-a-z, whence *wer-a-z, Go. vair-s (man), 
 O.E. and O.H.G. wer, surviving in E. and G. wer-wolf, in E. 
 world and G. welt 2 ; I.-E. root sed (to sit), whence a compound 
 *ni-zd-6-s (home, nest), Sk. nida~s, L. nidus, but Preg. *nistds 
 becomes *nestds, G. and E. nest. This law, however, is not yet 
 well defined and remains subject to some alternations, so that 
 we are able to point to many such contrasts as : E. to lick 
 and G. lecken (I.-E. root Ugh, in Sk. rih, " to lick," Gr. Xetx-oy, 
 L. ling-o) ; E. to live and G. leben ; E. liver and G. leber ; E. quick 
 (cf. L. vlvus = *gvlgv-o-s) and G. queclc (cf. the vb. er-quick-en, 
 "to vivify") in the compound queck-silber = ^j. quick-silver. 
 
 II. The I is even more consistent than the % : Pregermanic 
 keeps it unchanged; Gothic pronounces it I, though spelled ei 3 ; 
 in English and German, unless it has undergone later shorten- 
 ing, it is sounded ay, thongh written respectively i and ei. 4 
 I.-E. *su-s (a hog), Sk. su-, Gr. v-s o-D-s, L. sic-s, O.E. sih and 
 E. sow, O.H.G. sit and G. sau ; hence, a secondary I.-E. adjective 
 *su-ino-s *sw-ino-s, used as a substantive in Germanic, Preg. 
 *swina-z, Go. svein-s, O.E. swin and E. swine, O.H.G. swln and 
 G. schwein. 
 
 § 3. Short and long u. 
 
 (28) I. I.-E. # = Sk. u = Gr. v = L. # = Preg. u: I.-E. kun- 
 (dog), Sk. cun-, Gr. kuW kuv-os, Preg. *hunb-a-z, b whence Go. 
 
 1 Because it steams. Thus, F.fumier (dung), which comes from L. fimus, 
 is modified by analogy with the vb.fumer (to smoke). 
 
 2 Supra 21, and the note on weralt. 
 
 8 Ulfilas borrowed this symbol from Byzantine Greek, where the old 
 diphthong et had contracted to I. 
 
 4 This spelling is due to the M.H.G. i having first passed through the 
 sound of ey, still preserved in some dialects, before it came to be pro- 
 nounced ay. 
 
 5 The $ added in Germanic is of doubtful origii 
 
 T3 N 1 
 
56 ENGLISH AND GEEMAN GRAMMAR. 
 
 hund-s, O.N. hund-r, O.E. hund and E. hound, 1 O.H.G. hunt and 
 G. hund ; I.-E. root dhubh, with a vague meaning of infirmity 
 or bodily defect, Gr. tv<f>\6<; (blind) = *0v<£-Ao-s, Preg. *6~'umb-a-z, 
 whence Go. dumbs (dumb), O.N". dumb-r, O.E. and E. dumb, 
 O.H.G. tump turn and G. dumm (stupid); I.-E. *yuwen- (young), 
 Sk. yuvan-, L. juven-i-s, whence a derivative I.-E. *yuwn-ko-s, 
 Sk. yuvaqds, L. juvencus, Preg. *yuww\%ds contracted to *yui\%ds, 
 Go. jugg-s, O.E. geong and E. young, O.H.G. and Mod. G.jung, 
 etc. 
 
 Under the same conditions under which Go. I becomes e 
 (written ai), % Go. u becomes o (written au) : I.-E. *dhur-6-m 
 (door, cf. Sk. dur- and Gr. Ovp-a), whence Preg. *6"ur-d-m, but 
 Go. daur. 
 
 "We hare seen that a following a or o changes i to e. Under 
 the same restriction, — that is to say, everywhere but before a 
 group of nasal and consonant, 3 and moreover with a serious 
 number of exceptions, — Preg. u becomes o, if the following 
 syllable has a or o: Preg. *d~ur-d-m, 4: Go. datir, 5 E. door, G. 
 thor; I.-E. root yug (to join), *yug-6-m (yoke), Sk. yugdm, Gr. 
 £vyov, L. jugum, Preg. *yukdm (Go. yuk), whence *yokdm, O.E. 
 geoc and E. yoke, O.H.G. joh and G. joch ; I.-E. root ghu (" to 
 pour out," cf. Gr. x«-<»> x v * T °~ s > an ^ " ^° °^ er a libation," cf. Sk. 
 hu-td-s, " offered in libation "), whence a participial form I.-E. 
 *ghu-to-s, " he who is worshipped through libation, a god," Preg. 
 *iu-8d-s, then ♦jo-S'a-s, Go. gups (God), E. God, G. Gott, etc. 6 
 
 II. I.-E. «=Sk. w = Gr. i3 = L. « = Preg. « 7 : I.-E. root pu 
 
 1 With regular lengthening before nd, supra 20, 4 B. 
 
 * Supra 27, I. — These phenomena are of no direct interest for E. or G., 
 since they occur only in East Germanic ; but they must nevertheless be 
 briefly mentioned, or else the student would be unable to follow the phonetic 
 correspondences. 
 
 8 For the preservative effect of a nasal, compare the participles swollen 
 and geschwollen, from swell and schwellen, with the participles bound and 
 gebunden, found and gefunden, from bind and binden,find and finden. 
 
 * The quality of the vowel is proved by Preg. *Sur-i-, whence O.H.G. turi 
 and G. thiir with metaphonical U : supra 22. 
 
 8 Thus, in Gothic, there are two reasons for the u appearing as S. 
 
 8 It is obvious that, in spite of appearances, the regular metaphony of o 
 must be il (cf. supra 22 in fine) : if g'ottin were an early word, it would be 
 *giittin ; the b' shows it to have been a later derivation from gott. 
 
 7 Whence, in English and German, a process of diphthongizing which 
 runs parallel to that of i : supra 18, 1. 
 
VOWELS AND DIPHTHONGS. 
 
 57 
 
 (to be foul), Sk. puy (id.), Gr. 7ri5-0(o (id.), L. pus, in a Preg. 
 derivation *fu-lds, whence Go. fills, O.E. fill and E. foul, 
 O.H.G. /wZ and G. faul; I.-E. root ru ("free space," cf. L. rus, 
 "country"), in a Preg. derivation *ru-mds, whence Go. rums 
 (space), O.E. rum and E. (exceptionally not diphthongized) 
 room, O.H.G. rum and G. raum. Further compare : E. house 
 and G. Turns = O.E. and O.H.G. hits = O.N", hies = Go. *hus (only 
 in gud-hus, " God's house, temple ") ; E. thousand and G. tau- 
 send = O.E. ftiisend and O.H.G. ditsunt = Go. J>usundi. We have 
 already found ew for au in G. deuchte. It is less easy to explain 
 in G. enter — O.H.G. utar (E. udder with shortening), cf. L. 
 uber = Gr. ov6ap = Sk. udhar (udder). 
 
 § 4. Diphthongs of short e. 
 
 (29) To e correspond the diphthongs ey and ew. Now, ey 
 has been seen to become i before a consonant and y before a 
 vowel. 1 The treatment of ew is far more intricate. It might 
 be said, in theory, to continue in the form eu, with the excep- 
 tion that Gothic, changing e to 1, naturally changes eu to iu 2 : 
 I.-E. root gus and gews (to examine, taste, choose, cf. L. gus-tus, 
 G. kos-t-en), Gr. yevto = *ycuo--a), theoretically Preg. *keus-anam 
 (to appreciate, choose), then Go. kius-an. But, in fact, each of 
 the elements which form the diphthong obeys exactly the same 
 law that it would obey when isolated ; in other words, as e 
 becomes %, so also eu becomes iu, if the following syllable con- 
 tains i or j ; and, as u is changed to 6, so eu is changed to eo, 
 if the following syllable contains a or 0; which is as much as 
 to say that eu will hardly ever appear in its original form. 
 
 Afterwards, in English, eu and eo are sounded eo, whereas iu 
 becomes ie and y 3 ; in German, eo becomes io, and lastly ie, 
 which ends by contracting to i, 4 whereas iu contracts to long 
 
 1 Cf. supra 26, I. 1, 2. 
 
 2 It is seen here, and will still better be seen below, with what mathe- 
 matical regularity the application of a phonetic law takes place throughout 
 a language. 
 
 3 Metaphonical vocalisms already mentioned, supra 22 in fine. 
 
 4 Afterwards eventually shortened. The old diphthong is still very per- 
 ceptible in South Germany : Alsatian tief, Hep, lie'cht, etc. ; cf. G. tief (deep), 
 lieb (dear), licht (light), pr. tlf, lip, llcht, etc. 
 
58 ENGLISH AND GERMAN GRAMMAR, 
 
 it, which in Central Germany is diphthongized to eu. 1 Thus, 
 an T.-B. root dhub (hollow, deep, cf. Lith. dub-u-s "hollow") 
 gives the I.-E. derivate *dhewbos, Preg. *Seupaz, which becomes 
 *&eopaz : hence, Go. diup-s, O.N". djop-r, O.E. deop and E. deep, 
 O.H.G. tiof and G. tief. The same root before an t or a j, in 
 Preg. *t?eup-janam (to dive), becomes *&iup-jan, O.E. dyfan and 
 E. dive, cf. also E. to dip. 2 These complex correspondences are 
 still more complicated by the influence of analogy ; for the 
 two diphthongs eo and iu were too much alike not to be easily 
 confused, so that the metaphonical form and the pure one 
 largely usurped each other's legitimate place, though the pure 
 one as a rule proved the stronger. The vb. *keusanam quoted 
 above, for instance, ought to be in the infinitive *keosanam 
 (O.E. ceosan, and O.H.G. chiosan), and in sg. 3 pres. indicative 
 *kiusit (O.E. ciest and O.H.G. chiusit); but Mod. E. actually 
 knows no other stem but choose (the substantive choice being 
 re-borrowed from French), and Mod. German, no other but 
 kies-en. So again, a noun *leod-u- had a pi. *liud-i ; but O.E. 
 extends to the plural the non-metaphonical form, thus leode 
 (people) ; whilst O.H.G. carried into the sg. Hut (a people) 
 the metaphonical form which is regular only in the pi. liuti, 
 G. leute ; and there are many similar cases. 
 
 With this remarkable treatment of the Preg. diphthong eu 
 may be classed numerous alternations ; as, for instance, G. 
 licht = O.H.G. lioht = O.E. leoht, 3 and G. leuchten (to enlight) ; 
 G. ziehen (to draw) = Go. tiuhan (cf. L. ducd — *deuco), and G. 
 zeugen (to produce) ; G. biegen (to bend) = Go. biugan, and G. 
 beugen (to inflect). 4 Lastly, the reader may compare : O.H.G. 
 
 1 Just as u is diphthongized to au ; for long u and eu are the metaphonical 
 vowels corresponding to the pure ones u and au. The old vowel is kept 
 by the Southern dialects, which also preserve the I and u free from diph- 
 thongizing : Swiss t'liit (long ii) and Alsatian t'lit=G. die leute (the people). 
 
 2 G. tuafen=Go. ddup-jan, which has acquired the technical sense of 
 "christening" (thus Gr. pairrlfriv, "to dip, to christen"), shows, as a 
 causative verb, the deflected grade of the same root. If G. had retained 
 a neutral vb. " to dive " in this series of words, this would be Alamannic 
 *tiifen with long u, and classical *teufen. Cf. also O.E. /#r=E. fire, and 
 O.H.G. fiur=M.H.G. viur=Q. feuer, with Gr. vvp. 
 
 3 On the E. i in light, see above 21, note. 
 
 4 These verbs are causatives : thus, they had in I.-E. a diphthong ow, 
 which has become Preg. au (infra 32) ; but the result is the same, since this 
 
VOWELS AND DIPHTHONGS. 59 
 
 diota (people) and O.E. &eod = Go. J>iuda = J.-~E. *tewtd, further 
 O.H.G. diutish (translated into L. theotiscus), now Teutsch and 
 Beutsch (Alamannic tiitsch and titsch), with the non-metaphonical 
 forms still surviving in M.H.G. diet (people) and proper nouns 
 as Dietrich (L. Theodoricus, F. Thierry). And above all should 
 be mentioned the archaic conjugation of the vb. fliegen (to fly), 
 viz. ich fliege, du fleugst, er jieugt, wir fliegen, which will be 
 recalled at its proper place. 1 
 
 § 5. Short a and o, and their Diphthongs. 
 
 (30) I.— 1. I.-E. a = Sk. a = Gr. a = L. a = Preg. a: I.-E. 
 *sdld (salt), Gr. aA-s = *<raA-?, L. sdl = *sald, Preg. *salt, whence 
 Go. salt, E. salt, G. salz; I.-E. *ddkru (a tear), Gr. Sdxpv, L. 
 dacru-ma (lacrima), Preg. *tdhru *tayu, whence Go. tagr, O.E. 
 *teahr and tear (broken vowel), E. tear, O.H.G. zahar and G. 
 zdhre with metaphony ; I.-E. *agro-s (field), Sk. ajra-s, Gr. 
 dypo-s, L. ager, Preg. *akrd-s, whence Go. akr-s, O.E. cecer and 
 E. acre, O.H.G. acchar and G. acker; L. scSh-ere (to scratch), 
 Go. skab-an (id.), E. £0 shave, G. schaben, etc. 
 
 2. I.-E. o = Sk. a o = Gr. o = L. 0, but Preg. a: I.-E. *ghorto-s 
 (turf, enclosure), Gr. x°P T0? 5 ^. hortus, Preg. *jaro"a-s, whence 
 Go. gardrs (enclosure), and cf . E. gard-en and G. gart-en, E. ?/ard ; 
 I.-E. *ghosti-s "a foreigner," whence either "an enemy" or 
 "a guest," L. hosti-s (enemy), Go. gast-s (guest), E. guest, 2 and 
 G. gast ; I.-E. *oktow (eight), Sk. astdu, Gr. okto>, L. oc£o, Go. 
 ahtdu, O.E. eaftfo broken, and E. et^/ii, G. acht ; I.-E. *ozdo-s 
 (twig), Gr. o£os, Go. osi-s, G. ast, etc. 
 
 (31) II. Hence it must follow that the two primitive diph- 
 thongs ay and by unite in the Preg. ai. This is well kept in 
 Gothic. 3 But in O.E. it has become a, which in English is 
 
 au has again become eu through metaphony before the causative suffix -jan. 
 Cf . supra 26, I. 2. 
 
 1 Infra 206 B.— Of this development the reader may, if he choose, for 
 the present, retain only the general conclusion in the following formulas 
 which at first sight seem contradictory : 
 
 Preg. eu changed to eo appears in Mod. Or. as ie, i ; 
 „ eu „ iu „ „ eu. 
 
 2 With the metaphony we also meet with in the G. pi. gdste. 
 
 3 Here, as already observed, we always write it ai, in order to distinguish 
 it from the false diphthong Go. ai, which is simply an open e. 
 
60 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 changed to 6. In Old German ai has only undergone the 
 metaphony to ei, while M.H.G. ei has gradually returned to 
 the 'sound of ay, so that it in no way differs, — excluding, of 
 course, the Southern dialects, — from the ay, likewise spelled ei, 
 which proceeds from O.H.G. I. 1 As early however as O.H.G., 
 the diphthong ai contracted to e, when followed by r, h or w ; 
 see below lehren and ewig, and add here : Go. sdivala (soul), 
 O.E. sdwul and E. soul, but O.H.G. se(w)la and G. seele ; Go. 
 sdir-s (pain), O.E. stir and E. sore, but O.H.G., M.H.G. sir (pain), 
 the etymological sense of which is preserved in G. ver-sehr-en 
 (to damage), and much weakened in G. sehr (very). 2 
 
 1. I.-E. ay = 8k. e = Gr. cu = L. ai ce = Preg. ai: I.-E. *slaywo-s 
 (left, left-handed, awkward), Gr. Aaio's = *o-A.aifd-s, L. Iwvo-s, 
 O.E. slaw and E. slow, cf. O.H.G. sleo ; I.-E. *ayw- (time, age, 
 century), Gr. alwv = *alF-wv, L. azv-o-m, Go. div-s (time, eternity), 
 whence the secondary G. ewig (eternal) = O.H.G. ew-ig ; O.N". 
 heit-r (warm), O.E. htit and E. hot, O.H.G. hei$ and G. heisz ; 
 M.E. bothe and E. both = G. beide ; L. cced-ere (to cut), Go. 
 shdid-an and G. scheiden (to divide), etc. 
 
 2. I.-E. ot/=Sk. e = Gr. oi = L. oi (u or i), but Preg. ai: I.-E. 
 root uritf (to see), whence a pf. *woyd-e (" he has seen," so " he 
 knows"), Gr. oTSe = FotS-c, L. vid-it, Sk. ved-a, Preg. *wait-e, Go. 
 vatY, O.E. wa£ and E. (he) wot, O.H.G. wevz, and G. (er) u>et'sz ; 
 I.-E. *oy-n6-s (one), Gr. otvo-s, L. unus = oino-s, Preg. *aind-s, 
 whence Go. din-s, O.E. tin and E. one, O.H.G. ein and G. ein; 
 compare the same ti mutated and shortened in O.E. dtnig and 
 E. any, cf. G. einige; Preg. root Us (Go. *leis-an 3 "to learn"), 
 forming a causative vb. which would be I.-E. Hoys-eyo, " I 
 teach," 4 whence Go. Idis-jan, O.H.G. leren and G. lehren, natur- 
 ally the root we also find in O.E. ieornian and E. to learn, 
 
 1 In other words, classical G. pronounces ein like scheinen, whereas 
 Alamannic says ayn for the one and Bne for the other, etc. Dutch (Low 
 German) in the same way distinguishes een from schijnen. Cf . supra 17, 3, 
 and 18, 1. 
 
 * The successive meanings would be: "with pain— with struggle — with 
 intensity — very." 
 
 8 Cf. the Go. pf. Idis (he knows), which would be I.-E. *lbys-e, formed 
 like *w5yd-e. 
 
 * Cf. supra 26, I. 2. 
 
VOWELS AND DIPHTHONGS. 61 
 
 O.H.G. lernen and G. lernen, which would be Go. *liz-n-jan 
 (to learn). 
 
 (32) III. A further consequence is, that the two I.-E. 
 diphthongs aw and bw unite in Preg. au. This is still retained 
 in Gothic. 1 In Old English, strangely enough, it has become 
 ea, still mostly written ea in Modern English, but sounded 
 either I or open e, as we have seen above, according as it stands 
 in an open or in a close syllable. In German, au remains un- 
 changed, apart from a slight ou- sound, and usually so spelled 
 in O.H.G. writing; before a dental, however, as before r, I, h, 
 and occasionally in other positions, O.H.G. au contracts to b. 
 These processes may be illustrated by the following instances. 
 
 1. I.-E. dw = Sk. o = Gr. av = L. aw = Preg. au: I.-E. root awg 
 (to increase), Gr. av$dvw, L. aug-eo, etc., Preg. advb. *auh 
 (besides, also), whence Go. auk, O.E. eac and E. eke, O.H.G. auh 
 and G. auch-,* I.-E. *ka{w)p-et- (head, cf. L. cap-ut), Preg. 
 *haub-ij>, whence Go. hdubif>, O.N", haufujj, O.E. heafod and E. 
 head, O.H.G. houbit and G. haupt, etc. 
 
 2. I.-E. ow = Sk. o = Gr. ov = L. ou u, but Preg. <iw. I.-E. root 
 rudh (red, cf. Sk. rudh-ird-s — Gr. €-pu0-po-s = L. rub-er), whence 
 a deflected derivative *rbwdh-6-s, L. ruf-u-s, Preg. *rau&-d-s, 
 Go. rdu}>-s, O.E. read and E. red, O.H.G. rot and G. roth ; 
 I.-E. root dhiibh (bodily defect), 3 whence a similar derivation 
 *dhowbh-6-s, Preg. *&aub-d-s, Go. dduf-s (stupid), O.E. deaf and 
 E. deaf, O.H.G. toup and G. taub,* with contracted vowel in 
 toben (to deafen). 
 
 We need not go back farther than Germanic itself in men- 
 tioning: O.E. dead" and E. death, G. tod = Go. ddu)>-u-s, and 
 
 1 Written du, as di above, in order to distinguish it from the open which 
 is spelled ml, supra 28, I. 
 
 2 Thus, primitive au and primitive u are no longer distinguishable from 
 one another in classical German ; but, like primitive ai and primitive i, they 
 are distinguished in the Southern dialects. Alamannic au (High Alsatian 
 oy) = G. auch, but A1./mI = G. faul (foul) ; and so also, paue poye = bauen (to 
 bxxil&),fraufroy=frau,lauchloych = lauch (leek), etc. ; but sur=G. sauer = 
 E. sour, hus = O.H.G. hus (haus, house), sufe = G. saufen (to drink, from 
 animals). The correspondence is constant, though High Alsatian has soy = 
 G. sau (a sow). 
 
 8 Already seen in rv<f>\6s and dumb = dumm, supra 28, I. 
 * Compare the vocalism in High Alsatian tdyp (deaf), and tup (long U) = 
 G. taube = Hi. dove. 
 
62 ENGLISH AND GEEMAN GRAMMAR. 
 
 likewise dead = todt, great — grosz; O.E. hleapan and E. to leap, 
 O.H.G. loufen and G. laufen (to run) = Go. hldupan; O.E. leaf 
 and E. leaf, O.H.G. hub and G. laub (foliage); O.E. leak and E. 
 leek, G. lauch, the former shortened in the second term of the 
 compound gar-lie, the latter shortened likewise in the dialectal 
 knobloch = knoblauch (garlic); Go. ga-ldub-jan (to believe), O.E. 
 (metaph.) gelyfan and E. belief, to believe, O.H.G. giluuben and 
 G. glauben; 1 Go. duso (ear, cf. Gr. oxk, L. auris), O.E. eare and 
 E. ear, O.H.G. bra and G. oTir; Go. tduh (he drew), pf. of 
 tiuh-an, that is to say, I.-E. *dowk-e deflected from a root dewk, 
 which is reproduced in L. duc-ere, G. pf. (er) zog from the same 
 vb. zieh-en, etc. 
 
 § 6. Long a and o. 
 
 (33) The long 6 which comes in Pregermanic from both 
 I.-E. a and 0, remains 6 in O.E. and O.H.G., and then under- 
 goes in Modern English and German a similar modification. 2 
 
 1. I.-E. a = Sk. a = Gr. d (Ionian- Attic , »7) = L. a, but Preg. 0: 
 I.-E. *bhdgo-s. (a certain tree), Gr. <£dyos <f>rjyos (oak), L. fdgus 
 (beech), Preg. *t>okd-s (beech), whence Go. bok-s, O.E. boc-treow, 
 which would have become E. *book-tree, s and E. book,* O.H.G. 
 buohha and G. buche (beech), buck (book) ; I.-E. root sthd (to 
 stand), derivative I.-E. *sthd-lo-s (firm, anything that stands), 
 Sk. sthd-ld-m (porringer), Gr. arTrj-Xrj (pillar), Preg. *sto-ld-s, 
 whence Go. stol-s (throne), O.E. stol and E. stool, O.H.G. stuol 
 and G. stuhl ; I.-E. *mdter- (mother), Sk. mdtd, Gr. fxanqp firjrrjp, 
 L. mater, O.E. modor and E. mother, O.H.G. muotar and G. 
 mutter; so also, L. /rater = E. brother =G. bruder; Gr. fjy-eo-fji.ai 
 (to lead), L. sdg-lre (to be acute), Preg. *sok-jana-m (to search), 
 whence Go. sokjan, pf. sokida, O.E. pf. sohte and E. sought? 
 O.H.G. suohhan, pf. suohte, G. suchen suchte gesucht, etc. 
 
 1 Contrast High Alsatian kloyvS (to believe) with Muuc (long ii) = G. 
 klauben (to scratch) = O.H.G. chlubon. 
 
 * Cf. supra 18, 2. 
 
 8 The extant form beech is metaphonical, O.E. sg. bde, pi. 6ec, cf. /cet from 
 foot, etc. But the pure form, shortened, appears in the compound buck-mast 
 (actually " the food from a beech "). 
 
 4 Cf. the meaning of G. buch-stabe (" a letter," actually "a beech-stick "), 
 because the first Bunic writings were made of sticks. 
 
 s The infinitive seek = O.E. secan is metaphonical like beech. Cf. also the 
 inf. beseech, with pf. besought. 
 
VOWELS AND DIPHTHONGS. 63 
 
 2. I.-E. 6 = Sk. a = Gr. a> = L. o = Preg. o: I.-E. V^" s ( foot )> 
 Sk. pad, Gr. 7rovs ttw? = *7ro>S-s, Preg. *fdt-s (Go. fot-u-s, O.N. 
 fot-r), whence O.E. /o£ and E./ooi, O.H.G. /W03 and G. fusz; 
 I.-E. root 6&Z0 (to bloom, cf. L. flo-s), in derivation Preg. 
 *hlo-man-, whence Go. bloma (flower), O.E. bloma and E. bloom, 1 
 O.H.G. bluomo bluoma and G. blume ; cf . the metaphony, in G. 
 bluhen (to bloom) = Preg. *hlojanam, and the shortening in E. 
 blossom = O.E. blostma; I.-E. root jpZo (fluid), Gr. 7tA<o-to-s 
 (swimming), Preg. *flo-b~u-s (flood), O.E. flod, and E. flood 
 shortened, O.H.G. flbt float fluat fluot, and Gr.flut, etc. 
 
 Without going beyond Germanic, we may quote : O.E. blod 
 and E. blood, 2 O.H.G. bluot and G. blut ; E. stud and stud-book, s 
 G. stute (mare); E. boy = *bdf-ig(?) diminutive to O.E. *bof= 
 M.H.G. buobe and G. bube. 
 
 Section III. 
 
 VOWELS IN FINAL SYLLABLES. 
 
 (34) By accurately observing the preceding laws, the student 
 will find himself easily able to account for almost any vowel or 
 diphthong of either language, and to identify it with the cor- 
 responding sound of the other language, or even of Greek and 
 Latin. But the sounds which in primitive Indo-European stood 
 in a final syllable underwent a peculiar treatment, which took 
 place after Germanic had thrown back the word-accent to the 
 first syllable. 4 The result is comparable to the process we may 
 historically trace in the transition from Latin to French, as in 
 portum (haven) becoming port, porta or portam (door), or portat 
 (he carries) becoming porte, and portds (doors, or thou carriest) 
 becoming portes. We must distinguish the primitive vowels 
 according as they were either themselves final or followed by 
 a final consonant. 
 
 1 Flower, of course, is borrowed from French. 
 
 2 Also shortened. The metaphonical vowel appears in the vb. bleed — 
 Preg. 'Muf-janam ; but the G. vb. bluten is taken from blut without 
 metaphony. 
 
 B The metaphony in «teed = O.E. steda. 
 * On this recessive accent, see below 65, II. 
 
64 ENGLISH AND GERMAN GRAMMAR. 
 
 § 1. Final vowels. 
 
 I. Short. — Every short final vowel is kept in Preger- 
 manic, and then disappears in all the dialects descended from 
 it, without leaving any trace of its existence but the metaphony 
 it may have caused. 1 Final ti survives only in Gothic and Old 
 High German, but is dropped everywhere else : L. pecu (cattle) 
 = Preg. *fehu, whence G-o. faihu, O.E. feoh, O.H.G. f'ehu fihu 
 and G. vieh ; Gr. 7roXv (much), Go. filu, G. viel, etc. 
 
 II. Long. — Every long final vowel remains unchanged in 
 Pregermanic, and then becomes shortened in Gothic and 
 West Germanic ; 2 whereupon, in the further evolution of the 
 latter, it either gives place to a weaker sound, or is quite 
 dropped ; 3 lastly, the vowel, if preserved, has become even 
 weaker or vanished altogether in Modern English and German. 4 
 Thus, the final a in feminine nouns of the so-called 1st Greek 
 declension {r^fiipa. x^P 5 -) w iH be seen to have become Preg. 6, 
 Go. a — d, O.H.G. a, O.E. we (a sound duller than a), Mod. G. e, 
 and even completely dropped in Mod. English. 
 
 III. Diphthongs. — Primitive ey (as has been said) becomes I. 
 On the other hand, primitive ay and aw, if final, are sometimes 
 retained in Gothic and spelled di and du, but elsewhere they 
 contract to e and o. The consequence is, that in West Germanic 
 every final diphthong is changed to a long vowel, and treated 
 accordingly. The applications of this law will be met with on 
 a large scale in the study of declension and conjugation. 6 
 
 1 Examples have been given above : 22, 26, I., 4, etc. — In point of fact, 
 the law is more intricate than as here given, at least in West Germanic, and 
 it depends on the position of the accent ; but, as it was afterwards affected 
 by many analogical influences, it will be better not to follow it in its arbitrary 
 applications. 
 
 s This shows that the vowel was the same in Pregermanic ; for, if it had 
 been shortened there, the subsequent languages would have lost it altogether. 
 
 8 A dull vowel, in a dissyllable with short first syllable ; no vowel at all 
 anywhere else : this is the rule, apart from analogical influences. Cf . infra 
 143, II., and 157, § 2. 
 
 4 Cf. supra 19, 1. 
 
 5 As, for instance, in the dat. sg. (152), in the nomin. pi. (143, II.), in the 
 pres. subj. (192), in the pf. subj. (193, sq.), etc., etc. 
 
VOWELS AND DIPHTHONGS. 65 
 
 § 2. Non-final vowels. 
 
 I. Short. — Every short vowel in a Pregermanie final syllable 
 is treated as if itself final, without distinction, whether the 
 following consonant is itself dropped, or is preserved : l I.-E. 
 *agro-s (field), ace. *agro-m, Sk. djra-s djra-m, Gr. dypo-s aypo-v, 
 L. ager agru-m, Preg. *akrd-s *akrd, whence Go. akr-s akr, O.E. 
 cecer and E. acre, O.H.Gf. acchar and G. acker; I.-E. *sunu-s 
 (son), ace. *sunu-m, Sk. sunu-s sunu-m (cf. Gr* v-lo-), Preg. 
 *sunu-s *sunu, i whence Go. sunu-s sunu, O.E. sunu and E. son, 
 O.H.G. sunu and G. sohn, etc. 
 
 II. Long. — A non-final long vowel in a final syllable survives 
 in Gothic and West Germanic and further undergoes in English 
 and German the usual process of weakening : I.-E. *wlqos 
 (wolf), pi. wlqos, Sk. vfkas vfkas, Preg. *wulfaz *wulfoz, Go. 
 vulfs vulfos, E. wolf wolves, etc. 3 
 
 III. Diphthongs. — The law for a long vowel holds good for 
 any diphthong. 4 
 
 1 On the treatment of final consonants, cf. infra 39, 1-2, 46 and 63. 
 s The short vowel, instead of the long one in Sk., is probably due to a 
 different grade of root su (to bring forth). 
 
 3 In monosyllables, the long vowel or diphthong, whether final or non- 
 final, always remains long : cf . the declension of the definite article, infra 
 129 sq. 
 
 4 The investigation of Preg. finals is leading to new results. The reader 
 to whom it is of interest may be referred to Hanssen's and Hirt's essays, in 
 Kuhn^s Zeitschrift, xxvii. p. 612, and Indogermanische Forschungen, i. pp. 1 
 and 195. But, as their leading principle, namely, the distinction between 
 the acute and the circumflex accent on the final long syllable, cannot be 
 said to find as yet any sure application except in Gothic, it need not be more 
 than briefly mentioned here. 
 
CHAPTER III. 
 
 SEMI-VOWELS AND CONSONANT-VOWELS. 
 
 (35) The correspondences of the semi-vowels and consonant- 
 vowels being all of the utmost simplicity, it seems expedient to 
 trace them at once from the Indo-European period down to our 
 days. Having thus established the continuous descent of these 
 mixed sounds, we shall further enquire into the conditions 
 under which they alternately played, in the Indo-European 
 speech, the part of consonants or vowels. This point will be 
 discussed in the section on Vowel-Gradation. 
 
 Section I. 
 
 SEMI-VOWELS. 
 
 I.-E. had the two semi-vowels of i and u, viz. y and w, which, 
 in certain positions, were able to form a syllable by themselves, 
 inasmuch as they became changed to their respective vowels. 
 Pregermanic, we have seen, represents the two vowels by i and 
 u ; we have also seen how it represents y and w, when following 
 a vowel and forming a diphthong ; the consonants y and w, so 
 far as they survived, Pregermanic represented by j=y (spelled 
 Go. and G. j, E. y) and w (sounded v in Mod. German). 
 
 § 1. Semi-vowel y. 
 
 (36) I. I.-E. initial y = Sk. y = Gr. £ or rough breathing = 
 Lat. ^' = Preg. j: we have already quoted yoke and joch, young 
 and jung, etc. ; you and euch will appear among the pronouns ; 
 here we may add I.-E. root yer yor (a space of time), Zd. ydre 
 
SEMI-VOWELS AND CONSONANT-VOWELS. 67 
 
 (year), Gr. &pos wpa (time, season, year), Preg. *jer-a-, whence 
 Go. jer, O.E. gear 1 and E. year, O.H.G. ^'ar and Gr. jahr. 
 
 II. Medial.— 1. Between vowels, I.-E. y is kept in 
 Gothic and O.H.G. (as also in Sk.), but dropped in O.E. and 
 M.H.G. (as also in Greek and Latin), and thus vanishes, be- 
 traying itself only by the occasional metaphony : O.H.G. 
 bluo-jan, G. bliihen (to bloom); Go. *kiu-jan (to chew), O.E. 
 ceow-an and E. chew, O.H.G. chiuw-an and G. kauen, with the 
 metaphony preserved in the compound wieder-kduen. 
 
 2. After a consonant, whether primitive, or arising from 
 the primitive diphthong ey, 2 the y is preserved in Gothic ; but, 
 in the later dialects, it causes the preceding consonant to be 
 doubled, and then unites with it and disappears. 3 I.-E. *ten-yo 
 (I spread out), Gr. tcivoj = *t«v-t/co, O.E. ben-ian (to spread out) 
 and ftennan, O.H.G. dennen and G. dehnen (to extend). 4 Go. 
 hlah-jan (to laugh), O.H.G. hlahhan and G. lachen, O.E. 
 (metaph.) hliehhan and E. laugh. I.-E. root stheg teg (to cover), 
 Gr. o-T€ya> crreyos rcyos (covering, thatch), L. teg-o tog-a, etc. ; 
 hence, a derived noun, Preg. *J>ak-d-s = I.- 1 &. *tog-6-s, reproduced 
 by Go. *J>ak = ~E. thatch — G. dach ; from this again is derived a 
 denominative vb., Preg. *J>ak-jan-am, Go. J>akjan (to cover), 
 O.E. oeccara, O.H.G. decchan 6 and G. decken (the latter three, of 
 course, are metaphonical). The causative vb. of the E. to wake 
 and G. wachen points to a Preg. *wak-jan-am, whence Go. 
 vakjan (to rouse from sleep), O.E. weccean, O.H.G. wecchan and 
 G. wecken. To these may be added : O.E. settan (E. to set, G. 
 setzen) = Go. sat-jan (to cause to sit); and Go. mid-ji-s ( = L. 
 med-iu-s), O.E. midd and E. mid, O.H.G. mitti and G. mitte. 6 
 
 1 For this g, see infra 50, II. 
 
 2 See above, 26, I., 2. 
 
 8 As in Greek the X in dy7^XXw = *d77eX-y-a>. 
 
 4 The group nn shortened to n, whereupon the preceding vowel became 
 long. Group 11 in Go. vil-ja, E. will, G. wille. 
 
 5 More exactly decchen ; for it is a law in O.H.G. that j changes a follow- 
 ing o toe; but analogy has often made the law ineffective, that is to say, 
 decchan was restored on the model of such infinitives as sehan (to see, now 
 sehen) = Go. saihvan, where the a had no j before it. See Braune, Ahd. Gr., 
 §58. 
 
 6 It is not only before &j, but in several other positions, especially before 
 a nasal or a liquid, that a preceding consonant is liable to be doubled in 
 German words; further, as in the course of declension or conjugation 
 
68 ENGLISH AND GERMAN GRAMMAR. 
 
 § 2. Semi-vowel w. 
 
 (37) I. Initial. — 1. Before a vowel, I.-E. w = Sjc. « = Gr. F 
 (dropped in Ionic and Attic) = L. v = Preg. w : we have already 
 cited wine and wein (borrowed from Latin), he wot and er weisz, 
 wind and wind, etc. ; we may add water and wasser (Slav, voda), 
 I.-E. *wlqo-s = Sk. vfka-s = Go. vulf-s = JZ. wolf—G. wolf. 
 
 2. Before a consonant, the w, still sounded in Preg. and 
 Gothic, is lost in German ; in English it is nothing more than 
 a written survival : Go. wits (a scratch, an engraving), G. ritz 
 (a scratch), O.H.G. rijan and G. reiszen (to tear), risz (a rent), 
 O.E. wrlt-an (to scratch), whence E. to write. 
 
 II. Medial. — 1. Between vowels, w is retained, but unites 
 afterwards with the preceding vowel, which thus assumes a 
 slightly labialized sound. In such cases it may well happen 
 that the w is no longer visible, even in spelling, as in hue 
 instead of *hew, Tuesday, 1 etc. Further instances are : E. true 
 = O.E. treowe, and G. treu (faithful) = O.H.G. {gi)triuwi, con- 
 trasted with the deflected form of the same root in E. to trow = 
 G. trauen ; E. snnw = OM. sndw, and G. schnee = O.H.G. sneo, cf. 
 Go. sndiv-s, pointing to a Preg. *snaiw-d-s (L. nix niv-em) ; E. 
 tree = Go. triu, related to Gr. Spvs (gen. opvos = *8pvF-6<s) and to 
 Russian derevo (tree). 
 
 2. After a consonant, primitive w is rather uncommon, 2 
 but generally survives: I.-E. *dwo- (two) = Preg. *twd-, whence 
 Go. tvdi, E. two, s G. zwei ; I.-E. root kwit (to shine), derived 
 I.-E. *kweyt-o-s, Sk. Qvet-d-s (shining, white), Preg. *hwit-a-z, 
 whence Go. hveit-s, O.E. hwit and E. white, O.H.G. hwlT, and G. 
 weisz. The group nw, however, is assimilated to nn: I.-E. 
 *mdmt- (man), whence the secondary *manw-6-s (human, man), 4 
 Preg. *mann-d-s (cf. Go. manna), E. man — *mann, G. mann. 5 
 
 adjacent sounds may vary (thus, nomin. knabe, gen. knaben with a nasal, 
 etc.), the consonant is liable to appear alternately simple and double ; 
 whence arise such doublets as knabe (boy) and knappe, rabe (raven) and 
 rappe (black horse), etc. 
 
 1 Cf. supra 17, 5. 
 
 2 The common initial group sw will be referred to below, 59, II. 
 8 Of course the ?c melts into the sound of the following u. 
 
 4 The u, before a vowel, must become a consonant (w). The prototype 
 manu- is from Sk., where the progenitor of all men is called Manu. 
 
 & The treatment of w may sometimes appear very inconsistent; but it 
 
SEMI-VOWELS AND CONSONANT- VOWELS. 69 
 
 Section II. 
 
 CONSONANT-TOWELS. 
 
 (38) The consonant-vowels, i.e. the nasals and liquids, as 
 sounded in the Indo-European speech, were either consonants 
 or vowels. In either case, their correspondences are very 
 simple. 
 
 § 1. Nasals. 
 
 (39) The regular tendency of the nasals, throughout 
 the Indo-European family, is to become partly assimilated 
 to the following consonant, that is to say: the nasal is 
 labial if followed by a labial, dental if followed by a dental, 
 etc. A language, therefore, must possess as many nasals as it 
 has classes of explosives; for instance, four in I. -E., three in 
 Preg., Gothic, English and German. 1 Moreover, if in the 
 course of its evolution the articulation of any explosive happens 
 to be changed, the preceding nasal is strictly compelled to vary 
 in accordance with this change. Thus, the nasal was velar in 
 I.-E. *pei\qe (five), and has remained guttural in L. qulnque; 
 but it has become labial in Gr. (-ZEolic) ire/iTr* and Go. fimf, and 
 dental in common Gr. wevrc and G. fiinf. 2 This main point 
 being settled, the nasals exhibit an exact correspondence in all 
 languages. 
 
 I. Consonants. — 1. Labial: I.-E. m = Preg. m everywhere 
 retained, save in primitive finals. Such instances as mid — 
 mitte l room = raum, etc., have already occurred. The reader 
 may add: comb and hamm = Sk. jdmbha-s (jaw) = Gr. y6/x<f>o-s 
 
 should be borne in mind, that its position was not constantly the same in 
 the same word, because in declension or conjugation it was alternately 
 medial and final. Thus, the w is final in the O.E. nomin. yeolo = *geolw, 
 but medial in the gen. geoho-es, so that E. yellow appears as a compromise 
 between the two forms. O.H.G. shows likewise a nomin. gelo and a gen. 
 gelw-es, whence the doublet gehl (Alamann.) and gelb ; cf. L. helo-o-s 
 (yellow). And a similar change of w to b will be observed in such double 
 words as fahl and falb (fallow), in wittib (widow), in hieb (a blow) and pf. 
 er hieb, as opposed to witwe (L. vidua) and vb. hauen (to strike), etc. 
 
 i Cf. infra 52. 
 
 9 The difference between Gothic and German arises from the /, as this 
 letter is bi-labial in Gothic, but denti-labial in German. 
 
70 ENGLISH AND GERMAN GKAMMAE. 
 
 (a peg), etc.; E. meai = O.E. meodo, and G. me£7i = O.H.G. metu 
 = Gr. fxeOv (wine) = Sk. mddhu (a sweet and intoxicating drink) ; 
 Go. qiman, E. to come, G. kommen ; E. to swim, G. schwimmen ; 
 E. lamb, G. lamm, etc. 
 
 Primitive final m became Preg. n (Gr. v). The fact is seen 
 only in the uncommon cases where this n has been preserved 
 by the support of an affix : G. ace. den = Go. pan-a = Gr. tov = 
 Sk. tdm=^li. (is-)tum, and turn advb. accusative. 
 
 Apart from these exceptions, every final nasal is dropped in 
 Pregermanic, and then the preceding vowel is treated as final : 1 
 Go. ace. akr (field) = Gr. ayp6v = Jj. agrum. 
 
 Again, medial m, when made final by the disappearance of 
 the final vowel, becomes n in the O.H.G. period, whenever it 
 existed as a grammatical ending: 2 dat. pi. O.H.G. tag-um, later 
 tag-un, G. tag-en, from tag (day). A like change took place in 
 the passage from O.E. to M.E. 
 
 2. Dental : I.-E. n = Preg. to, always preserved, except when 
 originally final. We have already instanced night and nacht, 
 needle and nadel, snow and sclinee, one and ein, son and sohn, man 
 and mann. To these we may add E. can, ken, land, sand, etc., 
 and G. kann, kennen, land, sand, etc. 
 
 3. Guttural: I.-E. 7^ = Preg. ft, written grin Go., 3 elsewhere 
 to : I.-E. root sifiq (cf . Sk. siiic-ati " he pours out "), Go. sigq-an 
 (to dive), E. to sink, G. sivken ; L. long-u-s, Go. lagg-s, E. long* 
 G. lang. Add E. to drink (Go. drigk-an), to hang, finger, tongue, 
 and G. trinken, hangen, finger, zunge. 5 
 
 1 Cf. supra 34, § 2, I. 
 
 2 In the other cases, the m is generally preserved for the sake of analogy, 
 since it was not final everywhere : thus, arm (arm) has not become *arn, 
 because there existed a gen. arm-es, and nim (take) could not be changed to 
 *nin, because it belonged to the same series as the infinitive nem-an. We 
 have, nevertheless, the contrasted forms: G. 6oden = E. bottom; G. faden 
 (thread) = E. fathom ; and the like. 
 
 8 A spelling borrowed from Greek by Ulfilas. 
 
 * A nasal group changes a preceding O.E. a to o : thus, we have man and 
 mon (now dialectal), hand and hond, land and lond. This pronunciation, 
 however, was not Mercian, and therefore it has disappeared from classical 
 English, leaving only such scanty survivals as long, tong = G. zange, song = 
 G. savg, etc. 
 
 5 The correspondences are very numerous. But E. angel and G. engel 
 should not be included among them : the latter is Go. aggilus, borrowed 
 
SEMI-VOWELS AND CONSONANT- VOWELS. 71 
 
 (40) II. Vowels. — An Indo-European nasal vowel becomes 
 simply a in Sk. and Gr. (a) ; Latin vocalizes the m to em (im) 
 and the 9 to en (in) ; the Preg. vocalization results respectively 
 in um and un, 1 whereupon the group thus obtained naturally 
 follows all the laws which have been already stated as to u, m 
 and n. 
 
 1. I.-E. m. — The I.-E. language had a root sem, signifying 
 unity, which we recognize, for instance, in L. sem-el and Gr. 
 els = *cr€/i-s ; now, when this root happened to lose the e, there 
 remained sm, as in the derived pronoun *smm-6-s (a certain, 
 whoever), which has become Sk. sam-a-s, Gr. d/x-o-s, Go. sum-s, 
 O.E. sum and E. some. The same is the case with an I.-E. root 
 gem (to go), Gr. flawa} = *gm-yo, and L. venio = *gvem-yo : losing 
 its e, it gave rise to *gm-ti-s (marching), Sk. ga-ti-s, Gr. /3d-crt-s, 
 etc. ; and this form again occurs in Go. (ga-)qum-]>-s (arrival, 
 meeting), and G. (aus-) (ein-) (zu-)kunft, in contrast with G. 
 kommen = Go, qiman = L. venid. Gr. ap^L (around) shows that 
 the related Sk. abhi (towards) must have contained a latent 
 nasal, thus I.-E. *mbhi faithfully reflected by O.H.G. umbi 
 (around, towards), M.H.G. umb, G. um. Final um naturally 
 becomes un : I.-E. *dekm (ten), Sk. ddca, Gr. 8e/ca, L. decern, 
 Preg. *tehum, whence *tehun, Go. taihun, E. ten, G. zehn ; and 
 the same with Go. sibun = Hj. seven — G. sieben. 2 
 
 2. I.-E. n. — The negative I.-E. particle was *ne, Sk. nd : the 
 e being dropped, according to a general rule, when the word 
 entered into a compound, the syllable was reduced to *n-, Sk. 
 a-, Gr. d- (both privative), L. in- (negative), lastly the well 
 known negative Germanic pi'efix un-, as in E. un-fair, G. un- 
 rein, etc., exact parallels to Gr. d-o-flcvrjs (unf orcible), L. in-firmus, 
 etc. An I.-E. word *snter (apart, without), mainly illustrated 
 by Gr. drcp (without), assumes in Germanic the form Go. sundro 
 
 itself from Gr. &yye\os (observe the metaphony produced in the a by the 
 following i) ; whilst angel is O.F. dngele = L. dngelum, this also borrowed 
 from Gr. If angel were a Germanic word, it would not be pronounced enzl, 
 but avgl ; cf. E. angle=G. angel. 
 
 1 And, as a matter of course, v, if vowel, becomes un : cf. E. to drink and 
 drunk, G. trinken and getrunken, infra 181. 
 
 * The n, which should have been dropped as Preg. final, was preserved by 
 the analogical influence of the ordinal numbers : Go. tibun-da (seventh). 
 
72 ENGLISH AND GEEMAN GRAMMAR. 
 
 (apart, but), O.E. sundor and E. (a-)sunder, O.H.G. suntar, G. 
 sonder with its many derivates. Further instances are : I.-E. 
 *ndhero-s (inferior), Sk. ddhara-s, L. inferu-s, Go. undar (under), 
 E. under, G. unter; I.-E. *mn-to-, L. rnen-tu-m (chin), Go. 
 mun-J}-s (mouth), O.E. mub — *mun-b, x and E. mouth, G. mund, 
 etc. 
 
 § 2. Liquids. 
 
 (41) The liquids, r, Z, whether consonants or vowels, are re- 
 produced with considerable exactness, both in Germanic and in 
 the other members of the I.-E. family. 
 
 1. Consonants. — 1. I.-E. r = Sk. r = Gr. o = L. r=Preg. r, 
 preserved in the later languages. 2 We need but remind the 
 reader of red and roth, read and rathen, ride and reiten, brother 
 and bruder, etc. Let him further add : I.-E. *der-o, Gr. Sep-a 
 (I flay), Preg. *ter-o, whence Go. {ga-)tair-an (to tear), E. to 
 tear, G. (ver-)zehr-en (to devour) ; L. hester-nu-s adjective, Go. 
 gistra-(dagis) "yesterday," E. yester-(day), G. gester-n; I.-E. 
 *priy-6-s (dear, cf. Sk. priy-d-s, id.), whence a derived Germanic 
 vb.. *frijdn (to love), Go. frijon, of which the pres. participle is 
 frijpnds (friend), O.E. freond and E. friend, O.H.G. friunt and 
 G. freun-d 3 ; Gr. /ca/>7r-o-s (fruit), L. carp-ere (to gather fruit), E. 
 harv-est, G. herbst (vintage). 
 
 2. I.-E. Z=Sk. r = Gr. A = L. Z = Preg. I, preserved in the 
 later languages. I.-E. root luk lewk (to shine), Sk. root rue 
 (id.), cf . Gr. \evK-6s (white), L. lux luc-em, etc. : hence a derived 
 noun, Go. liuh-a]>, O.E. leoht and E. light, O.H.G. lioht and G. 
 licht. I.-E. root klu klew (to hear), whence Sk. cru-td-s (heard, 
 famous) and qrdv-as (glory), Gr. kAv-to-s (famous) and kAcos = 
 ♦kAcF-os (glory), L. (in-)clu-tu-s, etc. ; the same word in the 
 Germanic compound Hlut-hari (Merovingian proper name). 4 
 From the same root klu, after assuming the lengthened form 
 klu, is derived a Preg. participle *hlu- , bd-s (heard), whence: 
 
 1 O.E. compensatory lengthening, supra 20, 4 A. 
 
 8 Observe, however, that the r's now extant are not all descended from a 
 primitive r ; for a great many of them are to be traced to a rhotacized s (z), 
 infra 61, I. 
 
 3 O.E. pure ; the other forms show metaphony. 
 
 4 Actually " he who has & famous army " (G. heer " army "). 
 
SEMI-VOWELS AND CONSONANT-VOWELS. 73 
 
 O.E. hlud and E. loud, O.H.G. liit and G. laut (loud), laut 
 (sound) ; cf. the Swiss vb. losen (to listen). Gr. a\cro<s (forest) 
 = *FoAtFo9 (?), G. wald, O.E. weald and E. wold, which latter 
 became blended with wood = O.E. wudu = OB..G. witu. 1 This 
 example, however, lacks certainty. But Go. fugl-s (bird), O.E. 
 fugol and E. fowl, O.H.G. fogal and G. vogel, admits of no doubt. 
 
 (42) II. Vowels. — The I. -E. liquid vowels, r, I, only remain 
 as such in Sanskrit (f). In the other languages they are ac- 
 companied by a slight indefinite vocalic sound, whence arise 
 the various groups : Gr. ap, pa, and aA, Xa ; L. or, ur, and ol, ul ; 
 Preg. ur, ru, and ul, M, respectively, according as the vocalic 
 epenthesis is inserted before or after the liquid. 2 The u after- 
 wards undergoes the modifications required by the phonetic 
 laws of each later language : thus, we have regularly Go. aur 
 and ul, E. and G. either or and ol, or ur and ul, as determined by 
 the nature of the following syllable 3 ; and eventually further 
 lengthening. 
 
 1. I.-E. r. — I.-E. has a root mer (to die), Sk. mar, which is 
 reduced to mr in such derivations as Sk. mr-td-s (dead) = L. 
 mor-tuu-s, and Sk. mf-ti-s (death) = L. mors (gen. mor-ti-s); 
 another and no less correct formation would be *mr-tro-m } 
 with the meaning of " death-instrument," 4 and this is repro- 
 duced by Go. maur-pr, O.E. morbor and E. murther murder ; cf. 
 G. mord = *mr-to-s. Again, I.-E. has a root her, with the sense 
 of " head " and " horn," cf . Gr. /ccp-as : the root is reduced to 
 kr in Gr. /cap-d (head) and Kpd-vo-s (helmet), L. cor-nu (horn), 
 which latter (*Tcrnu) is nearly related to Go. haurn, E. and G. 
 horn. I.-E. root t8rs trs (to be thirsty), Sk. trs-ti-s (thirst), Go. 
 
 1 Lost in Mod. German, though still visible in wiedehopf " whoop " (bird) 
 = O.H.G. witu-hopfo, as we should say a " wood-hopper." 
 
 2 The latter case is by far the rarer, and belongs to analogy rather than 
 to phonetics. Let us consider an I.-E. root bhreg (to break), cf. the L. pf. 
 freg-l : its reduced form bhrg has become Preg. bruk, not burk (Go. bruk-an-s 
 
 = E. brok-en — G. (ge-)broch-en) ; but the main cause of this difference is that 
 the consonant and the vowel naturally tended to shift themselves to the same 
 place they respectively occupied in the present and in the perfect (Go. brik-a 
 brak, E. break brake, G. brech-en brach). 
 
 8 Cf. supra 28, I. 
 
 * See Derivation, especially 79 (xiii). All these words have taken the 
 sense of "violent death," while another stem assumed— th*-m©re . general 
 meaning of " death." ^^^£St' 
 
74 ENGLISH AND GERMAN GRAMMAR. 
 
 j?aurs-tei, O.E. (metaph.) tyrst and E. thirst, G. durst, without 
 metaphony ; but compare the metaphony in G. diirr (dry) = 
 O.H.G. durr-i = l.-~Ei. *trs-i-. L. porca (the ridge between the 
 furrows) = I.-E. *prkd, QM.furh and E. furrow, G. furche, etc. 
 
 2. I.-E. Z. — I.-E. *wlq-o-s (wolf), Sk. vfk-a-s, L. lup-u-s, Preg. 
 *wulf-a-z, Go. vulf-s, E. woZ/, G. woZ/. I.-E. root peZ (to fill), 
 reduced to pZ = Gr. 7rAa in nifji-TrXa-fiev (we fill), and likewise in 
 the derivate *pl-no-s (full), Sk. pur-nd-s, L. ple-nu-s l : hence 
 we must infer a Preg. *ful-nd-s, assimilated to *fullds, Go. full-s, 
 E./«ZZ, G. voZZ. I.-E. root mel (to grind), Sk. mar (to grind), re- 
 duced to mr : Gr. fivX-r} (millstone), L. mola, O.E. myln (metaph. 
 instead of *mul-in) and E. mill, O.H.G. mulin mull and G. 
 mulle. 2 
 
 Section HI. 
 
 INDO-EUBOPKAN VOWEL-GRADATION (APOPHONT). 
 
 (43) Whilst stating and exemplifying the forms of the I.-E. 
 semi-vowels and consonant- vowels, we have been constantly led 
 to suppose that they could appear, in the same word, either as 
 consonants or as vowels, which is as much as to say that an 
 Indo-European root could alternately assume several 
 aspects. We have now to determine the laws of these varia- 
 tions. This will be the final effort of our inductive method, and 
 the highest point we are able to reach in the prehistoric evolution 
 of I.-E. vocalism : earlier, indeed, the terms of comparison are 
 wanting ; and later, the vocalic correspondences are obscured 
 by the phonetic laws of each particular language. Hence it is 
 only Indo-European which can give us the key to the vocalic 
 change which is usually called Vowel- Gradation (F. apo- 
 phonie, G. ablaut) ; a change which, simple as it is in principle, 
 has produced many intricate consequences throughout all the 
 languages of our family, inasmuch as almost every particular 
 of their grammar is influenced by it, and especially the con- 
 
 1 The long vowels in Sk. and in L. may be attributed to an I.-E. alternation 
 between short and long I. 
 
 2 These words, however, may as well have been directly borrowed from 
 Late Latin mollnum (F. mouliri). 
 
SEMI-VOWELS AND CONSONANT-VOWELS. 75 
 
 jugation of the so-called strong verbs in the Germanic branch 
 remains quite unintelligible unless we trace it back to this 
 origin. 
 
 § 1. The Principle of Vowel- Gradation. 
 
 (44) I. We must begin by examining the principle in the 
 easier cases, which are likewise the more frequent, namely, 
 when the primitive syllable contains an e. Every such syllable, 
 whether a root or a suffix, may successively exhibit three 
 main grades of vocalism, which we denote by the terms 
 "normal," "reduced" (weak), and "deflected grade." 
 In the first grade, the syllable keeps the e ; in the second, the 
 e vanishes entirely ' ; in the third, the e is replaced by an 0. 
 Supposing, for instance, a root pet (Sk. pdt-ati, L. pet-ere) with 
 the sense of "falling" and "flying": if normal, it is pet, Gr. 
 TrtT-ccrOai (to fly), in the present ; if reduced, it becomes pt, Gr. 
 TTT-icrOai (to fly), in the aorist of the same verb, and ttl-ttt-w (I 
 fall) ; lastly, if deflected, it becomes pot, as in Gr. TroT-do/xcu (I 
 flutter). So also, root bher (to carry, Sk. bhdr-ati, Gr. <£cp-o>, 
 L. fer-o), viz. : normal in <pep-u> <ptp-eiv ; reduced in (Si-)<£p-o-9 
 (seat for two persons); deflected in <j>op-6-<s <j>6p-o-<s <pop-d </>op-e'w. 
 
 It is the province of morphology, whether derivative or 
 grammatical, to determine the formations in which each of 
 these three grades should regularly appear. Some few in- 
 stances, however, of their occurrence will be stated below. As 
 for the cause of the process, nothing could be more obvious ; at 
 any rate, as far as the normal and the reduced grade are con- 
 cerned: a syllable is normal if accented, whereas it becomes 
 reduced when the accent is shifted to another syllable, as may 
 be inferred at once from the contrast in Trkrzo-Qa.1 and 7rre'o-#ai. 2 
 
 1 If the syllable forms a consonantal group which would be unpronounce- 
 able without an intermediate vowel, the e is retained ; in which case, of 
 course, the reduced grade no longer differs from the normal. 
 
 2 The Sk. language is by far the most instructive in this respect, because 
 it has kept the original word-accent almost unchanged : for instance, the 
 conjugation of e-mi (I go), i-wds (we go), gives the reason for root ey being re- 
 duced to ?, whereas Greek, having thrown back its accent throughout (fl-m 
 Ulna) merely preserves the mutation, the principle of which has disappeared. 
 Some of these alterations, in accent and vocalism, occurred as far back as 
 Indo-European, so that we find in it, here and there, reduced and accented 
 
76 ENGLISH AND GERMAN GRAMMAR. 
 
 The change of e to o in the deflected syllable may perhaps also 
 depend on accent, but this point has not yet been decided. 
 
 The instances quoted as yet assume that the variable syllable 
 contained a single e between consonants : if the e is dropped, no 
 other sound remaining that is able to support the syllable, the 
 latter disappears as a syllable, and its consonants become sup- 
 ported by the adjacent vowels. But, if the e was followed by 
 any semi- vocalic sound, then, of course, as soon as it disappears, 
 the following semi-vowel or consonant- vowel becomes changed 
 to a pure vowel, in order to support the syllable which other- 
 wise would be altogether unpronounceable ; in other words, 
 
 to a normal grade ey, ew, there corresponds a reduced grade i, u, 
 „ „ em, en, „ „ m, n, 
 
 >> j> ^"i ^*> j> >> /"» y> 
 
 the corresponding deflected grades still remaining, naturally, 
 6y, ow, dm, on, or, 61, with e changed to 6. 
 
 Thus, the I.-E. semi-vowels and consonant-vowels are conso- 
 nants when they follow or precede a vowel, and vowels when 
 they support a group of consonants. 
 
 It is in Greek and, though not so clearly, in Latin that these 
 gradations can be best observed. 1 Let us then quote a single 
 Greek example of each of the six types described above : pres. 
 \eiir-uv (to leave), aorist \iir-elv, pf . Ae-Aot7r-a ; fut. cA€u(0)-o"o-/i.ai 
 (I will go), aorist rj\v6-o-v, pf. (xk-rjXovO-a; ets (one) = *o-c/x.-s, d- = 
 *sm- in a-7ra£ (once), and 6fx-6- (same) = *som-6- ; irivd-os (pain), 
 aorist 7ra0-eiv (to suffer) = *TrnO-elv, pf . iri-TvovQ-a. ; pres. SepK-o-fiai 
 (I see), aorist e-SpaK-o-v (Sk. d-drg-a-m) , pf. Se'-SopK-a ; pres. oreA.- 
 Ao) (I send), passive aorist i-crrdX-rj (he was sent), o-toA-o-s (ex- 
 pedition). Similar instances occur in Latin : thus, teg-o and 
 tog-a; dlc-6 (I say) = deic-o, and (causi-)dic-u-s (barrister); fug-id 
 and pf. fug-i-t (he fied) = *foug-i-t ; pendo (I weigh) and pond-us 
 (weight). Further, a few Latin roots appear even in all three 
 
 syllables, or unaccented and normal syllables, just as Germanic has shown 
 us metaphonical syllables which ought to be free from metapbony, and the 
 reverse (supra 22). 
 
 1 Sanskrit clearly illustrates the normal and the reduced grades ; but, since 
 it changes both the e and the o to a, it has almost everywhere obscured the 
 passage of the normal into the deflected grade. 
 
SEMI-VOWELS AND CONSONANT-VOWELS. 77 
 
 grades, as they would in Greek : normal fid-o=feid-o (" I trust," 
 cf . Gr. TreiO-0-fx.aL " I believe ") ; reduced fid-es (faith), as in the 
 Gr. aorist i-wL6-6-fxr]v (I believed); deflected foed-us (treaty) = 
 *foid-os, as in the Gr. pf. ire-Troid-a. 
 
 II. The second case where the normal syllable has no £, but 
 a long vowel, o, e, o, is perhaps nothing more than a variation 
 of the preceding. 1 However this may be, the vowel-gradation 
 has assumed in it a different aspect. The three grades are as 
 follows, the reduced and the deflected grade being respectively 
 the same for the three classes, whilst in the third class the 
 deflected cannot be distinguished at all from the normal grade : 
 normal o, reduced a, 2 deflected 6 ; 
 
 „ o, ,, a, „ o. 
 
 § 2. Germanic Vowel- gradation. 
 
 (45) I. The syllable has an e. — 1. The e is isolated, the 
 type being pgt — %>t — pot : the Pregermanic correspondences 
 would be fej> — ffi—fap, 3 whereupon each of these syllables fol- 
 lows, in each later language, its own laws of evolution. Now 
 Greek has shown us that the radical vowel in the sg. of the pf. 
 regularly assumes the deflected state, in contrast with the pre- 
 sent, which has usually the normal root ; on the other hand, 
 the past participle of strong verbs, when regularly derived, 
 reduces the root : thus, these three forms, viz. present infini- 
 tive, past participle, and perfect in the singular, appear respec- 
 tively the most convenient to exemplify the normal, reduced 
 and deflected grade of any root. 
 
 Here, however, at the first step, we have to state an irregu- 
 larity : roots of this kind are not capable of reduction in Ger- 
 manic. In imitation of the few cases in which the reduced 
 grade elsewhere kept the e, Germanic has preserved it every- 
 
 1 Inasmuch as the long vowel may proceed from some older contraction 
 of the radical 8 with another vowel. 
 
 * The sound which has become i in Sk., and a anywhere else, tupra 25. 
 
 3 Since I.-E. o results in Preg. a, and so forth. The reader will kindly 
 refer to the vocalic correspondences stated above. Still more accurately the 
 three grades would be /eb — /(—/ad" or/ab. But the text illustration is sim- 
 pler for our present purpose. 
 
E. (to) see, 
 
 G. seh-en ; 
 
 „ seen, 
 
 „ (ge-)seh-en ; 
 
 „ (he) saw, 
 
 „ (er) sah. 
 
 „ (to) give, 
 
 „ geb-en ; 
 
 „ giv-en, 
 
 h (ge-)geb-en; 
 
 „ (he) gave, 
 
 „ (er) gab. 
 
 78 ENGLISH AND GEEMAN GRAMMAR. 
 
 where, so that its vowel-gradation, as far as regards isolated e, 
 ■only admits of two grades. Examples : 
 
 Normal : Go. saihv-an (to see), 
 
 Reduced : „ saihv-an-s (seen), 
 
 Deflected : „ sahv (he saw), 
 
 Normal : ,, gib-an (to give), 
 
 Reduced : „ gib-an-s (given), 
 
 Deflected : ,, gaf (he gave), 
 
 2. Now to a root where the e forms a diphthong with a y, 
 I.-E. type steygh — stigh — stoygh (to go up), the Preg. corre- 
 spondences are stl% — sti^ — stai%, as may be easily verified : 
 
 Normal : Go. dreib-an (to drive), E. (to) drive, G. treib-en ; 
 Reduced: „ drib-an-s (driven), „ driv-en,„ (ge-)trieb-en ; 
 
 Deflected : „ drdif (he drove), „ (Tie) drove. 1 
 
 The German deflected form would have become *treib, and 
 would not have differed from the normal grade. The question 
 how German came to lose this form must be reserved for 
 further examination. But it still persists in some grammatical 
 survivals, which were secured by their isolation from any ana- 
 logical influence: G. er weisz = Hj. he wot = 0.1$. wdt= Go. vdit = 
 = Gr. fot8-c (oTSe). In this perfect the plural quite regularly 
 passes to the reduced grade 2 : G. wir wissen = Go. vit-um = Gr. 
 Fi8-fxev (ISficv). 
 
 3. If the e forms a diphthong with a w, it gives us the I.-E. 
 series ew — u — ow, to which corresponds a Pregermanic series 
 eo or iu — u or o — au, as follows : 
 
 Normal : Go. *lius-an (to lose), E. (to) lose, G. *lier-en ; 3 
 
 Reduced: „ *lus-an-s (lost), „ (for-)lor-n „ (ver-)lor-en ; 
 Deflected:,, Hdus (he lost) „ „(er)(ver-)lor. 
 
 1 O.E. drdf. Observe the difference in quantity between driven, and (to) 
 drive = O.E. drlfan. In the G. getrieben the i is lengthened in an open 
 syllable. 
 
 2 It will be seen that, in the plural of the perfect, the primitive accent 
 was thrown forward to the ending, and the root consequently reduced ; but 
 in German and English, as in Greek, the vocalism of the plural became 
 almost everywhere assimilated to that of the singular, or the reverse : infra 
 176, 3, and 177, 3. 
 
 8 As conjugated in the well-known compound verb, Go. fra-liusan, O.E. 
 for-leosan, G. ver-lieren. 
 
SEMI-VOWELS AND CONSONANT-VOWELS. 79 
 
 English has lost the deflected grade in conjugation, but 
 shows it in the O.E. derivation leas 1 (loose) = G. los. In Ger- 
 man, owing to the later lengthening, the reduced has been 
 blended with the deflected form; but they are still distinguish- 
 able in many other words, as in ver-lus-t (loss), and in gieszen 
 (to melt, Go. giut-an, normal grade), reduced gusz (melting), 
 deflected er gosz (he melted), etc. 
 
 4. The e is followed by a nasal : I.-E. en — n — on ; Preg. en 
 or in — un or on — an ; and these again further modified in West 
 Germanic. As examples may be chosen the two I.-E. roots 
 bhendh (to bind) and nem (to take, Gr. vlfi-io). 
 
 Normal : Go. bind-an (to bind), E. (to) bind, G. bind-en ; 
 Reduced: „ bund-an-s (bound), ,, bound,,, (ge-)bund-en ; 
 
 Deflected : „ band (he bound), O.E. bond* „ (er) band. 
 Normal : „ nim-an (to take), ,, nehm-en ; 
 
 Reduced: „ num-an-s (taken), „(ge-)nomm-en; z 
 
 Deflected : „ nam (he took), „ (er) nahm. 
 
 O.E. preserves the corresponding tenses nim-an — num-en — 
 nam, whereas E. has lost the verb altogether. But let the 
 student again compare, form with form, the three verbs : 
 
 Go. qim-an (to come), qum-an-s (come), qam (he came), 
 
 E. (to) come, come, (he) came, 
 
 G. komm-en, (ge-)komm-en, (er) ham, 
 
 wherein both G. and E., — though they have obscured the nor- 
 mal grade by introducing into the infinitive the labial vowel 
 of the participle, a process favoured moreover by the labial 
 nature of the preceding consonant, — exhibit the deflected form 
 as clearly as Gothic itself. 
 
 5. Lastly, the e is followed by a liquid : I.-E. er — r — or ; 
 
 1 Whence the E. suffix -less, infra 110, VIII. 
 
 2 With o instead of a before a nasal, xupra 39, 3, note. English has 
 bound, through confusion with the past participle ; but it has kept in the 
 derivation the old word bond (a chain) = G. band, and the corresponding 
 *band at least in the metaphonical form bend ; E. band is borrowed from F. 
 bande, which is itself taken from G. band. 
 
 8 The double hi, of course, as well as the h in nehmen nahm, is but a 
 device to denote that the vowel in the one case is short, and in the other 
 long. 
 
80 ENGLISH AND GEKMAN GRAMMAR. 
 
 whence Preg. er — ur or or — ar, and the same series for I. Lei 
 us consider the two roots, I.-E. wert (" to turn," Sk. vdrt-ati 
 " he turns," L. vert-ere, the Germanic meaning being " to be- 
 come "), and Preg. help " to help." 
 
 Normal : Go. vairfc-an (to become), G. werd-en ; 
 
 Reduced : „ vaurp-an-s (become), „ (ge-)word-en ; 
 
 Deflected : „ var}> (he became), „ (er) ward. 
 
 Normal : „ hilp-an (to help), E. (to) help, „ helf-en ; 
 
 Reduced : „ hulp-an-s (helped), „ (ge-)holf-en ; 
 
 Deflected: ,, halp (he helped), „ (er) half. 
 
 English has lost the former and the gradation in the latter ; 
 but O.E. has the corresponding forms : weorban — worden — 
 wearb, helpan — holpen — healp. In the plural perfect (reduced 
 grade) Gothic has, also quite regularly, vaurj>-um = O.E. wurd- 
 on = G. (wir) wurd-en, which form, transported by analogy to 
 the sg., occasions a German type ich wtirde, existing side by 
 side with the primitive ich ward. 
 
 II. If the syllable, instead of e, has a long vowel a, e, 5, 
 then the correspondences in vowel-gradation will theoretically 
 be: 
 
 I.-E. 
 Preg. 
 
 Here, the second class is kept apart from the first and third 
 classes, which latter become blended together, and further, do 
 not distinguish the deflected from the normal grade. Thus 
 the three are reduced to two, namely the sixth and seventh 
 classes in our Germanic terminology. 
 
 6. The vowel is an I.-E. and Preg. e, as in the I.-E. root dhe 
 (to set, to do) : normal Gr. 6-fj-o-w (I will set) and L. fe-c-it (he 
 did) ; reduced #e-To-s (placed) ; deflected 6o>-fx.6-s (heap). Preg. 
 e — & — o. 
 
 Normal : Go. let-an (to let), E. (to) let, G. lass-en ; 
 Reduced : „ lat-s (loose, idle), O.E. Icet, „ lasz ; x 
 
 Deflected : „ (lai-)ldt (he let). 
 
 1 The vowel-gradation is here much obscured by later processes in each 
 language. We need only observe that the short a in Icet corresponds to a 
 
 a 
 
 a 
 
 
 
 e 
 
 a 
 
 
 
 
 
 a 
 
 
 
 6 
 
 a 
 
 
 
 e 
 
 & 
 
 6 
 
 
 
 a 
 
 
 
SEMI-VOWELS AND CONSONANT- VOWELS. 81 
 
 Here G. and E. have no longer any trace of the deflected 
 grade. But it may be seen in the offspring of the I.-E. root 
 dhe quoted above, viz. E. (to) do = OM. do-n, and G. thun = 
 O.H.G. tuo-n, in contrast with the normal grade we had already 
 occasion to point out in the E. deed = Gr. that. 
 
 7. The normal vowel is an I.-E. d or o, Preg. 6 in either 
 case : normal Gr. oTa-crou cm^-cra) (I shall set), reduced otS-to-s 
 (placed) ; normal Gr. Su-crco (I shall give), reduced So-to-s 
 (given). The Germanic correspondences are here irregular, in- 
 asmuch as the vowel in the normal grade has become the same 
 as in the reduced forms, in order to maintain, as it were, with 
 the utmost clearness the peculiar characteristic of the perfect 
 tense, which otherwise would exhibit no difference in vocalism 
 from the present. Thus, instead of 6 — a — o, we have Preg. 
 a — a — 6. 
 
 Normal: Go. slah-an (to smite), E. (to) slay, G. schlag-en; 
 Reduced: ,, slah-an-s (smitten), „ slai-n, „ (ge-)schlag-en ; 
 Deflected: „ sloh (he smote), „ (he) slew, ,, (er) schlug. 
 
 German schlug = O.H.G. schluog,iB quite regular, and English 
 slew (pr. slii) is nothing but a conventional spelling for oo — o 
 (O.E. slog). 1 
 
 Germanic a, whereas the long <§ in Icetan is a Preg. e {supra 26, II.), now 
 shortened in let, just as the a in O.H.G. la^an appears now shortened in G. 
 lassen. Thus the regular correspondences are quite obvious. 
 
 1 This section, it must be remarked, belongs to the study of conjugation 
 almost as much as to phonetics; wherefore, important as it is, it cannot be 
 thoroughly understood, unless we subjoin to it the theory of strong perfects 
 and past participles, as explained below 175-184. 
 
CHAPTER IV. 
 EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 
 
 (46) The first view of the consonantal system in both 
 English and German shows us the impossibility of studying 
 the explosives and the spirants apart from one another ; for we 
 often find an English spirant corresponding to a German explo- 
 sive, or the reverse. It is, therefore, convenient to compare 
 together, in the same chapter, the explosives of both 
 languages, and the spirants of either where these are 
 merely substitutes for explosives of the other. Of 
 course, such sounds as exhibit the character of spirants in both 
 form a class by themselves. 
 
 Here, at the beginning, a general restriction is necessary : 
 none of the laws that will be stated below can apply either to 
 a final explosive or to its substitute, because every Indo- 
 European final consonant (properly so-called), with the 
 single exception of s, is dropped in Pregermanic, and 
 leaves no trace of its former existence in the later languages 1 ; 
 whereupon the vowel preceding the lost consonant is treated 
 in its turn as a final vowel. 2 As for such consonants as be- 
 came final after the loss (in Germanic) of a final vowel, we 
 scarce need observe that they were primitively medial, and 
 consequently are comprised as such in the following exposition. 
 
 1 The same is known to happen in Greek ; cf. Henry, Comp. Gr. of Gr. 
 and Lat., no. 65. Occasionally the consonant is kept, in the very rare case 
 when it was covered by an affixed vocalic particle: cf. G. dasz =E. that = Go. 
 )>at-a = Sk. tad, in the declension of the article, infra 130. 
 
 » Cf. supra 34, § 2. 
 
 83 
 
EXPLOSIVE CONSONANTS AND THEIE SUBSTITUTES. 83 
 
 Section I. 
 
 THE SECOND CONSONANTAL SHIFTING. 
 
 (47) In considering, in English and German, such instances 
 as to think and dunk-en, thumb and dawn-en, thin and dilnn, thorn 
 and dom, which might be indefinitely multiplied, we see at 
 once that German uniformly represents the English voiceless 
 spirant x by a voiced explosive of the same order. 
 
 Now, if we glance at the other languages of the Germanic 
 family, we find in them, with equal consistency, the same 
 spirant as in English : Go. J?agk-jan (to think) ; O.N. pum-all 
 (thumb), which would be Go. *pum-al-s ; O.K. fiunn-r (thin), 
 which would be Go. *}>unnu-s ; Go. paumu-s and O.N. porn 
 (thorn), etc. Thus, at the outset, we are led to the supposition 
 that English has preserved an old and Pregermanic consonantal 
 state which became in some way modified in High German. 
 
 Looking for further contrasts, still in the same order of con- 
 sonants (dentals), and at the beginning of a word, we find, for 
 instance, death and tod, deed and that, day and tag, etc., that is 
 to say, a German voiceless explosive corresponding to an English 
 voiced explosive ; and here, again, the related languages are at 
 one with English, Go. ddupu-s (death), defr-s (deed) and O.N. 
 dad, Go. dags and O.N. dag-r (day). 2 Further, we find tongue 
 and zunge, timber and zimmer (room), 3 ten and zehn, namely a 
 voiceless simple consonant in English corresponding to a voice- 
 less double one in German ; and again it is the English sound 
 we find in Go. tuggo (tongue), ga-timr-jan (to build, to timber), 
 taihun (ten), O.N. tunga, timbr, Swede tiu (ten), etc. 
 
 If we pass from the dentals to the two other orders of conso- 
 nants, gutturals and labials, a series of similar phenomena 
 occur, though with less exactness and regularity, so that the 
 whole process here described may be summed up in a brief and 
 
 1 Occasionally voiced (the = dU, that = das, thou = du), in consequence of 
 an English process of softening which will be explained, infra 49, I. 2. 
 
 * In this and the following case, Low German agrees with English. 
 
 8 The etymological meaning is still seen in zimmer-mann (carpenter) and 
 similar words. 
 
84 ENGLISH AND GERMAN GRAMMAR. 
 
 provisional phonetic law, which our further analysis will define 
 with greater precision, as follows : 
 
 The Pregermanic consonantal state seems to have 
 been preserved almost unaltered in East and North Ger- 
 manic, and likewise in the English and Low German branches 
 of West Germanic, whereas High German shows a 
 general tendency to substitute, in each order, a voiced 
 explosive for a voiceless spirant (d for J>), a voiceless 
 explosive for a voiced explosive (t for d), and a voice- 
 less spirant for a voiceless explosive (z = ts for t). 
 
 This principle was discovered by Grimm, though he assigned 
 to it some effects which are now known to stand beyond its 
 reach. It is the second Grimm's law or Second Conso- 
 nantal Shifting (zweite Lautverschiebung) , so called in opposi- 
 tion to the first consonantal shifting, which is common to all 
 the Germanic dialects, and must, therefore, be deemed much 
 earlier. We cannot go back to the latter, until we have traced, 
 throughout each order of English and German consonants, a'l 
 the consequences of the former. 
 
 § 1. Labials. 
 
 (48) I. Since our formula has been said to be of a provi- 
 sional and approximate character, we cannot illustrate it better 
 than by particularizing a case in which it proves ineffective, 
 German having kept without alteration the Germanic conso- 
 nantal type. This is the case of the voiceless labial spirant /, 
 which appears common to all the Germanic dialects, though it 
 will be seen further on to proceed- from an Indo-European ex- 
 plosive. 
 
 1 . Preg. /= E. /= G. / (often written v). 1 
 
 Examples are : Go. fadar, E. father, G. vater ; Go. fisk-s, O.N. 
 fish-r, ~E.fish, Q.fisch ; O.N. flesh, E. flesh, G. fleisch; Go. fugl-s, 
 E. fowl, G. vogel; Go. fimf-taihun and fimf tig jus, ~E.fif-teen and 
 
 1 This is merely conventional spelling, the actual pronunciation being 
 always /. As to the origin, however, of the double orthography, the reader 
 may consult Wilmanns' D. Gr., § 93, 94 : Preg. / primitively differed from 
 German /= Preg. p, ivfra HI. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 85 
 
 fif-ty, G. funf-zehn and fiinf-zig ; O.N. stif-r, E. stiff ( = O.E. $Hf), 
 G. steif; Go. vulf-s, E. wolf, G. wolf; O.E. advb. scfte = E. soft 1 
 = G. sanft, etc. 
 
 2. E./ between two vowels becomes voiced and softened to 
 v : cf. the plural wolv-es = O.E. wulf-as, and G. wblf-e ; E. oven 
 = O.E. o/en = O.N. o/« = G. q/e» ; E. s7ioveZ = O.E. sceq/Z = G. 
 schaufel. 2 This process of softening extends to all spirants. 
 
 II. From the voiceless spirant we now pass to the voiced 
 spirant (b), written v in English. Here the German sound no 
 longer appears the same : it is an explosive, the voiced b as in 
 G. geben = ~E. give, G. schaben (to scrape) = E. shave. German, 
 however, exhibits some remains of an earlier b, inasmuch as the 
 spirant has been kept, though unvoiced, before a voiceless con- 
 sonant, in such derivatives as G. gif-t (poison) =E. gif-t, 3 G. 
 schaf-t = ~E. shaf-t* etc. Thus, here as elsewhere, and accord- 
 ing to the testimony of German itself, the English spirant is the 
 original consonant, viz. : 
 
 1. Preg. b = E. v = G. b. 
 
 Examples: Go. gib-an, O.K gef-a, O.E. gif-an'° and E. (to) 
 give, O.H.G. geb-an and G. geb-en : Go. skab-an (to scratch, to 
 scrape), O.N. skaf-a, O.E. sceaf-an and E. (to) shave, O.H.G. 
 
 1 The word *suvfte has become O.E. *sonfte under the influence of the 
 nasal, and then softe through compensatory lengthening, lastly soft as short- 
 ened before a group of consonants. Here the o could not return to a (as 
 it did in O.E. mon which is now E. viaii), because it had been lengthened in 
 the meantime. 
 
 2 Even in other positions, the student will find, in English as elsewhere, 
 some irregular processes of softening or strengthening, which are very easy 
 to explain. Such a word as O.E. fif, if occurring before a voiced consonant, 
 as for instance in the collocation fif dag as (five days), could not but be 
 sounded as fiv, which form extending throughout became E. five. These 
 assimilating shiftings of a voiceless to a voiced consonant, or of a voiced to 
 a voiceless consonant, are occasionally to be met with in any spoken lan- 
 guage, especially in familiar speech, so that this observation will not need 
 repetition. 
 
 8 Strange as it is, the German euphemism becomes quite intelligible when 
 compared with F. and E. poison, which means really nothing more than 
 "drink" Lat. potionem. 
 
 * Literally " [a thing] scraped or planed." 
 
 s Go. b between vowels is a 15, as noted above. O.N. and O.E. /, in the 
 somewhat clumsy orthography of these languages, is an ambiguous symbol 
 which may correspond either to the voiceless or the voiced spirant. Here 
 no doubt it stands for a v. 
 
86 ENGLISH AND GEEMAN GRAMMAR. 
 
 seab-an and G. schab-en; Go. ibn-s (plane), O.E. efn and E. even, 
 O.H.G. iiban and G-. eben ; Go. liuf-s (gen. liub-is), O.N. Ijuf-r, 
 O.E. leof and E. Z« e /, O.H.G. Hob and G. Ziefc (dear) ; Go. Jcalb-o 
 (a female calf), whereof the msc. would be *kalf-s, O.N. Tcalf-r, 
 O.E. ceaZ/ and E. calf, O.H.G. chalb and G. halb. From the last 
 two instances we may infer that, in contrast with the change of 
 intervocalic / to v = b, original b becomes shifted to/ when final 
 in English; compare the voiced consonant kept in E. love = G. 
 liebe, E. pi. wives 1 of sg. wif e = O.T&. wif=G. weib, E. to live = Gr. 
 leben, but substantive life, etc. 
 
 2. If under the influence of an adjacent voiceless consonant, 
 primitive b has been unvoiced (/), it then naturally obeys the 
 same law as /, that is to say, it is preserved in German as 
 well as English : compare the instances quoted above, and add 
 G. half-te derived from */<aZb = E. half=G. halb. 
 
 3. Preg. b, when initial or medial after a nasal, was 
 changed to b in the Pregermanic period, wherefore it appears 
 as such in all the later languages. 2 Hence in this respect 
 again English and German are in full agreement. We have 
 already given : book — buch, beech — buche, bull = bulle, to bind — 
 binden etc. Here we may add: E. to bite = G. beiszen (Go. 
 beitan), and E. bit = G. bisz; E. buch = G. bock (he-goat); E. 
 brother = G. bruder (Alam. pruader) ; E. bride (with metaphony) 
 = G. braut ; E. bone (O.E. bdn) = G. bein = ~Preg. *bain-; E. to 
 blow = G. metaph. blahen (cf. O.E. bldwan) ; E. beard = G. bart ; 
 and a great many others. After a nasal : E. comb = G. Jcamm; 
 E. womb — G. wamme ; because the group mb, though kept in 
 English spelling, here becomes nearly assimilated to mm, and 
 completely assimilated to it in German writing and pronuncia- 
 tion. 
 
 1 German also has a general tendency to unvoice every voiced final con- 
 sonant, a process which even influenced the orthography of M.H.G., such 
 words being there preferably spelled Hep, kalp, etc. (but lieb-en, kelb-er). The 
 spelling has now been made uniform, and literary German spells lieb, hand, 
 tod, though the actual sound is rather Zip, hant, tot, etc. 
 
 2 But Alamannic, as we have often remarked, changes it to p. This is the 
 reason why Grimm, considering that Alamannic here underwent a " shifting " 
 (verschiebung) which he missed in High German, wrongly deemed that dialect 
 to be the only true representative of the Hochdeutsch, and favoured it with 
 the title of Strengalthuchdeutsch. Supra 47 in fine. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 87 
 
 III. We have now traversed the whole system of Preger- 
 manic labials, with the exception of the voiceless explosive, 
 English p. 
 
 1. E. p never corresponds to G. p, unless it stands in the 
 group sp. Compare : E. to spew and G. spei-en ; E. to spin and 
 G. spinn-en ; E. spear and G. speer. 
 
 2. Everywhere else we find the correspondence already men- 
 tioned in our general formula, namely, the G. voiceless spirant 
 instead of the E. voiceless explosive, thus : 
 
 Preg. p = E. p = G. pf or ff or /. 
 
 The High German change is as follows : the p, first, devel- 
 oped after itself an aspirate, ph ; then, the aspirate ph became 
 pf, 1 which remained when initial. But, when medial, the ex- 
 plosive combined with the following spirant, the group thus 
 resulting in ff : 2 Go. vairp-an (to throw), E. to icarp, O.H.G. 
 w'erpf-an werfan, G. werfen ; Go. (ga-)skapjan, O.E. scyffan and 
 E. to shape, O.H.G. scepfen and scaffan, G. schopfen and schaffen ; 
 E. sleep, G. schlaf, schlafen, and schlaff (lazy). Initial p is very 
 rare in Pregermanic, 3 and can scarcely be found anywhere but 
 in the early borrowings from Latin adopted by the West Ger- 
 manic speech, but here at least it is quite common : L. patina, 
 E. pan, G. pfanne ; L. palus, E. pale, G. pfahl ; L. pondo, E. 
 pound, G. pfund ; L. pavo, E. pea(-cock), G. pfau; L. planta, E. 
 plant, G. pflanze ; L. Pentecosta " fiftieth [day]," G. Pfingst-en 
 (Whitsunday) ; Late L. paraveredus (cf. P. palefroi and E. pal- 
 frey), O.H.G. pfarij rid, G. pferd (horse), etc. 
 
 IV. Putting together the results of the last paragraph and 
 those we had already obtained 4 in regard to the Pregermanic 
 
 1 Cf. supra 14, 3 A. 
 
 2 Reduced to simple / after a long syllable. In certain dialects initial pf 
 likewise became /, whereas on the contrary, in certain others, medial pf sur- 
 vived ; lastly, the Low German dialects, which keep the p unaltered, intro- 
 duced a few words into High German : hence the existence of such double 
 forms as Jiaum and pflaum (Lat. pluma), schaffen (to work) and schopfen (to 
 create, a false spelling for the regular metaphonical schepfen), pabst (pope) 
 and pfaffe (priest) = Late Lat. papa. Further, for these insignificant varia- 
 tions, we must always take into account the process of spontaneous redupli- 
 cation in Germanic explosives, as stated above, no. 36 note. 
 
 8 The reason will be seen below (57, L). 
 4 Cf. supra 37. 
 
OO ENGLISH AND GERMAN GRAMMAR. 
 
 labials, we may summarise the correspondences between these 
 consonants as follows : — 
 
 explosives 4 \ ; \ 
 
 C voiceless: „ p = „ pf,ff,f,p; „ p = „ p (rare). 
 
 spirants ( Voiced : »*=>»&./. w 5 » « = » »■ 
 
 (. voiceless : „ /= „ /, 6 j „ /, /, 23/= „ /, p. 
 
 § 2. Dentals. 
 
 (49) I. "VVe have already stated the general correspond- 
 ence : 
 
 1. Preg. > = E. > (written O.E. o, E. ifc) = G. «2. 
 
 In other words, the voiceless dental spirant, kept everywhere 
 else, first became in German a voiced spirant (0) in all posi- 
 tions, and then this spirant was changed to the corresponding 
 explosive, 1 just as b was shifted to b. Farther instances are : 
 O.E. burh and E. through, O.H.G. duruh and G. dutch ; Go. 
 J>usundi, O.E. busend and E. thousand, O.H.G. dusunt tusunt 
 and G. tausend 2 ; E. path — G. pfad ; Go. anfiar (other, second) 
 = Sk. dntara-s, whence O.E. *anber, *oriber, lastly ooer = E. other, 
 O.H.G. andar and G. ander ; Go. 6ro/»ar = L. frdter, O.E. brobor 
 and E. brother, O.H.G. bruoder pruoder and G. bruder ; O.N". 
 fjopr, O.E. feber and E. feather, O.H.G. fedara, M.H.G. vedere 
 and G. feder, etc. 
 
 2. From the last three examples it follows that, in English 
 and between two vowels, just as / becomes v, so ]> is voiced and 
 softened to &, leaving, however, the spelling unchanged. Now, 
 initial th also may occasionally become voiced, either owing to 
 the weak accent of the little words in which this softening is 
 seen, 3 or because, — since these words became one with the pre- 
 ceding word, in such collocations as by thee, to them, do that, 
 etc., — the th actually stood between two vowels. Analogy after- 
 
 1 A similar shifting is to be observed in some dialectal or bad pronuncia- 
 tions of Modern English, de = the, dey = they, dat = that, etc., and in Low 
 German throughout. 
 
 2 An apparent irregularity in Modern German, owing to a later alternation 
 between d and t, as shown by the hesitation in O.H.G. spelling. The irregu- 
 larity E. father = G. vater, E- mother =G. mutter, etc., is more puzzling ; but 
 here English is answerable for the discord, infra II. 2. 
 
 8 On the effect of atony, see below 66, II. 4. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 89 
 
 wards extended the voiced pronunciation to other cases, such as 
 thou, they, their, and so forth. 
 
 3. The Germanic group lp had become Id in O.E., as later in 
 German : whence the complete identity in E. gold = G. gold = 
 Go. gulp ; E. wild = G. wild = Go. vilfiei-s, etc. 
 
 II. Preg. b has undergone the same changes as Preg. b, 
 with the only difference that the former passed through a 
 phonetic grade more. Initial, or medial after a nasal, it became 
 Preg. d, but was kept as ordinary medial b, and is still seen as 
 such in the Gothic survivals ; in West Germanic, this b again 
 was changed to d in all positions, a grade in which it continues 
 throughout in English ; lastly, High German shifts the d to t, 
 in any position 1 except when following a nasal. 2 Hence the 
 following simple correspondences : 
 
 1. Preg. b = E. d=G. t when initial. Examples as above, 
 and also : Go. dauhtar, O.E. dohtor and E. daughter, O.H.G. 
 tohter and G. tochter ; Go. dius (wild beast), O.E. deor and E. 
 deer, O.H.G. tior and G. thier 3 ; Go. dreiban, O.E. drifan and E. 
 to drive, O.H.G. triban and G. treiben ; O.N. draum-r, E. dream, 
 O.H.G. troum and G. traum, etc. 
 
 2. Preg. b = E. d = G. t when medial: Go. fadar* O.N. 
 faber, O.E. feeder, O.H.G. fater and G. vater ; O.N. mober, O.E. 
 modor, b O.H.G. muotar and G. mutter ; Go. nadr-s, O.N. nabr, 
 
 1 It must of course be understood that the whole process had taken place 
 before — and perhaps a very long time before — the existence of the new voiced 
 spirant 6, which has been seen to arise in German from a Preg. >>, and then 
 not to have advanced beyond the grade of d ; else, the latter would have 
 become blended with the former, and both would have advanced to the grade 
 of t. The consonantal shifting is unintelligible, if supposed to have taken 
 place " at once," as it was described in Grimm's times : on the contrary, it 
 should always be borne in mind that this shifting is only the main result of 
 many little causes, which, though similar in their principle, wrought their 
 effects successively and at long intervals. The importance of chronology in 
 phonetics (supra 20, note) can be nowhere more deeply felt. 
 
 2 As a matter of fact, the d became O.H.G. t even after a nasal (bintan, 
 blinter) ; but the group nt was generally softened to M.H.G. nd (binden, 
 blinder) : so that the result is the same. Cf. however E. under =G. unter. 
 
 3 The original meaning still seen in thier-garten. And so too, E. deer 
 with the general sense of " game," down to the time of Shakespeare (once, 
 King Lear, III. 4, 144). 
 
 * The reader must not forget that Go. intervocalic d is a ff . 
 8 Here, Mod. E. has the Go. and O.N. consonant, father, mother ; this, 
 however, is not a survival, but a restoration, and a very late one, being hardly 
 
90 ENGLISH AND GERMAN GRAMMAR. 
 
 O.E. nceddre and E. adder, 1 O.H.G. natara and G. natter, cf. also 
 middle = mittel, saddle = sattel, etc. ; Go. mods (anger, gen. mod- 
 is), O.E. mod and E. mood, O.H.Gr. muot and G. muth (conrage) ; 
 cf. also E. tide = G. zeit (time), E. seed = Gt. saat, etc. 
 
 3. Preg. b, then cZ = E. d = Gr. d, after a nasal : compare hand 
 and hand (O.H.G. hant, M.H.G. hant owing to the unvoiced 
 final), land and land, hind and binden, blind and blind (O.H.Gr. 
 Mint), etc. 
 
 III. The shifting of the voiceless dental explosive is iden- 
 tical with that of the voiceless labial : initial, in German, we 
 have ts instead of t (as pf instead of p) ; medial it becomes ss 
 (like ff in the former case) . 
 
 1. Preg. t = ~E. t = G. t, only after s, ch (E. gh) and /, which 
 consonants preserve the t in German : E. stand (position), G. 
 stand ; E. first, G. fiirst ; E. metaph. fist, G. faust ; E. daughter, 
 G. tochter ; E. flight, Gt.flucht ; E. oft oft-en, G. oft, etc. 
 
 2. In any other position, Preg. t, retained in E., first 
 became aspirated to G. th, which further became a t accom- 
 panied by the corresponding spirant s : the group ts, spelled z, 
 survived when initial and, occasionally, when medial in 
 dialectal forms, but generally when medial it became a mere 
 sibilant, written O.H.G. 33,2 and now sounded like a voiceless 
 s (spelled ss or sz, according to a well known distinction). — 
 Examples : — When initial : no. 47 above, and Go. ga-tam-jan (to 
 tame), O.N. tam-r (tame), E. tame, G. zahm and zdhmen; E. 
 
 found before the year 1500. Mr. Skeat (Principles, I. p. 147) seems to ex- 
 plain it solely by the analogy of the regular brother ; but, since we find it 
 also in E. weather = 0.~E. weder = G. wetter, which is not a word of relation- 
 ship, it is more likely to be attributable to some obscure influence of the 
 following, r, though both causes may have combined, and though no such 
 has taken place in adder = natter, fodder =J hitter, udder =eutcr; for, in the 
 latter words, the explosive had been doubled, as following a short vowel. 
 On two other and very momentous irregularities, fifth =fiinfte and similar 
 cases, and he hath (now has) = erhat and similar cases, cf. infra 124 and 202, 
 3. In forth=G. fort = Go. *faur)> (the comparative being fauri>-is = E. 
 furth-er), the G. consonant is really a d (O.H.G. would be *ford), but unvoiced 
 as final, supra 48, II. 1. 
 
 1 The word-group a *nadder was separated as an adder (as in F. ma mie = 
 m' amie), whence the irregular dropping of an initial n. Cf. also E. apron 
 borrowed from M.F. naperon (napkin). 
 
 2 The precise value is unknown, but it was no doubt a voiceless sibilant. 
 After a long vowel, or final, it is spelled 3. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 91 
 
 tin, two, tale, to tell, and G. zinn, zwei, zahl, z'dhlen; — When 
 medial: Go. batiza (better), O.E. metaph. betera and E. better, 
 O.H.G. be^iro, M.H.G. beefier and G. besser ; Go. vat-o, O.N. 
 vat-n, O.E. wcet-er and E. water, O.H.G. wa^ar and G. wasser, 
 cf. to bite = beiszen, and bit = bisz; Go. sut-s, O.N. soet-r, O.E. 
 swete and E. swee£, O.H.G. suoji and G. s-iisz; E. sweat = G. 
 schweisz, and E. to sweat! = G. schwitzen, showing the alternation 
 between is and ss. But, after a nasal or a liquid, the substitute 
 for Preg. t regularly remains ts : Go. hairt-6, E. heart, G. herz ; 
 E. salt, G. salz ; L. moneta, borrowed in E. mint, and in G. 
 miinze. 
 
 IV. Recapitulation of dental correspondences : 
 
 explo- C voiced 
 sives I voiceless 
 
 E. d «G. «, d ; 
 
 G. d =E. 
 
 th, d. 
 
 „ t = „ z, sz, ss, t ; 
 
 >> ■ ~ jj 
 
 d (th), t 
 
 spi- ( voiced 
 rants 1 voiceless 
 
 „ th= „ d (0 ; 
 
 i 
 
 
 „ th = „ d; 
 
 „ f, sz, ss= „ 
 
 t. 
 
 § 3. Gutturals. 
 
 (50) I. The Preg. voiceless spirant (h) is in some cases a 
 mere aspirate (written h) ; in other cases, it is a true spirant 
 (written G. ch) ; while, in several positions, it has been dropped 
 altogether. 
 
 1. Preg. 7i = E. h=G. h when initial. — If, however, it 
 precedes I, r or n, both languages agree in ceasing to sound or 
 even to write it. The same is the case in German before w, 
 whilst English still spells the group (inverted) toh, but can 
 hardly be said to preserve it always in pronunciation. 2 — Ex- 
 amples : — Go. hauh-s, O.E. heah and E. high, O.H.G. hoh and 
 G. hoch ; O.N. hdr, O.E. holr and E. hair, O.H.G. hdr and G. 
 haar ; O.N. hagl, O.E. hagel and E. hail, O.H.G. hagal and G. 
 hagel; cf. further to hate = hassen, hand = hand, to hold = halten, 
 hard = hart, hundred = hundert, etc. ;— Go. hlaj>an, O.E. hladan 
 and E. to lade, O.H.G. hladan ladan and G. laden; cf. to leap = 
 laufen, to laugh = lachen, loud = laut, etc. ;— O.N. hring-r (circle), 
 O.E. hring and E. ring, O.H.G. hring and G. ring; O.N. hrafn, 
 
 1 Neither the G. voiced spirant (s pronounced E. z), nor the E. spirants s 
 and z belong to this chapter. 
 
 2 In fact E. ich is an unvoiced w. 
 
92 ENGLISH AND GERMAN GRAMMAR. 
 
 O.E. hrcefn and E. raven, O.H.G. hraban and G. rahe, etc. ; — 
 O.N. knot, O.E. hnutu and E. nut, O.H.Gr. nwz, and G. nusz, etc. ; 
 — Go. hvditei-s, O.N. hveite, O.E. hwcete and G. wheat, O.H.Gr. 
 wei^i and Gr. weizen, cf. likewise white — weisz; O.N. hvefja, O.E. 
 hwettan and E. to wTiei, O.H.G. wezzen and G. weteew; initial of 
 relative pronouns, what = was, etc. 
 
 2. Medial Preg. 7i = E. grfr = G. ch or simply 7L 
 The spirant h was retained in O.E. (h) and in E. (spelled 
 gh), wherenpon, at least when final, it was changed to the 
 labial spirant (/), while' in other positions it was dropped by 
 the same process that affected the voiced spirant j ; but all 
 these changes usually left the spelling unaltered. The gh, 
 moreover, happened occasionally to be sounded in certain 
 positions where it ought to have become mute. This was due 
 to the analogy of words in which it was regularly sounded, as 
 in laughter pronounced like to laugh. 1 Or it became mute 
 where it should have been sounded, as in through, owing to the 
 numerous cases where this word was followed by an initial 
 consonant. 2 We need but compare, on the one side, to laugh 
 and lachen, rough and *rauch ; 3 and, on the other, might and 
 macht (power), night and nacht, neighbour and nachbar, light 
 and licht, light and leicht, etc. 
 
 Gr. h is here seen to remain a spirant (spelled Mod. Gr. ch) 
 before a consonant or when final ; but, between two vowels, it 
 was reduced to the simple aspirate (h), which moreover has 
 now become a mere graphic symbol : compare, for instance, the 
 varying sound in hoch and hohe hoher (O.H.G. hoh " high," hohl 
 "height," hohir "higher"), in nach (E. nigh) and n'ahe ndher 
 (E. near, where the spirant has vanished as in toe = G. zehe), in 
 
 1 Cf. slaughter, which was not subject to a similar analogy, so that the 
 gh in it has become mute, in contrast with the G. vb. schlachten. The 
 dialects still distinguish sg. enough = M.E. inoh, and pi. enou; = M.E. inow-e ; 
 cf. infra II. 5. 
 
 2 gh is always mute after a palatal vowel (cf. infra II. 4) : high, nigh, 
 thigh, to neigh, etc. In some words, as in E. /<ee = O.E. feoh = G. vieh, the 
 mute gh even ceased to be spelled. 
 
 8 This is the true and original form, — still preserved in Alamannic ruch 
 (rough), and in class. G. rauch-werk (rough skins, peltry), — of the word 
 which, when isolated, has become rauh through a confusion that will be 
 explained below. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 93 
 
 sicht and sehen (Go. saihvan)=TZ. sight and (to) see, etc. Of 
 Course one of the two forms will occasionally be found to have 
 exerted an analogical influence upon the other ; thus some 
 dialects have hocher as comparative of hoch, and classical 
 German has an adjective rauh (rough) formed from the com- 
 parative rauher, etc. 1 
 
 II. The voiced guttural spirant (3) may be treated on the 
 same lines as the other voiced spirants, though they differ in 
 some particulars. Preg. 3 was a true spirant in any position, 
 even when initial, and had never become g, unless it followed a 
 nasal ; in the latter case this Preg. 3 was changed to G. g and, 
 usually, to E. y. 
 
 G. g is kept in the spelling, but dialectally assumed, between 
 two vowels or after a liquid, a spirant sound which partly 
 invaded even the classical speech. 2 
 
 English has no g, except as initial, in which case it is some- 
 times spelled gu (guest = G. gast) or gh (ghost = G. geist), so as 
 to insist upon its explosive value. Even when initial, if 
 followed by e or i, the g had usually become a mere y as early 
 as the Anglo-Saxon period. 3 When medial, it was gradually 
 adapted to the character of the preceding vowel. It was 
 written y if the vowel was i, e or a, and w if the preceding 
 vowel was u or 0, and combined with the vocalic sound in such 
 a way as to disappear throughout, sometimes even in spelling 
 (chiefly after an i). On the contrary, when it had become 
 final, it underwent occasionally, and chiefly when preceded by 
 u or 0, a strengthening to gh, as illustrated by the typical 
 instances: borough=Gr. burg*; dough = G. teig, and enough = G. 
 genug. 5 
 
 1. Preg. 3 changed to g after a nasal = E. g = &. g: Go. vb. 
 
 1 So also schuh = 'E. shoe, for *schuch, from the pi. schuhe, and er sah 
 (he saw), for er *sach, from the pi. sie sahen, etc. 
 
 2 It is known to hesitate, for instance, between legn and le%n = leg>n, 
 often with a preference for the latter. Some dialects go as far as the grade 
 of y : Germans are well known to make fun of the Berlinese pronunciation 
 of G. g. 
 
 3 The fact is proved by the O.E. spelling g (5) the initial of some words 
 which in I.-E. began with a y, and in consequence certainly never had a true 
 initial g. Cf. supra 36, I. 
 
 * Cf. O.E. burg and bnrh, dative lyrig = *burg-i. ^^"Z?~ ■ 
 
 * Cf. however infra 55. /r^" . ~' c 
 
 (UNI 7Z •— j.7 "■>" 
 
94 ENGLISH AND GERMAN GRAMMAR. 
 
 huggr-jan, O.N. hungr, O.E. hungor and E. hunger, O.H.G. 
 hungar and G. hunger ; Go. figgr-s, E. finger, G. finger ; E. tongue, 
 G. zzmoe ; E. Zowa, G. Zawgr. Observe moreover that the group 
 fig, when swiftly uttered in ordinary speech, is barely dis- 
 tinguishable from a simple A. 
 
 2. Initial Preg. j = E. o = G. g: Go. aas£-s ( = *gasti-s, L. 
 hosti-s), E. gwesi, G. <7as£ ; Go. gdcZ-s, O.N. add-r, O.E. god and 
 E. flood", O.H.G. guot and G. ow£; cf. likewise God and 6?ott. 
 Even sometimes before e or i': 1 to give = geben ; gift = gift 
 (poison) ; £o gild=OM. gyldan, cf. G. ooZo". 
 
 3. Initial Preg. 2 = E. y = G. g: Go. gairn-jan (to wish 
 for), O.N. g^'am and O.E. georn (covetous), E. to yearn, O.H.G. 
 gerno and G. gem (willingly) ; O.E. geolo and E. yellow, O.H.G. 
 gelo (gen. gelw-es) and G. gelb ; E. yester(-day) — G. gester-n ; E. 
 fo yield = G. gelten, etc. The Go. prefix oa- (O.E. ge- = G. ge-) 
 is still spelled ?/- in M.E. ; then it disappears, without leaving 
 any trace but, occasionally, an ill-defined vowel, which appears 
 as a in asunder = G. gesonder(t) (severed), and e in enough = G. 
 genug. 
 
 4. Medial Preg. j = E. y = G. g. — Following a or e : Go. 
 dags, O.N. dag-r, O.E. doeo and E. day, O.H.G. tac and G. tagr ; 
 Go. vig-s, O.E. wea and E. way, 2 O.H.G. wee and G. weg ; Go. 
 dugo, O.N. awga, O.E. eage and E. eye, O.H.G. ouga and G. auge ; 
 Go. fagr-s (convenient, able), O.E. fceger and E. fair, cf. G. 
 fegen (to sweep) s ; Go. magap-s, O.E. maigd and E. maid, O.H.G. 
 magad and G. mago" (dialectal maid). Compare also: E. to 
 lay = G. legen ; E. lain = G. (ge-)legen ; E. raw, wai'Z, 7uw7, sat'Z, 
 £o say, and G. regen, nagel, hagel, segel, sagen, etc. — Following i, 
 the a seems to have been exceptionally retained in E. twig = G. 
 zweig. But, everywhere else, it becomes y, which combines with 
 
 1 It is easy to verify the fact that g appears only before a, o, v: guest is 
 the metaphony of gast ; to gild has by analogy the consonant of guld gold ; 
 to get, that of got; to give, that of gave, and so forth. The word guilt 
 guilty comes from the root of the O.E. vb. geldan (to compensate, G. gelten, 
 to be worth, and vergelten, to compensate), which formed a pf. gald, whereas 
 the regular E. form is the vb. to yield. 
 
 * For *wey, the sound being the same. 
 
 8 Originally "to put in order," then "to cleanse, to purify"; cf. the 
 latter meaning preserved in fegefeuer = M.H.G. vege-viur (purgatory). 
 
 * E. egg**G. ei is borrowed from O.N. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 95 
 
 the i and disappears altogether : Go. ligan, O.E. licgan and E. 
 to lie, O.H.G. lichen ligen and G. liegen ; Go. liugan (to tell a 
 lie), O.E. leogan and E. metaph. to lie, O.H.G. lugin (a lie) and 
 G. luge lugen ; L. tegula, whence the early borrowed word, O.E. 
 tigel and E. tile, O.H.G. ziagal and G. ziegel ; further compare 
 to fly to fliegen, a fly to fliege, and add G. hiigel (hill) 1 ; in 
 unaccented finals, E. fifty = O.E. fiftig = G. fiinfzig, E. holy = O.E. 
 hdlig = Q. heilig, etc., etc. 
 
 5. Medial Preg. j = E. w = G. g. — Following o or u : Go. 
 biugan, O.E. bugan and E. tfo bow, O.H.G. biogan and G. biegen, 
 cf. also a tow and bogen; E. fowl=G. vogel ; E. Zow, from the 
 same root as the vb. £o lie = G. liegen; E. frow (hill), from the 
 same root as G. hug-el, etc. — Sometimes even after a 2 : Go. 
 dragan, O.N. draga, O.E. dragan and E. to draw, 3 O.H.G. tragan 
 and G. tragen (to carry) ; O.N. eiginn, O.E. ogrera and E. own, 
 O.H.G. eigan and G. eigen. 
 
 6. Preg. j, when followed by w, is subject to a peculiar 
 treatment. When initial, it was reduced to simple w in the 
 Pregermanic period : 4 Go. varm-jan (to warm), O.E. wearm and 
 E. warm, G. warm. Medial %w survived, and is occasionally 
 still visible in Gothic ; but, according to the nature of the 
 preceding vowel, it became afterwards a single sound, either g, 
 or w, whereupon analogy arbitrarily confounded the one with 
 the other series of forms. 5 Here we need but mention : Preg. 
 *snai%wa-s (" snow," chiefly attested by L. ningu-it, " it snows," 
 and Slav, sneg-u, " snow "), Go. sndiv-s, O.N. snce-r, O.E. sndw 
 and E. snow, O.H.G. sneo and G. schnee; cf. schneien = O.H.G. 
 
 1 The same root as in high and hoch, by Verner's law, infra 54, III. 1-2. 
 But not E. kill, which is akin to L. collis. 
 
 2 This may be due either to a dialectal mixture, or to a peculiar deepen- 
 ing of the a-sound, or else to the analogy of cases in which the g actually 
 followed o or u : thus, O.E. dragan had a regular pf. drog (G. er trug), 
 whence E. drew, and analogically to draw in the present. 
 
 8 The doublet to drag is reborrowed from the Scandinavian. 
 
 4 Hence it is only the Indo-European speech that can betray it, infra 56, 
 IV. B. 
 
 5 The %w is very difficult to distinguish, the more so because under cer- 
 tain conditions, as noted above, simple g may become E. to : Go. maurgin-s 
 (morning), G. morgen (morning, to-morrow), O.E. morgen, E. morn and 
 derivative morn-ing, but O.E. dat. to morive and E. to-morrow ; cf. like- 
 wise to follow = G.folgen, tallow = G. talg, and supra 37 in fine. 
 
96 ENGLISH AND GERMAN GRAMMAR. 
 
 sniwan. Further compare the treatment of the Preg. group hw 
 in : Go. saihvan, G. sehen and E. to see, and Go. sahv, G. (er) 
 sah, E. (he) saw ; O.E. mdwan, E. to mow and G. mahen. 
 
 7. The group gg, -which proceeds from assimilated gj, is 
 changed to E. dz (spelled dg), but hardened to G. kk : O.N. 
 egfjf, from a Preg. *a%jo (cf. L. ac-ies), O.E. ec<7 and E. erf^e, 
 but O.H.G. ekka, G. ec&e ecfc (corner); E. hedge = G. ftecke. 
 E. gr never assumes the sound of dz but in this one position, 
 or else in the many words borrowed from Norman French, 
 classical French or Latin. 
 
 III. The correspondences of the voiceless explosive (k) are 
 quite clear : English keeps the k (written k or c), except when 
 it is changed to ts (written ch) in certain well-defined positions ; 
 German likewise keeps the k (ck) when initial or following a 
 consonant, but shifts it to the voiceless spirant (written ch) 
 after a vowel. 1 The group sk and the gi*oup Jew should be 
 considered apart from the general cases. 
 
 1. Initial Preg. k = K. k (c) = G. k: Go. kald-s, O.N. kald-r, 
 O.E. ceald cald and E. cold, G. kalt ; Go. kaurn, O.N. kom, O.E. 
 E. corn, O.H.G. chorn and G. kom ; O.N. ky-r, O.E. cu and E. 
 cow, O.H.G. chuo kuo and G. kuh ; O.N. konung-r, O.E. cyning 
 with metaphony, syncopated to cyng, and E. king, O.H.G. chim- 
 ing and G. metaph. konig ; Go. kniu, E. knee, G. knie. The 
 latter example shows that initial E. k before an n has become 
 merely a graphic symbol. 
 
 2. Preg. & = E. k = G. k after a consonant : Go. Jjagk-jan, 
 E. to think, O.H.G. denchen and G. denken, cf. the O.E. noun 
 oanc (gratitude) and the E. thank-s, O.H.G. danc and G. dank : 
 Go. vaurk-jan, with the reduced grade of the root which appears 
 normal in G. werk, cf. E. work = O.E. weorc, and E. to work. 
 
 3. Preg. 7c=E. & = G. eft after a vowel: Go. wakan (to 
 sit up), O.E. wacian waeccan, E. fo wake, a-wake, O.H.G. wahhen 
 and G. wachen ; cf. tfo make = machen, book = buch, strike = streich, 
 etc. The group efts has reverted to the sound of ks : G. wachs 
 = E. wax. The law moreover admits a number of apparent 
 exceptions, owing to some Low German words, which had 
 
 1 In O.H.G. also often when initial or after a consonant. As to the actual 
 sound, cf. supra 14, 3 A. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 97 
 
 kept the k unchanged, having been introduced very early into 
 High German. 1 Thus we find side by side ; class. G. backen, 
 but Alamannic regularly bachen = O.H.G. bahhan = ~E. to bake; 
 O.H.G-. acchar and G. acker=K acre; G. dach (thatch), but 
 decken (to cover) = O.H.G. decchan; O.H.G. lecchon and G. 
 lecken (cf. the G. derivate lechzen)=E. to lick, etc 
 
 4. Preg. k (G. either k or ch) =E. ch, owing to a palatal 
 assimilation, when preceding or following a palatal vowel 
 (e, i), even before the e which proceeds from O.E. breaking, 
 and before the e of verbal endings in -an when changed to the 
 duller sound -en. 2 Initial : Go. kiusan, O.E. ceosan and E. to 
 choose, G. kiesen ; Go. kinnu-s (cheek), E. chin, G. kinn (chin) ; 
 O.E. dele and E. chill, s the root being the same as in cold, 
 cf. G. kiihl (cool) ; L. calcem, whence O.E. cealc and E. chalk, 
 G. kalk (lime). Medial: Go. svaleik-s, O.E. swiZc and E. such, 
 O.H.G. sulih and G. solch; E. Wc/i = G. re/cft; O.E. stearc 
 (strong) = G. stark, but E. noun starch, together with the adjec- 
 tive stark ; and the doublets to wake and watch, book and beech, 
 to bake and batch, to seek and to beseech* etc. 
 
 Probably on the analogy of such endings as -age in words 
 borrowed from French (courage), this ti has become voiced (dz) 
 in the suffix -ledge, as in knowledge = M.E. know-leche, borrowed 
 from the Scandian -leiki. 
 
 The k, like the g, is mute in unaccented finals, as in earth-ly = 
 O.E. eorb-lic, cf. the G. suffix -lich. 
 
 5. The group sfc was kept free from any shifting either in 
 English or German, 5 and underwent in both a separate though 
 parallel treatment, whereby it gave place to the simple cacu- 
 minal spirant s (written E. sh, G. sch) : Go. and O.N. skip, E. 
 ship, G. schiff; Go. (af-)skiuban (to remove), O.E. scufan and 
 E. to shove, O.H.G. scioban and G. schieben, cf. shovel = schauf el, 
 
 1 Frankish, though a dialect of High German, keeps it unshifted when 
 doubled (ck = klf=kj, supra 36, II. 2). 
 
 2 It is thus that L. k has become F. ch=§ in cheval = cabdllum, chien = 
 cdnem. — It seems scarcely necessary to observe that E. ch survived even 
 after the vowel from which it proceeded had been regularly dropped. 
 
 3 In king explained above, kin = Go. kun-i, kiss metaph. = O.E. coss = G 
 knsz, the k, in fact, did not stand before an i. 
 
 4 Cf. the identity of the participles sought and besought. 
 
 5 Just like sp and st, supra 48, III. 1, and 49, III. 1. 
 
98 
 
 ENGLISH AND GEEMAN GRAMMAE. 
 
 to shape = schaffen, sheep = schaf; Go. fisk-s, O.N. fisk-r, E. fish, 
 G. fisch ; L. discus, borrowed in O.E. disc and E. dish, in G. 
 tisch (table), cf. rash = rasch, flesh (O.N. flesk) =fleisch, fresh = 
 frisch (O.H.G. frisc). In the rare cases which show the group 
 sk preserved in English (O.E. ascian and E. to ask = O.H.G. 
 eiscon and G. heischen), the word is probably borrowed from 
 Old Norse. 
 
 6. The group kw is retained in English, but it may lose its 
 w in German : Go. qiva- 1 (alive, cf . L. vlvu-s), E. quick, G. queck 
 (in quecksilber and erquicken), but also keck (bold). When 
 preceding the vowels u or o = u, w is dropped even in English : 
 E. come = G. (ge-)kommen = Go. quman-s, and, analogically, also 
 in the other tenses, to come = kommen, instead of Go. qiman, 
 and (he) came = (er) karri = Go. qam. 
 
 TV. Recapitulatory scheme for guttural correspondences : 
 
 explo- /voiced : E.gr =G. g, k; 
 
 Gt.g =E.gr, y. 
 
 sives ( voiceless : „ k (c) = ,, k (ck), ch ; 
 
 „ k = „ k, ch, g. 
 
 spirants I yoice * * » * = " ? 7 ; % 
 
 ( voiceless: „ gh = „ ch, h, g 
 
 » 9 = n y- 
 
 „ cA = „ grfc, fc, ch 
 
 palatal (voiceless) : „ ch = „ k, ch; 
 
 >> >5 
 
 cacuminal (voice- / , ■, 
 less): 2 . . . A" Sh -" S ° ' 
 
 „ sch= „ sh. 
 
 aspirate : „ h = „ h ; 
 
 „ h = „ fe, jf*. 
 
 Section II. 
 
 
 THE FIRST CONSONANTAL SHIFTING. 
 
 (51) Having thus stated the correspondences of the English 
 and German consonants with the Pregermanic system from 
 which they have arisen, we now proceed to trace back the 
 latter to the consonantal system of the Indo-European language, 
 as exhibited to us by the comparison of the classical languages. 
 
 1 Bemember that Go. q is a symbol for kv. 
 
 8 The voiced palatal does not exist in English, apart from the type edge. 
 German has neither the voiceless nor the voiced palatal. German also lacks 
 the voiced cacuminal & (F.J), which occurs in English only as a peculiar 
 modification (palatalization) of the z in such words as glazier, pleasure, 
 occasion. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 99 
 
 Let us, for an instant, confine our attention to Latin : con- 
 sidering, for example, the initial consonants in brother = bruder, 
 father = vater, guest = gast, and comparing them respectively 
 with the initials in L. frdter, pater, hostis, we become immedi- 
 ately aware of some symmetrical contrasts which resemble 
 more or less closely those we have met with in the direct 
 comparison of English and German; in other words, we are 
 irresistibly led to think that we shall find in the passage from 
 Indo-European to Pregermanic, a coherent system of con- 
 sonantal mutations, a system prior indeed, but obviously 
 similar to the shifting which characterizes the passage from 
 Pregermanic to West Germanic, or from West Germanic to 
 English and to German. This Pregermanic series of phenomena 
 was discovered by Grimm, and was called by him erste Lauf- 
 verschiebung. Its principle may be very shortly explained and 
 easily understood. 
 
 In point of fact the First Consonant Shifting differs 
 essentially from the second, in that it extends in the same 
 manner over the whole series cf Indo-European explosives, and 
 is therefore capable of being comprised in a single formula or 
 law. It does not follow that all the changes occurred at the 
 same time, or were due to the same cause, as it were, to a 
 single turn of the wheel. 1 Far from this : many circumstances 
 agree to prove that these changes were made at different dates 
 and at long intervals, occasioned perhaps by various and quite 
 dissimilar causes. But the law for p holds good for t and k, 
 the law for b holds good also for d and g, so that, as far as the 
 mere final result is concerned, it is as if the whole series of 
 Indo-European explosives had changed places under a common 
 and simultaneous impulse. 
 
 With one exception, however; for, if we no longer confine 
 our comparison to initials, but take into account the contrast 
 between L. frdter and pater, in which the medial is the same, 
 and G. bruder and vater, in which the medial differs, we a^ain 
 become aware of a gap in the parallelism, the cause of which is 
 to be sought for in some secondary law crossing, as it were, and 
 
 1 It is almost superfluous to observe how obsolete and unnatural such a 
 view of phonetic phenomena would be. 
 
100 
 
 ENGLISH AND GEEMAN GEAMMAE. 
 
 modifying the normal effects of Grimm's law. This was dis- 
 covered and formulated only fifteen years ago by Dr. K. 
 Verner. 3 
 
 § 1. Grimm' $ and Verner' 's Laivs. 
 
 (52) In order to understand these two laws it is necessary 
 to be familiar with the whole Indo-European consonantal 
 system, consisting of sixteen explosives, which the comparison 
 of Sanskrit, Greek, and Latin has enabled us to restore, 
 namely : the four orders of explosives, labials, dentals, palatals, 
 and velars ; and, in each order, the voiceless and the voiced 
 explosive, either simple or aspirated. 2 
 
 Labials. 
 
 Dentals. 
 
 Voiceless . . 
 Voiceless asp. 
 Voiced . . . 
 Voiced asp. 
 
 Voiceless 
 Voicel. asp. 
 Voiced . . 
 Voiced asp. 
 
 I.-E. 
 
 Sk. 
 
 Gr. 
 
 L. 
 
 I.-E. 
 
 Sk. 
 
 Gr. 
 
 P 
 
 P 
 
 IT 
 
 P 
 
 t 
 
 t 
 
 T 
 
 ph 
 
 ph 
 
 <t> 
 
 f 
 
 th 
 
 th 
 
 e 
 
 b 
 
 b 
 
 P 
 
 b 
 
 d 
 
 d 
 
 8 
 
 bh 
 
 bh 
 
 4> 
 
 /,& 
 
 dh 
 
 dh 
 
 e 
 
 
 Palatals 
 
 
 
 Velars. 
 
 I.-E. 
 
 Sk. 
 
 Gr.| L. 
 
 I.-E. 
 
 Sk. 
 
 Gr. 
 
 k 
 
 Q 
 
 k ' k, c 
 
 3 
 
 k, c 
 
 ir, r, k 
 
 kh 
 
 kh 
 
 x k> c 
 
 qh 
 
 kh, ch 
 
 X 
 
 9 
 
 3 
 
 y 
 
 9 
 
 3 
 
 9> J 
 
 0,8>y 
 
 gh 
 
 h 
 
 X 
 
 h 
 
 gh 
 
 gh, h 
 
 <M,x 
 
 L. 
 
 t 
 
 t 
 
 d 
 
 f,d,b 
 
 L. 
 
 gu, c 
 gu 
 gu, v, g 
 
 Sanskrit alone, it is seen, has preserved the whole system. 
 Generally speaking, the voiced and the voiceless aspirates in 
 each order have been confused : Greek, especially, represents 
 both by voiceless aspirates. Latin is even more corrupted : 
 here, most aspirates have become spirants. No consonant, 
 however, has passed from one order to another, apart from the 
 velars, which in Greek are often represented by labials or 
 dentals. 
 
 This is also just what happened in Pregermanic : the first 
 
 1 See his admirable article, in Kukri's Zeitschrift, XXIII. p. 97. 
 
 2 Cf. supra 14, 1 and 3 A, and, as to the particulars in Greece-Latin cor- 
 respondences, my Gramvi. of Gr. and Lat., no. 57-60. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 101 
 
 three orders in our present scheme strictly correspond to the 
 three orders we have distinguished in primitive Germanic, so 
 that we may write the formulas : 
 
 I.-E. labial = Preg. labial ; 
 „ dental = „ dental ; 
 „ palatal = „ guttural. 
 
 In regard to the velars, it must be first observed that, as 
 early as the I.-E. period, they were liable to a peculiar in- 
 fluence (which, however, did not modify them in every case 
 any more than it followed them throughout every branch of 
 the I.-E. family). They often became accompanied by a labial 
 semi- vowel, so that q, qh, g, and gh were respectively sounded 
 almost like kw, khw, gw, and ghw. This phenomenon is com- 
 monly termed labialization, and consequently we give the 
 term labializing languages to those which have kept some 
 traces of it, viz. Greek, Latin and Germanic, as opposed to 
 Slavonic and Sanskrit. 1 
 
 In Germanic, the non-labialized velar cannot be distinguished 
 from the ordinary guttural, and therefore it is included in the 
 last of the three orders given above. The labialized velar, on 
 the contrary, falls under a double law : in some cases the I.-E. 
 group is simply transferred into Pregermanic, which in conse- 
 quence exhibits the ordinary guttural followed by a w, respec- 
 tively hw, kw, and %w ; or else, similarly to the Greek process, 
 the sound of the parasitic consonant partly absorbs the sound 
 of the original one, and the whole group then results in an 
 ordinary labial consonant, respectively/, p, and b. 2 
 
 Thus it is that in every way the four orders of Indo- 
 European explosives resulted in the three orders of 
 Germanic consonants preserved in English and German. 
 
 (53) The unerring and strictly parallel course of this pro- 
 
 1 These labialized velars are changed to Gr. -k or r, /3 or 5, <p or 6, whereas 
 pure velars remain k, 7, x- The same is the case in Latin with the contrast 
 between qu and c, gu (v) and g. 
 
 2 The latter case is by far the rarer, but this is all that may be stated 
 about it, its determining circumstances remaining as yet obscure. See 
 however Brugmann, Grundriss, I., § 444. 
 
102 ENGLISH AND GERMAN GRAMMAR. 
 
 cess will be illustrated by the following four main formulas, 
 wherein the whole of it is contained : 
 
 A. The voiceless aspirates became completely con- 
 founded with the voiceless explosives, whereupon both 
 were naturally shifted in the same way. Since, moreover, 
 voiceless aspirates are very rare in I.-E., and sometimes doubt- 
 ful, this single mention will suffice for them. We need but 
 compare : I.-E. *woyd-e (he knows), Sk. ved-a, Gr. oI8e = FoiS-e, 
 Go. vdit, and Gr. (er) weisz; with I.-E. *woyt-tha (thou knowst), 
 Sk. vet-tha, Gr. oT(r6a — Fota-6a, Go. vais-t, 1 and G. (du) weisz-t. 
 
 B. The voiceless explosive, when immediately preceded 
 by s, escapes the effect of Grimm's Law, that is to say, it 
 remains in Germanic just what it was in Indo- 
 European : L. spu-ere (to spew), Go. speiv-an, E. to spew, G. 
 spei-en; Sk. spac pac (to look out), L. *spec-io, G. spdh-en (to 
 spy); Sk. sthd (to stand), Gr. tcrTrj/xi = *cri-crrd-/Ai (to place), 
 ora-To-s (set), L. std-re si-st-ere sta-tu-s, E. to stand and he stood, 
 G. steh-en and er stand, etc. ; Gr. o-Tct^-co, Go. steig-an, E. to sty, 
 G. steig-en; Gr. cna-a (shadow, reflect), Go. skei-nan, E. to 
 shine, G. scheinen. The same is the case with t following any 
 other I.-E. voiceless explosive : Sk. napti (daughter, grand- 
 daughter), L. neptis (granddaughter, niece), O.E. nift, O.H.G. 
 nift* \ Gr. 6ktw, L. odd, Go. ahtdu, O.E. eahta and E. eight, 
 O.H.G. ahto and G. acht ; L. rectus, Go. raiht-s, O.E. reoht and 
 E. right, O.H.G. reht and G. recht ; Gr. tt€ix7tto-<;, L. quinctu-s, 
 Go. fimfta, M.E.fift, G.funfte, etc. 
 
 C. Grimm's Law. — In each order of I.-E. explosives, the 
 voiceless explosive becomes shifted to the Preg. voice- 
 less spirant, and the voiced aspirate becomes shifted 
 to the voiced spirant, whereas the voiced non-aspirate 
 remains an explosive, but becomes shifted to the voiceless 
 explosive, thus : 
 
 1 From this example the reader might infer that the primitive group V, 
 assimilated from dt, has become st in Pregermanic as well as in Greek. 
 This however would be a false view : assimilation changed tt to ss ; but, 
 here as in many other cases, analogy restored the group st (cf. the impf. 
 O.H.G. tcissa, M.H.G. wisse and iciste, G. wuszte, on the pattern of konnte 
 mochte), so that the law, inexact as it is, may suffice for practical purposes. 
 
 2 E. niece is borrowed from F., and G. nichte is taken from Low German, 
 where ft is changed to cht. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 103 
 
 I.-E. _p(jjft) = Preg. /; I.-E. bh^Vreg. b ; I.-E. 6 = Preg. p; 
 „ t (th)= „ >; „ dh= „ 5"; „ rf = „ t; 
 „ h(kh)= „ h; „ gh= „ 5; „ g= „ fc. 1 
 
 D. Verner's Law. — The above shifting having taken place 
 in a medial syllable, and further supposing the immediately 
 preceding syllable not to be marked with the I.-E. accent, 2 
 then the voiceless spirant which followed the unac- 
 cented syllable (/, J>, h) is softened to a Pregermanic 
 voiced spirant (b, 6", g), and therefore, in the whole course 
 of its further evolution, it is no longer distinguishable from 
 such voiced spirants as have regularly sprung from I.-E. voiced 
 aspirates. 3 
 
 This maybe illustrated from the example already mentioned : 
 I.-E. *bhrdter- (L. f rater) quite regularly became Preg. *tyro}>er-, 
 whereupon no further change could take place, since the first 
 syllable of the word was accented ; but I.-E. *pater- (L. pater), 
 after it had become Preg. *faj>er-, according to Grimm's law, 
 could not help becoming Preg. *f after-, through further soften- 
 ing, according to Yerner's law, because the preceding syllable 
 was unaccented in Pregermanic ; and thus it is that the word 
 assumes in the Germanic speeches (Gothic, English, German, 
 etc.) the same aspect it would exhibit if it had arisen from a 
 non-existing I.-E. form *padher-. 
 
 It follows from Verner's law, that each I.-E. voiceless explo- 
 sive — but, of course, these only — has in Germanic two repre- 
 sentatives, according as it was or was not immediately preceded' 
 by the I.-E. and Preg. accent. 
 
 1 It must, of course, be understood, that, if in the I.-E. period a primitive 
 voiced explosive had already become voiceless by being adapted to a follow- 
 ing voiceless consonant, then its Pregermanic treatment would not differ at 
 all from that of a primitive voiceless explosive : thus, the I.-E. groups gt ght 
 and bt bht become respectively Preg. ht and/t, and cannot become anything 
 else, since in I.-E. they were already changed respectively to ht and^f. Gr. 
 f£py-o-t> = 'E. work; but the E. pf. of the vb. to work is wroupht = O^E. 
 it>or/i-<e = I.-E. *we-wrk-tay, wherein the radical g is assimilated to the 
 following t. See some other Germanic instances, supra 24. 
 
 2 On this primitive accent, which is altogether lost in later languages, 
 but survived in Pregermanic, cf. infra 65, I. 
 
 3 The groups sp, st, sk, ft, ht, and subsidiarily ss, fs, ht, which are not 
 subject to Grimm's law, also escape the effects of Verner's law. 
 
104 ENGLISH AND GERMAN GRAMMAR. 
 
 A further consequence is, that, if the movable accent hap- 
 pened to he displaced in a given word, in the ordinary course 
 of accidence, ■whether in derivation, declension, or conjugation, 
 the medial consonant of this word changed accordingly and 
 gave rise to such alternations as are illustrated by the three 
 German types : hof (court) = Preg. *hufa-z, but hiibsch (" beauti- 
 ful, select," now, as it were, hbfisch " courtlike ") = O.H.G. 
 hubesch = Preg. *huhiskd-s ; ich siede (I seethe) = Preg. *seufi-o, 
 but gesotten (sodden) = Preg. *su5an&-s; ich ziehe (I draw) = 
 Preg. *teuh-o, but gezogen (drawn) = Preg. *tuyind-s, 1 etc. 
 
 It remains to point out the application of Grimm's and 
 Verner's laws in detail throughout the Germanic consonantal 
 system. 
 
 § 2. Primitive Voiceless Explosives. 
 
 (54) I. Labial. — 1, Initial, or following an accented syllable 
 (Grimm's law) : L. pisci-s, Go. fisk-s, etc. 2 ; Sk. pita, Gr. Tra-n^o, 
 L. pater, Go. fadar; Sk. pri (to love), priyd-s (dear), Go. 
 frij-ond-s, E. friend, G. freund ; Sk. ply (to blaspheme, to hate), 
 Go. fij-and-s (enemy), E. fiend, G. feind; Sk. ndpdt- (son, 
 grandson), Gr. vc7ro8«s (offsprings), 3 L. nepos (grandson, 
 nephew), O.E. nefa and E. nephew* O.H.G. nefo and G. neffe ; 
 Gr. kwtt-7} (handle), L. cap-ere (to seize), G. haf-t msc. (a tie), 
 haf-t fm. (capture), hef-t (handle), O.E. hcef-t (prisoner) = L. 
 cap-tu-s, E. haft (handle), etc. 
 
 2. Medial and following an unaccented syllable (Grimm's law, 
 and then Verner's law, p shifted to/, and/ softened to b) : the 
 same root kap in Go. haf-jan (to lift), O.N. hef-ia, O.H.G. 
 heffen, but E. to heave and G. heben, where the b extended 
 
 1 This is what Germanists call grammatische wechsel. Compare the E. 
 contrast in seethe and sodden. Of course the alternation was afterwards 
 liable to disappear under the influence of analogy : cf. infra 55, 179, 184, etc. 
 
 2 As for the English and German evolution of words already mentioned, 
 the reader must trust to his memory, or refer to the preceding section. The 
 English and German words have been repeated in the Indexes even when 
 they are not quoted as such under the following numbers (54-57). 
 
 3 Always observe the Sk. and Gr. accent. 
 
 4 Pr. nehyu (medial / softened to v), but the ending is borrowed from F. 
 neveu. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 105 
 
 throughout the conjugation, after the analogy of such cases as 
 would require it regularly, e.g., Preg. *hab-and-s (taken) * ; Gr. 
 airb, L. ab = *ap, Preg. *aba (but Go. af with the final un- 
 voiced), E. of (pr. ov, but off when actually final), G. ab ; Sk. 
 updri (over), Gr. iirep, hence a Preg. *uberi, cf. E. over and 
 a-b-ove, G. iiber and oben ; L. aper = *ap-ro-s, G. efeer (boar), 2 etc. 
 
 II. Dental. — 1. Initial, or following an accented syllable : 
 Sk.tf-na-m (a blade of grass), Go. )>aur-nu-s (thorn), etc.; Sk. 
 root tan (to extend), Gr. rav-vw, L. ten-ui-s (expanded, thin), 
 O.N. fiunn-r (thin), etc. ; Sk. tum-rd- (thick, strong), zd. tum-a- 
 (strong), L. tum-ere (to swell), O.N. }>um-al-(fingr) "the thick 
 finger, thumb," etc. ; Sk. bhrdtar- (brother) and Preg. *bro/>er-, 
 etc.; I.-E. root pet (to fly), Gr. 7rcT-o-/ta.i, Sk. pdtram (wing) = 
 *pdt-tra-m, Preg. *fej>ram, etc. 
 
 2. Medial, when following an unaccented syllable (t shifted 
 to p, and then^» softened to 5) : Sk. pitdr- (father), and Preg. 
 *faj>er-, afterwards *fa&er-, etc. ; Sk. mdtdr- (mother), and 
 Preg. *mo}>er-, afterwards *mob"er-, etc. ; Sk. damitd-s (tamed), 3 
 Preg. *tamij>ds, whence *tamid"ds, Go. (ga-)tamida, E. tamed, 
 O.H.G. (gi-)zemit and G. gezahmt, and so also with every 
 ending of the past participle in so-called weak verbs. 
 
 III. Palatal. — 1. Initial, or following an accented syllable : 
 Gr. Kax^-qi (little stone) and Preg. *hd%l-a-z (hail), etc.; Lith. 
 kauk-a-s (knob) and Preg. *hduh-a-z (high), etc. ; Gr. ndpr-a 
 (vehemently), xapr-cpo-s and Kpar-v-s (strong), Preg. *harft-u-s 
 (hard), etc. 4 ; Gr. Kop-a£, L. cor-vu-s (raven), Preg. *hra-hnd-s, 
 etc. ; Sk. cvetd-s (white) and Preg. *hwitd-s, etc. ; Sk. pd$u 
 nt. (cattle), L. pecu, Preg. *fehu nt., etc. ; Lith. rauka-s (rough) 
 and Preg. *rduha-z, etc. 
 
 2. Medial, when following an unaccented syllable (k shifted 
 to h, and then h softened to g) : the same root I.-E. kewk (high) 
 in Preg. *hu%ild-s (hill), etc. ; I.-E. *dekm (ten), Gr. Scxa, L. 
 decern, Go. taihun ; but, the accent being moved, as in Gr. 
 Se/cas (tithing), Preg. *te%un-, whence Go. *tigu- in fimf tigjus 
 
 1 Go. haf-an-s is due to the same though inverted analogy. 
 
 2 E. York = O.B. Eofor-wlc " borough of the boar." 
 
 8 .Remember the constant accent of Greek verbal nouns in -t6-j. 
 4 Hence we see that E. to call has nothing to do with Gr. /caXeiv. 
 
106 ENGLISH AND GERMAN GRAMMAR. 
 
 (five tithings, fifty), etc. ; Sk. aksi- (eye) and Lith. aM-s (cf. 
 SI. oJc-o, Gr. overt "both eyes " = 6/c-t/c, L. oc-ulu-s), Preg. *auhi-, 
 whence *au%i-, etc. ; I.-E. root Ik (to be master), Sk. ic-and-s 
 (sovereign, master), Preg. *aihand-s, whence *aiyind-s (own), 
 etc. 
 
 IV. Velar. — A. Non-labialized (confounded with the pala- 
 tal). — 1. Initial, or following an accented syllable: I.-E. root 
 qel (to act, to deal with, to manage), Sk. car (to move, to act), 
 Grr. 7T€A-o-)u,ai (to exist), likewise labialized in cu-7ro'Ao-s (goat- 
 herd), but without labialization in (3ov-k6\o-<s (cow-herd), again 
 labialized in L. (in-)quil-inu-s (inhabitant, planter) and col-6 = 
 *quel-6 (to cultivate), but without labialization in Go. hal-dan 
 (to keep), E. to hold, G. halten; I.-E. *noq-t-s (night), Sk. 
 ndk-ti-s, Gr. vv£ vvk-t-cl, L. nox noc-t-em, Go. nah-t-s, etc. — 2. 
 Medial, when following an unaccented syllable : L. vinc-o = 
 *vinqu-o (to vanquish) and Go.veih-an (to fight) = I.-E. *weyq-6, 
 but O.H.G. wlgant (warrior) and G. weigand, G. weigern (to 
 resist), O.E. wlgend (warrior) lost in Mod. English. 
 
 B. Labialized, Preg. either hw or yjo. — 1. Initial, or following 
 an accented syllable : Sk. cud (to whet), from an I.-E. root qed 
 qod = kwed kwod, Preg. *hwat-ja-, E. to whet, G. wetzen; I.-E. 
 pronoun *qo-, Gr. 71-0-, L. quo-, Preg. *hwd-, etc. ; I.-E. root seq 
 (to follow), Gr. €7r-o-/i,at, L. sequ-o-r, Preg. *sehw-o, whence Go. 
 saihv-an (to see), etc. — 2. Medial, when following an unaccented 
 syllable: perhaps I.-E. root merq mrq (dull), SI. mrak-u (dark), 
 mruk-na-ti (it is dark), whence a Preg. derivate *murpjoena-, 
 O.N", morgunn (morning), etc., though Go. maurgin-s shows no 
 apparent labialization. 
 
 C. Labialized, Preg. either / or b. — 1. Initial, or following 
 an accented syllable : I.-E. *penqe (five), Sk. pdnca, Gr. ttcvtc 
 7re/A7re, L. qulnque, Preg. *fernfe, etc. ; I.-E. wlqo-s (wolf), Sk. 
 vfka-s, Gr. Au'ko-s, L. lupus, Preg. *wulfa-z, etc. — 2. Medial, 
 when following an unaccented syllable: I.-E. *wlqi (she-wolf), 
 Sk. vrki, Preg. *wulfi, whence *wulfoi, preserved exclusively in 
 O.H.G. wulpa and M.H.G. wiilpe. 1 
 
 1 We should expect *wiilbe ; but the b has been doubled (supra 36, II. 2), 
 and the group bb becomes G. pp, just as gg becomes ck (supra 50, II. 7). 
 Mod. G. woljin has analogically restored the medial consonant of wolf. 
 
EXPLOSIVE CONSONANTS AND THEIK SUBSTITUTES. 107 
 
 (55) V. We have already stated the grammatical alter- 
 nations which proceed from the application of Verner's law : 
 they are easily seen in the earlier Germanic languages, and, 
 though analogy has levelled them in the later languages, they 
 again become obvious through the comparison of one tongue 
 with another. Thus, the contrast of enough with G. genug may 
 be due to an alternation of this kind : a G. *genuch, the only 
 form which would regularly answer E. enough, is to be restored 
 from Go. ganoh-s (sufficient), ganauh-a (sufficiency), in O.H.G. 
 ginah (it was sufficient), ginuht (sufficiency), in M.H.G. genuht- 
 sam (sufficient), etc., so that we may suppose the form with h 
 and the form with 3 to have been normally interchangeable 
 according to the accent, whereas analogy preserved the latter 
 in German and the former in English ; hence the divergence 
 in the two related languages. 
 
 It will be seen, in fact, that the I.-E. accent often changed 
 its place, especially in conjugation. For instance, an I.-E. vb. 
 *dewk-o (I lead, I draw, L. duc-o), in the present indicative, 
 had the root in the normal grade, with accented radical syllable, 
 thus : Preg. *teuh-6, Go. tiuh-a, G. ich ziehe. In the pf . sg. the 
 root was deflected, but it kept the accent : I.-E. *de-dowJc-a, Go. 
 tduh, O.H.G. zoh, M.H.G. ich zoch (I drew). But in the pf. pi. 
 the accent was transferred to the ending and the root was 
 reduced : I.-E. *de-duJc-mme, whence a Preg. *tu%-ume and G. 
 wir zogen ; and so also in the past participle, Preg. *tu%ana-s 
 (drawn) and G. gezogen. The analogical change in Mod. G. 
 is quite plain : owing to wir zogen and gezogen, the sg. has 
 become ich zog, instead of *zoch. German, however, has pre- 
 served the curious alternation seen in ziehen and gezogen, which 
 cannot be understood unless it is traced back to Pregermanic. 
 Gothic, though older, is much more corrupt : here the h ex- 
 tended throughout. On the pattern of such regular forms as 
 tiuhan and tduh, Gothic irregularly has tauhum (we drew), 
 instead of *tugum, and tauhan-s 1 (drawn), instead of *tugan-s, 
 which forms, if reproduced in a Mod. German conjugation, 
 would give us the paradigm ziehen, ich *zoch, wir *zohen, 
 
 1 Naturally, with au, instead of w, before h, supra 28, I. 
 
108 ENGLISH AND GERMAN GRAMMAR. 
 
 *gezohen. Observe, moreover, the contrast of G. zucht with G. 
 zug, all these word's being derived from the very same root. 
 
 Such analogical processes could not but take place sooner or 
 later, and in time give the regular variability of the Germanic 
 consonantal' system a fictitious and seemingly satisfactory unity. 
 No Germanic tongue is free from them, Gothic least of all, 
 old as it is. We have just seen Mod. German to be purer 
 even than Gothic, and the same is the case with Old English, 
 as the conjugation of this very verb shows, teon (to draw), 
 togen (drawn), etc. Gothic also shows the medial consonant 
 unchanged throughout the whole conjugation of the vb. hafja 
 (I take)=L. capio, and pf. hofum (we took), whereas O.H.G. 
 keeps the alternation in heffu and huobum, which latter gave 
 rise in M.H.G. to the vb. heben (to heave, to lift). Sometimes, 
 however, Gothic has preserved an alternation which has been 
 lost everywhere else : thus, the Go. conjugation of the pf. t>arf 
 (I need) is in the pi. J>aurbum = Preg. *fiurh-ume, 1 a remarkably 
 pure survival, whilst O.E. Surfan, pf. sg. 1 ftearf, pi. 1 Surfon, 
 and G. (metaph.) durfen, z pf. sg. 1 darf, pi. 1 diirfen, both 
 placed the voiceless consonant where only the voiced would 
 be regular. 
 
 § 3. Primitive voiced aspirates. 3 
 (56) I. Labial. — I.-E. root bheyd, Sk. bhinad-mi (I split) 
 and bhid-yd-te (it is split), L. find-o, Preg. *fclt-d (I bite), Go. 
 beit-a, etc. ; Gr. (prjyos, L. fdgus, G. buche, etc. ; I.-E. root bhld 
 (to blow), L. fid-re, E. to blow, G. blahen ; I.-E. root bhendh 
 (to bind), Go. bind-an, etc. ; I.-E. root lewbh (to delight in), 
 Sk. lubh-ya-ti, L. lub-et lib-et, Preg. *leuh-a-z (dear), whence 
 Go. liuf-s, etc.; L. scab-o (I scrape), E. to shave, G. schaben; 
 Sk. garbhd-s (an embryo, a young one), and Preg. *kalh-d-s 
 (calf), etc. 
 
 II. Dental. — I.-E. *dhughater- (daughter), Gr. Ovydrqp, Sk. 
 duhitdr-, Go. dauhtar = Preg. *d'uhter-, etc. ; I.-E. root dhe (to 
 
 1 I.-E. "te-tdrp-a and *te-trp-mme respectively. 
 
 2 For the meaning, compare bedvrfen. 
 
 3 From this point to the end of the chapter, we have of course no longer 
 to deal with Verner's law. 
 
EXPLOSIVE CONSONANTS AND THEIR SUBSTITUTES. 109 
 
 set, to do), Sk. dhd, Gr. 6rj (ri6r]p.i), E. do, G. thun; I.-E. root 
 bhendh and Preg. *bind, above ; I.-E. *medh-yo-s (placed in the 
 middle), Sk. mddh-ya-s, Gr. fieVos = [itcrcros = *fiW-yo-<s, L. med-iu-s, 
 Go. mid-ji-s, etc. ; I.-E. root dhegh (to burn, to shine), Sk. 
 ddh-a-ti (it burns), whence a nominal derivation *dhogh-6-s, 
 Preg. *6"ajd-s, Go. dagr-s (day), etc. 
 
 III. Palatal. — I.-E. *ghosti-s (foreigner, enemy, guest), L. 
 hosti-s, Preg. *%dsti-z, Go. gast-s, etc. ; I.-E. ghans- (swan, goose), 
 Sk. hans-d-s (swan), Gr. xv v (g oose )> L. anser = *hans-er, Preg. 
 *gans-, whence G. gfaws, O.E. *gons gos, and E. goose ; I.-E. root 
 dhigh (to mould), Sk. di/i. (to besmear), Gr. 0iy-eTv (to touch), 
 tci^-os (wall), L. fing-ere, Preg. root dig, whence Go. deig-an 
 (to mould), E. dough and G. £et<7, etc. ; I.-E. *ghyes- (yesterday), 
 Sk. hyds, Gr. x$ £ 's> k. ^er-i hes-ternu-s, Go. gis-tra-(dag-is), etc. 
 
 IV. Velar. — A. Non-labialized, confounded with the palatal : 
 I.-E. *dhogho-s (day), as above II. ; L. helvu-s (yellow), E. yellow, 
 G. gelb. 
 
 B. Labialized, Preg. %w : I.-E. root snigh sneygh (to be 
 moist), Sk. snih, Gr. ace. vi<£-a (snow), vci</>-« (it snows), L. ace. 
 nivem = *nihv-em, Go. sndiv-s, etc.; Sk. gharmd-s (warm), Gr. 
 6ep/x6-s, L. formu-s, Preg. *%warmd-s, afterwards *warmd-s, etc. 
 
 C. Labialized, Preg. b : no certain examples. 
 
 § 4. Primitive voiced explosives. 
 
 (57) I. Labial. — No certain examples, apart from I.-E. 
 root dhub dhewb, Lith. dub-u-s (deep), perhaps Preg. *&eup-a-z, 
 Go. diup-s, E. deep, G. tie/. 1 
 
 II. Dental. — Sk. jihvd (tongue )= *dihvd, L. lingua = dingua 
 = I.-E. *dnghwd, whence a Preg. *tufi.%d-, etc.; I.E. *dekm 
 (ten) and Preg. *tehun; I.-E. *dem-o (I build), Sk. ddm-a-s 
 
 1 I.-E. 6 is an extremely rare sound, and the very few Greek or Latin words 
 which might bear testimony of it happen not to have any corresponding 
 forms in Germanic, and reciprocally we do not find elsewhere but in Germanic 
 such Preg. roots as slap (E. to sleep, G. schlafen, cf. however SI. slab-u = G. 
 schlaff, " lazy"), skup (E. to shape, G. schaffen), etc., which would presuppose 
 an I.-E. b. The latter, perhaps, might be identified with L. scab-o (to scrape) ; 
 but, according to the resemblance in meaning to E. shave and G. schaben, 
 the primitive form skabh seems more probable for the L. vb. In Preg. werp 
 (to throw) the p has arisen from a velar, infra IV., C. 
 
110 ENGLISH AND GERMAN GRAMMAR. 
 
 (house), Gr. Bifx-ia S0//.-0-S, L. dom-u-s, Go. (ga-)tiwi-r-jan (to 
 build), cf . E. timber and G. zimmer ; I.-E. root bheyd, as above, 
 56,1.; I.-E. *swddu-s (sweet), Sk. svddu-s, Gr. rj8v<: = *aFa.8v-s, 
 L. sudvis = *sudd-ui-s (cf. the vb. sudd-ere), 1 E. sweet, G. swsz; Sk. 
 svid (to sweat), Gr. ioYw = *o-FiS-ia), L. sudor (sweat) =*svoid-os, 
 E. sweat and io sweat, G. schwitzen and schweisz, etc. 
 
 III. Palatal. — Sk. j'awtt (knee), Gr. yoVu, L. #en«, Go. &m'«, 
 etc. ; L. grdnum = *gr-no-m, Go. kaurn, etc. ; Gr. <£wy-a) (to roast), 
 E. to 6afce, G. back-en bach-en; Gr. dyoo-s, L. agrer, Go. akr-s, etc.; 
 &k.jus (to choose, to accept), Gr. yeuo/Aai = *yei;V-o-/ J t,cu, (to taste), 
 L. gus-tu-s (taste), Go. kius-an (to prove, to choose), etc. 
 
 IV. Velar. — A. Non-labialized, confounded with the palatal: 
 I.-E. root gel (cold, frost), L. gel-u, Go. kal-d-s (cold) ; I.-E. 
 stheg teg (to cover), Sk. sthag-a-ti (he covers), Gr. crrcy-co (I 
 cover), o-Tcy-rj (shelter), T€y-os (thatch), L. teg-o tog-a, Lith. 
 stog-a-s (thatch), O.E. 5cec and E. thatch, G. dach and decken, etc. 
 
 B. Labialized, Preg. few : I.-E. *gdw-s (cow), Sk. gdu-s, Gr. 
 (3ov-<s, L. feo-s, E. cow and G. A;«7i ; I.-E. root gem (to go), Sk. gam 
 (to go), Gr. fiaLvu), L. ven-io, Go. qim-an, etc. 
 
 C. Labialized, Preg. p: I.-E. root wer£ (to throw), cf. SI. 
 vrug (id.), whence a Preg. root werp, Go.vairp-an, O.E. weorp-an 
 and E. to warp, O.H.G. werf-an and G. werf-en. 
 
 1 "To make [something] sweet [to somebody]," is as much as "to per- 
 suade him of it." 
 
CHAPTER V. 
 
 SIBILANT CONSONANTS. 
 
 (58) The Indo-European speech had two sibilants, voiceless 
 s, voiced z ; but, the voiced sibilant having only arisen from the 
 process of adapting the voiceless sibilant to a following voiced 
 consonant, — as, for instance, in a group sd changed to I.-E. zd, — 
 the two were originally but one, and it will be better to keep 
 them together. 
 
 The I.-E. sibilant may be either initial, or medial, or 
 final. In all these positions it has been remarkably well pre- 
 served in Pregermanic, and the slight modifications it has 
 undergone are to be assigned to the later languages, so that 
 its evolution may be traced at once from the beginning down 
 to the present time. 
 
 Section I. 
 
 INITIAL SIBILANT. 1 
 
 (59) I. Before a vowel. — All the I.-E. languages, including 
 Germanic, have faithfully preserved the primitive initial s, 
 with the single exception of Greek, 2 where it is changed to h 
 (rough breathing) : I.-E. root su (to bring forth), whence Sk. 
 su-nu-s (son), Gr. vios = *<tv-l6-<s, Lith. sunu-s, SI. synu, Go. sunu-s, 
 O.N. sun-r, O.E. sunu and E. son, O.H.G. sunu and G. sohn; 3 
 L. sat sat-is (enough) sat-ur (fed), Go. safi-s (fed), OJST. saft-r, 
 
 1 The z seldom occurs as initial, and is never final. 
 
 2 More exactly, the same remark applies to Avestic (Zend), and in part 
 to Celtic. 
 
 3 We need scarcely remark that this s, as well as the medial intervocalic s, 
 has become voiced in the correct pronunciation of classical German (sohn, 
 pr. zon). 
 
 ill 
 
112 ENGLISH AND GEEMAN GRAMMAE. 
 
 O.E. seed (fed) and E. sad ("made to loath "), O.H.G. sat and 
 G. satt; I.-E. root se (to throw, to sow), Gr. rj-fia (a throw), 
 L. se-men (seed), 1 Go. sai-an (to sow), O.E. sdw-an and E. to sow, 
 O.H.G. sd-en and G. sden, further E. seed — G. saat; Gr. dA-s, 
 L. sdl = *sald, E. salt, G.salz; Sk. sdd-as (seat), Gr. I8-os, L. 
 sed-es and sed-ere (to sit), E. to sit, G. sitz-en; Gr. £0o9 = *e#-o9, 
 Go. sid-u-s, O.E. sidtt, lost in E., O.H.G. situ and G. si^e 
 (custom, behaviour) ; SI. sirebro (silver), E. silver, G. silber, etc. 
 
 II. Before a semi-vowel. — The initial group sy is rare and of 
 no importance, since the y is dropped : E. to sew and G. sdu-le 
 (awl), cf. Go. siu-jan = Gr. (Ka-)cro-v-a) = L. su-o. The initial 
 group sw, on the contrary, is extremely common and generally 
 survives, though High German now changes the s to s, as it 
 does before any other consonant. Thus, E. swine = G. schwein, 
 E. swim = G. schwirnmen, E. sword = G. schwert, E. swea£ = G. 
 schweisz, etc., have been already mentioned. Here we may add : 
 L. sudsum (a dark colour), Go. svart-s (black), O.N. svart-r, 
 O.E. sweart and E. swart, O.H.G. swarz and G. schwarz ; Gr. 
 o-ty?7 (silence) = *<rfZy-a., O.E. swig-ian and G. schweig-en (to be 
 silent); Sk. svdsar- (sister), L. soror — *svesor, Go. svistar (O.N. 
 syster), O.E. sweostor, O.H.G. sw'ester and G. schwester, etc. The 
 exceptions to this law are only apparent : E. sister is borrowed 
 from the Scandinavian ; the two words E. swamp and G. sumpf 
 probably contain two different grades of the same root ; as for 
 G. stts2 = E. sweet, we need but go back to the O.H.G. form 
 suoji swuo?fii, where the w and the u are seen to have combined, 
 owing to an absolute identity in sound. 
 
 III. Before a nasal or a liquid^ Greek and Latin usually drop 
 the s, but Germanic keeps it intact. 
 
 1. Before m, n, I (E. s, G. s changed to s) : Sk. root smi (to 
 laugh, to wonder), L. ml-ru-s (wonderful), ml-ro-r (I admire), 
 E. smi-le, cf. G. schmei-cheln (to caress); E. small = G. schmal 
 (thin) ; Sk. root snih, Gr. vi<j>-a, L. niv-em, E. snow, G. schnee ; 
 L. lub-ricu-s, E. to slip, G. schliipf-en ; L. lac-er (rent, put to 
 pieces), E. to slay, G. schlag-en, etc. 
 
 2. Before r, Germanic and Slavonic insert a t : I.-E. root 
 sru srew (to flow), Sk. srdv-a-ti (it flows), Gr. 
 
 1 Cf. Gr. trifju = *<ri-<rr)-iu, and L. seio = si-$o. 
 
SIBILANT CONSONANTS. 113 
 
 from this root, a derived noun with deflected root would be 
 *srow-mo-s (current), which has become Preg. *strau-md-s, O.E. 
 stream and E. stream, O.H.G. stroum and G. strom. 1 Cf. Russian 
 o-strov-u (island), similar in meaning to a G. word um-ge-strom-t 
 (surrounded by a current). 2 
 
 IV. Before an explosive, we have seen that the s is remaining, 
 and the explosive is not affected by Grimm's law. 3 The group 
 sk becomes s; for instances, see above (50, III. 5), and add: 
 G. schreiben (to write), O.E. scrlfan, borrowed from L. scrlb-ere; 
 E. shrine = G. schreinf etc. In the case of sp and st, Mod. 
 German is well known to admit of both pronunciations, sp st, 
 and sp st. 
 
 The greater part of such roots as began with the group 
 in question exhibited already in the I.-E. period a peculiar 
 alternation : in some conditions, which are not yet precisely 
 defined, 5 they were liable to lose their initial s, a process fully 
 illustrated by the Sk. doublet spaq, and pag, (to see). 6 The 
 same alternations are to be observed between the Germanic 
 and the related languages : thus we have already mentioned 
 Gr. o-Tcyo) and G. dach decken ; reciprocally, we have Gr. /coco 
 (to remark) = *Kof-€w, and L. cav-ere (to take care), without 
 initial s, whereas the s reappears in the whole Germanic branch, 
 especially in O.E. sceaw-ian (to consider) and E. to show, O.H.G. 
 scouw-on (to spy) and G. schau-en (to look at). 
 
 Section II. 
 
 MEDIAL SIBILANTS. 
 
 (6o) Verner's Law governs the voiceless dental Preg. 
 spirant (s), since it is a general law for all Germanic voiceless 
 
 1 We should expect *straum, but cf. supra 32. 
 
 2 This insertion is constant. In stride and schreiten, either we have two 
 different roots, or, in any case, G. has kept unaltered an initial group skr, 
 whilst E. has accidentally changed the k to a t, the more so since O.E. has 
 a vb. scridan (to stride). 
 
 8 Cf. supra 53 B. 
 
 * Both borrowed from L. scrlnium (a box). 
 
 5 Cf. my Gramm. of Gr. and Lat., 68, 4. 
 
 6 Then, of course, in Germanic, the explosive being thus uncovered again 
 falls under Grimm's law. 
 
114 ENGLISH AND GERMAN GRAMMAR. 
 
 spirants. The voiced spirant (z), on the other hand, falls 
 indirectly under the application of Grimm's Law. 
 
 § 1. Voiceless Sibilant. 
 
 (61) I. Between vowels. — General formulas of Verner's law. 
 
 A. Medial s, if preceded by the I.-E. and Preg. accent, is 
 kept unchanged throughout the whole Germanic branch, 
 with this difference only that Mod. E. and G. now sound it 
 mostly as z, though the spelling s is usually retained (always 
 spelled s in German, occasionally z in English). 1 
 
 B. But medial .9, when following an unaccented syl- 
 lable, is changed to a voiced spirant, Preg. z, which is kept 
 in Gothic ; West Germanic again changes this z to the 
 liquid r, thus displaying a process of rhotacism which is 
 similar to the Latin mutation in arbos (tree), ace. arbor-em, 
 and gen-us (birth), gen. gen-er-is = *gen-es-is — S)s.. jdn-as-as = Gr. 
 *yev-€cr-os, then ye'veos and yevovs. 2 
 
 This new and constant application of Verner's law cannot 
 be better illustrated than by certain alternations between s 
 and z, which are distinctly preserved in some archaic conjuga- 
 tions, though more or less obliterated by analogy in the later 
 languages. 3 Thus, for instance, the Preg. vb. *keus-6 (Go 
 kius-a, Gr. ycv((r)-w) shows the s in the O.E. infinitive ceos-an 
 (E. to choose), where the accent is known to have rested on the 
 first syllable or normal root, while its participle is O.E. cor-en 
 (chosen) = Preg. *kusand-s, regularly changed to *kuzand-s, 
 because here the accent is thrown on to the suffix; and, in 
 the same way, archaic and literary German has an infinitive 
 er-kies-en, the past participle being er-kor-en ; but Mod. English 
 created chosen after the analogy of the infinitive choose. The 
 E. vb. to lose = Go. lius-an had likewise an O.E. participle lor-en, 
 still preserved in Mod. E. for-lor-n, an old word which became 
 
 1 Cf. G. glan, pi. glas-er, but E. glass, derivative glaz-ier, wherein the 
 group z+y takes the sound of 2. 
 
 2 The position of the vocal apparatus is the same for lingual r as it is 
 for z. If we arrange it to utter a 2, a very slight quivering of the tongue will 
 give us a trilled r. Cf. the F. doublet chaise (chair) and chaire (cathedra), 
 though here, contrariwise, the r has been changed to 8 (z). 
 
 8 Cf. supra 55. The parallelism is absolute. 
 
SIBILANT CONSONANTS. 115 
 
 severed from its original stock, and therefore escaped the 
 analogical influence of the other conjugated forms, whereas a 
 new participle los-t was created on the pattern of lose. Now, 
 in German, we have exactly the reverse of the English pro- 
 cess : ver-lor-en was kept, and the infinitive, which should be 
 *ver-Ues-en = O.H.G. vir-lios-an, adopted the participial r and 
 became ver-lier-en, so that the regular s is only seen in the 
 derivative ver-lus-t (loss). The same is the case with E. frost = 
 G. frost, contrasted with E. to freeze = G. frier en, etc. 
 
 1. After an accented syllable : I.-E. s = Preg. s = E. and G. s. — 
 Sk. mils- (mouse), Gr. fivi, L. mils and ace. mur-em, 1 O.N. mus, 
 O.E. mus and E. mouse, O.H.G. mus and G. maus ; Sk. ndsd 
 (nose), L. ndsus ndres, O.E. nasu nosu and E. nose, O.H.G. nasa 
 and G. nase; Preg. *les-6 (I gather), Go. lis-an, O.N. les-a, 
 O.E. les-an and E. to lease, O.H.G. les-an and G. les-en, etc. 
 The I.-E. word "*haso-s (hare) probably varied in accentuation, 
 inasmuch as O.H.G. haso = Q. hase supposes a derivation from 
 a Preg. *hdsa-z, whereas O.N. here, O.E. hara and E. hare do 
 not admit of any other primitive but a Preg. *hazd-s, which 
 is moreover confirmed by the Sk. ca$d-s. 
 
 2. After an unaccented syllable: I.-E. s = Preg. z = Go. z = E. 
 and G. r. — E. hare has just been mentioned. To Sk. mdhlyas- 
 (greater), L. mdjus major mdjor-em, must correspond a Preg. 
 *md^is-, changed in consequence to *md%iz-, whence : Go. mdiza 2 
 (more), O.E. mdra and E. more, O.H.G. mero and G. mehr; and 
 so also with every comparative, Go. bat-iza, E. better, G. besser. 
 The I.-E. root wes (to dwell, Sk. vas) has a pres. sg. 1 Preg. 
 *wes-o, which is regularly reproduced, with the s unaltered, 
 in Go. vis-an (to remain, to be), O.E. wes-an, G. wes-en (an 
 infinitive used as a substantive) ; but, in the pi. of the pf., the 
 accent is known to have passed from the root to the ending, 
 whence a Preg. *wes-mme and *wez-mme, E. we were, G. wir 
 waren. 3 The I.-E. word *6ws- (ear) shifted the accent in 
 
 1 The I.-E. 8 is medial everywhere but in the nominative sg. 
 
 2 Putting out of question the puzzling loss of the medial z. The com- 
 parison, of course, is confined to the sibilant. _^____ 
 
 8 Cf. the E. sg. I was. In German the r has intru&jd£tfi|oThe sg.;,- ich 
 tear. In Gothic, on the contrary, which theoretically raDuTa have sg.'iws, 
 
 (nan -V :n y 
 
 V _ Or y 
 
116 ENGLISH AND GERMAN GRAMMAR. 
 
 declension to some of its endings, as shown by the Gr. alter- 
 nation oSs euros : hence, as opposed to L. aur-i-s = *aus-i-s (cf . 
 aus-culto, "I listen"), we may conceive a Preg. doublet *dus-a- 
 and *auz-d-, the former of which is reproduced only by Go. 
 (ius-o, all the remaining members of the stock showing the 
 rhotacized z, viz. O.N. et/r-a, O.E. ear-e and E. ear, O.H.G. or-a 
 and G. ohr. 
 
 II. Befoi'e or after a semi-vowel. This case seldom occurs, 
 and is of no importance. 
 
 III. Before or after a nasal or liquid. — Verner's law holds 
 good, unless the s has been assimilated to the neighbouring 
 sound : thus, sn is changed to zn, and, subsequently, to rn, cf. 
 G. lern-en (to learn) and lehr-en (to teach) ; but rs and rz give 
 rr ; Is, Iz, si and zl become 11 ; and lastly, sm and zm give mm. 
 In the group sr, we have the same inserted t as when initial, 
 whence sir, in E. sister = G. schwester, cf. the Sk. dat. svdsr-e 
 (to the sister). 
 
 IV. Before or after an explosive : the treatment is the same 
 as when initial. 1 
 
 § 2. Voiced Sibilant. 
 
 (62) In the I.-E. groups zb, zd, zg, zg, the explosive becomes 
 respectively Preg. p, t, k, according to Grimm's law ; where- 
 upon assimilation unvoices the z, and the final result is sp, st, 
 sk. We need but quote one typical instance : the I.-E. root sed 
 (to sit, to dwell), losing its accent and being reduced in a 
 derivative, formed an I.-E. word *ni-zd-6-s 2 (dwelling, nest), 
 Sk. nidd-s, L. nldu-s ; this word in Pregermanic became *nistd-s 
 *nestd-s, E. and G. nest, cf. G. nist-en (to nestle). The same is 
 the case with Sk. hed-as (anger), compared with E. ghost and 
 G. geist (mind). 
 
 pi. "vezum, the s has been carried on throughout the whole tense, and we 
 have vas vesum. 
 
 1 Nothing indeed can be clearer than that Verner's law is unable to 
 govern such groups as sp, st, sk, or even such as ps, ks, changed re- 
 spectively to fs, hs ; for, the s being here accompanied by a voiceless con- 
 sonant, even supposing that it could have become z, this z would again have 
 reverted to the sound s by assimilation, infra 62. 
 
 2 For the meaning, compare G. ein-sitz-en. 
 
SIBILANT CONSONANTS. 117 
 
 In the I.-E. groups zbh, zdh, zgh, zgh, the Preg. explosive 
 became respectively b, 6", g, and therefore the z was kept un- 
 changed ; then West Germanic regularly altered it to r: I.-E. 
 *mizdho-s (salary), Sk. mldhd-s, Gr. /wo-0o'-s, Go. mizdu, O.E. 
 meord, cf. E. meed and G. miethe (rent). So also zn has become 
 m. Everywhere else, the z is assimilated {supra 61, III.)., 
 
 Section III. 
 
 FINAL SIBILANT. 
 
 (63) I.-E. has no final sibilant but s. It is obvious that in 
 Pregermanic this s must either remain s, or become z, accord- 
 ing as it is or is not preceded by the Preg. accent, so that 
 *kasos becomes *hazds, while *kdsos gives *hdsaz, supra 61, I. 1. 
 Unfortunately, clear as it is, the fact cannot be verified in 
 Gothic, where the final sibilant is always spelled s. In O.N. 
 we find exactly the reverse : final z, though regular only at the 
 end of an unaccented syllable, was extended throughout, and 
 analogically replaced the final s; it was then rhotacized, and 
 thus we find the -r to be the regular ending of the nomin. 
 sg. in a great many words quoted from this language. West 
 Germanic is no less corrupted, but, as it were, without a 
 decided preference for either sound ; for, in some cases, the s, 
 and, in some other cases, the z was kept and extended throughout. 
 Further, final s remained (G. gen. vater-s, sohn-es), occasion- 
 ally however becoming voiced to z (E. gen. father's, sons, E. pi. 
 fathers, sons); whilst final z was dropped altogether, though 
 kept and, of course, changed to r in those little words which 
 became united in pronunciation with the following word, so 
 as to form a single group of sounds. Thus, we have G. voir 
 (we) = Go. veis, because, in such a locution as Go. veis bindam = 
 G. wir binden (we bind), the two words were blended into one, 
 and the final z was treated as medial. 1 
 
 1 E. we has lost the z even in this case. Further applications of these 
 laws will occur frequently in the accidence, infra 137, 139, 143, 150, 212, etc. 
 
CHAPTER VI. 
 ACCENT. 
 
 (64) Rhythm is as natural and essential an element of 
 human speech as the words themselves. It is only by an artifi- 
 cial method of abstraction that we can sever these two essential 
 factors of any spoken language. Even in the most common- 
 place sentence the tune, as it were, accompanies the words. 
 Though, of course, these tunes are less marked, less modulated, 
 and, therefore, less artificial, in language, than they are in 
 music, the ear, nevertheless, easily recognises them in respect of 
 both measure and melody. The principal time in each measure 
 is what we call a stress-accent (expiratory accent, emphasis, 
 intensity), whereas the high note in the spoken melopcea is 
 said to bear a pitch-accent (musical, chromatic, tonic accent). 
 
 Stress and pitch are combined, though in very unequal pro- 
 portions, in every human language, and the character of the 
 speech varies according as the one or the other prevails. Thus, 
 for instance, our European tongues admit very little of a sing- 
 ing tune, at least, in the accurate and conventional pronuncia- 
 tion of educated people 1 ; Swedish, however, has developed a 
 pitch-accent, which is the more noticeable because it is often 
 separated from the stress-accent, some syllables being sung on 
 a high note without stress, and others being emphatically 
 uttered though in a lower tone. The languages of Oriental 
 Asia are well known to have developed this musical power of 
 spoken language, so that in Chinese and, above all, in Annamite, 
 
 1 For rural dialects and even provincial languages are not at all free from 
 musical intonations, which often sound strange and unpleasant to an unac- 
 customed hearer. 
 
 118 
 
ACCENT. 119 
 
 a given monosyllable may assume the most various meanings, 
 according to the note to which it is sung. But, even in the 
 tongues of Europe, the sentence-accent, if not the word-accent, 
 allows an attentive ear to perceive some musical intervals. 
 
 In this province, however, and especially in English and 
 German, the stress-accent decidedly prevails. Thus, an Eng- 
 lishman can scarcely conceive the extreme difficulties under 
 which young people in' France long labour before they have 
 mastered the energetic stress of the German accent, which 
 usually rests on the first syllable of the word : scharf, scharfe, 
 scharfsinnig, scharfsinnigkeit. The English, in fact, differs from 
 the German accent only in degree : though the syllables are 
 uttered less emphatically, their respective value is very nearly 
 the same in both languages, whereas in French the stress- accent, 
 as resting regularly on the ending of each word, is much duller 
 to the ear, and, moreover, seems to have been of late in process 
 of transformation. 
 
 From this word-accent, which lays stress on one syllable, 
 and — in the languages we are studying — always on the same 
 syllable of a given word, we must carefully distinguish the 
 sentence-accent, which, in a given phrase, according to its 
 meaning and to the respective place of the word-accents con- 
 tained in it, lends a cadence to the whole speech, as the speaker 
 lays stress upon one syllable, slowly drawls out another, and 
 swiftly gets rid of the remainder, thus indefinitely varying his 
 tones according to his special purpose. 
 
 Section I. 
 
 WORD-ACCENT. 
 
 (65) I. In contrast with the fixed and expiratory accent 
 in our modern languages, the Indo-European accent was essen- 
 tially musical and moveable : a pitch-accent, for it consisted of 
 an alternation of grave and acute tones, which had been pre- 
 served even in Greek, as we learn from the testimony of ancient 
 grammarians ; a moveable accent, for it was shifted from place 
 to place, throughout the whole system of derivation, declension, 
 
120 ENGLISH AND GEEMAN GRAMMAR. 
 
 and conjugation, resting at one time on the root, and at another 
 passing to the suffix or to the ending ; and these alternations 
 of the accent have already been seen to provoke such strength- 
 enings and weakenings of syllables as are comprised under the 
 general name of I.-E. vowel-gradation, a process still surviving 
 after the death of the living accent which gave rise to it. 1 
 
 It remained musical in Sanskrit and Greek ; musical also in 
 Latin, but here it became fixed. Then, at the outset of the 
 Middle Ages, it changed its character and became a stress- 
 accent both in Modern Greek and the Romance languages. 
 
 The same change had been accomplished, even in the prehis- 
 torical period, throughout the Germanic family : here, the 
 primitive accent survived in the place it had occupied in the 
 Indo-European words ; it was still moveable, as shown by the 
 numerous and delicate effects of Verner's law 2 ; but it had 
 become a stress-accent. For these effects are inexplicable, 
 unless we suppose the accented syllable to have been pro- 
 nounced with a greater intensity and a strengthened expiratory 
 breath, so that the following spirant was maintained with its 
 full value, whereas the same spirant, when following an unac- 
 cented syllable uttered with less energy, became weakened, and 
 assumed a voiced sound. 
 
 II. This Pregermanic accent, however, is no longer to be 
 recognised except in its effects ; for it has disappeared in the 
 historical languages. * As early as the Pregermanic period 
 another accent had been developed, namely, a stress-accent, 
 invariably resting on the first syllable of each word ; the same, 
 perhaps, as the stress which may be observed in Latin, where 
 it produces such well-known processes of vocalic degradation as 
 are best illustrated by the L. types f actus confectus, lego eligo, 
 ccedo occido, and claudo secludo. This new initial accent gradu- 
 ally prevailed over the old one, and finally destroyed it. It has 
 now long been the governing accent. Let us briefly state the 
 general principles of this Germanic stress. 
 
 1. As a rule, every word, unless it is either enclitic or pro- 
 clitic, 3 has one, and only one accented syllable. 
 
 1 Cf. supra 44. 8 Cf. supra 53 D, 55 and 61. 
 
 3 Cf. infra 66, II. 
 
ACCENT. 121 
 
 2. From this we are not to infer that all unaccented syllables 
 should be uttered in the same way : what they have in common 
 is, that all are less marked than the accented syllable ; but 
 some are more marked than others. This may be easily veri- 
 fied by experiment with such words as E. opportunity and G. 
 unsterb lichkeit. 
 
 3. Especially in compound nouns, the chief accent always 
 rests on the first syllable of the first term, but the first syllable 
 of the second term also retains its accent, though there it has 
 only a secondary stress : E. black-bird, contrasted with the 
 sentence, the raven is a black bird ; E. yellow-hammer, corrupted 
 from *ammer = O.E. amore = O.H.G. amero and G. ammer, cf. G. 
 gold-ammer ; Gr. sonnen-finsternisz (eclipse of the sun) ; E. bride- 
 grobm = O.E. bryd-guma, and G. brauti-gam = O.H..Gc. bruti-gbmo, 1 
 etc. 
 
 4. The same is the case with compounds formed of prefix and 
 verb, so far at least as prefix and verb remain present to the 
 speaker's mind as two distinct words, each keeping its own 
 and original meaning, nay, even still separable from one 
 another in German : E. to over-look, G. ein-setzen, etc. If, how- 
 ever, the prefix has become blended with the verb into one 
 compound inseparable in meaning, then, of course, the prefix 
 loses its accent, and the accent of the verb is alone retained : 
 we need but compare G. uber-setzen (to get over) and ubersetzen 
 (to translate). Under these conditions the prefix is reduced to 
 a dull syllable (Go. fra-lusan-s, E. for-lorn, G. ver-loren; Go. 
 bi-gitan, E. to be-get, G. be-kommen, " to receive," be-quem, 
 "handy"; Go. ga-vaurpan-s, G. ge-worden, "become," etc.); 
 occasionally in current speech, and even in the written lan- 
 guage, to a mere consonant (G. b-leiben, g-lauben ; G. zwar — 
 O.H.G. zi-wdre, " forsooth," which in Mod. G. would be *zu 
 wahr, cf. G. fiir-wahr and HZ. forsooth). 
 
 5. The peculiar energy of this initial stress, which causes 
 every syllable but one to be more or less slurred over, explains 
 quite satisfactorily, in the Germanic as well as in the Romance 
 languages, the well-known and forcible syncopes, which often 
 
 1 Literally " the bride's man," cf. Go. guma='L. homo. The r inserted in 
 English proceeds from the false analogy of groom (a youngster). 
 
122 ENGLISH AND GEEMAN GKAMMAE. 
 
 shortened, so as to scarcely allow of their recognition, both 
 original Germanic words, and those which were early introduced 
 into English and German from Latin or other tongues. Some 
 examples have already occurred : E. world and G. welt ; G. 
 pferd, Pfingsten, etc.; G. pfalz = h. palatium; E. tninster and G. 
 miinster =L. monasterium. Here may be added a few more, 
 selected from the oldest compounds : E. neighbour and G. nach- 
 bar, whereof the second term is bur (planter, peasant), still 
 visible in the G. doublet Nachbaur (proper name) 1 ; G. arm- 
 brust, corrupted by a popular etymology from arm (arm) and 
 brust (chest), really the L. word arcu-balista, cf. E. arbalist = ¥. 
 arbalete, Italian balestra; E. marshal and G. marschall= O.H.G. 
 marah-scalc (horse-groom), 2 also latinized to mariscalcus, F. 
 marechal; G. Samstag (Saturday) = O.H.G. samba^-tac = *sam- 
 bata-, cf. Go. sabbato dags ; E. hussy = house-wife ; E. lady = E. 
 hldf-dige, literally " she who kneads the bread," 3 and E. lord 
 = O.E. hldford=*hldf-weard "loaf- ward"; E. sheriff = O.E. 
 sclr-gerefa " officer of the shire," cf. E. reeve, etc. Lastly, in 
 English, compare the word borough with the final in Canter-bury 
 = G. burg, and the word home with the final in Notting-ham = 
 G. heim- (abode). 4 
 
 6. The tendency to regressive accent is so strong and so 
 universal in the Germanic languages, that it has survived down 
 to the present day. Not only such French words as were 
 brought into English through the Norman conquest, mutton, 
 coward, danger, reason, etc., underwent this general influence ; 
 but, even in Mod. English and German, it often modifies the 
 
 1 Alamannic nochpr, with the second term completely reduced. 
 
 * Cf. the modern words E. mare, G. mature (mare), Low Breton viarc'h 
 (horse). 
 
 8 This being deemed the essential function of a landlady. For the terms 
 of the compound, see the words loaf and dough. 
 
 4 Still more exactly, -bury — OJEi. by rig is the dative of borough=burg. 
 The most extreme reduction is no doubt to be observed in G. messer (knife) = 
 M.H.G. messer, corrupted from me^eres, which was mistaken for a genitive, 
 and this again representing an O.H.G. me^iras=me^irahs, with regular 
 rhotacism, for *me^i-sahs, literally " a blade for eating." For the first term, 
 compare Go. mats, O.H.G. ma%, O.E. mete (meal), the latter surviving with 
 a restricted sense in E. meat (cf . Fr. viande " meat " = L. vivenda " victuals ") ; 
 for the second term, L\ sax-u-m (stone), O.H.G. sahs and O.E. seax (blade, 
 sword, knife), which trace back the birth of this curious expression to the 
 stone-age. 
 
ACCENT. 123 
 
 learned words introduced, not excepting some nearly contem- 
 porary borrowings: E. nation, country-dance = F. contreddnse ; 
 G. ocean, leutnant = Y. lieutenant} But these processes, curious 
 as they are, it is the province of the dictionary, not of the 
 grammar, to describe. 
 
 Section II. 
 
 SENTENCE-ACCENT. 
 
 (66) The various and almost infinite shades of expression 
 comprised under the general name of sentence-accent would, 
 even were they capable of classification, naturally stand far 
 beyond the reacb of the present work. We have here to deal 
 with two cases only : the pitch-accent, illustrated by the well- 
 known contrast between question and answer ; and the stress- 
 accent, as affecting the relative atonic character of certain little 
 words, called either proclitics or enclitics. 
 
 I. The difference between an interrogative or exclamatory sen- 
 tence and a mere affirmation never escaped the most inaccurate 
 ear, the less so because it is often from the tone only that we 
 are able to learn the purpose of the speaker. The characteristic 
 pitch in the former was remarked by Greek grammarians, and 
 was expressed by an invariable acute accent on the interroga- 
 tive pronoun ns, whereas the indefinite pronoun tis always 
 remains unaccented. In English, we need but repeat in the 
 usual key such short phrases as " they are here," and " are they 
 here ? " the concluding syllable in each being sung respectively 
 on the lowest and on the highest note of the spoken melopoea. 
 The intervals, descending or ascending, may in this instance be 
 estimated at the average value of a fifth at most, though in- 
 creasing whenever the speaker is labouring under strong 
 emotion ; and the same may be observed, though with even 
 more vigour, in similar German phrases. 2 A further combina- 
 tion of word-accent and sentence-accent will even reproduce 
 
 1 If, however, the word is a long one, the accent is hardly ever thrown 
 back beyond the antepenult : E. opportunity, etc. 
 
 s Cf. the triple notation of the phrase " er geht fort " (he goes off), in Be 
 haghel, die Deutsche Sprache, p. 145. 
 
124 ENGLISH AND GERMAN GRAMMAR. 
 
 the four musical tones of the peculiar Chinese accentuation in 
 a single English question and answer, as : " You wish to stay 
 two days, do you ? — I do." * 
 
 II. In current speech, such auxiliary words as articles, pre- 
 positions, conjunctions, etc., are unaccented and closely united 
 to the words by which they stand. In the English sentence, 
 " he was there as I came," the hearer actually perceives but two 
 words, each of three syllables, whilst grammar finds in it six 
 monosyllables. This case is termed proclisis, i.e., the unac- 
 cented words are leaning forward upon one following. In the 
 contrary case, enclisis takes place, that is, the unaccented 
 words are leaning backwards upon a preceding word as in the 
 sentences, E. go on, yes Sir, tell it me, G. er schlagt ihn, 2 schame 
 dich, was mdchst du ?, gieb es mir, etc. 
 
 It follows from the definition that the so-called category of 
 enclitics and proclitics is at once large and indefinite ; for it 
 depends entirely on the speaker, who may glide over a word 
 upon which in any other sentence stress would be laid, — as in 
 dont play the fool, where the verb to play is almost lost in the 
 command given, — or, on the other hand, he may dwell upon 
 some monosyllable which otherwise would appear insignificant, 
 as in " you shall do it, this won't do, that's too bad," and so forth. 
 Here we must confine ourselves to certain main general prin- 
 ciples. 
 
 1. The so-called demonstrative, especially when weakened 
 to the vague meaning of the definite article, is proclitic and 
 leans upon the following noun : E. the man, Gr. der sohn. To 
 this unaccented article are to be referred many syncopes in 
 
 1 Viz. to, short and brusque ; two, raised and monotonous ; 1. do, raised 
 and ascending ; 2. do, raised and descending. In other words, an affirma- 
 tive sentence is a completed tune, which concludes on the tonic, or, at least, 
 on the dominant note (authoritative tone), whereas an interrogative sentence 
 is an uncompleted tune, which lacks a conclusion (the answer), and is wait- 
 ing for it. 
 
 * Of course, er is proclitic, and the three monosyllables make one word. — 
 In these alternations of accented and unaccented syllables, M.H.G. possesses 
 some very delicate laws of secondary accentuation, which might be illustrated 
 by a thorough study of the metrical sources displayed in the Nibelungenlied ; 
 but here we are unable to enter into such minute details. Gf. Paul, Mhd. 
 Gramm., p. 8. 
 
ACCENT. 
 
 125 
 
 both, languages : G. am = an dem, zur = zu der, even (Alamannic) 
 s puech = das buch; vulgar E. tother = *thet other, where the 
 article has lost everything but the final t, the same that still 
 appears in the demonstrative that. 1 As to the indefinite article, 
 Ave have the contrast between G. ein mdnn " one man " and G. 
 ein mdnn " a man," and this becomes more marked in popular 
 speech, as in Swiss e ma. The same difference has caused the 
 spelling of the word to diverge in English, the O.E. an (one) 
 having become one in the former case, and an a in the latter. 2 
 Any other determinative word with slight signification may 
 undergo a similar degradation: we need but compare the swift 
 and short utterance of the possessive, in mylord, mylady, with 
 its emphatic and diphthongized pronunciation in my Lord 
 and Father ! 
 
 2. A preposition is proclitic upon the following noun. The 
 two G. words zu (too) and zu (to) are both from O.H.G. zuo = 
 Preg. *to : the sole difference lies in the accentuation, which, 
 although well marked in German, appears even more clearly 
 in the English spelling, too and to. s The same is the case with 
 the contrast between the E. and G. preposition in and the G. 
 prefix ein- : the former was always unaccented ; the latter 
 occasionally had the stress laid upon it. Thus every preposi- 
 tion appeared under a double form, which some of them still 
 retain in current speech, though in the literary language one 
 of the two usually prevailed and made the other disappear: E. 
 by = G. bei, both strong, as opposed to the same word reduced 
 to an inseparable prefix, E. be-, G. fee- ; E. out = G. aus, invari- 
 ably strong 4 ; but E. up, invariably weak, may be contrasted 
 with G. auf, which is no less constantly strong ; and there are 
 many such. 
 
 1 Cf. infra 130, 1. — Usually but wrongly spelled t'other, as if the word 
 actually proceeded from an elision of the other. 
 
 a An before a vowel, a before a consonant (cf. supra 19, 2, note), and con- 
 sequently also before initial u, since u is sounded yu, and y is a consonant. 
 But, of course, an urn, etc. 
 
 8 Even zu (too) and too may become proclitic if we do not lay stress upon 
 their meaning. The contrast is well marked in Alamannic, proclitic tse, 
 accented tsue. 
 
 4 The weakened vocalism would be E. *ut, G. *us. Cf. E. b-ut (apart 
 from, however), where b- is be-, E. utt-er comparative, and ut-most superla- 
 tive to out, further to utter (to bring out), etc. 
 
126 ENGLISH AND GERMAN GRAMMAR. 
 
 3. The pronoun, when subject, is almost always proclitic 
 npon the verb, unless the speaker means to lay some stress 
 upon the pronoun, a process which is familiar in popular elocu- 
 tion, as in Alamannic te vays s vol "you are aware of it," but 
 yd tit vays s " of course you are fit for it " (ironically) . 
 
 4. Under the same reservation, the pronoun, when object, is 
 generally enclitic — or proclitic if placed before — upon the verb 
 upon which it depends. It is owing to this loose and unac- 
 cented pronunciation, that in E. him and her the initial h is no 
 longer heard, nor even written in i£ = O.E. hit. 1 From this 
 also many syncopes arise, more or less violent : E. I told 'em = 
 I told them, 2 G. i kenris = ich Jcenne es. 
 
 5. The auxiliary verb is rather proclitic — or enclitic if fol- 
 lowing — upon the principal verb. The process is quite regular 
 in English pronunciation : Fve, Tm, he'd, etc., 3 and here, at 
 any rate, much more constant than in German, because G. 
 syntax often severs the auxiliary from the principal verb, so 
 that they cannot lean on one another. But in both languages, 
 at least in the great majority of cases, the verb " to be," even 
 when it does not play the part of auxiliary, is either unac- 
 cented, or at most has a secondary accent. 
 
 6. Little conjunctions, E. and, if, when, G. und, wenn, wann, 
 etc., are mostly proclitic. Hence the reduction of O.H.G. also 
 " quite so " to G. als and (dialectal) as, and of the same word 
 O.E. eal-swd to E. as, whereas the non-proclitic type still sur- 
 vives in E. also and G. also. 
 
 These summary remarks must suffice for the theory of Eng- 
 lish and German accentuation. A more particular view may 
 be derived from the study of special grammars, or, better still, 
 from the practice of the two languages. 
 
 1 Hence also the weakening of a voiceless to a voiced spirant in thee, 
 them, where, owing to enclisis, the th was really medial ; it then followed by 
 analogy in the nominatives, thou, they, and also in this, that, the, and the 
 like proclitic monosyllables. The same is the case with of and off: the 
 former is proclitic, and never final, thus of him, of the, pr. evim, eiffS; the 
 latter is always final and has the secondary accent, as in he goes off, pr. 
 hlgowzof. Cf. supra, 49, I., 2, and 54, I., 2. 
 
 1 Strictly speaking, however, 'em is not syncopated from them, but from 
 the old regular plural of him, infra 166, II. 
 
 8 Compare the pronunciation of it 's done and 't is done, according as the 
 enclisis obliterates either the auxiliary or the pronoun. 
 
SECOND PART. 
 WORDS. 
 
 (67) It is a trite and common-place remark, that, however 
 rich a language may be, the whole bulk of its words, if sub- 
 mitted to analysis, amounts to very few. Let us, for example, 
 take such a series as to respect, respectable, respectful, respectful- 
 ness, all depending on the same word, respect. It will only 
 require a slight power of reflexion, even without any knowledge 
 of Latin, to add aspect, to suspect, and many others, including 
 their derivatives. Further, any one who is but slightly ac- 
 quainted with the history of language, may find, as it were, 
 from every part of the dictionary, such distant words as species, 
 spice, bishop, respite, etc., all coming to rank under the same 
 head. Shortly, designating by the conventional term " root " 
 the irreducible syllable which either appears or is concealed 
 in each one of these words and yields it its general meaning, we 
 might enumerate more than a hundred English words which 
 more or less directly have arisen from the Latin root spec (to 
 look). 
 
 Yet, at the outset, there is a fundamental difference to be 
 noted between these offsprings of a common ancestor. Some of 
 them are formed according to a process of which every speaking 
 person still remains vaguely conscious, and which may at any 
 time be used anew to enrich the tongue with new words at 
 command. In this century, indeed, in this very decade, steamer 
 and boycotter have sprung from steam and to boycott, just in the 
 same way as spice and to respect had formerly produced spicer 
 and respecter. Others, on the contrary, are old and fixed forms, 
 which are no longer capable of reproduction, inasmuch as his- 
 
128 ENGLISH AND GERMAN GRAMMAR. 
 
 torical or prehistorical research would be required in order to 
 discern their primitive elements : thus, the word respect, through 
 L. re-spec-tu-s, goes back to L. re-spic-ere, and claims a place in the 
 numerous family of such Latin nouns as were created by adding 
 to the root a syllable -tu- ; but it can no longer serve as a 
 pattern for new English formations, since, as a matter of course, 
 people who now-a-days are speaking English know nothing of 
 the word respectus, or of the way in which it once arose from 
 respicere. If, then, we designate by "suffix" the formative 
 element subjoined in each of these words, we may say that 
 the suffix -er in boycott-er, etc., is a living suffix, still re- 
 cognised and used as such by all English speakers, whereas the 
 final t in respect, though representing a Latin suffix -tu-, is a 
 dead suffix, a fossil, as it were, dug out of the ground by 
 linguistic exploration. 
 
 This distinction likewise holds good for the prefix, or ele- 
 ment added before a root. It is true that re- in re-spect is no 
 longer recognised as a prefix to root spec ; but it is still recog- 
 nised as such in re-turn, re-place, re-flow, and many others, so as 
 to render possible the creation of such verbs as re-bellow, re- 
 borrow, re-photograph, the meaning of which would be plain to 
 any hearer. On the other hand, though the collective prefix 
 co- may be recognised in a great many English words, and 
 though it is still living in the learned derivation of some 
 technical terms, viz. to co-habit, co-heir, co-partner, etc., to be 
 sure no member of a group would think of saying : " We were 
 *co-photographed." And, if he did, he would scarcely be under- 
 stood. 
 
 All spoken languages, perhaps English and German more 
 than any others, suggest these and similar observations. 
 
 First, they exhibit, quite visible and still endowed with a 
 great amount of vitality, the process of composition, which 
 is well nigh lost in French. Naturally the two terms in a 
 compound, unless it is very old and has undergone in the course 
 of time considerable phonetic reduction', 1 remain almost always 
 clear to the mind's eye, so that new compounds may be formed 
 
 1 See the instances, supra 65, 5, and infra 114 sq. 
 
WORDS. 
 
 129 
 
 just as they are needed, the old ones serving as models for 
 further imitation. 
 
 Next, some relatively recent derivations are really 
 compositions, with but this difference, that their second 
 term is no longer used as a single word : thus we have E. 
 child-hood and G. kind-heit, whereas E. *hood and G. *heit have 
 now no meaning by themselves. They are dead words, but 
 living suffixes, so intensely living indeed, that the speaker is 
 fully sensible of the specific function he attaches to them, inas- 
 much as no other derivative from E. child and G. kind would 
 represent exactly the same shade of meaning. 1 
 
 The case is somewhat different, if we consider, for instance, 
 E. child-ish and G. kind-isch. Although the suffix -ish and -isch 
 is a living one and used in forming new words, its meaning is 
 no longer so distinct as in the former case : in other words, a 
 little reflection will show that some other word, — let us say 
 *child-ly, had usage sanctioned it, or kind-lich (cf. mann-lich) , 
 had it not assumed a different shade of signification, — would 
 suit as well. Here we have a suffix properly so-called : per- 
 haps, long ago, this suffix also was a word in itself, like *hood 
 and *heit, or like *ly and *lich, which it resembles in force. 2 
 But it no longer bears any trace of the fact, and the meaning 
 which is now attached to it, — though still accessible to the 
 speaker's mind, and, therefore, still capable of reproduction in 
 some new formation, — is so vague and yielding as to allow half 
 a dozen periphrases to be equally fit for replacing it. 
 
 Further, the same suffix -ish or -isch is to be found in E. 
 Welsh, from Wales, or in G. men-sch, from mann. Here we 
 may say, if we choose, that it is still perceived, though less 
 
 1 And therefore the creation of such new terms as fenian-ship (cf. gypsy - 
 dom, G. Eliot, Mill on the Floss, XL) or brahmaventhum is always possible, 
 though they are now becoming uncommon. It is because the illiterate 
 have rarely occasion to exert their skill in this way, whereas cultured people 
 are accustomed all their lives to the use of Greek and Latin suffixes 
 (fenian-ism, brahvian-ismus, etc.). 
 
 2 Common sense indeed would suggest that there was a time, — however 
 far back, — when every suffix had a real and distinct meaning ; thus every 
 suffix was once a word, and every derivation, strictly speaking, is a compo- 
 sition. However, such deductions are of little practical jg&am-ki the philo- 
 logist. ' ^-p^SE LlBRv?) 
 
 [UNivE^-j :■■?: 
 
130 
 
 ENGLISH AND GERMAN GEAMMAE. 
 
 clearly. But it is dead and buried, no reproduction of it being 
 at all conceivable. A German would never forge such a word 
 as *g'6tt-sch " a being of a divine nature," nor an Englishman, 
 such a word as *Burm-sh " an inhabitant of Burma." And, if 
 this is the case with an element still visible though phonetically 
 weakened, how much more with a great many prehistorical 
 suffixes, which are often reduced to a single consonant, or have 
 even disappeared altogether, leaving but an accidental change, 
 a metaphony to bear witness of their presence, or even vanish- 
 ing without a vestige to betray them ! 
 
 (68) The distinction of dead and living suffixes is important 
 not only in respect of their actual use in a given language : it 
 ought also to be considered in the light of such accidental 
 alterations as current speech might impose upon these unstable 
 elements. A word is an organism : while living, it is struggling 
 for life ; when dead, it becomes inert and is readily corrupted. 
 Supposing, for instance, that, after the Norman conquest, our 
 word tree had been replaced by arhre, though the old compounds 
 apple-tree, plum(e)-tree, nut-tree, etc., had been preserved, what 
 would fain have become of these survivals ? Since the word 
 tree was dead, it would no longer be recognised in such com- 
 pounds, except as an arbitrary element added to the name of 
 the fruit. Thus, the two i's in nut-tree being easily blended in 
 rapid utterance, the word would be likely to become *nutree. 
 Now, as the speaker would be aware that the difference between 
 nut and *nutree consisted only in an affixed syllable -ree, he 
 might well be induced to affix the same syllable to apple, cherry, 
 etc., and thus to form such tree-names as *appleree, *cherryree, 
 etc. Further, as the final would be unaccented and no longer 
 defended by any outer protection, the dictionary might now ex- 
 hibit something like *nutry, *applery, etc., wherein the primitive 
 tree would be completely lost. 1 Now, such blunders as these, 
 
 1 This is just what historically happened to O.N. \>HSjuvg-r (a third, a 
 division by three), which was pronounced riding in North- thriding, whence 
 the names of the two other thirds, East- and West-Riding, so that now the 
 territorial divisions of Yorkshire appear to have something to do with eques- 
 trian exercise. A reversed accretion to a primitive element is likewise pos- 
 sible and even more frequent : thus, from the same cause, Go. -assu- became 
 E. -ness and G. -nisz, G. -heit produced -keit, etc., as stated below. 
 
WORDS. 131 
 
 more or less extended or restricted as to their effects, are quite 
 common throughout the history of every language, so that even 
 the most complete etymological repertories cannot pretend to 
 register them all in every particular. 
 
 Since derivation is liable to such irregularities, it follows 
 that we cannot study it methodically except by tracing it back 
 to its origins ; for there, and there only, must it appear as free 
 as possible from perturbing influences. But there is a further 
 reason : living suffixes, however recent they may be, however 
 degraded by successive sound-reductions, however increased by 
 subsequent accretion, differ from the earliest derivative syllables 
 only in form, and not in substance ; the processes in language 
 may become more and more intricate, but their nature remains 
 unaltered. A creation out of nothing would be as inconceiv- 
 able for a suffix as for a word : there is not a derivative element 
 in English or German which may not claim descent from the 
 whole Indo-European past, and the structure of such a modern 
 verb as veroffentlichen is based on the same principle as that of 
 a radical verb like heb-en = Go. haf-jan = h. cap-id. It is a slow 
 and uninterrupted process of analogical breeding, that en- 
 riches a tongue whether in the case of primitive monosyllables 
 or modern polysyllables ; and thus we are compelled to begin 
 with the root, as being the foundation of the building, in order 
 to distinguish and classify such materials as have been brought 
 together and added to it in successive accretions. 
 
 (69) We must, therefore, study briefly Primitive Deri- 
 vation, which gave West Germanic, as well as Gothic or any 
 other Indo-European language, its first stock of pattern-words. 
 Prom this first stock, English and German formerly drew, and 
 are still drawing new words, by applying to them the same 
 processes which served to form the old ones : this is what we 
 call English and German Derivation. To these deriva- 
 tive resources, which had become inadequate owing to the 
 phonetic reduction undergone by many suffixes, English and 
 German subjoined some others, such as accumulation of 
 several suffixes, or borrowing from a foreign derivation, 1 or 
 
 1 As a matter of course, the following pages deal only with Germanic 
 
132 ENGLISH AND GERMAN GRAMMAR. 
 
 wide extension of a so-called false suffix. This last-named 
 had been first a word in itself, had then become the second 
 term in a compound, and finally losing its meaning, had but 
 the value of a nought at the end of a number. And the last of 
 these processes establishes an easy transition from derivation 
 by suffixes to Composition properly so-called, a process trans- 
 mitted likewise, without noticeable change, from the early Indo- 
 European to the contemporary English and German languages. 
 
 suffixes, while those which were borrowed from French, Latin, and Greek 
 require at most very brief mention. Here we may quote: (1) E. -ess (F. 
 -esse = li. -issa), which serves to form the feminine (mistr-ess), even for nouns 
 of Germanic origin (quak-er-ess) ; (2) E. -able, with the same meaning as F. 
 -able, borrowed from L. -abilis, and similarly extended (eat-able, read-able), 
 as well as its derivative -ability ; (3) E. -y, G. -ei, which is simply the O.F. 
 -ie, widely extended, E. butcher-y, grocer-y, then amplified, as in yeoman-ry, 
 G. arzen-ei, reiter-ei, then further amplified, as in sklave-rei, etc. ; (4) Gr. 
 suffix -itrnjs, L. -ista, F. -iste, E. and G. -ist, in common use, in spite of its 
 learned origin, E. novel-ist, G. artiller-ist, hobo-ist, horn-ist, etc. Further 
 add, in English, the Latin or Romance Buffixes, all of them more or less 
 naturalized, of such words as fulftl-ment, slumber-ous (F. -eux), hindr-ance, 
 starv-ation, Siam-ese (F. -ois), etc. As far as derivation is concerned, Eng- 
 lish might be deemed almost as much a Eomance as a Germanic tongue. 
 
CHAPTER I. 
 PEIMITIVE DEEIVATION. 
 
 (70) The combination of a root with a single suffix, 1 
 being the simplest element in spoken language after the pure 
 root, is called a stem, or, more accurately, a primary stem. 
 To the primary suffix may be added a secondary one ; to this, 
 a third, and so on indefinitely. 2 Since, however, the derivative 
 process remains uninterrupted and invariable, we need but 
 distinguish two degrees in the scale of derivatiou, namely 
 primary and secondary suffixes, the latter including any degree 
 whatever beyond the primary, observing, moreover, that the 
 so-called secondary suffixes are for the most part either primary 
 suffixes, or accumulations of primary suffixes, which have been 
 analogically extended from primary to secondary formation. 
 
 Section I. 
 
 PRIMARY SUFFIXES. 
 
 (71) The primary stem derived from a root may be either 
 nominal or verbal, that is to say, adapted for playing in any 
 proposition the part either of a noun or a verb, according as it 
 is capable of receiving, by accretion to its characteristic suffix, 
 the peculiar endings of declension or conjugation. This funda- 
 
 1 The I.-E. language knew nothing of derivation by prefix. The element 
 which has become, for instance, G. ge- or ver-, was originally a word in 
 itself, and indeed quite as independent a word as may now be G. auf or 
 durch ; which is as much as to say that in Germanic the prefixation is 
 primitively a true composition. 
 
 2 We may instance, starting from an I.-E. root pu (to be foul) : — primary 
 stem, I.-E. *pu-lo- = O.E. fu-l = E. fou-l ; — secondary stem, O.E. fy-l-ff 
 (*=*fu-l-id r ), E. fi-l-th; — tertiary stem, E. fi-l-th-y ; — quaternary stems, 
 T&.fi-l-th-i-ly, fi-l-th-i-ness, — and so forth ; for, if a vb. to *filthify ever hap- 
 pened to be formed, the quinary *filthification would ensue without difficulty. 
 
 1SS 
 
134 ENGLISH AND GERMAN GRAMMAR. 
 
 mental distinction governs the whole of grammar, 1 and no other 
 could here better suit our purpose. 
 
 § 1. Nominal Steins. 
 
 (72) I. Root-Stems. — In some rare cases, the bare I.-E. root, 
 without any suffix, may be declined and therefore plays the 
 part of a noun. Thus, an I.-E. root, normal ped, deflected pod, 
 was declined with the nominal meaning "foot," whence, after 
 a vocalic lengthening which often appears in the nomin. sg., 
 I.-E. *ped-s or *pod-s, Sk. pad, Gr. irovs = ttws = *7ro>8-s, L. pes = 
 *ped-s, etc. ; these forms, in Germanic, become O.N. fot-r, O.E. 
 jot and E. foot, O.H.G. fuoj and G. fusz. As, however, the 
 stems belonging to the classes given below, after they had 
 regularly dropped the final vowel, were no longer distinguish- 
 able from stems formed without any suffix-vowel at all, analogy 
 could not fail to group together these apparently similar words, 
 so that the root-stems were as a rule not distinguished from 
 further derivations. 2 
 
 II. Stems ending with a short vowel: -#- (-e), 3 -£-, -u-. — It 
 is of course impossible to recognise, in the nomin. sg., either in 
 West Germanic or Gothic, a stem formed by addition of a bare 
 suffix -0- or -£-. 4 But the vowels reappear, in Gothic, in the 
 other cases ; and, even in English and German, they may be 
 sometimes inferred from the declension, especially from a meta- 
 phony which the % had produced before it was dropped. In 
 these formations, the root generally assumes either the normal 
 or deflected grade. 
 
 1. Stems in -0-, masculine or neuter, never feminine : I.-E. 
 
 1 Speech consists of no other rvords than nouns and verbs : the adjective and 
 the pronoun are but varieties of the noun ; invariable words, as far as they 
 admit of an etymology, ought to be traced back to ancient cases of nouns, 
 adjectives, and pronouns, which survived from a forgotten declension. 
 
 2 Thus, G. fusz (pi. fiisz-e) is declined as if it were derived from a Preg. 
 *fot-i-, and, even in Gothic, the regular *fot-s has become nomin. fot-u-s, 
 through the analogy of ace. fot-u = *fot-u-m (cf. L. ped-em) = I.-F>. *pod-vi. 
 
 8 It should always be borne in mind that any syllable containing I.-E. 0, 
 may change this o to i, and vice versa, supra 44 : Gr. nomin. sg. iirir-o-s, 
 voc. sg. tirir-e. As a matter of faot, however, the grade o (Preg. a), 
 favoured by analogy, was the only one retained in suffixes ending with a 
 vowel ; the alternation between o and e, when followed by a consonant, will 
 occur below, 74 and 80. 
 
 4 Cf. supra 34, § 2. 
 
PRIMITIVE DERIVATION. 
 
 135 
 
 root dhegh (Sk. dah, " to burn, to shine "), whence a msc. stem 
 *dhogh-6-s (day), Preg. *^a^-d-s, Go. dags, O.N. dag-r, O.E. 
 dceg and E. day, O.H.G. tac and G. tag ; I.-E. root yewg (Sk. 
 yuj, "to join, to team"), whence a nt. stem *yug-6-m (yoke), 
 Sk. yugdm, Gr. £vy6V, L. jugum, Preg. *yiik-d-m, Go. juk, O.E. 
 #eoc and E. yoke, O.H.G. joh and G. joch. 
 
 2. Stems in -*-, in the three genders : I.-E. root bhid (to split, 
 cf. L. find-ere), whence msc. *bhid-i-s (the act of splitting or 
 biting), Preg. *hit-i-z, Old Saxonic bit-i and E. bit, G. bisz; so 
 also, Go. slah-s (a blow), O. Sax. sleg-i, G.schlag 1 msc; Sk. 
 fm. *jdn-i-s (woman), Go. qen-s, E. queen ; L. nt. mare (sea) = 
 *mar-i, Go. mar-i- only in composition, E. mere, and G. meer 
 (sea), where the i has caused metaphony. 
 
 3. Stems in -u- (the u is kept in Gothic, but is dropped 
 everywhere else, so that the class has become obscured) : Gr. 
 KodV-v-s (strong), Go. hard-n-s, E. hard, G. hart ; Go. hand-u-s 
 (hand), E. and G. hand, the G. pi. being hiind-e, as if the 
 declensional stem were *hand-i-. 
 
 (73) III. Stems ending with a long vowel -a, essentially 
 feminine. — This class, which corresponds to the so-called first 
 declension in Greek and Latin, is of considerable importance 
 in Indo-European, and yet of little significance in modern 
 Germanic languages, final -a (Preg. -0) having become Go. -a, 
 and then being changed to the dull vowel, or even completely 
 dropped in later languages (Go. airfia, 0.7$. jor&, O.E. eor<Se and 
 E. earth, O.H.G. erda, M.H.G. erde and G. erde erd) ; whereupon 
 the declension, which had lost its characteristic feature, became 
 the sport of analogy. 2 But we must here lay stress upon the 
 old and constant parallelism between the fm. d, and the 
 msc.-nt. which has been illustrated above. From this it 
 follows, that any suffix, viz. -0-, -mo-, -no-, -to-, etc., ending 
 
 1 Thus, in the sg., G. tag and schlag are alike; but in the pi. we have 
 tag-e = Go. dagos = I.-E. *dhoghos contracted from *dhogh-d-e.i, and metaph. 
 schlcig-e which would be Go. *$lageis = 'slag-iy-t>s, so that the suffix- vowel 
 becomes visible. — Observe moreover that it is not at all necessary that a 
 given derivation should always exhibit the same suffix throughout the I.-E. 
 family : the word which, for instance, is L. pisc-i-s (fish), is Germanic 
 *fisk-a-z (Go. sg. fisk-s, pi. fiskds), whence we may infer two distinct I.-E. 
 forms, viz. *pisk-i-s and *p/sA-o-«. 
 
 2 Infra 142, 4, and 150, 1. 
 
136 ENGLISH AND GEKMAN GKAMMAR. 
 
 with o and forming a masculine or nenter stem, may likewise 
 end with d, thus -d, -ma, -nd, -td, etc., and represent under this 
 shape a correlative feminine. In particular the nouns of three 
 genders, that is to say the adjectives, exhibit the three well- 
 known endings: Sk. msc. priy-d-s (dear), im.priy-d, nt. priy-d-m; 
 Gr. Si/ca-io-s 8tKa-td 8Ua-io-v, /caA-o-s KaX-rj kcl\-6-v ; L. bon-u-s 
 bon-a bon-u-m, etc. This process still appears very clearly in 
 Gothic : lagg-s lagg-a lagg, liuf-s liub-a liuf, etc. And though, 
 of course, the German alternation, lang lang-e, lieb lieb-e, can 
 no longer be said to reproduce it, 1 yet there is still some re- 
 semblance between the two processes ; indeed, the latter proceeds 
 indirectly from the former, inasmuch as the final in lang-e is 
 borrowed from a primitive demonstrative stem which originally 
 formed its feminine in that way. 
 
 IV. Stems with suffixes -yo- (-yd) and -wo- (-wd). — The 
 laws of Germanic sounds have completely altered these once 
 important classes : I.-E. *medh-yo-s *medh-ya, Sk. mddh-ya-s, 
 Gr. [xeo-os — fieo-<ro<; = *iAc6-yo-<s, L. med-iu-s med-ia, Go. mid-ji-s, 
 whereof the metaphony survives in E. mid, G. mit (by means of, 
 with), and G. fm. mitte (middle) ; I.-E. root ro rd (to rest), 
 Gr. ipwrj (rest) = *pw-Fd, Preg. *ro-wo, whence O.H.G. ruowa 
 and G. ruhe. 2 
 
 (74) V. Stems with suffixes -en- (reduced -n-, deflected -on-) 
 and -yen- (red. -in-, defl. -yon-). — Here at length we have come 
 to an end of the suffixes which could not escape being com- 
 pletely defaced in Germanic phonology. Every one of those 
 which follow contained at least one element capable of resisting 
 this perverting influence ; and, in the case of the present nasal 
 suffix, the n not only subsisted, as it regularly ought to do 
 except when final, but even spread out on a large scale, a 
 process which in grammar constitutes what we may call the 
 encroachment of the weak declension. 3 In order to account 
 for it, we must observe first, that the vowel in this suffix is 
 lengthened in the nominative, and further that it passes in 
 
 1 Cf. infra 155 and 157. 
 
 4 It is the same root which, with another suffix and in a different (reduced) 
 grade, became Go. ra-sta (staple), E. rest, G. rast. 
 s Cf. infra 140-142, 149, 150, 156, etc. 
 
PRIMITIVE DERIVATION. 137 
 
 declension through the three vocalic grades : Gr. ku-wv (dog), 
 but gen. kv-v-os; L. hom-o (man), but gen. hom-in-is. Now 
 the I.-E. n in this suffix, when final, that is to say only in 
 the nomin. sg., might either be kept or disappear. 1 All these 
 alternations, variously distributed, are to be found again in 
 Germanic stems : thus, Gothic answers the L. hom-o, quite 
 regularly, with nomin. sg. gum-a ; it answers the L. gen. 
 hom-in-is with its gen. gum-ins, wherein the vowel is altered, 
 and the n likewise reappears ; and lastly, like Gr. kv-wv, Go. 
 has the nomin. sg. tuggo, with a long final vowel irresistibly 
 pointing to a Preg. *tui\ym (tongue), inasmuch as the o, if 
 uncovered, would have been shortened. 
 
 1. Stems in -on-. — A. Nomin. sg. in -o changed to -a, msc. 
 nouns: Sk. uks-dn- ("bull," nomin. uks-d, gen. uks-n-ds), Go. 
 nomin. anhs-a and gen. aiihs-in-s, O.E. ox-a and E. ox (pi. ox-en), 
 O.H.G. ohs-o and G. ochs-e ochs (gen. ochs-en); so also, Go. han-a 
 (" cock," gen. han-in-s) and G. hahn, to which the L. correlative 
 would be *can-o (singer) with gen. *eo,n-6n-is ; Go. gard-a 
 (enclosure), gen. gard-in-s, cf. E. yard and G. gart-en, which 
 would be L. *hort-6 *hort-in-is (cf. hort-u-s), etc. 
 
 B. Nomin. sg. in -on changed to -o. — (a) With further exten- 
 sion of the long vowel to the oblique cases, feminine nouns : 2 
 Go. tugg-6 (gen. tugg-on-s), 3 O.E. tung-e and E. tongue, O.H.G. 
 zung-a and G. zung-e (pi. zung-en) ; Go. azg-o (gen. azg-on-s), 
 O.E. cesc-e and E. ash, O.H.G. asc-a and G. asch-e, etc. — (b) 
 Without this extension, neuter nouns (very few, but important) : 
 Go. dug-o (gen. dug-in-s), O.E. eag-e (gen. eag-an) and E. eye, 
 O.H.G. oug-a and G. aug-e (pi. aug-en) ; likewise, Go. hairt-6, 
 E. heart and G. herz, Go. dus-o, E. ear and G. ohr. 
 
 2. Stems in -yon-. — A. Nomin. sg. in -yd changed to -ja, 
 
 1 Greek keeps it throughout, Latin never. 
 
 2 Thus, Germanic is seen to have made the most of the old distinction 
 between short and long vowel, by employing it subsidiarily to distinguish its 
 masculine from its feminine stems. 
 
 3 Sk. jihv-d and L. lingu-a>=*dingu-a appear without any nasal suffix; 
 but, owing to the fact that both suffixes Preg. -5 ( = I.-E. -a) and -On ( = I.-E. 
 -on) had assumed a feminine function, the latter, as being better preserved 
 and, in consequence, more characteristic, showed an early tendency to 
 replace the former. This may be deemed the first step in the expansion 
 of the weak declension. 
 
138 ENGLISH AND GEEMAN GEAMMAE. 
 
 masculine nouns : Go. arb-ja (" heir," gen. arb-jin-s), O.E. yrf-e, 
 replaced in E. by heir = 0.¥. hoir from L. heres, G. erb-e (pi. 
 erb-en) ; Go. frdu-ja (lord), cf. O.H.G. fro, the derivative base 
 of an adjective frohn, preserved, for instance, in frohn-dienst 
 (husbandry service). 
 
 B. Nomin. sg. in -yon changed to -jo, feminine nouns : Go. 
 arb-jo (" heiress," gen. arb-jon-s) ; likewise, G. frau (" lady," 
 pi. frau-en) = O.H.G. frouw-a, which presupposes a Go. *frdu-jo 
 fm. of the msc. frdu-ja above. 
 
 C. Nomin. sg. in -In (changed to -%), due to a transfer by 
 analogy, to the nominative, of the suffix-form which originally 
 belonged to the oblique cases : thus, a fm. abstract noun Preg. 
 *hauh-jon (height) had a regular gen. *hauh-in-az ; analogy 
 gave a nomin. *hauh-in, whence *hauh-i, Go. hduh-ei (gen. 
 hduh-ein-s), O.H.G. hoh-i and G. hbh-e 1 (pi. hoh-en) ; similarly, 
 Go. manag-ei (multitude), O.E. menig-o, O.H.G. menig-i and 
 G. meng-e, cf. G. marec/i=O.H.G. manag = ^l. many. The whole 
 German derivation of abstract nouns, as tief-e, lang-e, hiirz-e, 
 as opposed to lief, lang, hurz, is based on the addition of this 
 suffix, 2 which is further to be found in such Latin types as 
 leg-id (gen. leg-ion-is), ob-sid-io (siege), etc. 
 
 (75) VI. Stems with suffix -mo- (-ma), quite common : 
 I.-E. root gher (heat), whence an adj. *ghor-mo-s (warm), Sk. 
 ghar-md-s (Gr. Oep-fio-s), L. for-mu-s, Preg. *%war-md-s, Go. 
 *var-m-s, E. and G. war-m ; I.-E. root dhe (to place, to do), 
 whence a noun *dhd-mo-s, Gr. 6<a-ix6-<s (heap), Preg. *ftd-md-s 
 (business, judgment, jurisdiction), Go. do-m-s, O.E. dom and E. 
 doom -dom, O.H.G. tuom and G. (suffix) -thum; 3 I.-E. *bhudh-mo-s 
 (bottom, cf. Gr. irvd-firjv), O.E. botm and E. bottom, O.H.G. 
 bodam and G. boden ; similarly, E. strea-m and G. stro-m = I.-E. 
 *srow-mo-s, from a root sru srew (to flow) ; E. roo-m and G. 
 ra«-wi = I.-E. *ru-mo-s ; L. cul-mu-s (stubble), O.E. heal-m. 
 O.H.G., G. and E. hal-m, etc. 
 
 VII. Stems with suff. -men- (-mn-, -mon-), reproducing exactly 
 
 1 Observe the constant and regular metaphony. 
 
 2 English replaced it by another derivation, infra 90 (V. 2), but still pre- 
 serves it in such alternations as proud pride = O.E. prut pryte. 
 
 8 Cf. infra 109, II. 
 
PRIMITIVE DERIVATION. 
 
 139 
 
 the alternations of suff. -en- : compare Gr. rrvB-^v 7rv0-pci'-os, 
 Sai-fuov Sai-/xov-o<s, and (nt.) ovo-fia ( = *6vo-mn), Sk. nd-ma 
 ( = *«o-??i?i, gen. nd-mn-as), L. no-men no-min-is, etc. — Msc. Go. 
 6Zd-ma = Preg. *t>ld-mo (flower) =I.-E. nomin. sg. *bhld-md, from 
 a stem *bhlu-men- and a root bhlu (to bloom), O.E. blu-ma andE. 
 bloom, O.H.G. bluo-mo msc, and G. blu-me now fm. 1 Further 
 compare: O.H.Gr. msc. sd-mo and G. sa-me, to L. se-men (seed), 
 from root se (to sow, G\ sa-en) ; Go. nt. na-mo, O.H.G. msc. 
 na-mo and G. na-me (gen. na-men-s), E. no-mp, to L. nt. no-men, 
 etc. The double suff. -mri-to- in the L. nt. type co-gno-men-tu-m, 
 etc., is represented by G. leu-mund (renown) = *hliu-mun-d = 
 I.-E. *klew-mn-to-, from root itZ# (to hear), cf. Gr. kAv-to-s and 
 L. in-clu-tu-s (famous). 
 
 (76) VIII. Stems with suff. -ro- (-rd) and -lo- (-Id), -ri- 
 and -U-, -ru- and -Zu-. — Of these six liquid-classes, which have 
 been more or less confused with one another and have lost a 
 great number of their examples, only three can be recognised 
 in Germanic. 
 
 1. Stems in -ro- : we need only recall the words already 
 quoted, E. acre = G. acker (Gr. dy-po-s ), E. j/ear = G. jahr (Gr. 
 w-pcl, w-po-s), E. timber = G. zimmer — Preg. *tem-rd- (Gr. Sep-w, 
 "to build "), and add such adjectives 2 as E. sour and G. saner — 
 O.N". su-r-r = I.-E. *su-ro-s, Go. fag-r-s (apt, convenient), O.E. 
 /oegr-r and E. /air, cf. G. /e#-erc ("to sweep," originally "to 
 purify, to adapt"), etc. 
 
 2. Stems in -lo- : I.-E. root pu (to be foul, cf. Gr. ttv-6o), L. 
 j3W-s, etc.), whence adj. *pu-lo-s, Preg. *fu-ld-s, Go. fu-l-s, O.N. 
 /ft-W, O.E. fu-l and E. /omZ, O.H.G. fu-l and G. /awZ; I.-E. 
 root syu (to sew, Gr. Ka-crorvaj, L. sttd, etc.), whence a noun of 
 instrument I.-E. *syu-ld (awl), O.H.G. siu-la and G. saule. In 
 some cases the suffix is preceded by an s of doubtful origin : 
 L. dla (wing, arm-pit) = *axla = *ak-sld (cf. L. axilla "arm-pit"), 
 O.H.G. ah-sala and G. ach-sel (shoulder), O.E. eaxl lost in E. 
 Further, an accessory i, which was developed by the resonance 
 of the liquid, gave birth to a Germanic suffix -ila, which plays 
 
 1 With the n re-appearing in the Go. gen. bld-viins, and in the G. pi. 
 blu-mev. 
 
 2 Bemember the numerous Gr. adjectives in -pb-s. 
 
140 ENGLISH AND GERMAN GRAMMAR. 
 
 an important part in the diminutive derivation, as will be seen 
 below (no. 88 and 103). 
 
 3. Stems in -ru-: O.E. flo-r and E. floor, M.H.G. vluo-r and 
 G.flur, from a msc.-fm. Preg. *flo-ru-s = I.-'E. *pld-ru-s, root pld 
 (even, plane, cf. L. pld-nu-s) ; I.-E. nt. *ddk-ru (tear), Gr. Sa/c-pv, 
 L. lac-ru-ma, Go. tag-r, O.N". tdr = *tah-r, O.E. tear = *teah-r, 
 and E. tear, O.H.G. zahar and G. ziihre. 
 
 (77) IX. Stems with suff. -no- (-no), -ni- and -nu-. — 1. The 
 I.-E. suff. -no- is a common one, as also in Germanic : I.-E. root 
 pel (to fill), whence Sk. pur-nd-s and L. ple-nu-s, Preg. *ful-nd-s, 
 Go. fulls — *ful-n-s, E. full, G. voll; I.-E. root deyk (to show, 
 cf. Gr. 8eiK-vv-[u and L. dic-ere), whence a derivation *doyk-no-s 
 (sign), Preg. *taik-na-z, Go. tdik-n-s, O.E. tdc-n and E. tok-en, 
 O.H.G. zeihh-an and G. zeich-en (cf. the G. vb. zeig-en, "to 
 show ") ; but these scanty nominal survivals give but little 
 idea of the extent to which in one particular form this suffix 
 has become used. 
 
 As early as the I.-E. period, there existed a suff. -eno- or -ono-, 
 which, inasmuch as it could be declined in the three genders 
 (-ono-s -ond -ono-m), was directly added to the root by way 
 of forming certain stems of middle and passive participles. 
 Let us consider a root wert (to turn), Sk. vdrt-ati (he turns), 
 L. vert-ere: the pf. middle participle was *we-wrt-ono-s (Sk.vd-vrt- 
 dnd-s), with the threefold meaning "turning, turning himself, 
 turned." Now, the literal transliteration of this *we-wrt-ono-s, 
 •omitting the reduplication, is the Go. participle vaiirj>-an-s (be- 
 come), G. (ge-)word-en. Thus, nothing could be clearer than the 
 whole formation, both in English and German, of the so-called 
 past participles (in -en) of strong verbs. 1 The same suffix, again, 
 inasmuch as it was invariable in a neuter form -ono-m, was a 
 characteristic ending for neuter nouns of action : thus, from an 
 I.-E. root bher (to carry, <p€p-a), fer-o), an I.-E. word *bher-ono-m 
 (the act of carrying), Sk. bhdr-ana-m, Preg. *fter-ana-m, whence 
 Go. bair-an. Now, this bair-an, having lost its whole declension, 
 merely serves as an infinitive to the verb bair-a (I carry) =L. 
 
 1 It should be noticed that the suffix is accented and, in consequence, 
 
 duces the root. As to this process, the reader is referred to the study of 
 
 owel-gradation, as stated above and illustrated by this derivative, supra 45. 
 
PRIMITIVE DERIVATION. 141 
 
 fer-o; and so also with any Go. infinitive: steig-a (I go up), 
 steig-an (to go up), G. (ich) steig-e and steig-en, etc., etc. In 
 short, the ending -an, O.E. -an, dropped in M.E., 1 O.H.G. -an, 
 G. -en, is a survival from the formative suffix of an ancient 
 neuter noun, which signified " the act of doing " what was 
 meant hy the verbal root. Hence, this single suffix -ono- gives 
 us the key to two grammatical categories of great importance, 
 and, at the same time, we are taught that in Germanic, as also 
 in every other representative of the I.-E. family, 2 neither the 
 infinitive nor the past participle is a mood of the verb, 
 both being in reality nominal formations, respectively, a 
 noun denoting an action and a noun denoting a state. 
 
 2. Stems in -ni- : a few adjectives. I.-E. had a root shaw 
 (to look at, to spy), Lat. cav-ere, G. schau-en: hence proceeded 
 a word *skdw-ni-s (conspicuous, beautiful), O.H.G. sco-ni, repro- 
 duced without metaphony in G. schon, 3 with metaphony in O.E. 
 scy-ne and E. sheen, M.H.G. schoz-ne and G. schon. Compare 
 also: root kri (to discern, to purify), L. cri-bru-m (sieve), 
 whence a word *kroy-ni-s (pure), Preg. *hrai-ni-z, Go. hrdi-n-s, 
 G. rein.* 
 
 3. Stems in -nu- : I.-E. root su su (to bring forth), Sk. su-nu-s 
 (son, cf. Gr. vios= *<ru-io-s), Go. su-nu-s, E. son, G. sohn ; Sk. nt. 
 tf-na-m (blade of grass), Go. J>aur-nu-s, E. thorn, G. dorn msc. ; 
 L. nt. cor-nu, Go. haur-n, E. and G. horn nt. In the last two 
 instances, suff. -nu- is seen to alternate with suff. -no-. 
 
 (78) X. Stems with suff. -to- {-to), -ti- and -tu-. — Any suffix 
 with initial t will usually present itself in three Germanic 
 forms, according to a distinction stated above, 5 namely : t (after 
 s, etc.) ; i> (Grimm's law), and fir (Verner's law). Recalling 
 this once for all, and leaving it henceforth to the reader, we 
 need only observe that suff. -to- of past participles is regularly 
 accented in Greek and Sanskrit (-ta-), and so also, in Sanskrit, 
 with the suff. -ti- of nouns denoting an action. 
 
 1 Cf. supra 19, 2. 
 
 2 Henry, Gramm. of Gr. and Lat., 115 (5), 117 and 125. 
 8 The meaning is " fairly," whence " already." 
 
 4 E. clean ( = G. klein, supra 17, 6) is similar in meaning and formation, 
 but the root is unknown. 
 8 Cf . supra 53. 
 
14*2 ENGLISH AND GERMAN GRAMMAR. 
 
 1. Stems in -to-. — This very extensive formation essentially 
 consists of words of three genders, derived from a verbal root 
 and having the sense of passive participles l : Sk. root ric 
 (to leave), rik-td-s (left) ; Gr. Acwr-w, A«7r-rd-s ; L. linqu-o, 
 lic-tu-s, etc. In Germanic primary derivation, it is chiefly 
 represented by certain adjectives no longer connected with any 
 verb : thus, L. rec-tu-s (right) may still be recognised as the 
 participle of the vb. reg-ere (to guide), whereas Go. raih-t-s, E. 
 righ-t and G. rech-t now appear quite isolated. In the same 
 way a great many similar English and German adjectives were 
 originally participles of an obsolete verb, as : Go. kal-d-s (cf . 
 L. gel-u), E. col-d, G. kal-t; Go. vun-d-s (hurt) and G. wund, 
 Preg. fm. *wun-^bd (wound), E. wound, G. wun-de ; Go. ddu-Jj-s, 
 E. dead, G. todt; E. tigh-t = G. dich-t ; E. sligh-t = Gr. schlecli-t 
 (bad), the primitive sense being "flattened"; E. lou-d = Gr. 
 lau-t, already contrasted with Gr. kAu-to-s (heard of) ; E. un- 
 couth, the negative to O.E. cw-6"=G. kun-d (known), cf. Gr. 
 yva)-To-9, etc. 
 
 As a participial suffix, the primary element -to- occurs with 
 equal frequency : here we may recall though-t and ge-dach-t, 
 brough-t and ge-brach-t, 2 E. weep wep-t, sleep slep-t, lose los-t, etc. 
 But these are analogical formations : cf. G. ge-schlaf-en, ver- 
 lor-en. Originally primary -to- could scarcely resist the almost 
 overwhelming influence of suff:. -ono- = Germanic -en, which had 
 assumed the same meaning. It will be seen, on the other hand, 
 to have prevailed in weak verbs, that is to say, in secondary 
 derivation. 
 
 2. Stems in -ti-. — This suffix, also very common, formed 
 many nouns of action or of object, in gender almost con- 
 stantly feminine : Sk. ga-ti-s (march), Gr. #ct-Ti-s (speaking, 
 word), L. ves-ti-s (garment), etc. Both meaning and gender 
 are well preserved in Germanic : compare G. komm-en and 
 -kunf-t (with an / phonetically inserted, cf. Go. ga-qum-p-s 
 " meeting "), G. kenn-en and kun-st (art), the suff. here receiving 
 
 1 On the ordinal function of this suffix, cf . infra 124. 
 
 2 On the phonetic relation between think and thought, etc., cf. supra 24 
 and 53 C. 
 
PRIMITIVE DERIVATION. 
 
 143 
 
 an initial accretion, E. to do and dee-d, G. thu-n and tha-t, E. to 
 give and gif-t, G. geb-en and gif-t (poison, now msc), E. to see 
 and sigh-t, G. seh-en and sich-t, E. he may and migh-t, G. mog-en 
 and mach-t, G. denh-en (to think) and an-dach-t (piety), E. 
 flight = Gr.fluch-t, etc. 
 
 3. Stems in -tu-. — Rare in Greek, though very common in 
 Sk. and Latin (sta-tu-s, gus-tu-s, fliic-tu-s, and all the so-called 
 supines in -tu-m and -tu), these are likewise nouns denoting an 
 action, nearly synonymous with the former class, but strictly 
 masculine. They are still seen in Gothic: thus, from an I.-E. 
 root pld (Gr. ttAw-to-s " navigable "), an I.-E. stem *plo-tu-s, 
 Preg. *flo-b~u-s, Go. flo-du-s (flood). But, after the u had been 
 dropped in the later languages, these stems no longer remained 
 distinguishable from those in -ti-, so that both classes were 
 blended together: O.E. flbd and E. flood, O.H.G. fluot and G. 
 flut fm. ; similarly, Go. luf-tu-s (air), but O.E. lyft, which is 
 shown by its metaphony to be *luf-ti-, and G. fm. luft with 
 metaph. pi. liif-te. 
 
 (79) XI. Stems with suff. -t- and -nt-. — The former is 
 rare, though well marked in I.-E. *noq-t- (night), Gr. vv£ (gen. 
 vvk-t-o's), L. nox (gen. noc-t-is), Preg. *nah-t-, Go. nah-t-s, O.E. 
 neaht and E. night, O.H.G. naht and G. nacht. The latter is 
 the regular suffix for active participles, in which function it 
 will be illustrated by the secondary derivation. It seems to 
 have been kept as primary only in I.-E. *do-nt- red. *d-nt- 
 (tooth), Sk. dd-nt- d-at-, Gr. 68ois (gen. 686-vt-os), L. dens (gen. 
 d-ent-is), Preg. *ta-n)>- *t-un}>-, whence Go. tunfiu-s, 1 O.E. *ton<5 
 toft and E. tooth, O.H.G. zand zan and G. zahn. 
 
 XII. Stems with suff. -ter- (-tr-, -tor-). — This formation is 
 common in Sanskrit, Greek and Latin, and comprises a con- 
 siderable number of so-called nouns of agent, either primary or 
 secondary : Sk. da-tar- " giver," Gr. So-rfc Sw-rwp, L. da-tor. 
 In Germanic it has been supplanted by another derivation of 
 secondary origin, 2 the only survivals being the nouns of re- 
 
 1 The second m proceeds from the analogy of the ace. tun\w = L. dentem = 
 I.-E. *dnt-m. But this corruption is an exclusively Gothic one, the O.E. 
 plural being (metaph.) teff= *to\>- , iz = *ton\>-iz = *dnt-es. „ . ■ _ 
 
 2 Cf . infra 102. . --^t SE L I B R -c J 
 
 / UNI V ERS1 T "2 
 
 V OF . , , 
 
144 ENGLISH AND GEEMAN GKAMMAR. 
 
 lationship " father, mother, brother, daughter," which have 
 often occurred in our exposition. 
 
 XIII. Stems with suff. -tro- (-trd) and -tlo- (-tld), homo- 
 logous to the preceding stems, but essentially nouns of 
 instrument, very common : I.-E. root pa (to feed, L. pd-sco, 
 pd-bulu-m "food"), whence I.-E. *pd-tro-m, Preg. *fo-d~rd-m, O.N. 
 fo-dr, O.E. fodor and E. fodder, 1 O.H.G. fuotar and G-. f utter ; 
 noun of action, E. laugh-ter and O.H.Gh hlah-tar, whence G. 
 collective ge-ldch-ter ; E. slaugh-ter, etc. ; I.-E. root we (to spin, 
 to sew, cf. L. we- re, Gr. vrj-fia, G. ndh-en — O.H.G. wa-ew = Go. 
 *ne-jan), whence Preg. stem *ne-J>lo, Go. ne-J>la, O.E. ncedl and 
 E. needle, O.H.G. nddal and G. nadel; G. beil (hatchet) = O.H.G. 
 bl-hal, 2 from the I.-E. root bhid (to split), etc. 
 
 XIV. Stems with suff. -tero- (-terd). — This element origin- 
 ally indicated an alternation between two things : Sk. ka-tard-s 
 " which of both ? " Gr. iro-Tepo-s, L. u-ter, al-ter, etc. It is dead, 
 but still visible in : E. ei-ther and G. je-der ; E. nei-ther and G. 
 we-der = O.H.G. ni-we-dar; O.E. wi-b"er (against) , whence E. with, 
 G. wi-der (against) and wie-der (again) ; perhaps E. fur-ther 
 and G. vor-der, etc. 
 
 (80) XV. Stems with suff. -es- (-os) : neuter nouns. — This 
 is a most important class of words in Sanskrit, Greek and 
 Latin : thus, from an I.-E. root gen (to beget), we have Sk. 
 jan-as (birth, race), gen. jdn-as-as, Gr. ycv-os, gen. ycv€os = *y€v- 
 ecr-os, L. gen-us gen. generis = *gen-es-is ; and the like. Consider- 
 ing then a Sk. stem rdj-as (dark space) = Gr. cpeyS-os, its exact 
 correspondent is Go. rig-is (darkness), gen. riq-iz-is, pi. riq-iz-a, 
 etc. Now, in every other Germanic dialect, the final -is in the 
 nomin. sg. is dropped, so that this case is no longer distin- 
 guished by any sign ; but, in the other cases, and especially 
 in the plural, the medial syllable -iz- reappears in its rhotacized 
 form (-*r-)> and the reason becomes obvious why such G. 
 neuter nouns as chalb (calf), lamb (lamb), etc., are in the plural 
 number chalb-ir-(u) = Go. *kalb-iz-a, lemb-ir-(u), Mod. G. 
 kdlb-er, lamm-er, etc. An old formative suffix, being lost 
 
 1 The same suffix in the form -stro-, in O.E. foster (food), E. to foster. 
 
 2 Still in Bavarian beichl. 
 
PRIMITIVE DERIVATION. 145 
 
 everywhere but in the plural, has been changed to a plural 
 ending. 
 
 XVI. Stems with suff. -yes- (-is-, -yds-) : comparatives, 
 cf. L. mel-ius mel-ior (better) ; further, by adding the suff. -to- 
 to the reduced syllable, a suff. -isto-, characteristic of super- 
 latives, Gr. /Acy-toro-s (the greatest), etc. — This derivation is 
 of constant use for the formation of Germanic comparatives 
 and superlatives : Go. hduh-is-, in declension hduh-iz-a, O.H.G. 
 hoh-ir-o, E. high-er, and G. hoh-er, with the regular metaphony, 
 which survives in E. old eld-er eld-est l ; further, Go. hduh-ist-s, 
 E. high-est, G. hoh-est hbch-sb ; Go. bat-iz-a bat-ist-s (from a 
 positive which became obsolete in Preg., cf. Sk. bhad-rd-s 
 u happy "), E. better best (metaph.), G. besser beste ; E. last, cf. 
 the positive late, and G. letzt ; E. nigh and next = G. ndchst ; Go. 
 md-iz-a (cf. L. mag-is), O.H.G. mero and G. mehr, E. more, and 
 superlative E. most = G. meist. The reader will easily supply 
 further examples. 
 
 § 2. — Verbal Stems. 
 
 (81) I. Root-Stems. — Primitive verbal derivation, like 
 nominal derivation, had its root-stems ; that is to say, many 
 bare roots could be conjugated as such, without any suffix, and 
 by mere addition of the person-endings to the radical syllable. 
 The well-known Greek verbs in -fu clearly illustrate this simple 
 conjugational system, which has been quite lost in Germanic, 
 apart from the conjugation of the vb. " to be," and from the 
 formation of the perfects of strong verbs, both of which will be 
 found in their proper places. 2 The root was often preceded by 
 a reduplication, Gr. oY-8o>-/ai, ri-Ov-fii, and this element is still 
 visible in G. beb-en (to tremble) = Sk. bi-bM-mi (I am afraid), 
 from a root bhl (to fear) ; but this verb, no less than any 
 other, is now conjugated like the following classes. Hence 
 Germanic may be said no longer to possess any primitive verbs, 
 but the so-called verbs in -w, which correspond to the third 
 Latin conjugation. 
 
 1 In a special sense, whereas " I am older than you," after the analogy of 
 old. 
 
 * Infra 216-217, 175-184, and 208-209. 
 
 L 
 
146 ENGLISH AND GERMAN GRAMMAR. 
 
 (82) II. Stems with suff. -0- (-e-). — In every verbal suffix 
 which ends with this characteristic short vowel, the normal and 
 the deflected grade are substituted for one another according 
 to a regular alternation which has become slightly altered in 
 Latin, but was faithfully kept throughout Greek and Germanic : 
 in the first person sg. and pi., and in the third pi., the vowel is 
 6 ; it is e in the three other persons, without exception in 
 either case. One instance will suffice, namely, the present 
 indicative of the vb. Gr. vc/a-m (I divide) = Go. nim-a (I take) : 
 
 vc/x-o) vt/i-et-s vc/x-ei vefx-o-jxev 1 vc/a-c-tc ve/i-o-rri 
 nim-a nim-i-s nim-i-J* nim-a-m nim-i-p nim-a-nd* 
 
 This point being understood, the stems with suff. -0- assume 
 three distinct aspects. 
 
 1. The radical syllable is in the normal grade (by far the 
 most common case): I.-E. root ed (to eat), Gr. «?8-u>, L. ed-5, Go. 
 it-an, E. to eat, G. ess-en ; I.-E. root bher (to carry), stem 
 *bher-o (I carry), Gr. <f>ep-w, L. fer-o, Go. bair-an (to carry) and 
 ga-bair-an (to bring forth), O.E. ber-an, ge-ber-an, and E. to 
 bear, O.H.G. gi-ber-an and G. ge-bar-en ; I.-E. root bhid (to 
 split), stem *bheyd-o (I split), Go. beit-an (to bite), E. to bite 
 and G. beisz-en ; I.-E. root der (to tear), Gr. Scp-w, Go. ga-tair-an, 
 E. to tear, G. zerr-en and ver-zehr-en; I.-E. root bhendh (to bind), 
 Go. bind-an, E. to bind, G. bind-en ; I.-E. root steygh stigh (to 
 go up), Gr. <rr€tx-o)> Go. steig-an, E. to sty, G. steigen; I.-E. root 
 leyq (to leave), Gr. AaV-a), L. lingu-o, Go. leihv-an (to lend), 
 O.E. leon (cf. E. loan and to lend), G. leih-en; I.-E. root duk 
 dewk (to lead), L. duc-o = *deuc-o, Go. tiuh-an (to draw), G. 
 zieh-en; I.-E. root giis gews (to taste, to appreciate, to select), 
 Gr. ycvia = *y€v<r-u>, Go. kius-an, E. £0 choose, G. kies-en, etc. 
 
 2. The radical syllable is in the reduced grade, because the 
 original accent was shifted to the suffix : 3 I.-E. root skabh (to 
 
 1 Here Latin has the e : leg-i-mus*=*leg-e-mus, instead of *leg-o-mus = Gr. 
 \iy-o-fiev. 
 
 2 The identity of these stems with the Greek and Latin verbs being thus 
 put beyond question, it will be convenient to quote them hereafter in the 
 shape of the infinitive (supi-a 77, 1), viz. Go. nim-an, G. nehm-en, etc. 
 
 8 Observe the contrast, in Greek accentuation and vocalism, between the 
 types tpefryeiv and Qvytiv, Xehreti' and \nreu>, etc. 
 
PKIMITIVE DEKIVATION. 147 
 
 scratch, to scrape), L. scab-o, Go. sJcab-an, E. to shave, G. schab- 
 en ; I.-E. root bhdg bhdg (to cook), Gr. <£<oy-a>, but O.E. bac-an 
 and E. to bake, O.H.G-. bahh-an and G. back-en (pf. buck) ; simi- 
 larly E. to take, pf. took, borrowed from Scandinavian; I.-E. 
 root md weyd (to know), Go. vit-an, G. wiss-en; the radical 
 grade is doubtful in Go. faran, E. to fare, G. fahren, contrasted 
 with the Gr. root irtp 7rop (to pass over). 
 
 3. The radical syllable becomes nasalized (compare Gr. 
 \afx.(3-dveiv and \afi-tLv, L. find-o linqu-o above from roots bhid 
 Uq, etc.) : I.-E. root pdg p&k, Gr. iray-r; (machinery, trap), L. 
 pang-6 (to drive in), G. fang-en (to catch), cf. the root without 
 nasal infach (compartment) ; I.-E. root siq siitq (to pour), Sk. 
 sinc-d-mi (I pour out), sik-td-s (shed), Go. sigk-an, E. to sink, 
 G. sink-en ; Gr. o-Trtpx-o-fAcu (I ran), E. to spring, G. spring-en ; 
 E. to cleave (to cling) and G. kleb-en, but, with an inner nasal, 
 E. to climb and G. klimm-en, etc. 
 
 (83) III. Stems with suff. -yd- (-ye-). — This class is no 
 less considerable than the preceding, 1 and admits of a similar 
 division. 
 
 1. The root is in the reduced grade (the more frequent case) : 
 I.-E. root stig (to sting), Gr. o-ti£o> = *o-TLy-yio (cf. fat. crri£a) = 
 *<rTiy-(ru>), E. to stitch and G. stick-en; I.-E. root svnd (to sweat), 
 Sk. svid-yd-mi, Gr. iS-iw, O.H.G. swizz-en and G. schwitz-en; 
 I.-E. root kap (to seize), L. cap-id, Go. haf-jan, O.E. hebban = 
 *heb-jan, and E. to heave, O.H.G. heffan and G. heben ; I.-E. root 
 bhidh bhgydh (to persuade), Gr. iretd-<i) = *<f>etd-<i), but Go. bid-jan 
 (to pray), E. to bid, G. bitt-en; Go. Jnigkjan (to seem), E. fo 
 think, G. dunken; Go. bug-jan, E. £0 6wi/, etc. 
 
 2. The root is in the normal grade : I.-E. root sed (to sit), 
 Gr. !£o/xcu = *crcS-t/o-/jtai (cf. L. sed-ere), O.N. sit-ja, E. £0 si£, G. 
 siizew ; I.-E. root Ze#7i (to lie, cf. Gr. Ac'^-os, L. lec-tu-s " bed "), 
 O.N. ligg-ja, O.E. licg-an and E. £0 lie, G. liegen. There are 
 probably two different grades of root werg (to act, to work, cf. 
 Gr. fipy-o-v " work ") to be recognised in E. to work = Go. vaurk- 
 jan = I.-E. *wrg-yo, and G. wirken = Preg. *werk-yo. 
 
 IV. Stems with saff. -eyo- (-eye-) : causative verbs. — It 
 
 1 Even still more, if we take into account the important part it plays in 
 secondary derivation, infra 92-93 and 106-107. 
 
148 ENGLISH AND GERMAN GRAMMAR. 
 
 seems advisable to place this class immediately after the -yd- 
 class, for the sake of the contrast between them in snch notable 
 pairs as to lay and legen, to set and setzen, as opposed to to lie 
 and liegen, to sit and sitzen. As early as the I.-E. period, a suffix 
 -eyo-, added to the deflected root, formed a great many cansative 
 verbs: thus, from root bher, a verb *bhor-eyo (I cause to carry), 
 Sk. bhdr-dyd-mi, Gr. <f)op-€w (I carry). 1 Similarly, from root 
 legh above, an I.-E. vb. *logh-eyo, which would be Gr. *Aox-€a>, 
 meant " I cause to lie down." Now, I.-E. *logh-eyo, becomes Go. 
 lag-ja (I lay), the suffix here being no longer distinguishable 
 from an original suffix -yo-, whereas the radical grade remains 
 clear and characteristic. Further, lag-jan, with the regular 
 metaphony, results in O.E. lecgan and E. to lay, O.H.G. lecken 
 legen and G. legen. In the same way, to set and setzen should 
 be traced back to Go. satjan = I.-E. *sod-eyo (I cause to sit 
 down) ; and the process may be further illustrated by E. to fell 
 and G. fallen, O.E. feran and G. fuhren (to lead) = *for-jan, O.E. 
 hangian 2 and G. hangen, E. to raise = Go. (ur-)rdis-jan, G. lehren 
 (to teach) = Go. Idis-jan, etc., in contrast with E. to fall and G. 
 fallen, E. to fare and G. fdhren, E. to hang and G. hangen, E. 
 to rise = Go. (ur-)reis-an, G. ler-n-en, etc. 3 
 
 (84) V. Stems with suff. -nd- (-ne-) : Gr. m-zLv (to drink) and 
 pres. TTL-vd), L. ll-nb (I besmear), supine li-tu-m ; I.-E. root ski 
 (to shine, to appear, cf. Gr. cna-a "shadow, image"), Go. skei- 
 nan, E. to shi-ne, G. schei-nen ; I.-E. root gno gn (to know, cf. 
 Gr. yi-yvw-o-Ku), L. (g)no-sco, etc.), Go. kun-nan, O.H.G. chun-nan 
 and G. (metaph.) kbn-nen 4 ; Gr. <r<£aA-Ao> (I throw) and <r<f>d\- 
 
 1 Compare : root men (to think) and L. mon-eo (I cause to think, I warn, 
 G. mahn-en) ; root nik (to perish), L. nex " death," and L. noc-eo (I harm). 
 
 2 We should expect E. to *heng, cf. G. henk-er (hangman); but the intran- 
 sitive to hang acquired the active meaning. 
 
 8 Since, moreover, it is natural that a causative verb should govern both 
 the accusative of its direct object and the accusative of the person caused to 
 act, we may understand at once why G. lehr-en and the like, as well as L. 
 doc-eo, require a double accusative. This rule was far better observed in 
 M.H.G. than it is now, such regular expressions as ich verhehle ihn die sache 
 having been replaced by ich verhehle ihm, etc., after the analogy of verbs 
 governing the person in the dative case (sagen, geben). 
 
 4 E. can has the vocalism of the pf. can = G. kann, infra 224. E. ken= 
 G. kennen is the causative, Go. kannjan ; several German provinces are 
 known to confuse kennen and konnen. E. to know = O.E. cnaw-an is the 
 verb formed by suffix from the radical form gno. 
 
PRIMITIVE DERIVATION. 149' 
 
 ko-fxai (I fall), L. fal-lo (I betray) = */aZ-no, Go. *fal-lan = *fal- 
 nan, O.'N.falla, E. to fall, G. fallen, etc. 
 
 VI. Stems with suff. -to-, -dho- and -do- : Gr. ve-w and vrj-Ou 
 (I spin), L. ten-do, etc. ; I.-E. root gel (to behave, to manage, 
 Gr. 7r€'A-o-/Aai "I find myself," L. col-5 = *quel-o "I cultivate," 
 etc.), whence a vb. *gl-to, Go. hal-dan (to hold), E. to hold, G. 
 halten x ; I.-E. root pZw ^Zew (to flow, cf . Gr. 7rAe'f-u) irAev-o-co " to 
 ship "), whence a vb. *plew-do, which would be Go. *fiiu-tan, 
 O.Hj.fleotan and E. to fleet, O.H.G. flio^an and G.flieszen ; simi- 
 larly, Gr. \zf-w (to pour), L. fu-ndo, Go. giu-tan, G. gieszen, 
 etc. 
 
 VII. Stems with suff. -sk8- : Gr. /Ja-cr/cw, L. cre-sco, etc. ; 
 I.-E. root w#d iid (water), Preg. *wat-sko (I wash), E. to wash, 
 G. waschen; I.-E. root jpr£& (to ask, to pray, cf. L. prec-o-r), 
 whence a vb. *prk-sko, L. posed (I ask) = *porc-sco, Germanic 
 *forh-skan, but O.H.G. forscon and G. forschen, this verb having 
 afterwards assumed a form of secondary derivation, 2 as is the 
 case with many others of the same class, viz. E. to wish = O.E. 
 wyscean, and G. wiinschen = O.H.G. wunsken, E. to ask = OM. 
 dscian, and G. heischen (with inorganic h) — O.H.G. eiscon, the 
 latter perhaps related to L. cerusco (to beg), etc. 
 
 (85) VIII. Stems with suff. -e-. — The primitive I.-E. language 
 probably possessed a number of verbal stems with reduced 
 root-form and invariable suffix -e-, with a general intransitive 
 meaning as opposed to some transitive verbs belonging to the 
 preceding classes : thus, Gr. e-Twr-i|-s (thou wast struck), but 
 TV7i--Tei-s (thou strikest) ; L. jac-e-s (thou liest), 3 but jac-i-s 
 (thou throwst), etc. Similarly, we find in Germanic : O.H.G. 
 leb-e-n, G. leben, E. to live, contrasted with a root lip shown by 
 Gr. \.l-7r-apri<s (clinging to), cf. the compound bleiben (to remain) 
 = Go. bi-leib-an ; O.H.G. hab-e-n, G. haben, E. to have, which 
 would have been I.-E. *k&p-e-, from the root that produced 
 L. cap-id and G. hebenf etc. Very early, however, stems with 
 
 1 Of course the suffix might equally well be -dho-, with the accent 
 optionally on the radical or the suffix-syllable. 
 
 2 Cf. infra 92, 93 and 174. 
 
 8 The original meaning must have been " thou art cast down." 
 
 4 An etymology which would sever E. to have**Q. haben from L. Jiabere. 
 
150 ENGLISH AND GERMAN GRAMMAR. 
 
 final long vowel were conjugated in Germanic, by adding, at 
 least in some of their personal forms, the suff. -ja- = I.-"E. -yo-, 
 to the vowel : just in the same way as we have in Latin a first 
 person jaceo from jaces, formed after the analogy of moned 
 mones, though the two words originally belong to quite different 
 verb-classes, so Gothic corrupted the conjugation of the type 
 *hab-e-, to such a degree as no longer to exhibit its characteristic 
 e ; and, though this e is still extant in O.H.G., later languages 
 blend these primitive formations with the much more recent 
 verbs which have arisen from secondary derivation. This 
 point here need only be hinted at, as it will appear more 
 clearly in the further development of the system of verbal 
 derivation built on the foundations thus briefly laid. 
 
 Section II. 
 
 SECONDARY SUFFIXES. 
 
 (86) We must not expect to meet, in the secondary derivation, 
 with every one of the suffixes we have studied in the primary 
 formation ; on the other hand, we may meet with some new 
 suffixes. Some of these appear to be of a later origin, having 
 spread out at a time when the language had already created 
 nearly all its primary stems. Further, the suffix-accumulations, 
 being longer and more cumbrous than mere syllables, were in 
 less danger from the action of phonetic laws, and therefore had 
 a far greater power of duration and multiplication. Lastly, a 
 lawless and capricious analogy encumbered the language with 
 a swarm of neologisms and useless synonyms, which a clearer 
 and more sober speech was fain, in some way or other, to get 
 rid of. It is thus, that Germanic derivation, though extremely 
 rich, may be reduced, upon analysis, to a very small number of 
 primitive elements. 
 
 § 1. Nominal Stems. 
 
 (87) Owing to the reason just given, suffixes consisting of a 
 simple vowel, including also the suff. -yo-, which plays such a 
 momentous part in the derivative system of Sanskrit, Greek 
 and Latin (type pitr-ya-s ira.Tp-io-% patr-iu-s), barely occur in 
 
PRIMITIVE DERIVATION. 
 
 151 
 
 Germanic secondary stems. 1 A feminine snffix, however, 
 characterized by a long vowel, -I, and so rare in primary 
 derivation that it has not yet required onr attention, here occu- 
 pies a prominent place. 
 
 I. The I.-E. suff. -I has the special function of forming the 
 feminine of such stems as do not end with the suffix- vowel -o- 2 : 
 Sk. svdd-u-s (sweet), fm. svdd-v-i, Gr. ^8-v-s r)8-e-la ; Sk. bhdr- 
 a-nt- (carrying), fm. bhdr-a-nt-i, Gr. (pepwv <pe'pouo-a, 3 etc. But 
 this simple process, though extant in Gothic, 4 has not been re- 
 tained by its sister-tongues, being in these illustrated by a 
 single survival, which moreover is lost in English and corrupted 
 in German: G. nicht-e (niece) = I.-E. *nept-i, derived from 
 *nepot- (nephew). A suffix-accumulation, however, has pre- 
 served down to the present time the old I.-E. feminine 
 formation. 
 
 The -I, being occasionally added to a stem in -en-, 5 produced 
 a combination -en-l, which Germanic changed to -in-i. Thus a 
 derivative in -en- (-on-), from a root duk (to lead), was I.-E. 
 *duk-en- (leader), which in Latin would be nomin. sg. *duco 
 and appears strictly reproduced in O.H.G. heri-zog-o (chief of 
 an army, duke). Now, to this, the regular fm. (I.-E. *diik-en-l) 
 is naturally heri-zog-in-i, wherein the syllable -in- belongs to 
 the msc. stem. But, since this syllable no longer survives in 
 the nomin. msc. heri-zog-o, the speaker was led in thought to 
 separate from heri-zog-in-i the ending -ini, mistaking it for the 
 characteristic symbol of the feminine gender, and henceforth 
 to transfer it as such to new formations, as gottin from gott, 
 fiichs-in (O.E. fyx-en and E. vix-en) from fuchs, etc. 
 
 At the end of the O.H.G. period, a further addition changed 
 this ending to -inna, M.H.G. -inne, to which corresponds the 
 modern spelling -inn, now obsolete in the singular. 6 
 
 1 Though we have E. shep-herd (*shep=sheep) and G. hirt-e = O.H.G. 
 hirt-i^Go. haird-ei-s, from, as it were, a Preg. *her$-ja-z "which belongs 
 to the herd." The same suffix will occur below in a neuter collective G. 
 formation (96, 1.). 
 
 8 And even of these concurrently with -a, supra 73. 
 
 8 On Gr. -ta = -i, see my Grammar of Gr. and Lat., no. 112. 
 
 * Go. mag-us (boy), ma-v-i (lass), frij-ond-i (female friend), etc. 
 
 5 Cf . supra 74, and never forget that all parts of the derivative system are 
 intimately connected with one another. 
 
 8 English has lost this ending and replaced it by -ess borrowed from 
 
152 ENGLISH AND GERMAN GRAMMAR. 
 
 II. The same suffix -en-, which, by combining with the end- 
 ing -I, endowed German with such a remarkable element of 
 feminine derivation, also underwent, even when isolated, a 
 peculiar process of analogical extension : for it will be seen 
 that a great many nouns, which did not originally belong to 
 the so-called weak declension, passed to it by later analogy, 
 either in the plural, or in both numbers, and that the whole of 
 the weak declension of adjectives is also due to a blundering 
 use of this element -en-, which, though a derivative suffix, was 
 mistaken for a grammatical ending. 1 
 
 (88) III. The suff. -ro- was not productive ; but the corre- 
 sponding -lo- multiplied, especially as a diminutive. When 
 added to primary stems in -e-/-o-, it gives -Sid-, -old-, which in 
 Greek characterises some derived adjectives (ev-Tpa7r-e-Ao-s 
 "buxom," e?K-e-Ao-s "alike "), whereas in Latin it has a de- 
 cidedly diminutive meaning (parv-o-lu-s " tiny," agellus = *ag-ro- 
 los "little field "). These forms would be represented in Ger- 
 manic by -ila- and -ala-, which are actually exhibited by a 
 great many derivatives, viz. : ordinary adjectives, E. crijyp-le 
 and G. kriipp-el, cf. Gr. ypv7r-o-s (crump) ; names of animals, 
 E. weas-el and G. wies-el, cf. G. wies-e (meadow) ; nouns denot- 
 ing instruments, E. shov-el = G. schauf-el = 0.'H..Gc. scuv-ala, E. 
 spind-le (with inserted d) = G. spind-el = O.H.G. spinn-ila, G. 
 meiszel (chisel) = O.H.G. meij-il ; lastly and chiefly, diminutives, 
 with a regular metaphony pointing to the suffix-form -ila, 
 E. litt-le = O.E. lyt-el, and G. liitz-el 2 = O.H.G. luzz-il, O.H.G. 
 tur-ila (little door) and G. thur-el (valve), O.E. cyrn-el and E. 
 kem-el. s 
 
 (89) IV. The suff. -no- has spread widely. 
 
 1. Form -ono-s, Germanic -an-s : after the analogy of radical 
 verbs, those with suff. -to- and -no- form their past participle 
 
 French. In German, on the contrary, it is so common a feminine deriva- 
 tive form as to have been even added to feminine nouns which were bor- 
 rowed ready made from foreign languages: F. princ-esse, but G. prinz-ess-in ; 
 M.H.G. ebbet-isse (abbess) = L. abbatissa, but G. abt-iss-in. 
 1 Cf. infra 140 sq., 149, 150 and 156. 
 
 * Preserved in proper names : Liitzel-burg, Lutzel-ttein " Littlestone." 
 8 Further : E. thimb-le, " little finger," from thumb ; G. arm-el (sleeve), 
 " little arm." E. hill and G. hiigel have already occurred. 
 
PRIMITIVE DERIVATION. 153 
 
 according to the strong conjugation, namely : Go. -an-s, etc. ; 
 E. held, G. ge-halt-en ; E. fall-en, G. ge-fall-en ; even G. ge- 
 wasch-en, though the other verbs in -sko- belong to the weak 
 conjugation ; a few verbs in -yo- followed the same method, E. 
 sat, G. ge-sess-en ; the other verbal derivatives, chiefly those in 
 -eyo- and -e- (-eyo-), as well as those of secondary derivation, 
 admit only of the dental suffix, E. -d or -t — G. -et or -t. * 
 
 2. Form -ono-m, Germanic -an : verbs of every class have 
 been seen above to form the infinitive by adding the nasal 
 suffix, Go. -an, etc., to the present stem. 
 
 3. Form -i-no-, due to the combination of certain stems in -i- 
 with the adjectival suffix -no-, Germanic -ina-z, then -ins. — 
 This derivation is to be traced back to I.-E., in L. su-inu-s 
 (belonging to swine), Go. sv-ein-s (swine), E. swine, G. schwein. 
 It will also be found below in the possessive adjectives. 2 It is 
 common in Gothic in the formation of adjectives denoting 
 material : gulj>-ein-s (golden), airfi-ein-s (made of earth), 
 stdin-ein-s (of stone), etc. Phonetic reduction changed -in- to 
 -in-, whence the regular metaphony in the preceding syllable, 
 and further to -en : O.E. gyld-en, which has again become E. 
 golden, and G. gulden, now obsolete and replaced by gold-en ; 
 G. ird-en (made of earth), etc., etc. This -in-, when added to 
 astern in -es-, as in I.-E. *reg-es-ino-s = Go. riq-iz-ein-s (dark), 3 
 resulted in West Germanic in an ending -eren, which is seen in 
 such German words as stein-em (of stone), holz-ern (wooden), 
 etc. 
 
 (90) V. The suffixes consisting of ai and a vowel are kept 
 in one of the two languages, and occasionally in both. 
 
 1. Suff. -to- and -e-to-, whence in Germanic a dental conso- 
 nant, either alone, or preceded by an 1 which very early 
 assumed the duller sound of e. 
 
 (a) In the formation of the past participles of derived 
 verbs : Gr. vc/x-c-to-s (divided), L. gen-i-tu-s (born), dom-i-tu-s 
 (tamed) ; likewise, Go. ga-tam-i-da, from vb. tam-jan, E. 
 tam-e-d, G. gi-zem-i-t ge-zahm-t, etc., etc. 
 
 (b) In the formation of adjectives proceeding from nouns, 
 
 1 Infra 90, 173 sq. and 186 sq. * Infra 168. 
 
 8 From riq-is (obscurity), supra 80. 
 
164 ENGLISH AND GERMAN GRAMMAR. 
 
 as L. cord-d-tu-s (noble-hearted), crin-i-tu-s (hairy): Go. un- 
 qen-i-J>s (having no wife), O.N. hcer-&-r (hairy), O.H.G. gi- 
 stim-o-t (full of stars), M.H.G. ge-jdr-e-t (aged), G. be-jahr-t, 
 especially in E. compounds, such as fair-hair-ed, 'red-nos-ed, 
 hump-back-ed, etc. 
 
 (c) In the derivation of ordinal from cardinal numbers. 
 
 2. The fm. suff. -td, usually -i-td, serves to form abstract 
 nouns from adjectives: L. juven-ta (youth), Go. jun-da, O.E. 
 geogoft and E. youth, O.H.G. jugun-d and G.jugen-d ; Go. hduh-s, 
 whence hauh-i-fca (height), O.H.G. hoh-i-da ; G. heil (whole, 
 sound), whence heil-i-da (health), etc. This mode of deriva- 
 tion has become extinct in German, but still exists and has 
 been much developed in English, where it accounts for the 
 metaphonical contrast in such types as high heigh-t, whole 
 heal-th, broad bread-th, long leng-th, strong streng-th, foul fil-th, 
 merry mir-th, as well as for the shortening in wide wid-th, deep 
 dep-th, and, in short, for a considerable number of abstract 
 nouns. 
 
 3. The two suffixes -ti- and -tu-, after a long vowel, like the 
 Latin type sen-d-tu-s, may be recognised in such secondary 
 derivatives as: G. arm-ut (poverty) = O.H.G. aram-uoti ; G. 
 heim-at (fatherland) = O.H.G. heim-uoti ; G. ein-ode (solitude) 
 = O.H.G. ein-oti 1 ; O.E. mon-atS and E. mon-th, O.H.G. mdn-od 
 and G. mon-at, from stem mon- (moon). 
 
 VI. The suffix -nt-, which characterizes the words friend = 
 freund and fiend=feind, when added to the stem- vowel of an 
 ordinary verb, forms with it a combination, I.-E. -o-nt-, Go. 
 •a-nd-, G. -e-nd-, lost in E., the constant and regular ending of 
 present participles in all verbs, whether primitive or deriva- 
 tive : Go. nim-and-s bair-and-s, G. nehm-end ge-bar-end ; O.E. 
 Hcelend and G. Heiland "the Saviour," cf. G. heil-en = 'E. to 
 heal, the causative derived from G. heil=~E>. whole. 
 
 VII. The suffixes -tro- and -tero- find scarcely any applica- 
 tion in secondary derivation, 2 though we are probably right in 
 
 1 Primitively unconnected with Me (desert) = O.H.G. ot-i = Go. <£i4>-* = O.N. 
 auff-r (void), though corrupted under this influence. 
 
 2 The nouns of instrument now adopt the suffix of nouns of agent, infra 
 102 : E. a bore-r, a decant-er ; G. lauf-er " the runner " (mill-stone). 
 
PRIMITIVE DERIVATION. 155 
 
 tracing back to one of them, at least indirectly, the O.E. ending 
 -stre, which formed a great many feminine nonns, and still sur- 
 vives in E. youngster and spinster. 1 
 
 VIII. The suffix -es- of neuter nouns has been seen not to 
 survive, except as a grammatical ending for the plural of these 
 nouns, in the form -er, which belongs to the study of declen- 
 sion. 
 
 IX. The comparative (-yos-) and superlative (-isto-) suffixes, 
 in their modern form respectively -er and -est (-st), and with 
 the regular metaphony which almost always accompanies them 
 in German, 2 remain in both languages the specific and constant 
 form of comparison for all adjectives, primitive or derivative. 
 English, however, in modern and, especially, in learned adjecti- 
 val formations, preferred the periphrastic locution by more and 
 most, borrowed, no doubt, from French syntax, whereas German 
 adheres to the traditional formation, thus : merkwiirdig-er, wis- 
 senschaftlich-er, dynamisch-er. 
 
 (91) X. The I.-E. suff. -ho- is extremely rare in the primary 
 derivation of any I.-E. language, whilst, in all (Gr. <£i>o-i-*d-s 
 hnr-i-K.6-<i, L. civi-cu-s Hispdn-i-cu-s, etc.), it attains a prominent 
 place in secondary formations : a remark which holds good for 
 Germanic as for any other member of the family. 
 
 1. When added to a stem in -i-, the suffix becomes either 
 -i-ko- or -l-ho-. It is chiefly the latter, changed to Preg. -tja-, 
 that has survived and extended in adjectival derivation : Go. 
 gab-eig-s (open-handed) and G. (frei-)geb-ig ; Go. maht-eig-s, 
 O.E. meaht-ig and E. might-y (hand-y, filth-y, swarth-y, etc.), 
 O.H.G. maht-lg and G. macht-ig ; O.H.G. r'eht, whence a deriva- 
 tive riht-ig, now richt-ig (right, exact), etc. 
 
 2. The same suffix, when added to a stem in -en-, formed 
 with the stem-syllable a combination -eilko- or -7&fed-, Preg. 
 either -wlga- or -ufa^a- (O.H.G. arm-ing "poor man," edil-ing 
 
 1 Literally " a spinning person," whence " an unmarried damsel." 
 
 2 More accurately, Gothic has two comparative suffixes, rather arbitrarily 
 distributed, namely, -iz- the reduced grade of I.-E. -yos-, and -oz- of doubt- 
 ful origin. Thus, the German metaphonical and unmetaphonised types are 
 likely to correspond respectively to the former and the latter form. Further 
 particulars must be looked for in ordinary grammars. 
 
156 ENGLISH AND GEEMAN GEAMMAE. 
 
 " noble man," werd-unga " dignity," from werd " worth "), the 
 various forms and uses of which will appear below in the 
 modern languages. 1 
 
 3. In the form -isko-, the guttural suffix occurs in a number 
 of, mainly diminutive, Greek secondary derivatives, such as 
 ttciiS-io-ko-s, vcdv-iV/co-s, do-mS-io-xio-v, and in one of the most 
 widely extended forms of adjectival derivation in Germanic : 
 Go. piud-isk-s (popular), representing L. gentilis, O.H.G. 
 diut-isc and G. Deut-sch, etc. ; O.H.G. ird-isc and G. ird-isch 
 (earthly) ; O.E. engl-isc, wlel-isc, and E. English, Welsh, French, 
 etc. 
 
 § 2. Verbal Stems. 
 
 (92) The chief factor in I.-E. secondary verbal derivation 
 is the suffix -yo-, by means of which, from almost any nominal 
 stem, may be formed a corresponding verbal stem, a process 
 clearly illustrated by Sanskrit grammar, e.g. devd-s (god), 
 dev-a-ya-ti (he adores) ; go-pd-s (cowherd), go-pd-ya-ti (he 
 guards), etc. This suffix becomes rather more obscure in 
 Greek: Imr-cw-s (rider), linr-ev-Qi (I ride) =*i7r7r-cv-i/u) ; Tl-firj 
 (honour) = Tt-/xd, and rZ-fid-o) (I honour); </>tA.-o-s (dear), and 
 <f>iX-e-<D (I love) ; St}-A.o-s (clear), and S^-Ao'-w (I show), etc. It 
 is still less clear in Latin, owing to the regular contraction : 
 for-ma (shape), and formo (I shape) — *for-ma-yo. In Germanic 
 we must expect to find it equally obscured : in fact, it could 
 no longer be discerned without the help of linguistic com- 
 parison ; by this light, however, we are able to perceive that a 
 Go. vb. salbo (I salve, G. [ich~\ salbe) must have been formed 
 from a fm. noun salba (salve), 2 just in the same way as rlfxdwi 
 and formo were respectively derived from rlfia. and forma, or, in 
 other words, that an I.-E. fm. noun *solp-d, changed to Preg. 
 *salb-d, produced the verb by adding the formative suffix -yd-, 
 changed to Preg. -ja-, to which the person-endings were added 
 in conjugation. 
 
 But here a previous observation is necessary. We have 
 
 1 Infra 103. 
 
 8 Not Gothic, but found everywhere else : G. salb-e (salve). 
 
PRIMITIVE DERIVATION. 157 
 
 already seen that Germanic often added this suffix -ja- to 
 verbal stems originally ending in a long vowel, where there 
 was no reason at all for this accretion. 1 On the other hand, 
 the reverse also happened : even as early as I.-E., verbs with 
 snff. -yn- might be, at least in some forms, conjugated without 
 this element. Thus, side by side with the common Gr. form 
 <f>LX-€-w = *<f>iX.-€-yo) ) we have ^olic <f>ik-r]-fu, wherein the person- 
 ending is added directly to the nominal stem; and, even in 
 ordinary Greek, no form but that of present-stems ((^tX-^-o-w 
 i-<f>i\-r)-(r-a Tre-4>i\-7)-K-a) exhibits any trace of a formative suffix. 
 So, too, in Latin, though the first person albed (I become white) 
 is certainly *alb-e-yo, with the suff. -yo-, the second person albes 
 may as well — and perhaps better — be explained by alb-e-s, with 
 the person-ending immediately attached to the derivative base, 
 as it would be by a contraction of *alb-e-ye-s ; and so also for 
 every other person, and further for the conjugation of formo 
 formds. 
 
 The same alternation occurred in Germanic. Here, more- 
 over, the distinction of the forms which did or did not possess 
 the formative suffix, is often impossible and, at all events, 
 hardly more than theoretical guess-work, since the forms 
 which possessed it became, by dropping the intervocalic j and 
 contracting the remainder, exactly the same as the forms in 
 which the endings were immediately added to the nominal base. 
 Thus, it does not matter whether we resolve Go. salb-d-J? (he 
 salves) into salb-6-J> = I.-E. *solp-d-ti, or into *salb-o-ji-}> = I.-~E. 
 *solp-d-ye-ti, since either equally well explains both the his- 
 torical form and the actual E. (he) salv-e-th and G. (er) salb-t. 
 
 (93) This point being settled, we must distinguish the 
 
 1 Cf. supra 85. Tbis circumstance occasioned a confusion between the 
 conjugation of the e-class and that of the derived or denominative verbs 
 here referred to, a curious process in the older languages, though of no 
 importance for modern speeches, where all conjugational classes are now 
 blended together. Thus, O.H.G., purer in this respect even than Gothic, 
 still preserves hab-e-m (I have, Go. hab-a), and also, analogically, salb-o-m (I 
 salve, Go. salb-O). But Mod. G. conjugates hab-e, salbe, like nehm-e, fahr-e ; 
 and it is only from the lack of metaphony in the 2nd and 3rd persons of the 
 two former verbs, that we are able to see that they are not all of the same 
 origin. In English, the metaphony in the present of strong verhg-heing lost 
 altogether, the confusion is past remedy. ^<£^£SE LIBR ' Tt 
 
 (tfNIVERSITl 
 
158 ENGLISH AND GERMAN GRAMMAR. 
 
 denominative formation according to the nature of the 
 nominal stems to which it is applied. 
 
 1. The nominal stem ends in -a, Preg. -d. — The formation is 
 plain and regular, as shown above: salb-a (a salve), Go. 
 salb-o-n, B. to salve, G. salb-en ; O.H.G. dht-a (care), aht-o-n (to 
 take care), G. acht-en, etc. Then, the derivative verbs in -6- 
 being partly confused with the primary verbs in -e-, the class 
 also admitted of a secondary formation in -e-, which, of course, 
 is now no longer to be distinguished from the former: Go. 
 saurg-a (care, anxiety), vb. saurg-a-n like hab-a-n, O.H.G. 
 sorg-e-n and G. sorgen, but O.E. (with suff. -ja-) sorg-ia-n and 
 E. to sorrow. 
 
 2. The nominal stem ends in a short vowel, type <£iA-e'-a>. — 
 In this case, since both -e-yo- and -%-yo- become Preg. -ja-, while 
 -ti-yd- seems to have done the same through analogy, the form 
 of the verb is necessarily the same as if the nominal stem had 
 ended with a consonant, e.g. I.-E. *koy-lo-s (whole, safe), cf. 
 Sk. keva-la-s (whole), Preg. *hai-la-s (Go. hail-s, O.E. hot, E. 
 whole, G. heil "health"), whence Go. hdil-ja-n, E. to heal, G. 
 heilen. 
 
 3. The nominal stem ends in a consonant. 
 
 A. Normal formation. — The suff. -ja-, which still remains 
 visible in Gothic, is adapted to the final consonant of the 
 nominal base, whatever it may be ; next, the j vanishes in the 
 later languages, and nothing is left, but the obscure and 
 uniform ending, -e in sg. 1 pres. indicative, -en in the infinitive, 
 and so forth. Thus, supposing the base to end with t, it will 
 form with the suffix a Go. complex -t-ja-n, which in German 
 becomes -zen. After a final s, we should expect a Preg. -z-ja-n ; 
 but, analogy restoring the s, we find O.E. -sian and E. -se. 
 Lastly, these derivative syllables will be seen below to extend 
 their influence beyond their proper place, and to become in 
 their turn productive of new word-classes. 1 
 
 B. Analogical formation. — From a similar cause, the vowel 
 6, borrowed from verbs derived from stems in -a, was trans- 
 ferred, even in the Pregermanic period, to such as were derived 
 from consonantal stems. Just as Latin constructed on the base 
 
 1 Infra, 106-107. 
 
PRIMITIVE DERIVATION. 159 
 
 honor a vb. honor-d-re, as if there were a fra. noun *honor-a, 1 
 so, Germanic has formed a great many weak verbs, simply by 
 adding its derivative o to the final consonant of a stem ending 
 in -es- or -en- : Go. hat-is (hate), hat-iz-6-n (to hate) ; Go. 
 frau-ja (lord, gen. frdu-jin-s), frdu-jin-6-n (to be lord). 
 Further, these final compounds, -izon, -inon, becoming inde- 
 pendent, at once made their way as general forms for secondary 
 derivation : the former, no less than -sian above, might have 
 resulted in E. -se ; the latter was blended with the E. ending 
 -en, 2 as must be inferred from E. to fast-en, which is the equiva- 
 lent of O.H.G. fest-inon, derived from O.H.G. fest-i, now G. fest 
 (fast). Thus, the elements of English and German verbal 
 derivation have been confused and reduced to a very slight 
 number ; but these again, as will be farther seen, have been 
 extensively used. 
 
 These brief remarks will prove sufficient to summarize the 
 principles of Germanic verbal formation, the more so as our 
 modern endings have been made uniform by analogy and have 
 degenerated under the action of phonetic laws, to such a degree 
 that no stress should be laid upon the particular form of their 
 original components. 
 
 1 On the pattern of oper-a rl (and the like), which was traced back to opus 
 (gen. oper-is), whilst it really proceeded from the fm. noun oper-a. 
 » Cf . infra, 106, 1. 
 
CHAPTER II. 
 ENGLISH AND GERMAN DERIVATION. 
 
 (94) In the preceding pages the reader has met with a 
 large number of English and German Suffixes, which, 
 altered and shortened as they may be, yet clearly reproduce a 
 primitive suffix-type. The following pages are intended to 
 show, that, among all the suffixes as yet used and recognised 
 as such by a speaker in both languages, there is not one that 
 cannot claim the same origin. To such derivative resources, 
 however, as were afforded by its original stock of words, 
 Germanic added two new processes, which both arose from the 
 primitive process of composition or, at least, juxtaposition of 
 independent words, namely : the so-called Inseparable Pre- 
 fixation, 1 and the Suffixation of some nominal elements, 
 now void of any independent meaning. Thus, the present 
 chapter must be divided into three sections. 
 
 Section I. 
 
 PREFIXES. 
 
 (95) Prefixes, like suffixes, may be either nominal or 
 verbal. But it must be understood, at the outset, that any 
 
 1 We need scarcely recall the reason why separable prefixes are to be 
 excluded : in point of fact, they are still independent words ; inseparable 
 prefixes are the only ones that can be said to have descended to the grade of 
 a mere derivative element. Thus, in ein-tretev, aus-treten, etc., even in 
 ein-tritt, nach-druck, indeed, even in ein-kunft, zu-kunft, — though "kunft no 
 longer exists as a separate word, — the speaker distinctly perceives two words, 
 and a kind of compound ; whereas, in be-treten, contrasted with treten, he 
 simply perceives a word modified by a mere exponent of meaning, since be- t 
 by itself, means nothing at all. 
 
 160 
 
ENGLISH AND GERMAN DERIVATION. 161 
 
 verbal prefix is also a nominal prefix, inasmuch as every verb 
 derived by prefixation may again serve as a basis for further 
 substantival or adjectival derivation : thus, after the analogy 
 of the relation he perceives between werf-en and wurf, a speaker 
 naturally forms ent-wurf from ent-werf-en ; and, as soon as a 
 vb. be-weg-en (to move) has been formed, such derivatives as 
 be-weg-ung, be-weg-lich and be-weg-lich- keit easily follow. We 
 must, therefore, confine the name of nominal prefix to such as 
 characterize either nominal or adjectival formations which 
 appear to be independent of any verbal formation. There are 
 very few of them, and still fewer in English. 
 
 § 1. Nominal Prefixes. 
 
 (96) I. By far the most common of nominal prefixes is the 
 syllable which appears as ga- in Gothic. Its meaning, still 
 clearly copulative and .collective in a great many nominal 
 formations, points to an original connection with the L. pre- 
 position cum (with) = *com, as, for instance, L. par (equal) and 
 corn-par (pairing), cf. G. ge-paar-t : Go. sinjy-s (walking), whence 
 Go. ga-sinp-a (fellow- walker, companion), O.E. ge-sib", O.H.G. 
 neuter collective gi-sind-i (attendance) and G. ge-sind-e. But 
 it is the common fate of such derivative elements to have their 
 value gradually obscured and defaced 1 : no wonder, therefore, 
 if we often find the prefix ga- in words which have almost or 
 altogether dropped their old collective meaning. 
 
 In form, this unaccented ga- 2 is weakened to O.E. ge-, 
 O.H.G. gi-, M.H.G. and Mod. G. ge-. O.E. ge- was still spelled 
 y- in M.E. ; but it afterwards disappeared, except when it was 
 changed to an obscure vowel, as in E. e-nough = Q. ge-nug, and 
 E. a-ware = O.E. ge-wcer=G. ge-wahr, so that a modern English 
 speaker no longer recognises this formative syllable. 
 
 1 Cf. L. co-gnoscere nearly synonymous with the simple noscere. 
 
 8 The phonetic correspondence is irregular, as we should expect Preg. 
 */ia- = L. co-, according to Grimm's law. But the reader will observe that, 
 in the middle of a sentence, the initial might appear either as *ha- or *s«, 
 according as the last syllable of the preceding word was or was not accented : 
 so, supposing the form *£d- to have extended through analogy, the case 
 would fall under the head of Verner's law. 
 
 M 
 
162 ENGLISH AND GERMAN GRAMMAR. 
 
 As to the function, the copulative value is quite obvious 
 in: G. g-leich — Gco. ga-leik-s, and E. a-like, cf. E. like; G. 
 ge-mein = Go. ga-mdin-s, cf. L. communis = *com-moin-i-s, and E. 
 mean (vulgar) = O.E. ge-mcene ; G. ge-mahl (husband) =O.H.G. 
 gi-mahal-a, from mahal (assembly, contract) ; G. ge-fdhrt-e 
 (fellow-traveller), from fahr-t (journey) ; G. ge-vatter (god- 
 father), ge-sell-e (fellow), etc. The collective function is 
 likewise well marked in the numerous G. neuter nouns, like : 
 ge-birg-e = O.H.G. gi-birg-i, from berg (mountain), 1 ge-fild-e 
 (fields), ge-fieder (plumage), ge-stirn (constellation), etc.; in 
 O.E. hand-ge-weorc and hand-ge-crceft, now become E. handiwork 
 and handicraft, the collective prefix, reduced to a mere vowel, 
 is scarcely visible except to an etymologist. 
 
 II. An isolated prefix, which cannot be recognised in any 
 other I.-E. language, though common to the whole Germanic 
 domain, appears as Go. us-, with the meaning " out." 8 In 
 nominal prefixation, this syllable keeps its accent and its vowel 
 unshifted, O.E. and E. or-, G. ur-, as in O.E. or-dal (-judgment) 3 
 and E. or-deal, O.H.G. ur-tel and G. ur-theil, cf. E. deal = G. 
 theil. The function did not survive in English. But, in 
 German, either owing to the sense of " extraction " which was 
 contained in it, — e.g. in ur-sache (cause), ur-sprung (origin), — or 
 merely after the analogy of the word ur-alt, wherein the mean- 
 ing " of yore," though originally inherent in the adjective, has 
 become attached to the prefix, this particle assumed a specific 
 value as denoting origin and still survives as such, giving to 
 any noun or adjective to which it is affixed the force of an 
 epithet like "ancestral": ur-ahn (ancestor), ur-heimat, ur- 
 sprache (mother- tongue), Ur-germanisch, etc. 
 
 1 The suffix in these words is I.-E. -yo-, of. L. con-jug-iu-m (marriage). 
 It is sometimes added to primary stems in -eto-, whence in Germanic a 
 derivative complex -it>-ja. G. -de with metaphony : G. ge-mahl-de (marriage) 
 «= O.H.G. gi-mahal-idi; G. ge-bau-de (building) == O.H.G. gi-ba-idi. 
 
 2 It may, however, be identified with Russian voz- (Meillet). But, at any 
 rate, it must be carefully distinguished from the G. preposition au« = E. out 
 = Go. ut m Sk. ud, which is phonetically different and, moreover, has retained 
 an independent meaning and use. 
 
 3 Properly " distribution," then " judgment," and especially " God's judg- 
 ment." Low L. ordalium and F. ordalie are borrowed words. — On the un- 
 accented form of the same prefix (verbal), see below, 98 (II.). 
 
ENGLISH AND GERMAN DERIVATION. 163 
 
 III. The old preposition which is Gr. dvrl (against, in ex- 
 change), L. ante (opposite, before), was not kept as such in 
 Germanic, bnt still survives as a prefix, which, if accented, is 
 O.E. and- and G. ant-, if unaccented, G. ent- : O.E. and-swaru 
 and E. answer, cf. Sk. svdr-a-s (sound) and G. schwbren (to 
 take an oath) = O.H.G. swer-ien; G. ant-wort (answer) =O.H.G. 
 ant-wurti, cf. G. wort = ~E. word; O.E. and-wlita and G. ant-litz 
 (face), cf. Go. vlit-s (face) ; for the meaning, compare Gr. dVri- 
 Sutpov (mutual gift), etc. The particle is reduced to a single a 
 in E. a-long = O.E. and-long = Q. ent-lang, and in E. a-gain 
 (a-gain-st) = O.E. on-gean = G. ent-gegen (opposite), cf. G. 
 gegen. 1 
 
 IV. The negative prefix, E. and G. wn-, 8 has already been 
 explained as the reduced grade of the I.-E. negative particle 
 *n£, that is *»-. If this consonant happened to lean on a follow- 
 ing vowel, it remained a consonant, as in L. non (not) = *n-oino-m 
 (literally "not one"), e.g. : E. n-one (not one), and G. n-ein, 
 now a mere negation ; E. n-ought (cf. aught), likewise weakened 
 to a mere negation in the unaccented form not, and so also G. 
 n-icht = O.H.G. n-eo-wiht, cf. Go. ni vaiht-s (not a thing); E. 
 n-ever, n-either, n-or, etc.; G. n-ie ("never," cf . je "ever"), 
 n-immer =■■ n-ie rnehr, noch = L. ne-gue, nur (only) = O.H.G. ne 
 wdri (literally "were [it] not"), etc. This negative n is now 
 a dead element. But its counterpart un-, arising from the 
 cases in which it had become n by meeting with an initial 
 consonant, was retained and multiplied to such an extent as to 
 be considered the general and exclusive prefix denoting nega- 
 tion, both before a vowel or consonant : thus, E. un-even, as 
 un-like, and G. un-eben, as un-gleich ; cf . L. in-ermis, as in-firmus, 
 etc. Moreover, chiefly in German, it is currently associated 
 even with Romance words : E. un-just, G. un-sicher, etc. 
 
 V. The pejorative prefix mis-, in E. mis-take, and G. misz- 
 fall (accident), etc., had . originally nothing to do with the 
 Romance pejorative prefix mes- (F. me-fait " misdeed," me-dire 
 
 1 M.E. again is commonly used for against. Its present meaning " anew," 
 presents no difficulty : cf. G. xoider (against) and wieder (again). 
 
 s This is not to be confounded with the inversive E. prefix un-, which is 
 an exclusively verbal prefix, infra, 98 (III.). 
 
164 ENGLISH AND GERMAN GRAMMAR. 
 
 "to speak evil") = L. minus, though, being similar in form and 
 meaning, they naturally became blended in English as soon as 
 French had crossed the Channel (E. mis-chief —O.F. mes-chief). 
 In Germanic, the prefix is really a past participle, that is to 
 say, a nominal stem with suff. -to-, seen in Gothic as missa- = 
 *mij}-td-, in missa-dep-s — ^j. mis-deed = G. mis-that, literally "a 
 missed deed," the root of the past participle being the same as 
 that of the L. vb. mitt-ere (of. d-mitt-ere "to lose"), the G. vb. 
 meid-en (to avoid), and the secondary derivative E. to miss = 
 G. missen. In consequence of this etymology, mis-, as well as 
 the preceding, ought to be an exclusively nominal prefix, but 
 analogy has extended it to verbal formation. 1 
 
 VI. Lastly, for want of a better place, we may mention here 
 the E. prefix a-, an unaccented form which represents a number 
 of prepositions, retained in English in their accented form, at, 
 of, on, viz. : E. a-do = at do (to do), at being the sign of the 
 M.E. gerund, as to is now of the infinitive; E. out a-doors = out 
 at doors or out of doors, cf . L. forls fords (out) and fores (door) ; 
 E. a-down = O.E. of dune; E. a-way = O.E. on-weg (on the way), 
 cf. G. weg (away) =M.H.G. en-wec; E. a-foot = on foot; E. a-kin, 
 most probably of kin, and many others. 8 
 
 § 2. Verbal Prefixes. 
 
 (97) I. The Go. pref. ga- is as widely diffused throughout 
 verbal derivation as con- in Latin, and its primitive value 
 frequently remains obvious ; for, just as the L. vb. con-struere 
 (to build) involves an idea of " assembling materials," so there 
 
 1 Thus, G. misz-grif (for the sense, compare fehl-griff) produced a vb. 
 misz-greifen; then followed misz-rlenten, misz-fallen (to displease), after the 
 analogy of ge-fallen (to please) and misz-fall (accident), etc. Still more so 
 in E., owing to the influence of F. vies- : to mistake, to mislead, etc. 
 
 2 The preposition at ( = L. ad) is O.H.G. 05, now lost in Mod. German, 
 though still visible in bis (till) = fci «5 (the first term being now frt = E. by). 
 — E. 011, with O.E. o=a before a nasal, is G. an. — Of course, prefixes of 
 Romance origin have been here omitted. An instance of them is O.E. cerce- 
 and E. arch-— O.H.G. erzi- and G. erz-=~L. archi- = Gr. apxi- (chief-): E. 
 arch-hishop, G. erz-bischof; prefixed even to German words, erz-herzog 
 (archduke). In English we have also the negative Romance prefix in-, 
 which is used concurrently with the native un-, and associated as a rule with 
 borrowed words, in-drfatigable, in-expedient, in-dejinite, but un-defined, etc. 
 
ENGLISH AND GERMAN DERIVATION. 165 
 
 can be hardly any donbt that a Go. vb. ga-timr-jan denotes the 
 act of " adjusting timber " ; and similarly with many others, 
 ga-hdit-an " to con-vocate," ga-nim-an " to con-ceive," ga-lis-an 
 " to col-lect," etc. Even in Gothic, however, a number of verbs 
 with pref. ga- in no way differ in meaning from simple verbs : 
 ga-bau-an (to dwell), like bau-an; ga-sai-jan (to place), like 
 sat-jan, etc. The same is the case with G. ge-wdhr-en, ge- 
 winn-en, contrasted with O.H.G. wer-en and E. to warr-ant, 
 O.H.G. winn-an (to win) ; and, in English, among the scanty 
 relics of this prefix ge-, we have to ajford = M.E. a-forth-en = 03£. 
 ge-fortS-ian (to put forward), as opposed to G. forder-n (to 
 exact), the former derived from forth, and the latter from its 
 comparative vorder. 
 
 Thus, at the outset, this prefix is seen to be wholly uncon- 
 cerned with the formation of the past participle : in other 
 words, a Go. vb. lis-an had a participle lis-an-s, whereas 
 ga-lis-an-s (G. ge-les-en) was the participle of vb. ga-lis-an ; and 
 so also with bau-an bau-ida, and ga-bau-an ga-bau-ida (G. 
 ge-bau-t), etc. But it happened that West Germanic speakers 
 took to using, instead of the past participle of the simple verb, 
 the past participle of the prefixal verb, -so that this syllable ge- 
 seemed and actually became the grammatical exponent of 
 the past participle. Such is its specific value in O.E. as 
 well as in German. In M.E. the initial syllable is reduced to 
 y- (y-clept "named," y-covered, etc.), 1 and then disappears, so 
 that found and had no longer differ from to find and to have. 
 On the contrary, in German, the rule is that every native verb, 
 unless already provided with an inseparable prefix, assumes in 
 the past participle, whether in -en or -et, the inseparable prefix 
 ge-. 
 
 A further consequence, paradoxical enough at first sight, of 
 the extended use of the prefix in German is, that, apart from 
 its grammatical function, it is now as dead in German as in 
 English where it has been phonetically suppressed ; for, as soon 
 as, for instance, ge-les-en, ge-zimmer-t were considered to be the 
 
 1 Still very common in Chaucer and contemporary writers : yclad (clad), 
 yborn, yfrdd, etc., though, probably owing to Danish influence, Mercian had 
 lost the participial prefix quite early. 
 
166 ENGLISH AND GEEMAN GEAMMAR. 
 
 regular participles of the verbs les-en, zimmer-n, etc., the exist- 
 ence of such verbs as *ge-les-en = ga-lis-an, *ge-zimmer-n = 
 ga-timr-jan naturally appeared to be somewhat strange and 
 contradictory, and thus they soon became obsolete ; and, in con- 
 sequence, the prefix ge- now characterizes but a very few G. 
 verbs : either those of which the unprefixal form had become 
 obsolete in early times, as ge-nes-en (to be healed), from a 
 primitive vb. *nes-an (cf. the Go. causative nas-jan " to heal "), 
 which is no longer to be met with ; or those in which the 
 compound had acquired a new meaning, apparently unrelated 
 to that of the simple verb, as ge-f all-en (to please), 1 as opposed 
 to fall-en ; or else, especially, such compounds as had assumed 
 a decidedly transitive or causative sense, viz. ge-steh-en (to 
 maintain, to own), ge-schweig-en (to silence), 2 contrasted with 
 steh-en (to stand), schweig-en (to be silent), etc. 
 
 (98) II. The pref. ur-, losing its accent when added to a 
 verb, is G. er- ; for there is no difference, except in the accent 
 of the prefix, between ur-theil (judgment) and er-theil-en (to 
 divide), ur-laub (permission) and er-laub-en (to allow), ur-kun-de 
 (document) and er-kenn-en (to acknowledge), 3 etc. ; and the same 
 unaccented prefix is concealed in the E. vb. to a-rise (cf . to rise) 
 = O.E. a-rls-an = Go. ur-reis-an (to arise). 4 
 
 III. The pref. and- ant-, when verbal and therefore unac- 
 cented, is changed to E. un-, G. ent-, assimilated to emp- before 
 an initial /: emp-fang-en = *ent-fang-en, emp-find-en, etc. 5 The 
 primitive sense of reciprocity or contrariety inherent in the 
 
 1 Literally " to fall in with." 
 
 s The intransitive meaning still preserved in ge-schweige (the more so), 
 literally " should we [even] forbear to mention." — Of course, the type 
 ge-sell-en (to accompany), though often met with, is here beside the question, 
 as not being formed by adding a prefix ge- to a verb, but derived from a noun 
 which already had this prefix, supra 96, I. 
 
 8 Though, as a matter of course, urtheil produced a vb. urtheilen, whereas, 
 on the other hand, erkevnen produced erkenntnisz, supra 95. 
 
 * Hence initial E. a- is seen to be the etymological representative of six 
 distinct particles at the very least. 
 
 8 Literally : " to take in exchange," whence " to receive " ; "to find or 
 acquire in exchange, to answer any outer fact with a corresponding and 
 adequate impression," whence " to feel," etc. — No modern verb, however, 
 but these two and emp-fehlen, preserves the form emp- : everywhere else, 
 ent-falten, ent-fiihren, ent-farben, the form ent- has been restored after the 
 analogy of the other members of this class. 
 
ENGLISH AND GERMAN DERIVATION. 167 
 
 I.-E. particle (Gr. dvri) does not always appear with clearness, 
 or even seems to have vanished in such German formations as 
 ent-schlafen (to fall asleep), ent-laufen (to run away), ent-flieszen 
 (to flow off), etc. But this meaning is quite plain in the 
 common use of the prefix, which may be termed inversive, 
 as involving the contrary action to that which is expressed by 
 the simple verb, in other words, the reverse of the function of 
 the G. prefix be- or ver-, and synonymous with L. dis-, F. de- 
 (in de-fair e, de-planter) : G. ent-decken (to discover), ent-binden, 
 ent-ehren (to dishonour) ; and similarly in English, to un-do, to 
 un-bind, to un-fold = G. ent-falten, 1 etc. 
 
 (99) TV. It has been already stated that the E. and G. 
 pref. be- is merely the unaccented form of the preposition, E. 
 by, G. bei, which may again be discovered in the final element 
 of the preposition, Sk. abhi (towards), Gr. a/x<f>l (around, 
 about). Though unaccented, it is still a preposition in such 
 adverbial locutions as E. be-fore, be-neath, G. be-hende (close at 
 hand, swiftly), etc., and, in the same form, it was very early 
 adapted to a large number of simple verbs, with a greater or 
 less signification of tendency or accession : Go. bi-git-an, E. to 
 be-get, cf. to get ; O.E. be-cum-an (to suit), E. to be-come, G. 
 be-kommen (to receive) ; O.H.G. bi-ginn-an (cf. Go. du-ginn-an) 
 and G. be-ginn-en = l&. to be-gin, etc. English rarely uses the 
 prefix in other senses, though it possesses some derived verbs, 
 with causative meaning, like be-dim, bespread, beset, even 
 developed from a French basis, as besiege, etc. ; but, in 
 German, the prefix is a living one, and has undergone a con- 
 siderable extension in a special direction. 
 
 As shown by the contrast in meaning between E. become and 
 G. bekommen, the prefix contains nothing to assign a transitive 
 value to any verb it modifies. But, though of course "to come 
 to " may merely result in " to become " (" to suit " or " to 
 grow"), it is no less capable of meaning "to reach, to get," 
 and of governing, as such, a complement in the accusative ; and 
 this is just the case with the G. verb. 2 Thus also, since sitzen 
 
 1 The development of the inversive meaning was carried very far in G. 
 ent-setzen " to unsettle — to overturn — to terrify." 
 
 2 And partly down to M.E., since to besiege is derived from a noun which 
 
168 ENGLISH AND GERMAN GRAMMAR. 
 
 means " to sit," be-sitzen, " to sit by " or " round," gains the 
 meaning " to occupy, to possess," be-fallen, " to fall upon," 
 becomes " to attack," * and so forth. Hence the first stage in 
 the evolution of prefixed fee-, namely, the possibility of con- 
 verting nearly every verb from intransitive to transitive 
 value, folgen be-folgen, schwbren be-schworen, streiten be-streiten, 
 siegen be-siegen, etc. 
 
 Secondly, no sooner had such verbs as bestreiten, bekampfen, 
 besiegen become current, than the speaker's mind took to associ- 
 ating them, no longer with the simple verbs whence they had 
 arisen, but directly with the nouns, streit, kampf, sieg, which 
 were respectively visible in each. Then, in consequence, it 
 seemed but natural to construct from any noun, or even 
 any adjective, by means of the prefixed be-, a transitive or 
 causative verb, which should involve an idea of " providing, 
 endowing with " the object or quality denoted by the noun or 
 adjective, viz. : bau (building), be-bau-en " to build upon [a 
 ground] " ; frucht, be-frucht-en (to fertilise) ; frei, be- jr ex-en (to 
 free) ; fahig, be-fahig-en (to habilitate) ; lastly, there arose a 
 fresh set of analogical creations, be-giinst-ig-en (to befriend) 
 on the basis gunst (favour, kindness), be-vollmacht-ig-en (to 
 authorize) on voll-macht (full power), etc. 2 
 
 V. The G. pref. ver-, no less widely diffused, and retained 
 also in a few E. compounds (for-lorn = ver-loren, to for-give — 
 ver-geben), represents not only the Go. fra, — which again might 
 correspond both to Gr. 7rp6 (before) and L. pro (for), — but also 
 reproduces the Go. particles fair faur, which represent L. per 
 (by, for, cf. B. for and G. fur), and probably, also, go back to 
 another I.-E. pref. *per- inasmuch as the latter is suggested by 
 Gr. Trapa (sideways, awry), and by the pejorative meaning of 
 such L. verbs as per-dere, per-ire, etc. In short, several words 
 
 is certainly borrowed from French. — By the way, observe the arbitrary sub- 
 stitution of tbese exponents with obliterated meaning : E. to beseech is 
 equivalent to G. er-suchen. — Observe also the peculiar sense in E. to be-head. 
 
 1 Cf. again the intransitive sense in E. to befall. 
 
 8 The same may be said of the following (ver-breiten, ver-einigen, ver- 
 nachlassigen), and also, to a certain extent, of the preceding prefix: er- 
 leichter-n, er-mannen, er-ledigen (there being no such verbs as *mannen, 
 *ledigen). 
 
ENGLISH AND GERMAN DERIVATION. 169 
 
 have been blended together in this nnaccented syllable, ■which 
 is successively O.H.G. for-, fur-, then far-, lastly fir- and fer-. 
 Thence proceed the manifold and various functions assigned in 
 G. to this prefix, which in one place appears decidedly signifi- 
 cant, and in another expresses but a slight shade of meaning, 
 perhaps intensive at the outset, but afterwards more or 
 less weakened: G. fr-essen = Go. fra-it-an 1 ; G. ver-lieren, E. 
 synonym to lose ; G. ver-dndern, synonymous with andern (to 
 change), etc. Under the pejorative head may be said to fall, 
 more or less directly : the idea of excess, in ver-brauchen (to 
 wear out), ver-bluhen (to shed its blossoms) ; the idea of error, 
 oblivion, in ver-gessen (cf. E. to get and to for-get), ver-fuhren (to 
 seduce), ver-legen (to lose) ; the idea of ruin, destruction, in 
 ver-gehen* which corresponds to L. per-ire, ver-geben-s (vainly) ; 
 that of a deed taken in ill part, ver-leumden (to defame, cf. 
 leumund " fame ") ; and even a completely negative meaning, 
 in ver-bieien (cf. E. to for-bid), ver-achten (to despise), etc. 
 Lastly, as used at present, the pref. ver-, like the preceding 
 one, may change an intransitive into an active verb, thus 
 dienen ver-dienen, 3 or from a nominal basis may form a corre- 
 sponding causative verb, with slight differences of meaning, 
 however, which must be learnt from a special study of German, 
 or rather from use and practice. 4 
 
 (ioo) VI. The dissociative prefix, O.E. to-, O.H.G. za-, 
 zi-, ze-, then subsidiarily zur-, zar-, zer, the latter being the only 
 one preserved in Mod. G., has certainly nothing to do with the 
 
 1 Intensive " to devour," and, therefore, confined to the feeding of animals, 
 or else used humorously. E. to fret, by passing through the meanings 
 "to devour — to bite — to gnash the teeth," has become used for strong and 
 outward emotion. 
 
 * It must not be confounded with E. to fore-go, which contains an accented 
 form, G. vor = Go. faura (before): Go. faura-gaggav, "to go forward, to 
 precede," whence " to outreach" and (E.j "to give up." 
 
 8 " To serve (dienen) on account of," whence " to earn " and " to deserve." 
 — On the contrary, from the active recken (to stretch, cf. L. por-rig-ere), we 
 have the pejorative participle ver-reck-t (stiffened, dead), whence the in- 
 transitive verb ver-reck-en " to die " (animals). 
 
 4 Observe both the transitive and the pejorative function cumulated in 
 such locutions as seine zeit vergeigen " to waste one's time by playing upon 
 the violin," and the contrasted meaning in kaufen (to buy) and verkaufen 
 (to sell), Go. kdupon meaning only " to trade." 
 
170 ENGLISH AND GERMAN GRAMMAR. 
 
 E. preposition to = G. zu, and seems, like the preceding, to be 
 of manifold and rather intricate origin, inasmuch as it corre- 
 sponds to two Go. particles : one, visible in Go. tvis-standan (to 
 stand apart), which at once recalls Gr. 8is = *8fis (twice), and 
 the well-known dissociative prefix, L. dis- 1 ; the other, of 
 pejorative force, Go. tuz- = Gr. 8vor- = Sk. dus- (ill-). The prefix 
 no longer occurs in English except in the old verb (pf.) to- 
 brake = G. zer-brach (he broke to pieces); but it remains in 
 common, specific and extended use in German, either as 
 modifying the sense of a simple verb (zer-reiszen, zer-stampfen, 
 zer-flieszen, zer-brockeln, etc.), or even, occasionally, as forming 
 a verb from a nominal stem (zer-fetzen, zer-fleischen, zer-nichten, 
 zer-lump-t " ragged," no vb. lumpen existing, at least with this 
 meaning). 
 
 From this contrasted study of E. and G. verbal prefixes, it 
 follows that both languages, though starting from the same 
 point and still preserving many identical formations, have now 
 reached quite a different stage in the evolution of their verbal 
 prefixes : all of them are still living in German, apart from 
 ge-, 2 whilst all are dead in English, with the single exception 
 of wn-. 3 
 
 Section II. 
 
 SUFFIXES PROPERLY SO-CALLED. 
 
 (101) English and German Suffixes, like those in I.-E. 
 and Preg., are either nominal or verbal. 
 
 § 1. Nominal Suffixes. 
 
 Nominal stems formed by suffixes fall under a double head, 
 according as they respectively play the part either of nouns 
 or of adjectives. 
 
 1 It is almost superfluous to state that whenever the prefix occurs in this 
 shape in English (to dis-honour, to dis-qualify), it is simply borrowed from 
 Latin. 
 
 2 Dead as a derivative element, though living as a grammatical exponent. 
 8 Thus, such a sentence as " we must un-boycott Ireland " would be con- 
 sistent with the genius of our language. 
 
ENGLISH AND GERMAN DERIVATION. 171 
 
 A. Nouns. 
 
 (102) I. E. -er, G. -er : nouns of agent. — The interesting 
 history of this extremely common and still wonderfully active 
 suffix, may be given in a few characteristic processes, by start- 
 ing from an original form, -ar-ja-, or -dr-ja-, that is to say, 
 I.-E. snff. -yo-, added to a stem of secondary derivation with 
 suff. -ro-, e.g. : Sk. rdtha-s (chariot), whence a vb. ratha-r-yd-ti 
 (he drives a chariot). 
 
 1. Latin is rich in adjectives in -driu-s, all of which are 
 derived from nouns : ferr-driu-s (of iron), pisc-driu-s (of fish), 
 etc. Now, these adjectives, when qualifying a person, acquire 
 a substantival meaning, viz. : (faber) ferrdrius, literally " iron- 
 smith " ; piscdrius (fisher), etc. The same is the case with the 
 Go. : vull-arei-s (woollen-draper), bok-arei-s (transcriber) ; and 
 the Go. suff. -arei- is to be recognised, though with an initial 
 long vowel, in the metaphonical forms, O.E. -ere, M.H.G. -cere 
 ( = O.H.G. -dri), then -er, E. and G. -er, the function of which 
 is identical in the types Jish-er (O.E. fisc-ere) and fisch-er, and 
 many similar, all derived from nouns, which will at once occur 
 to the reader's mind. 
 
 2. This element, of conrse, could be added to a nominal stem 
 in -en- as well as to any other : E. gard-en-er, and G. gdrtn-er = 
 M.H.G. gart-en-cere = O.H.G. gart-in-dri, cf. the E. name Gar- 
 diner; O.H.G. ohs-an-dri (cowherd), G. ochs-n-er, etc. Further, 
 the peculiar property of these stems in -en-, as stated above, is 
 to drop the n in the nomin. sg., so that, when compared with 
 a nomin. O.H.G. gart-o and M.H.G. gart-e, O.H.G. ohs-o and G. 
 ochs, etc., the n inserted in the derivative seems a part of the 
 second suffix : whence a fictitious suffix -ner, which appears in 
 G. bild-ner (statuary), 1 pfort-ner (porter), harf-ner (harper), 
 and several others. 
 
 3. It will be understood from these considerations that suff. 
 -er, like L. -drius, cannot, in theory, be added to a verbal 
 stem, and that its function, as clearly verified by the earliest 
 
 1 In the fourteenth century, bild-en-are, but before that the regular 
 bild-cere. — The E. forms in -ar (li-ar) and -or (sail-or=sail-er, warr-ior, cf. 
 law-yer) are mere varieties of spelling. 
 
172 
 
 ENGLISH AND GERMAN GEAMMAE. 
 
 known formations, is merely the derivation of nouns from 
 nouns, e.g. : O.E. hoc (book), whence O.E. boc-ere (scribe), cf. 
 L. libr-drius ; O.H.G. *fano (flag), 1 whence fan-eri (ensign- 
 bearer), M.H.G. venre, Mod. G. corrupt fahnrich, cf. L. vexill- 
 drius, etc. But it often happened that, side by side with such 
 a noun, there existed a verb, either identical in form, or at 
 least closely akin, and it was easier for the speaker to refer, 
 e.g., the words fisher and fischer, to the verbs to fish and fischen, 
 than to the substantives fish and fisch, whence they really pro- 
 ceeded. 2 The analogical and necessary consequence of this 
 confusion led to the derivation of a noun of agent from any 
 verb by the mere addition of suffix -er, a process still widely 
 used in both tongues : E. shave-r, hate-r, mdke-r ; G. kauf-er 
 (purchaser), sauf-er (drunkard), schneid-er (tailor), etc. This 
 occasioned moreover, in German, a number of doublets, which 
 were afterwards partly eliminated, either by one of them be- 
 coming obsolete, or else by each assuming a slightly different 
 meaning: O.H.G. sang-ari, a normal derivative from sang — T&. 
 song, now sdng-er, and G. sing-er, an abnormal derivative from 
 sing-en ; G. schnitt-er and schneid-er ; G. ritt-er (knight), cf. 
 ritt (ride), and reit-er (rider), from reit-en (to ride), etc. 
 
 4. The further use of the same element in the derivation of 
 nouns denoting inhabitants (E. London-er, G. Berlin-er) is con- 
 sistent- with its adjectival origin, though, of course, L. -arius 
 never assumed such a function. 
 
 5. This suff. -er ought not to be confounded with another, 
 identical in form, but quite different in meaning, and of very 
 rare use, which characterizes the names of some male animals : 
 E. cat and G. katze as generic name, but E. *cat-er preserved in 
 cater-waul, and G. kater=0.~H..Q. chat-aro; E. goose = *gons = 
 *gans, and G. gans, but O.E. gand-ra and E. gand-er, M.H.G. 
 gan^-er and G. ganserich. 3 
 
 1 In the compound gund-fano (battle-ensign), borrowed in F. and thence 
 in E. gonfanon. 
 
 1 Cf. supra 99, the reversed process as to bestreiten, which was referred to 
 streit, while it really came from streiten, and infra IX. (105). "Blunders " 
 of this kind are very common in all languages, and greatly increase their 
 derivative power. — All these words form their feminine in E. -ess (if at all 
 needful), and G. -in, supra 69, note, and 87, 1. 
 
 8 Corrupt like fahnrich above ; cf . enterich, tauberich. 
 
ENGLISH AND GERMAN DERIVATION. 173 
 
 (103) II. E. -ing, G. -ung : nouns of action. — These two 
 suffix-forms are not identical, but the difference between them 
 is almost immaterial, as they are Germanic substitutes re- 
 spectively for I.-E. -eH-ko- and -n-ko-, 1 especially as German 
 will be seen to show the form -ing in some cognate derivations, 
 whereas O.E., even in this very class, has the form -ung, now 
 lost in English: O.E. leorn-ung = G. lern-ung, but E. learn-ing 
 (as noun), etc. In both languages, in fact, it is the specific 
 function of this suffix to form either a noun denoting an action 
 or an abstract noun constructed on a verbal basis. 
 
 1. This is the only function it assumes in German, where it 
 has been developed on a considerable scale, but has remained 
 strictly derivative : O.H.G. hantal-on (to deal with), fm. hantal- 
 unga] man-on (to warn), fm. man-unga; G. handel-n handl-ung, 
 mahn-en mahn-ung ; further fuhr-ung, riist-ung, fest-ung, brech- 
 ung, etc. 
 
 2. In English, although it remains derivative (a yearn-ing, a 
 liv-ing, a hunt-ing party, a danc-ing master? etc.), it has also 
 become an important grammatical element. Final -end in the 
 present participle, at least in some parts of England, had come 
 to be sounded nearly like -ing, 3 so that a confusion of the two 
 was possible. On the other hand, O.E. had many such sen- 
 tences as ic wees on huntunge, where huntunge is the dative of 
 a noun denoting an action, and this sentence, becoming E. I 
 was a-hunting, could very easily be understood as I was hunting 
 (pres. part.), especially as a-waiting (and the like), in the sen- 
 tence I am a-waiting, might be mistaken for the pres. part, of 
 the vb. await, whereas it is really a noun of action preceded by 
 a preposition. From this arose in English a curious confusion 
 between the forms of the noun denoting an action, of the 
 gerund and of the present participle, so that, in a great 
 
 1 Cf. supra 91, 2. 
 
 1 Thus, and not by the present participle, — as in F. une matinie dansante, 
 etc., — are to be explained similar E. locutions, which are really compound 
 words in which the first term is no less clearly a noun than in the G. type 
 erlernungs-mittel, etc. The same is the case with the form in -ing when 
 governed by a preposition : on asking, i.e. " upon interrogation," of doing, 
 for playing, etc., etc. 
 
 8 Cf. supra 90, VI. The pronunciation -ing is attested by spelling as early 
 as 1200. 
 
174 ENGLISH AND GEKMAN GKAMMAR. 
 
 many cases the ending -ing has completely lost its nominal 
 character, the word thus formed being mistaken for a verb and 
 treated as such in syntax : for having done ; nature's chief master- 
 piece is writing well (Pope), etc. 
 
 III. E. -ing, -ling, G. -ing, -ling: diminutives.— 1. The 
 suff. -ing has no diminutive value in itself (cf. E. king = cyn-ing, 
 and G. konig = kon-ing, the primary stem being a word which 
 means "race, people"), though it is not to be denied that it 
 often occurs in the designation of little objects : E. shill-ing, 
 farth-ing (" one-fourth "), G. schill-ing, etc. It was uncommon 
 and is now dead. 
 
 2. Secondly, by combining with -ing the diminutive suffix 
 -ila, Germanic secured a diminutive suffix -iling, -ling, which, 
 though still tolerably well represented, hardly enters into any 
 new formations: O.E. gos-ling and E. gos-ling; E. darling = 
 *dear-ling, caressing diminutive ; E. young-ling and G. jung-ling 
 = O.E. geong-ling and O.H. G. junga-ling ; E. lord-ling, and even 
 lord-ing, contemptuous diminutive ; O.E. cedeling and G. edeling 
 (nobleman), without any precise diminutive value, etc. 1 
 
 IV. G. -lein and -chen (E. -kin) : diminutives. — 1. The 
 suff. -In, added likewise to the diminutive suff. -ila, formed a 
 complex -illn, -lin, which no longer appears either in English 
 or Low German, but is in very common use in High German : 
 O.H.G. fugil-ln (little bird), chezz-ilin, chind-ilin; G. voglein, 
 katzlein, kindlein, mdnnlein, etc. 
 
 2. The same -in, adapted to a stem with final guttural, 
 formed a complex -kin, unknown in pure High German, though 
 current in Low German : Dutch skipe-kln (little ship). English 
 borrowed this ending (manni-kin), without however further 
 developing it. 2 In Central German it became -chen and spread 
 widely : mann-chen, schaf-chen (little sheep), haus-chen. As far 
 as popular speech is concerned, suff. -chen is exclusively used 
 in the Northern, and suff. -lein exclusively in the Southern 
 countries of Germany; but the literary language is well known 
 
 1 Quite common as such in E. : weakling, fondling, changeling, hireling, 
 etc. Cf. also G. zdg-livg (pupil), havpt-ling (chieftain), fiiicht-ling (runa- 
 way). 
 
 1 Though we have lamb-kin, nap-kin (literally " little table-cloth," F. 
 nappe), and a few proper names (Jenkin). 
 
ENGLISH AND GERMAN DERIVATION. 175 
 
 to admit of both concurrently, and no Germanic formation is in 
 more constant use. 
 
 (104) V. E. *-sel, *-sle ; G. -sal, -sel : abstract nouns. — 
 The origin of this element is doubtful, owing to the many 
 phonetic alterations which it underwent from I.-E. down to Pre- 
 germanic. It may be explained in several ways, all fairly 
 plausible, though none can be deemed beyond doubt. The 
 simplest, however, would be to start from the I.-E. suff. -tlo-, x 
 and to suppose it to be added to a root ending with a dental 
 consonant, so that the primitive group became I.-E. -t-tlo-, Preg. 
 -ssla-, which was afterwards used as a whole in other deriva- 
 tions. 2 
 
 1. The suffix is still clearly visible in German, having partly 
 escaped the effects of the law which causes unaccented finals 
 to become obscured. Thus, it often keeps the vowel a : trub-sal 
 (affliction) =O.H.G. truob-i-sal, lab-sal, schick-sal, 5 etc. But its 
 power of extension is nearly extinct. Still more so, of course, 
 with its duller form -sel : wech-sel (exchange) = O.H.G. weh-sal, 
 from an I.-E. root uik meaning alternation, cf. L. vic-es, vic- 
 issim; raih-sel (riddle) = O.H.G. rdt-isal, from G. rath-en = 
 O.H.G. rdt-an (to guess) ; iiber-bleib-sel (remainder), etc. Most 
 of these nouns are true to their instrumental origin in keeping 
 the neuter gender. 
 
 2. In English, the suffix is hidden under a double veil. 
 Firstly, the word which is G. rdihsel is O.E. rvedels, the two 
 consonants having changed places. 4 So we should expect to 
 find E. a *riddles ; but the final s was mistaken for the sign of 
 the plural number, and dropped in consequence to form the 
 seemingly correct singular riddle. In the same way we have 
 E. burial = M.E. burials = O.E. byrg-els, derived from byrg-en (to 
 
 1 Of nouns denoting instrument, tuyra 79 (XIII.). 
 
 * The 8 was not changed to z, because it was double. Besides, a euphonic 
 vowel was inserted between the * and the I (O.H.G. -isal), which prevented 
 the former from becoming S . 
 
 8 In some words the syllable was probably mistaken for the second term 
 of a compound, such as -sam or -bar, and was therefore sounded with a 
 secondary accent, which preserved the vowel. 
 
 4 No process is more common, in any language, than the metathesis of 
 liquid sounds: cf. P. Passy, ChangemenU phcmitiqueijj^oJ^iji^ sq.^ 
 
 ' TJNIVEB.S1'- ' 
 
176 ENGLISH AND GERMAN GRAMMAR. 
 
 bury), cf. G. berg-en (to bury, to hide). We need scarcely ob- 
 serve that this method of formation has long been dead. 
 
 VI. E. -ness, G. -nisz: abstract nouns, also mostly neuter. 
 — This suffix, though far more important than the preceding, 
 and still living in both languages, has no less doubtful and 
 intricate an origin. The earliest form known to us is a msc. 
 Go. suff. -assu-s, wherein the final element is probably the I.-E. 
 msc. suff. -tu-. If so, the E. and G. initial n cannot belong to 
 the suffix : it must have arisen from cases in which -assu- was 
 added to a stem in -en-, e.g. Go. J>iud-a (people, gen. }>iud-in-s, 
 cf. L. Teut-o pi. Teut-on-es), whence a derivative J?iud-in-assu-s 
 (kingdom, in the Lord's Prayer), and several others. So also 
 with derivatives from participles ending in n, as O.H.G. for- 
 loran-issa, which should be E. *forlorn-ess (altered to forlorn-ness) 
 and G. *verlorn-isz. These words gave rise to a false suffix, E. 
 -ness, G. -nisz, which took the place of the true form : * E. holy- 
 ness = O.E. hdlig-ness, high-ness, sour-ness, forgive-ness ; O.H.G. 
 hart-nissa (hardness), got-nissa (divinity) ; G. teug-nisz (testi- 
 mony), bild-nisz (image), biind-nisz (alliance, league), etc. 2 
 
 B. Adjectives. 
 
 (105) VII. E. -en, G. -en, further -em: adjectives of 
 material. — From a Preg. suff. -ina- = I.-E. -ino-, supra 89, 3. 
 
 VIII. E. -ish, G. -isch: adjectives of extraction and 
 attribution. — This element is alive in English, though it has 
 generally given place to the suff. -ly 3 or the borrowed French 
 formation (Americ-an, Pruss-ian). It is largely used in German 
 (Franzos-iscK), where it forms even learned terms (psycholog- 
 isch) or is added to words of Romance origin (Latein-isch, 
 Amerik-an-isch, Chin-es-isch, Annam-it-isch) . For its earlier 
 form, Preg. *-iska- = I.-E. *-isko-, see above, 91, 3. 
 
 IX. E. -y ( = 0.E. -ig), G. -ig : adjectives of qualifica- 
 
 1 It is more active in English than in German ; for it is used not only 
 with native words, but indifferently, with words borrowed from French (coy- 
 ne$*, F. coi " quiet ") and learned adjectives imported from Latin (acid-ness). 
 
 s In gefangnisz (gaol), if analysed ge-fang-n-isz (cf. the part, ge-fang-en 
 " prisoner "), we might perhaps find the original form of the suffix. 
 
 8 Contrast, for instance, heaven-ly with himml-isch, earthly with ird-isch, 
 etc. 
 
ENGLISH AND GERMAN DERIVATION. 177 
 
 tion. — Still living and quite common : E. heart-y, mood-y, heal- 
 ih-y, fil-th-y, silver-y ; added to borrowed words, greas-y, juic-y, 
 etc. ; G. herz-ig, muth-ig, zorn-ig, zott-ig, woll-ig, gold-ig, etc. ; in 
 compounds, vier-fihsz-ig , weit-schweif-ig .* Its origin is discussed 
 above, 91, 1. 
 
 X. O.E. -Hit (lost in E.), G. -icht: with the same meaning, 
 though not at all common. — It looks like a compound formed 
 by an exclusively Germanic combination of the I.-E. suffixes 
 -ko- (-iko-) and -to- : O.E. stcen-iht (stony), G. stein-icht, woll- 
 icht, thor-icht (foolish), etc. The adjective has become a sub- 
 stantive in kehr-icht (dirt), derived from O.H.G. *chara, which 
 is the nominal basis of the vb. cher-ian, G. kehr-en (to sweep) . 
 
 § 2. Verbal Suffixes. 
 
 (106) We have already stated that the whole E. and G. 
 verbal derivation now rests on the ancient Germanic suffix -jan, 
 which moreover has been blended with the whole verbal end- 
 ings -en and -on : that is to say, it rests on the ending -en for 
 the German infinitive, and on no ending at all for the English 
 infinitive. But this general derivation will appear clearer, if 
 we first get rid of some less productive formations, which de- 
 pend on the addition of the same final -jan to a prior nominal 
 suffix, or on some similar combination. 
 
 I. E. -n, -en (rather common), G. -n-en (much rarer) = Ger- 
 manic -n-jan and -n-an: mostly causative verbs. — In prin- 
 ciple, in English as well as in German, the simple addition of 
 suff. *-jan to a stem is sufficient to give the derivative com- 
 pound the meaning of causality which is originally involved in 
 this verbal suffix 2 : thus, from warm, we have the vb. to warm, 
 as in G. warm-en = Go. varm-jan ; from better, the vb. to better = 
 G. besser-n, etc., etc. But in consequence of English having 
 
 1 Such adjectives as gldub-ig, streit-ig, irr-ig, etc., which really proceed 
 from nouns, glaube, streit, irre, etc., being afterwards referred to the verbs 
 glaub-en, streit-en, irr-en, etc. , analogy subsequently produced er-biet-ig (the 
 normal form is er-bot-ig from the noun er-bot), aus-find-ig, zu-lass-ig, ab- 
 hdng-ig, etc., from er-biet-en, aus-Jind-en, zu-lass-en, ab-hang-en, and many 
 more. Cf. supra 99, 102 (3), and infra 110 (I. 2, II., III.), etc. All these 
 processes fall under the same head. 
 
 2 See the so-called causative and denominative verbs in Sk., Gr. and L., 
 supra 83 (IV.) and 92. 
 
 N 
 
178 ENGLISH AND GERMAN GRAMMAR. 
 
 dropped the final n, the causative verb was occasionally no 
 longer clearly distinguishable, either from the word whence it 
 proceeded, or, in particular, from an intransitive verb derived 
 from the latter, so that the speaker naturally preferred a more 
 characteristic derivative process. Now, Germanic had formed 
 from its past participles in -an-, as well as from other adjectives, 
 either causative or intransitive verbs, the former with an infini- 
 tive in -n-jan, the latter with an infinitive in -n-an. These 
 afterwards, by the loss of the j, became more or less blended 
 together, assuming however as a rule the causative meaning. 
 Let us consider therefore a root wak, Go. vb. wak-an (to 
 be awake) : from this root came a part. Preg. *wak-and-s 
 (awake), and, from this again, an intransitive vb. *yi-wak-an-an 
 (to become awake) and a causative vb. *^a-wak-an-jan (to make 
 awake) ; the O.E. correspondent a-wcec-n-an still means " to 
 become awake "; but the E. offspring to a-wak-en hardly means 
 anything else but "to make awake," in opposition to the purely 
 intransitive to a-wake, which would be Germanic *yi-wak-an (to 
 become awake). 1 
 
 After this final -en had been thus obtained, it was deemed 
 th.e specific exponent for causative verbs, and was used to 
 modify any adjective (deep-en, sharp-en, soft-en, rough-en, 
 straight-en),, and even some substantives (height-en, length-en, 
 strength-en), always with the same causative value, which how- 
 ever concurrently admits of the intransitive use (to ripen is " to 
 become" and "to make ripe," to sick-en, "to fall" and "to render 
 sick," etc.). In German, since the formative element *-jan 
 still remained perceptible enough, it was felt to be sufficiently 
 clear, so that the verbal derivation by -n-en seemed useless, and 
 was not developed, though we find here, as in English, such old 
 and legitimate types as ler-n-en (E. to learn), leug-n-en (to deny, 
 O.E. lyg-n-an " to call [a thing] a lie "), war-n-en (E. to war-n), 
 as opposed to lehr-en, lug-en, wahr-en. 
 
 1 The argument seems somewhat far-fetched, owing to the fact that the E. 
 final syllable is dropped. The reader need only be reminded, that an E. 
 infinitive ending without an n ought to have one, and that therefore an E. in- 
 finitive ending with an n ought to have two n's: hence, for instance, E. sharp- 
 en and G. schdrf-en, though outwardly alike, are etymologically different ; 
 the G. verb corresponding to E. to sharp-en would be * scharf-n-en. 
 
ENGLISH AND GERMAN DERIVATION. 179 
 
 II. E. -se = -is-an and -s-ian {supra 93): causative verbs. — 
 From an adjective hreinn (pure) O.N. has a verb hrein-sa (to 
 wash), whence E. (borrowed) to rinse. From clcene (clean) 
 O.E. likewise derives clcen-s-ian, cf. E. clean and to cleanse. The 
 types to clasp = M.E. clap-s-en, and tj grasp = *grap-s-eu (cf. to 
 grapp-le) fall under the same head, having undergone a meta- 
 thesis. 
 
 III. E. -le, -I, -er ; G. -eln, -em : frequentative verbs. — 
 This simple suffix consists of a nominal stem in -ila or -ira, to 
 which is added the ordinary verbal element. The diminutive 
 meaning of the nominal suffix -ila satisfactorily explains the 
 frequentative force of the verbal ending thus obtained. 1 In- 
 stances are extremely numerous : E. to hurt-le, to snuff -le, to 
 draw-l, etc., compare to hurt, snuff, to draw 2 ; G. lach-eln (to 
 smile), schiltt-eln (to shake), klug-eln (to affect wisdom), nds-eln 
 (to speak through the nose), as opposed to lach-en, schiitt-en, 3 
 Mug, nase ; E. to glimm-er, to glitter = G. glitz-em, to flutter, 
 contrasted with gleam, *glit (Go. glit-mun-jan "to shine "), to 
 float, etc. ; Gr. er-schihtt-ern (to shake), folg-ern (to draw a con- 
 sequence), zog-em (to balance, to protract, cf. ziehen "to draw," 
 pf. zog), and so forth. 4 
 
 IV. G. -zen = -t-jan (supra 93) : frequentative verbs. — A well 
 known and common formation : schluch-zen (to sob, cf. schluck-en 
 "to swallow"), krdch-zen, grun-zen, jauch-zen, dch-zen, etc. In 
 seuf-zen (to sigh) = O.H.G. suft-on (cf. E. to sob), the z has in- 
 truded (M.H.G. siuft-en and siuf-zen) after the analogy of other 
 verbs with a similar regular ending. In du-tzen (to thou) the 
 exponent was also originally frequentative, but has become 
 reduced to the ordinary verbal value. 5 
 
 1 Thus, "to drink small draughts" maybe "to drink often," and so 
 forth. 
 
 2 From spark, a frequentative to spark-le, the only derivation. The mean- 
 ing is merely intransitive in to kneel. 
 
 3 Still in the sixteenth century, er-schiitt-en, " to throw down, to stagger." 
 
 4 The type ver-stein-ern must not be confused with the frequentative for- 
 mation ; it contains simply the suffix of adjectives denoting material, supra 
 105. So also, E. to girdle is the verb formed on girdle, not the frequentative 
 of the simple to gird. The G. ending -eln occasionally claims a descent 
 from some other suffix : handeln (to deal with) is O.H.G. hant-alon. 
 
 5 By combination of suffixes III. and IV., we have G. bren-z-eln " to have 
 a burnt smell." 
 
180 ENGLISH AND GERMAN GRAMMAR. 
 
 (107) V. No exponent at all in E. ; in G. the simple 
 ending -en: ordinary verbal derivation. — This important 
 formation is the modern form of the Germanic suffix -jan and 
 its equivalents, and may therefore form indifferently intran- 
 sitive, transitive and causative verbs. It is impossible here to 
 go into details, but the general features of the formation may 
 be shortly summarised. 
 
 1. Firstly, in form, it admits naturally of the regular meta- 
 phony (G. scharf scharf-en, *queck (keck) er-quicken " to 
 vivify," los Ids-en "to unbind," voll full-en, brut brut-en), but 
 this again may be defaced by analogy 1 : from gold, we have E. 
 to gild, whereas G. ver-giild-en is replaced by ver-gold-en ; and 
 vice versa, G. warm-en, but E. to warm. In fact, the O.E. 
 metaphony has become quite extinct : English now derives all 
 its verbs without any metaphony, whilst in German the pro- 
 cess, being later, still survives. 
 
 2. The G. derivation, as a matter of course, admits of every 
 verbal prefix which has been described above : be-fremd-en 
 er-hoh-en, ent-blosz-en, ver-dunkel-n, etc. On the contrary, in 
 E., where the use of prefixes has been lost 2 and the verbal 
 final dropped, the derivative verb became identical with its 
 derivative base, whatever the latter might be. After the 
 analogy of such cases was developed an extremely simple 
 derivation of verbs, which is peculiar to English, namely, the 
 power of forming a verb from any word by merely conjugating 
 it as it stood : to ink a pen, to pen a word, to word a thing, to 
 boycott a man ; and so forth. English may be said to have 
 thus regained the simplicity of structure of the so-called 
 
 1 No mutation being possible from still to still-en (to quiet), from siisz 
 (sweet) to ver-siisz-en, etc., the language was led to create, also without 
 mutation, trockn-en (to dry), er-starr-en (to stiffen), er-blass-en (to turn pale), 
 ver-dumpf-en, ver-faul-en, etc. Even German dialects are at variance in this 
 respect : thus niitz-en (to use) is Alamannic nutz-en. It has been stated 
 besides that the derivation by -jan was confounded with the derivation by 
 -en and -on, and the latter naturally do not require any mutation (supra 
 92-93). For English metaphony, compare full and to fill, doom and to 
 deem, brood and to breed, food and to feed, etc. 
 
 2 Apart from the inversive un- (supra 98, III.), and a number of Eomance 
 prefixes, to en-able, to dis-own, etc. 
 
ENGLISH AND GERMAN DERIVATION. 181 
 
 primitive languages, where nearly every word may play the 
 part either of a noun or a verb. 
 
 3. The function, as we have said, varies widely. Passive 
 intransitives : E. to ache, G. Jieber-n (to be in a fever). Active 
 intransitives : E. to $hip = G. schif-eu. Simple transitives: E. 
 to love = G. lieb-en. Causative transitives: G. schwdch-en, 
 schwdrz-en, krtink-en. Sometimes the meanings are cumulated : 
 E. to ripe = G. reif-en (to become ripe and to cause to ripen). 
 If more precision is wanted in marking the causative sense, 
 German uses its prefixes be-, ver-, and ent- (inversive), whilst 
 English uses its suffix -en. It is chiefly a matter of usage. 
 
 4. G. -ig-en. — A most important development of the causative 
 derivation remains peculiar to German. Such verbs as M.H.G. 
 eineg-en, noteg-en, schedeg-en, schuldeg-en, G. einig-en, nothig-en, 
 schddig-en, schuldig-en, krdftig-en, be-mdchtig-en, — which are 
 quite regularly derived from the adjectives einec, notec, schadec, 
 schuldec, respectively krdftig, mdchtig, etc., — having been erro- 
 neously traced back to the simple words ein, noth, schade, schuld, 
 kraft, macht, etc., the speaker formed, in the same way, from 
 ende and theil, for instance, the verbs end-ig-en and be-theil-ig-en, 
 though no adjectives *end-ig and *theil-ig actually exist. From 
 this false analogy has sprung the causative suffix -igen, which 
 is widely diffused : be-fehl-igen, be-schdft-igen, ver-ein-igen, ge- 
 nehm-igen, be-schon-igen, etc. 
 
 VI. G. -ieren, -iren. — It is only on account of the wide use of 
 this element, that it may be held to deserve a place in our list ; 
 for it is of Romance origin, though modified by the general G. 
 infinitive ending : G. spazier-en = L. spatidri (to walk) . It 
 characterizes chiefly borrowed verbs (rasonn-iren), or such as 
 are derived from borrowed words (stud-iren) . But, at the time 
 when it became general under the influence of the French 
 language (XVIIth-XVIIIth century), it was even attached to 
 some Germanic stems, as in stolz-iren (to strut), schatt-iren (to 
 shadow), buchstab-iren (to spell). 
 
18'2 ENGLISH AND GERMAN GRAMMAR. 
 
 Section III. 
 
 OLD WORDS CHANGED TO SUFFIXES. 
 
 (108) This subject has been seen above to belong properly 
 to the study of composition, inasmuch as, for instance, child- 
 hood and kind-heit, if referred to their origin, are no less truly 
 compound nouns than child-birth and Jcind-bett. The only 
 difference lies in the fact that birth and bett are still extant in 
 their respective languages, while *hood and *heit have long 
 been obsolete. Now, this difference, though theoretically quite 
 insignificant, is important in practice ; for, since these final 
 syllables have no longer any meaning by themselves and are 
 merely used as a vague derivative exponent, they have gradu- 
 ally acquired various significations, and have moreover under- 
 gone slight alterations which would have been prevented if the 
 isolated word had survived. 1 Therefore this peculiar Anglo- 
 German derivation, which might be termed derivative 
 composition, is well fitted to mark a transition from the 
 matters previously dealt with to the following subject. 
 
 The old words changed to derivative suffixes were either 
 nouns or adjectives, so that the result of derivative composition 
 is either a noun or an adjective. Some adjectives, more- 
 over, here and there became adverbial locutions. 
 
 § 1. Nouns. 
 
 (109) I. E. -hood and (if unaccented) -head = 0.'E. -had; G. 
 -heit = *hait = Go. hai-du-s (manner) : abstract nouns. — From 
 an I.-E. root qi (to see, to remark, to discern, Sk. ci) proceeded 
 an I.-E. stem *qoy-tu-, Sk. Tce-tu-s (sign), Preg. *hai-Su-s (sign, 
 quality, manner) : thus O.H.G. chint-heit means simply " the 
 quality, the state of a child," and so with any similar words, 
 English or German. 
 
 1. The E. form -hood is common, whereas the form -head is 
 rare: maiden-hood and maiden-head = O.E. oncegden-hdd. It is 
 added to borrowed as well as native words : O.E. biscop-hdd, E. 
 
 1 Cf. svpra 68. 
 
ENGLISH AND GEEMAN DEBIVATION. 183 
 
 priest-hood. But it forms scarcely any derivatives except from 
 substantives : brother-hood, neighbour-hood, man-hood, etc., 
 though we have false-hood. In German, on the contrary, -heit 
 may be added to adjectives as well as to substantives, forming 
 abstract nouns from both classes, indeed, the derivation from 
 adjectives is now far commoner: O.H.G. magat-heit, w,an-heit, 
 etc.; Mod. G. schon-heit, frei-heit, bos-heit (wickedness) = *bosi- 
 hait, 1 etc. 
 
 2. Whenever the syllable -heit was attached to an adjective 
 ending in -ec (Mod. G. -ig), which was a very common case in 
 M.H.G., the whole became a complex -cheit : thus, from milt-ec 
 (piteous), the abstract noun milt-ec-heit. Now, beside the 
 derivative milt-ec, there existed a simpler adjective milte (Mod. 
 G. mild = ~E. mild), and this, contrasted with miltecheit, natur- 
 ally created the false suffix -cheit, which was then used in 
 other cases : eitel-keit, lust-ig-Jceit, dank-bar -keit, freund-lich-keit, 
 etc. Both suffixes -heit and -keit have the same meaning as 
 well as the same origin, and their employment is merely a 
 matter of usage. 2 
 
 II. E. -dom = 0,~E. -dom; G. -thum (-turn) =O.H.G. -tuom : 
 abstract nouns (occasionally concrete nouns, as E. king-dom, 
 G. herzog-thum). — The word is the same as E. doom, which has 
 been seen to come from a root meaning "to do," cf. E. do, G. 
 thu-n. It forms in both languages several derivatives either 
 from nouns or adjectives: O.E. biscop-dom, hdlig-dom; E. earl- 
 dom, free-dom, wis-dom ; O.H.G. meistar-tuom, fri-tuom; G. 
 bis-thum (syncopated), heilig-thum, etc. It is a living suffix, 
 especially in German, where such formations as junkertum, 
 brahmanentum are still current. 
 
 1 The vowel was syncopated early enough not to have occasioned meta- 
 phony, cf. bo.ie = O.H.G. bosi. — In derivation from nouns, G. gave the 
 preference to suff. -schaft (compare bruder-schaft with brother-hood), while, 
 in derivation from adjectives, E. usually used suff. -dom (compare freedom 
 with freiheit, wisdom with weisheit). — Observe the variety of resources dis- 
 played : for the abstract noun from " high," E. and G. together have four 
 different derivatives (E. height and highness, G. hbhe and hoheit), of course 
 with various shades of meaning. 
 
 2 Again, from a further combination of -ig and -keit, arose a suff. -igkeit 
 (neu-igkeit, klein-igkeit, ge.nau-igkeit), which spread in the same way as 
 verbal -igen, supra 107 (V. 4). 
 
184 ENGLISH AND GERMAN GRAMMAR. 
 
 III. E. -ship = O.B. -scipe; G. -schaft = O.H.G. -shaft i 
 abstract nouns. — The two words, though not identical, both 
 mean " shape, form," inasmuch as they have sprung, by 
 different suffixes, from the root which produced to shape and 
 schaffen. Besides, O.H.G. still possesses a simpler suff. -skaf= 
 Preg. *skap-i-, which is even nearer to the E. form : O.H.G. 
 lant-skaf=~E. land-scape, the latter borrowed from Dutch, but 
 M.E. land-skip. The formation is common in both languages : 
 E. friend-ship, town-ship, worship — * 'worth-ship, court-ship; G. 
 freund-schaft, diener-schaft, 1 gesell-schaft, herr-schaft, burger - 
 schaft, etc. Exceptionally a few derivatives from adjectives : 
 gemein-schaft, eigen-schaft (" peculiar quality," contrast the 
 meaning with that of eigen-tum, " property, estate "). 
 
 IV. E. -red in kind-red (cf. kin) and hat-red (cf. to hate) 
 comes from the same root as E. read-y and G. be-reit. It 
 occurs rather often in O.E., but is nearly extinct in later times, 
 and seems never to have existed in German. As to E. hnnd-red 
 = G. hund-ert, see the chapter on Numeration. 
 
 V. E. -lock, in wed-lock, and -ledge, in knoiv-ledge, are two 
 forms (the latter being Scandinavian) of the same suffix of 
 obscure origin (O.E. -lac, M.E. -loc), which was once more 
 widely used and has now become obsolete. 
 
 VI. E. -ric, in bishop-ric, is O.E. ric-e = Go. reik-i, still exist- 
 ing as a separate word in G. reich (kingdom), cf. the compound 
 kaiser-reich (empire) . 
 
 § 2. Adjectives. 
 
 (no) I. E. -Zt/ = O.E. -lie, shortened (as unaccented) from 
 Zic«=E. like 2 ; G. -lich, shortened (as unaccented) from O.H.G. 
 -llh = Go. -leik-s ; cf . the derivative g-leich (alike) = Go. ga-leik-s : 
 adjectives denoting resemblance, manner or attribute. 
 — This class is extremely important. 
 
 1 . It is easy to see, at once, the formation of such adjectival 
 compounds as E. man-ly = G. mdnn-lich, namely "man-like"; 
 
 1 Acquired a concrete sense, and so also the following words : cf . E. her 
 ladyship, F. la domesticity (the servants), etc. 
 
 a The full form is to be seen in modern compounds such as court-like. 
 The original sense is " body, substance," cf. G. leiche (corpse). 
 
ENGLISH AND GERMAN DERIVATION. 185 
 
 so also, woman-ly and weib-lich, father-ly and vater-lich, hing-ly 
 and konig-lich ; again, when added to names of things or to 
 abstract nouns, ivorld-ly = weltlich, night-ly = ndcht-lich, hof-lich, 
 hdss-lich, lieb-lich, anfdng-lich, the latter type greatly developed 
 in German. 
 
 2. Secondly, these adjectives happening in German to be 
 referred to verbs instead of nouns (thus, to hass-en, lieb-en, 
 anfang-en, etc.), it was only natural that new ones should be 
 formed directly by derivation from other verbs, whence be- 
 greifen, empfind-en, beschreib-en, according to a process which 
 must now have become familiar to the student, produced such 
 adjectives as begreif-lich, empfind-Uch, un-beschreib-lich. 
 
 3. It is clear, from its original meaning, that suff. -lich might 
 as well modify an adjective as a noun, either by lessening in 
 some degree the adjectival sense (r'dth-lich, " red-like, reddish "), 
 or even without any definite value (reich-lich, " rich, plentiful "). 
 This is the case with E. silly = O.E. syl-lic, supra 23 E. 
 
 4. In order to understand the further evolution of the suffix, 
 it must be borne in mind that the West Germanic adverb is 
 merely a certain case — probably the instrumental case — of the 
 adjective, its ending being an -e, which had a tendency to 
 disappear : e.g. Go. hlut-r-s (pure, cf . L. lau-tu-s " washed ") 
 and O.E. hliitor, whence O.E. hlutr-e, O.H.G. luttar and G. 
 lauter (purely, merely), 1 etc. To the M.H.G. adjectives fro-lich 
 (gay), grd^-lich (great), etc., naturally corresponded the adverbs 
 
 fro-llch-e (gaily), gro^-llch-e (greatly), etc. Now, these forms, 
 if directly opposed to the simple adjectives froh, grosz, etc., 
 contained an ending -llche -lich, which was mistaken for the 
 characteristic of the adverbial function and transferred else- 
 where with this value by an obvious analogy. This process 
 was carried on with great energy in M.H.G., and, though now 
 extinct, has left some traces in the contemporary speech, several 
 formations with suffix -lich being even now exclusively used 
 with an adverbial meaning: frei-lich (to be sure), hiirz-lich (in 
 short), schwer-lich, hoffent-lich. 
 
 1 Final -e kept in G. lang-e (a long time), contrasted with the adjective 
 lang. 
 
186 ENGLISH AND GERMAN GRAMMAR. 
 
 5. English has gone the same way, but farther, and quite 
 consistently : E. -ly is no longer an adjectival suffix, except 
 •when added to a noun ; if added to an adjective, it forms an 
 adverb. Thus, the type reich-lich is essentially an adjective 
 in G., while its E. representative rich-ly can be nothing but 
 an adverb. This transformation gave English a specific 
 adverbial exponent, whereas German usually employs as such 
 the old instrumental case, now represented by the undeclined 
 adjective, as final -e has dropped long ago. The E. suff. -ly 
 modifies indifferently any adjective, whether primitive or 
 deinvative (wise-ly, idly = *id-le-ly = G. *eit-el-lich, form-er-ly, 
 sorrow-ful-ly), native or borrowed (veri-ly, 1 vacant-ly, glorious-ly) , 
 and even some nouns (night-ly, name-ly, purpose-ly) . 2 
 
 II. E. -som -some = O.E. -sum, the unaccented form of E. 
 same; G. -sam : adjectives denoting qualification or 
 aptitude. — The word is lost in German, but may be found in 
 O.N. scem-r (able) and traced back to Sk. sam-d-s — Gr. 6/x-o-s 
 (alike, same, cf. L. sim-ili-s). Its meaning being very nearly 
 that of the preceding word, both were equally used in deriving 
 an adjective from a noun or an adjective : Go. lustu-sam-s 
 (lustful) ; O.E. hyr-sum (obedient), long-sum (tiresome) ; 
 O.H.G. gihor-sam (from a noun gihor, "hearing, obedience"), 
 ein-sam (united), heil-sam (wholesome) ; E. hand-some (actually 
 "clever"), win-some, trouble-some, wearisome; G. muh-sam, 
 arbeit-sam, duld-sam, etc. Further, analogically, some adjec- 
 tives were immediately derived from verbs : folg-sam, erfind- 
 sam, bieg-sam, 3 etc. 
 
 III. G. -bar = *-bdr-i, from the same root as the vb. ber-an 
 (to carry), and with the same meaning as Sk. bhdr-d-s, Gr. <f>op- 
 6-<s, L. -fer&t the end of a compound: adjectives denoting 
 
 1 E. rert/ = O.F. verai, now vrai (true). The native adverb is sooth-ly = 
 O.E. soff-lice, wherein * 'sod '«=* sonff is the present participle of the vb. I am 
 (cf. Gr. &v 6vt-os, L. sons sont-is "culprit"). — Moreover, E. still exhibits 
 some adjectives used adverbially : he works hard, he speaks loud, pretty nigh, 
 and very itself. 
 
 2 So, to conclude, like-ly contains the same word twice over. 
 
 8 In buxom = 0.lt. buxum=*buh-sum, the vocalism shows a derivation 
 from a noun (cf. bough = OJZ. bog), whereas bieg-sam is formed from bieg-en 
 (to bend). 
 
ENGLISH AND GERMAN DERIVATION. 187 
 
 production and consequently aptitude. — The etymological 
 sense is still clear in such types as O.E. leoht-bcere (L. luci-fer 
 " carrying light ") and G. frucht-bar " bearing fruit," and it is 
 but slightly distorted in furcht-bar, wunder-bar, gang-bar, dank- 
 bar, etc. The suffix is lost in English, but has reached in 
 German a high degree of development, thanks to the analogical 
 process by which it was directly added to many verbal stems : 
 thus, streit-bar, really a derivative from streit, but erroneously 
 referred to streit-en, led to the creation of esz-bar, les-bar, nenn- 
 bar, annehm-bar, erreich-bar, and many others, the power of 
 creation being as yet by no means exhausted. 
 
 IV. G. -haft = Preg. *haf-td-s=h. cap-tu-s (taken), literally 
 " seized by," whence " endowed with," the exact force of the 
 Mod. G. derivative be-haft-et (no trace of it either in E. or 
 O.E.) : adjectives of qualification. — As required by its 
 etymology, this class is rich in derivatives from nouns (herz- 
 haft, tugend-haft, fehler-haft), but poor in derivatives from 
 adjectives (wahr-haft " sincere," 1 bos-haft), and much poorer 
 still in derivatives from verbs (schwatz-haft " prattler "). 
 
 V. E. -fast = O.E. -fast (the same as G. fest = O.B..G. fest-i, 
 " fast, firm "), a suffix resembling the preceding in function, in 
 O.E. hils-fcest, " having a house," kept only in stead-fast, literally 
 " having a stead, a foundation." 2 
 
 VI. E. -/oZd=O.E. -feald; G. -fait = Go. faljj-s: adjectives 
 of multiplication. — For the etymology, compare E. fold and 
 G. falte with L. sim-plec-s " having but one fold," duplex (two- 
 fold), etc. E. two-fold, twenti-fold, etc. ; Go. din-falp-s (simple,, 
 good, stupid), M.H.G. ein-falt (id.), whence the derivatives, 
 Go. fm. din-falj>-ei, O.H.G. ein-falt-l and G. ein-falt (stupidity), 
 O.H.G. ein-falt-ig, M.H.G. ein-velt-ec and G. ein-falt-ig, etc. 
 
 VII. E. -ful: adjectives of qualification.— Well-known 
 to be the E. full = G. voll. But whilst, in such a G. compound 
 as bewunderungs-voll (full of astonishment), the meaning of the 
 word voll still remains present to the speaker's mind, the case 
 
 1 Might be translated etymologically "endowed with truth," the adjective 
 wahr (true) being used as a substantive. 
 
 2 O.E. sceam-fcBst, M.E. sham-fast is thus derived, but has been changed 
 by popular etymology to shame-faced. 
 
188 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 is quite different with the element -ful in thank-ful, aw-ful, 
 wonder-ful, sorrow-ful, merci-ful, fanci-ful, etc. Even the spell- 
 ing emphasizes the difference : the syllable is unaccented, and, 
 being now added to any noun whatever, simply plays the part 
 of a convenient but vague and commonplace formative ending. 
 VIII. E. -less, unaccented form of O.E. leas (loose, void, 
 false); G. -los = O.H.G. lds = Go. Idus (void) : negative of the 
 preceding suffix. — G. los still exists as a single word. 1 Its 
 value as a suffix is quite clear: father-less = vater-los, life-less = 
 leb-los, god-less = gott-los, forming a negative adjective from 
 almost any noun. 
 
 § 3. Adverbs. 
 
 (in) Adverbs, which fall under this head of derivation, 
 are simply ancient cases of the original declension of nouns 
 or adjectives. The case, if it happened to be accusative or 
 instrumental, is no longer to be discerned at first sight, since 
 its ending has been dropped 2 ; but, if a genitive, it shows the 
 characteristic -s. Now West Germanic was very rich in 
 adverbial locutions in which the genitive was used, and the 
 later languages, especially German, still preserve a considerable 
 number of such expressions. We need but mention: O.E. 
 ■dceg-es " by day," niht-es " by night," lost in E. ; E. else (cf. 
 L. al-iu-s "other"), needs, after-wards, al-ways, 3 sin-ce = M.E. 
 .slth-ens (cf. O.H.G. sld, G. seit), on-ce, twi-ce* etc. ; G. tags, 
 nacht-s, falls and jeden-falls, rings, recht-s, unter-wegs (by the 
 way), augen-blicks (suddenly), and more complex constructions 
 like gerad-es weg-es, recht-er hand, all-es ernst-es, mein-es eracht- 
 ens, rein-es herz-ens, hungers sterben (to starve), etc., etc. Thus 
 the adverbial use of nouns or adjectives in this way is easily 
 explained. 
 
 I. E. -ly ( = G. -lich) : the ordinary adverbial suffix, supra 
 110, still adjectival in German. 
 
 1 And so also E. loose, but that is borrowed from Scandinavian. 
 
 2 See Declension, infra 149. 
 
 8 For this word, as historically traced, does not mean " by all ways," but 
 *' the whole way." 
 
 4 Cf. Numeration, infra 124, II. 
 
ENGLISH AND GERMAN DERIVATION. 189 
 
 II. E. -wise: a few adverbs, like-wise, no-wise. — It is the 
 noun Mn'se = O.E. wise = G. weise (manner), afterwards in French 
 written guise. German has the genitival locution, e.g. merk- 
 wiirdig-er-weise (remarkably), and a few modern compounds, as 
 theil-weise (partly), kreuz-weise, paar-weise. 
 
 III. E. -ling, G. -ling-s (gen.) : a few adverbs. — This suffix 
 seems to come from the gen. pi. of nouns in -ing * used adver- 
 bially : M.E. hed-l-ing, corrupted to head-long ; E. dark-ling, 
 side-ling and side-long, etc. ; Gr. blind-ling-s, seit-lings, ruck-ling-s 
 (backwards), schritt-ling-s (step by step), etc. 
 
 IV. E. -ivard (also adjectival) and -wards (advb. gen.) ; 
 G. -wart-s (id.) : adverbs of direction. — The O.H.Gf. form is 
 •wert-es, the genitive of an adjective which is Go. nomin. sg. 
 vair}>-s, from the same root that may be seen in the Go. vb. 
 vair}>-an (G. werden "to become ")=L. vert-ere (to turn) : E. 
 to-ward and to-ivard-s, back-ward, after-ward-s, sea-ward, lee-ward, 
 wind-ward, awk-ward (literally " turned awry "), etc. ; G. vor- 
 wart-s, riick-wcirt-s, auj '-wdrt-s; and, in further derivation, E. in- 
 ward-ly, G. wider-wart-ig (contrary), etc. 
 
 The remaining E. and G. derivations are exceptional and 
 may be neglected here, especially as the reader will find an 
 easy explanation for each of them in some one of the processes 
 described above. 
 
 1 Cf. the formation of nouns in -ung, -ing, -ling, supra 103. Of course in 
 G. the gen. is in the sg. 
 
CHAPTER III. 
 COMPOSITION. 
 
 (112) The process of compounding words with one 
 another is Indo-European : remarkably well preserved, and 
 even developed, in Sk. and Gr., less familiar in Latin, and 
 almost lost in the Romance languages, it was bequeathed to 
 Pregermanic, where it remained in full vigour, and is still 
 extremely active in English and German. This may be said 
 to be one of the most important points in which the Germanic 
 languages differ from the neighbouring Romance tongues, and 
 especially from French : such relations as the latter expresses 
 by a periphrasis (nuit d'ete) or a derivation (pomm-ier) are as 
 a rule, and clearly enough, given in English and German by 
 the mere juxtaposition of two stems, the first usually limiting 
 or qualifying the second : summer-night, sommer-nacht ; apple- 
 tree, apfel-baum. 
 
 The details of English and German composition belong to 
 ordinary grammar, or else must be sought in an etymological 
 dictionary. Here we have only to sketch the general outlines 
 of the classification and formation of those compound words, 
 which are found in the two languages as well as in the other 
 members of the I.-E. family. 1 
 
 Section I. 
 
 CLASSIFICATION OF COMPOUNDS. 
 
 (113) Compound words must be regarded in a double light, 
 according as we consider their form, or their meaning, the 
 
 1 Cf. Henry, Grammar of Gr. and Lat., 175-181. 
 
 190 
 
COMPOSITION. 191 
 
 former being usually quite independent of the latter, and vice 
 versa : thus, G. hund-fliege and hund-s-fliege (horse-fly) are 
 exactly synonymous, though their principle of formation is 
 quite different ; whereas B. red-lead and red-breast, though 
 identical in formation, widely differ in sense, the former merely 
 meaning " a red lead," and the latter, " having a red breast," a 
 distinction explained below. 
 
 § 1. Grammatical Classification. 
 
 (114) As to their grammatical formation, compounds, 
 whether I.-E., or English, or German, must be divided into 
 two classes, namely : syntactical and non-syntactical com- 
 pounds. 
 
 I. Syntactical composition occurs when two words assume 
 in respect to each other the form required by the ordinary 
 rules of grammar and syntax, e.g. the type, E. king's wife, G. 
 kbnig-s-sohn, where the first term appears in the genitive as 
 governed by the second. Such a locution, indeed, cannot be 
 termed a true composition, and strict logic would even banish 
 from this chapter the study of all syntactical compounds. They 
 only claim our notice from the fact that they influenced, chiefly 
 in German, non-syntactical composition, in such a way as to 
 render the study of the latter impossible to one who had not 
 considered the essential types of syntactical construction from 
 which these analogies arose. 
 
 1. The first term stands in the genitive and is governed by 
 the second term. 1 
 
 A. Genitive singular. — (a) Of so-called strong nouns (gen. 
 in -s). E. Tue-s-day and G. Diens-tag = zio-s-tac have already 
 occurred. Further instances are: E. Thur-s-day = 0. N. }>or-s- 
 dag-r (cf. the name of the god Thor), and G. Donner-s-tag, cf. 
 donner = thunder ; E. kin, and kin-s-man; E. daisy = * day s eye, 
 i.e. " [little] sun," the ending shortened because unaccented ; 
 O.H.G. gotes-hus and G. Gottes-haus, but Go. gud-hus (temple) 
 non-syntactical ; O.H.G. hundes-fliuga and G. hunds-fliege 
 
 1 We need scarcely observe that a compound may contain more than two 
 terms ; theoretically, however, the number of terms is wholly indifferent, the 
 rales being always the same for the relation of one to another. 
 
192 ENGLISH AND GERMAN GRAMMAR. 
 
 (horse-fly), but also hunt-fliuga and hund-fliege non-syntactical ; 
 G. land-s-mann (compatriot), bnt non-syntactical land-mann 
 (peasant), 1 etc. English possesses and still may create many 
 juxtapositions of this kind, though they are usually spelled in 
 two separate words, as {mid-summer) night's dream, new years 
 day, love's labour, etc., 2 having, in fact, kept here, and almost 
 nowhere else, the use of the genitival s for nouns which do not 
 denote living beings. 
 
 (6) Of so-called weak nouns (gen. in -en, -n). — This is the 
 type: O.H.G. ohs-in-zunga (the name of a plant), G. ochs-en- 
 zunge; O.E. el-n-boga, literally "arm's bow," also el-boga, E. 
 el-bow, which might represent either form, bnt Gr. elle-n-bogen 
 syntactical. English, having dropped its finals, has generally 
 lost this formation : thus, it answers ochs-en-zunge with the 
 seemingly non- syntactical Ox-ford ; cf . however the compound 
 maid-en speech. But in German, the process has been favoured 
 by the analogy of the composition with the genitive plural and 
 partly confounded with it. 3 
 
 B. Genitive plural. — (a) Of so-called strong neuter nouns 
 and strong adjectives (in -er) : G. all-er-erst, all-er-hochst, and 
 the like ; G. rind-er-schaar (a flock of oxen), rind-er-hirte, etc. ; 
 no further development. 
 
 (6) Of so-called weak nouns (in -en, -n) : O.H.G. Franch-ono- 
 lant and G. Frank-en-land, etc. ; M.H.G. vrouw-en-zimmer and 
 G. frau-en-zimmer, literally " women's hall," whence " female 
 attendance," and lastly, with concrete meaning (the only one 
 surviving), " waiting gentlewoman, young lady." 
 
 2. The first term governs the second term, usually in the 
 accusative, very seldom in any other case, occasionally with a 
 preposition between the two : an exceptional type. — G. stbre-n- 
 fried (trouble-feast), wherein the n is the syncopated form of 
 
 1 Compare likewise wasser-noth (scarcity of water) with wassers-noth (over- 
 flowing). 
 
 8 Spelled in one word, coxcomb <= cock's comb. 
 
 8 Infra B b. — Confounded in form, not in meaning; for it would be a gross 
 mistake, in most cases, to explain it by a plural instead of a singular in the 
 first term. The supposition is true for Frank-en-land, and may be true for 
 mensch-en-freund, biene-n-korb, etc. ; but it is obviously false for the type 
 todt-en-kopf, lind-en-baum, tinte-n-fasz, which has been indefinitely multi- 
 plied. 
 
COMPOSITION. 193 
 
 the accusative den, literally "trouble the peace," the noun, 
 however, showing no case ending ; G. vergisz-mein-nicht " forget 
 me not " (mouse-ear), which is a sentence rather than a word 1 ; 
 G. spring-ins-feld (light-headed youngster), etc. In English, 
 the general loss of unaccented endings does not allow us to de- 
 termine whether the similar type Shakespeare 2 ever possessed 
 an accusative ending. 
 
 II. The composition is said to be non- syntactical, when the 
 relation of meaning which nnites the terms, whatever this 
 relation may be, — namely, appositional (E. to ful-fil, G. still- 
 schweigen), possessive (E. hus-band, G. haus-herr), adnominal 
 (E. church-yard, G. hof-thor), locative or temporal (E. nightingale 
 = O.E. nihte-gale, G. nachti-gall, "sounding by night"), instru- 
 mental (E. steam-boat and god-father, G. dampf-schiff and hand- 
 haben), etc., etc., — lacks any peculiar exponent, and is denoted 
 by mere juxtaposition of the two terms, without the help 
 of a grammatical ending between them. This is true primitive 
 composition and would still be, as noted above, the only true 
 English and German composition, had it not undergone a 
 number of alterations caused by the intrusion of syntactical 
 types. Sometimes, however, though rarely, a composition is 
 only apparently non-syntactical, having sprung from some old 
 syntactical construction which has been misunderstood and 
 consequently transferred to places where it had no business to 
 be : thus, an O.H.G. sentence sie sind ein anderen ungelih* 
 which taken by itself is grammatically clear and correct, gave 
 rise to a false compound ein-ander, which was afterwards invari- 
 ably used, in any sentence or position, with a meaning of re- 
 ciprocal action. 4 
 
 § 2. Functional Classification. 
 
 (115) When the function is considered, compounds may be 
 divided into three great classes, namely : copulative, deter- 
 minative and possessive compounds. 
 
 1 The vb. vergessen formerly governed the genitive. Cf. infra 164, I. 3. 
 
 2 " He who shakes his spear," cf. the proper name Make-peace, etc., infra 
 117, in. 2. 
 
 ■ Literally " they are, one to other, unlike." 
 4 Cf. the simple juxtaposition in E. one another. 
 
 O 
 
194 ENGLISH AND GERMAN GRAMMAR. 
 
 I. The copulative is a compound, the meaning of which 
 would require between its two terms, if separated and gram- 
 matically construed, an inserted particle " and." The type is 
 rather rare : it appears in numerals (E. four-teen, G. funf-zehn), 
 in some proper names (E. Griqua-namaqua-land, G. Oestreich- 
 JJngarn) and in a few double adjectives (G. taub-stumm, but E. 
 deaf and dumb). In Sk., of course, it has multiplied, but may 
 be neglected elsewhere. 1 
 
 II. The determinative compound is equivalent to a locution 
 wherein either of the terms, and usually the second, would 
 govern the other and require it to take some case-ending. Ac- 
 cording as the governed term would be put in the same case as 
 the governing one, or in a different case, we have to distinguish 
 respectively attributive or appositive, and dependent com- 
 pounds. 
 
 1. Appositive compounds : Gr. type /icya\o'-7roXi-v (great 
 city). — E. wo-man = M.E. wim-man (compare the pronunciation 
 of the pi. women) =O.E. wlf-man, as it were "human being 
 of female sex." E. gospel = O.E. god-spel (god- shortened 
 instead of god), originally "good spell, good news," a literal 
 translation of Gr. cu-ayye'Aio-v. E. wer-wolf, wal-nut M nut from 
 Welsh (non-German) land," New-haven, pea-cock, she-wolf, hen- 
 canary-bird, etc. — G. grosz-vater, and, curiously enough, an 
 analogical formation grosz-sohn ; so also, E. grand-son from 
 grand-father. G. all-od, a word borrowed from Low L. allodium, 
 but this again borrowed from O.H.G. al-od, "whole or full 
 property." G. silsz-holz (liquorice) , wer-wolf ', filrst-bischof, Herr- 
 gott, Christ-kind, Neu-stadt, etc. 
 
 2. Dependent compounds : Gr. type av8p-a8cX<£o-s (husband's 
 brother, brother-in-law), L. luci-fer, E. bride-groom, G. pelz- 
 waar en-handler, etc. — Instances will occur to the reader in such 
 considerable number, that it would be superfluous to quote any 
 more ; indeed, the student may even form any new compounds 
 
 1 Compounds made of two adjectives are almost always determinative 
 compounds : bo roth-gelb does not mean " red and yellow," but " yellow with 
 a tinge of red," and so forth. In bitter-siisz and a few more, we may hesitate 
 between the two shades of meaning. — A true O.H.G. copulative compound 
 may be found in the fourth line of the Ilildebrandslied : sunufatarungo " the 
 son and the father." 
 
COMPOSITION. 195 
 
 he chooses, for the process is still alive in hoth languages, and 
 the hulk of their compound words are of very recent forma- 
 tion. 1 
 
 III. The possessive compound is only a form of the deter- 
 minative, inasmuch as here too one of the terms determines the 
 other ; bat the whole adds an additional idea to the meaning 
 of the two words separately : thus, Grr. p 080- SolktvXo-s and E. 
 red-breast do not mean simply, "rose-finger" or "red breast," 
 but " endowed with rose-fingers," " (a bird) showing a red 
 breast," and so forth. In short, the possessive compound is 
 always an epithet, which involves the existence of a subject 
 possessing the character or realizing the idea expressed by the 
 union of the two terms : E. light-foot, heart's ease, wag-tail, etc. ; 
 G. Roth-hdppchen, drei-fusz, lihgen-zunge, plage-geist, etc. This 
 composition is of essentially popular origin and goes hand in 
 hand with the preceding. 
 
 Section II. 
 
 FORMATION OF COMPOUNDS. 
 
 (116) The power of composition in either language is 
 theoretically unlimited. 2 But since, however numerous they 
 may be, the successive terms qualify one another in turn, any 
 compound may theoretically be reduced to two terms, as it 
 were, the first and the second, each again being either the 
 qualifying or the qualified term. 
 
 1 Further, the reader may be referred to supra 65, 5, and infra 117, I., II. 
 and III. 1. 
 
 * Three terms : E. pine-apple-juice ; G. haupt-zoll-amt, dudel-sack-pfeifer. 
 Four terms : E. midsummer night's dream ; G. eisen-bahn-liof-strasse, kupfer- 
 schmied-werk-zeug (here, instead of determining one another, the four terms 
 are divided into two groups, and the first group determines the second, as, in 
 eaoh group, the first term determines the other ; that is to say, kupferschmied 
 and werkzeug bear exactly the same relation as if each of them were a simple 
 word ; the case is very common). Five terms : alt-milch-ferkel-markt-platz 
 (where milchferkel is one word in relation to markt, and milchferkelmarkt one 
 word in relation to alt), etc. Humorous accumulations : the United Metro- 
 politan Improved Hot Muffin and Crumpet Baking and Punctual Delivery 
 Company (Dickens, Nickleby). 
 
196 ENGLISH AND GERMAN GRAMMAR. 
 
 § 1. Form of the First Term. 
 
 (117) The first term of a compound may be a nominal 
 stem (noun, adjective or pronoun), or a particle (invariable 
 word), or a verbal stem. Each of these cases will be separ- 
 ately examined. 
 
 I. The first term is a nominal stem. — The general 
 principle of I.-E. composition is, that any nominal stem, occur- 
 ring as the first term of a compound, must appear in the 
 shape of a bare stem, without any ending whatever. Let 
 us consider, for instance, the I.-E. stems in -0- : the term taken 
 from words like Sk. dgva-s, Gr. flco'-s, L. auru-m, is neither the 
 nominative, nor any other case in compounds like Sk. dqva- 
 yoga-s (whereto a horse is put), Gr. 6eo-<f>i\rjs (dear to the gods), 
 L. auru-fex (goldsmith), etc. ; it is neither the dual nor the 
 plural, though there may be two horses to the chariot, and the 
 gods actually are many in number ; it is simply the bare stem, 
 dcva-, 0eo'-, auru-, which, owing to the place it occupies, becomes 
 able to express either the instrumental, dative, or accusative 
 case, or indeed the singular, dual or plural number, together 
 with many other grammatical relations. Such is also the 
 general rule for English and German composition, the applica- 
 tions of, or exceptions to, which we shall now pursue throughout 
 the various classes of primitive stems. 
 
 1. If the first term ends with a vowel, the formation usually 
 conforms to the rule given above, save that the characteristic 
 vowel, being unaccented and consequently more or less altered, 
 is no longer seen as it is in Gr. or L. : E. and G. wer-wolf, re- 
 presenting an I.-E. word *wiro-iolqo-s, which would be L. *viri- 
 lupu-s=*viro-lupo-s (man-wolf) ; O.H.G. tago-sterno and taga- 
 stern (morning-star) ; O.H.G. wego-wlso and wega-wlso, cf. G. 
 weg-weiser, E. way-mark ; O.H.G. junc-frouwa (damsel) 1 and G. 
 jung-frau ; O.H.G. turi-sul and G. thur-saule (door-post), cf. Gr. 
 Kaai-yvrjTo-s (brother) ; G. viel-blatt, cf . Go. filu- (in composi- 
 
 1 Adj. junc-*c*ywflgo-, not *yuAgd, that is to say, the epithet does not 
 agree with the noun : cf. Gr. 'A^i-iroXtj instead of ixpa ir6\is. When the 
 noun is declined, the adjective undergoes no change. — With the second 
 (unaccented) term reduced, we have jungfer. 
 
COMPOSITION. 197 
 
 tion) = Gr. 7ro\v- (mucli) ; Gro. faihu-gaim-s (covetous), cf. L. 
 pecu- (cattle) and the G. type viek-zucht, etc. After such pat- 
 terns were created the invariable and common compounds by 
 mere juxtaposition (Gr. tag-lohn, friih-stilck, hand-habe, E. day- 
 break, way-farer, hand-mill), from which English hardly ever 
 deviates, apart however from the occasional insertion of the 
 genitival s if justified by the meaning. 1 In German, the same 
 insertion is often possible, or even necessary (tages-stern), and 
 is now always found in certain classes of words, especially those 
 in -heit (freiheits-liebe), whence it has even extended to those 
 in -ung (rettungs-ufer, wasserungs-graben),. though the latter 
 never really had or could have any genitive ending in -s. 
 German further admits of the insertion of the n which belongs 
 to the weak declension, especially for such words as end in 
 -e in the nominative singular, when involving in their com- 
 pounded state an idea of plurality 2 : gaben-reich, bienen-korb, 
 enten-teich, stunden-glas (but sohl-leder, rede-kunst, since the first 
 term may be referred to the verbs sohlen and reden), 3 etc. 
 
 2. Stems originally ending with any other consonant but n 
 and s follow the general rule without exception : E. foot-step, 
 mother-less, night-mare ; G. fusz-boden, vater-land, nacht-mahl. 
 Very often in English, and rather arbitrarily in German, the 
 •s is inserted, when required by usage or euphony. 4 
 
 3. Stems in n have been seen to assume a double form. The 
 simple form, without final n, is here the regular one, and con- 
 stantly appears in the earliest examples of Germanic composi- 
 tion : Go. guma-kund-s " of male sex," cf. the same shortened 
 stem in L. homi-cida (from homin-) ; O.H.G. hano-crdt (cock's 
 comb), namo-haft (having a name) and G. nam-haft (famous) ; 
 Go. duga-dauro "eye-door" (window), O.H.G. oug-brawa (eye- 
 brow), M.H.G. ouge-lit (eye-lid), etc. But, besides this ancient 
 type, Germanic also formed the compound with the first term 
 in the genitive sg. or pi., viz. ochsen-zunge and Franken-land ; 
 
 1 Supra 114, I. 1 A. 
 
 2 Supra 114, I. 1, and cf. the note. 
 
 3 Cf . the verbal compounds, infra III. 1. 
 
 4 Examples : sommernachts-traum, vaters-bruder and vaters-bruders-$ohn , 
 but vaten-bruder-frau ; and so also with other compounds, jahr-gang and 
 jahres-bericht, monat-schrift and monats-heft. 
 
198 ENGLISH AND GERMAN GRAMMAR. 
 
 and this, in German, quite overwhelmed the type nam-haft and 
 fried-los, which is now exceptional. The nsual form is as fol- 
 lows : hahnen-kamm, 1 ochsen-hirte, sonnen-schein, namen-buch, 
 augen-brauen, augen-lid, hasen-pfote, etc. ; secondly, analogi- 
 cally, christen-thum, studenten-blume, etc. ; lastly, even with a 
 double genitival exponent, namens-tag, friedens-richter, herzens- 
 wunsch? Nothing similar occurs in English : eye-brow, eye-lid> 
 sunshine (cf. Sun-day = Sonn-tag) , ox-lip = 0.~E. oxan-slyppe, etc. 
 
 4. Stems ending in s show a similar peculiarity. West Ger- 
 manic is known to drop the final s of the nomin. sg. : G. sieg = 
 Go. st'gris = Sk. sdhas (strength, victory) = I. -E. *seghos 3 ; so also, 
 ei (egg), kalb (calf), since the plurals eier and kalber reveal the 
 presence of a hidden s in the sg. stem. Now, this s naturally 
 also belonged to the compounded stem, cf. Gr. o-aK«r-iraAos 
 (shaking his shield, nt. nomin. o-a/cos, gen. craK£os = *o-a/c«r-os), 
 av0e<r-<f>6pos (carrying flowers, av6o<s avfoos) ; and, no less natur- 
 ally, being here medial, it was fain to survive. It was kept 
 therefore, according to a well-known distinction, either as s or 
 r : G. sieges-lohn = Go. sigis-ldun, and compare the proper names 
 which begin with Sigis- ; M.H.G. eier-vel (the thin skin in an 
 egg) and G. eier-gelb (yolk), etc. Of course, these forms 
 sieges-, eier- are the bare stems, and nothing more ; but, if con- 
 trasted with the nomin. sg. sieg, ei, they now look as if they 
 were declensional forms, namely either a gen. sg. or pi. ; and 
 thus it is they have contributed to the expansion of the syn- 
 tactical composition-type containing an apparent genitive either 
 in -s or in -er, according as the first term was understood to be 
 either singular or plural, e.g. kalbs-fleisch, rinds-fleisch, and 
 rinder-stelze (wag-tail), worter-buch (dictionary). 4 Here, as 
 elsewhere, English has only s : calf's leg. 
 
 1 Notwithstanding that the usual genitive to hahn is no longer *hahn-en, 
 but hahn-s. 
 
 * Generally, however, the simple form herz-. Cf. infra 150, 1 b. 
 
 8 Originally " neuter, cf. svpra 80 (XV.). For the form of such words in 
 composition, cf. the L. type veneri-vagus, where the s of Venus also became r. 
 
 * The latter is not widely diffused, though, in the case where it survives, 
 it remains true to its origin, inasmuch as it does not involve any necessary 
 idea of plurality : a plural would be absurd in eier-gelb, and kalber-magen is 
 strictly synonymous with kalbs-magen. 
 
COMPOSITION. 199 
 
 II. The first term is a preposition or invariable particle : 
 types, E. out-law, G. ab-fall, etc. — This very simple case re- 
 quires no explanation, but it may be of some interest to observe 
 that the particle thus used sometimes no longer survives as a 
 separate word in the present language, or has sometimes lost 
 its original meaning which is retained only in the compound : 
 E. mid-wife, where the first term is the same as G. mit (now 
 replaced as a preposition by with) = Gr. [xtTa, literally " a 
 woman that helps " ; G. after-wort (abuse), the first term being 
 identical with E. after, 1 which is no longer a G. word ; E. 
 alone = * all-one, where all would now be replaced by quite, 2 cf. 
 G. all-ein ; G. aber-glaube (superstition), the particle being the 
 same as Sk. apara-m (after) and consequently expressing a 
 pejorative meaning akin to the sense of after-, while isolated 
 aber is now merely an adversative conjunction. 
 
 III. The first term is a verbal stem. — This formation 
 is unknown in Indo-European etymology proper : in Sanskrit, 
 Latin, and even Gothic, no compounds are found except with 
 a noun or a particle as first term ; English and German, how- 
 ever, separately developed in their historical period many com- 
 pounds in which the first term has a verbal form and meaning. 
 These must be divided into two classes, according to the 
 different causes from which they arose. 3 
 
 1. The first is the determining term. — We have already 
 seen how such a compound as schiff-bar, which may be trans- 
 lated literally "bearing a ship," was wrongly supposed to con- 
 tain the vb. schiffen "to ship," and how this error led to the 
 formation of such erratic compounds (now called derivatives) 
 as esz-bar, trink-bar, les-bar, from the verbs essen, trinken, Zesen. 4 
 Now, since derivative composition is but a particular case of 
 composition, the same, as a matter of course, happened with 
 
 1 The change of meaning is " late — coming too late — improper — bad." 
 
 2 Afterwards pleonastically, quite alone ; and, by syncopating the initial 
 vowel, which seemed an insignificant element (cf. supra 96, VI., and 103, 
 II. 2), the words lone, lonely, loneliness. 
 
 3 Greek has likewise created such compounds, but only of one kind, the 
 second, namely Scuct-dvuos " biting the heart," rcpm-ictpavvos " shaking the 
 thunder," (paeaC-uPporos " enlightening the mortals." 
 
 4 Supra 110, III., and cf. supra 99, 102 (3), etc. 
 
 (UNI 
 
200 ENGLISH AND GERMAN GRAMMAR. 
 
 compounds properly so-called: thus, G. bet-haus = O.H.G. bete- 
 hus (prayer-house) being referred to vb. beten (to pray), on 
 this and the like were and still may be created, if needed, the 
 Gr. types wohn-haus, schiesz-pulver, mieth-kutsche, etc. In Eng- 
 lish the same process would have been possible ; but here it is 
 no longer to be recognised, because there scarcely exist any 
 verbs without homonymous nouns. 1 It must be observed, how- 
 ever, that English verbs, when they occur in the first term of 
 a compound, are always used in the present participle, which 
 is as much as to say that the first term is etymologically a 
 nominal form 2 : eating-house, frying-pan, racing-club, etc. 
 
 2. The first is the determined term. — Locutions in 
 which the first term is a verb governing the second term are 
 very common, chiefly as popular nicknames, which have often 
 become proper names, or are still used as epithets in familiar 
 speech. "We have already noted E. Make-peace, and Shake- 
 speare, which was sometimes parodied as shake-scene ; here we 
 may add turn-back, wag-tail, break-fast, spend-thrift, run-away, 
 would-be, pin-afore, round-about, and such possible nicknames as 
 Kill 'em all, Go-to-bed, etc. In German we have proper names, 
 as Bleib-treu (remain true), Lebe-recht, and familiar epithets, 
 tauge-nichts (good-for-nothing), sauf-aus (drunkard), wage-hals 
 (rash), stell-dich-ein (appointment), spring -ins-f eld, etc. As, in 
 such compounds, the verb always shows the bare stem-form, 
 they probably proceed, at least to a large extent, from familiar 
 exclamations which required the verb to stand in the imperative 
 mood : a few nicknames of this kind, being once created, might 
 easily serve as models for further imitation. 
 
 § 2. Form of the Last Term. 
 
 (118) The last term of any English or German compound 
 is always a nominal stem; in other words, every compound 
 properly so-called, in either language, is a noun or an adjec- 
 tive ; there are no compound verbs. Such verbs as are com- 
 
 1 As to ink was drawn from ink, so also, reversely, a noun drink was 
 formed from to drink, so that, in English, the originally nominal and verbal 
 stems are now scarcely distinguishable. 
 
 2 Like the G. type rettungs-ufer. Cf. supra 103 (II-)- 
 
COMPOSITION. 201 
 
 monly termed by this name, will be seen to belong to one of 
 these three classes : — (a) verbs with an inseparable prefix, 
 already analysed nnder derivation ; — (6) verbs with a separable 
 prefix (G. aus-gehen, E. to go out), well-known types of simple 
 juxtaposition, so loose indeed that they form two distinct words 
 in English and are often separated in German * ; — (c) verbs 
 derived from compound nouns or adjectives, an important class 
 treated below. 
 
 This point being settled, the rules for the last term of an 
 English or German compound are extremely simple. 
 
 1. As a general rule, the last term of a compound undergoes 
 no change, and does not differ from the isolated word, apart 
 from such mechanical alterations as may be caused by its being 
 unaccented and swiftly sounded. 2 
 
 2. This rule, of course, does not prevent a compound, when 
 formed, receiving, like any other word of the language, a 
 derivative suffix, and developing in this way new derivatives, 
 which might prove more generally used than the word from 
 which they sprang : thus, E. light-footed and G. vier-fiiszler 
 cannot be deemed true compounds, since there are no words E. 
 *footed and G. *fiiszler; they are derivatives from the com- 
 pounds light-foot and vier-fusz. Now, this process having once 
 become familiar, a language could easily construct such deriva- 
 tives on the basis of two simple words, without any inter- 
 mediate compound ever existing, e.g. E. gray-headed on gray 
 and head, 5 G. ein-augig on ein and auge, and so forth : this in 
 fact still constantly happens. 
 
 3. For the same reason it will be understood that a verb 
 may be derived from a compound as from any other word, and 
 that the process of derivation is the same in this case as for 
 ordinary verbal derivation : thus, from E. lord, we have the 
 intransitive to lord, and the transitive to unlord, and further 
 such verbs as to husband, to elbow, to mildew, to worship, to horse- 
 
 1 The inseparable type, E. to over-eat [o. *.], G. iiber-setzen (to translate), 
 is a peculiar and rarer variety of the same class. 
 
 2 See the numerous instances mentioned in their proper places : 65, 5 ; 
 114, I. 1 ; 115, II. etc. 
 
 8 There exists no word headed, in this acceptation at least, and the 
 existence of a word * gray -head is only a possibility. 
 
202 ENGLISH AND GERMAN GRAMMAR. 
 
 whip, etc. ; G. arg-wohn (distrust) has produced argwohnen, — 
 for there is no G. vb. *w'6hnen, — and, in spite of appearances,, 
 handhaben (to handle) is not a mere compound of hand and 
 haben, but a derivative from the nominal compound hand-habe- 
 (a handle). The same is the case with the types rechtfertigen. 
 wetteifern, lustwandeln, etc. ; similarly, with still more complex 
 derivatives, nothziichtigen from noth-zucht, bevolbnachtigen from 
 voll-macht * ; and thus there will be no difficulty in understand- 
 ing how, by imitating such patterns, German even sometimes 
 united into one verb two words which ought to remain syntac- 
 tically separated, and formed such faulty verbal compounds as 
 lobsingen and wahrsagen? 
 
 1 Cf. supra 107 (V. 4). 
 
 2 As a counterpart, E. has to vouchsafe, wherein the verb is to vouch, safe 
 being a mere apposition ; but a strange confusion caused the verb to be con- 
 jugated as a whole, so that the person-endings are now adapted to safe 
 instead of to vouch. 
 
CHAPTER IV. 
 THE SYSTEM OF NUMERATION. 
 
 (119) Germanic Numeration is an Indo-Enropean decimal' 
 system, depending on the general principles of derivation and 
 syntactical as well as non-syntactical composition previously 
 described. It therefore forms a suitable vehicle for a recapitu- 
 lation of the Second Part and, as it were, an exercise on the 
 rules we have stated. We must first distinguish the so-called 
 cardinal numbers, and the various numeral expressions which 
 have been derived from them. 
 
 Section I. 
 
 CARDINAL NUMBERS. 
 
 (120) The I.-E. units are conventional expressions, of un- 
 known origin, and without any visible etymological connection 
 with one another. The tens are either derivatives from, or 
 combinations of units. The numbers 100 and 1000 are isolated 
 expressions, which however show a prehistorical derivative 
 relationship with the tens. 
 
 § 1. Units and Sums of Units. 
 
 (121) 1. The primitive root of "one" is i, probably the 
 same that occurs in the L. demonstrative i-s i-d. Its numeral 
 value may be seen in the Homeric ?-a (one fm.). 1 But it ap- 
 pears but seldom in such a bare state : usually it is deflected 
 
 1 Not to be confounded with ets fda $v, since this proceeds from a different 
 root, namely sem, also in Latin sem-el (once). 
 
 203 
 
204 ENGLISH AND GERMAN GRAMMAR. 
 
 and followed by a derivative suffix, Sk. e-ka-s (one), Zd. ae-va-, 
 Persian ai-va-. In Latin, Germanic and Slavonic, the accom- 
 panying suffix is -no-: I.-E. *oy-no-s, L. u-nu-s — oi-no-s, 1 Preg. 
 *ai-na-z, whence Go. di-n-s, O.N. einn, O.E. an and E. one (an 
 or a when unaccented), O.H.G. ein and G. ein. 
 
 2. The primitive stem was *duo- and *dwo-, furnished with 
 the ordinary dual endings : nomin. msc. *duow, fm. *dudy, nt. 
 *duoy ; Sk. msc. dvdu, fm. nt. dve ; Gr. 8vw Svo invariable ; L. 
 duo, duos, duo. Gothic had partly adopted the plural-endings : 
 msc. tvdi, im.tvds, but nt. tva = *dwoy. West Germanic like- 
 wise altered the primitive endings : O.E. msc. twegen, fm. two, 
 nt. tu ; O.H.G. msc. zwene, fm. zwo, nt. zwei. 2 Whereupon the 
 ■declension was reduced to one form, the neuter serving for all 
 genders : E. two (which however keeps the old spelling of the 
 fm. two)=OM. tu; G. zwei. — In composition, the stem of 
 number 2 assumed a form *dwi-, Sk. dm-, Gr. 8t-, L. bi-, Preg. 
 *twi- eventually lengthened to *twl-, which latter are respect- 
 ively to be recognised in : E. twi-n, to twi-ne, and twi-light, i.e. 
 "" a double, doubtful light " ; G. zwie-fach (two-fold, double), 
 zwi-r-nen (to double, to twist), and zwei-fel (doubt) = O.H.G. 
 .zwi-fal. 3 
 
 3. I.-E. root tri, in the normal grade and with the pi. msc- 
 fm. ending *trey-es, Sk. trdy-as, Gr. and L. (the y being dropped 
 and the vowels contracted) Tpcis tres, nt. rpi-a tri-a. Similarly 
 in Germanic: Go. msc.-fm. preis (contracted from *J>rij-is), nt. 
 Jjrij-a ; O.E. tSrl and (Sreo, the latter being now E. three, the 
 former extinct ; O.H.G. drl drlo driu, whence G. drei. 4 - 
 
 4. I.-E. stem *qetwor- *qetwr-, Sk. msc. catvdr-as, Gr. Terraces 
 
 1 Thus we read OINO(M) in early Latin inscriptions. 
 
 2 The msc. twegen and zwene are as yet unexplained. The fm. two and 
 zwo reproduce Go. tvos. The nt. tu is Go. tva. As to the nt. zwei, it 
 probably contains the diphthong seen in the stem of the Gr. derivative 
 5oi-ol. — All these forms survived very late : down to the seventeenth century, 
 classical German still distinguished zween manner, zwo frauen, zwei kinder, 
 and some traces of the distinction persist even now in Alamannic, and in 
 •class. E. and G. in the number twen-ty and zwan-zig, infra 122, 20. 
 
 8 Literally " double case." — The diphthong is not the same in zwei and 
 zweifel, since, in Alamannic, we find tsvay, but tsvifh 
 
 4 Numbers 2 and 3 had a strong tendency to become invariable, because 
 numbers 5 and the following very early became, or even had always been, 
 invariable words. For number 4, cf. L. quatuor invariable. 
 
THE SYSTEM OF NUMERATION. 205 
 
 Te<r<rapes = *T€Tfap-€5, L. quattuor, Go. fidvor, O.N. fjor-, O.E. 
 feower and E. four, O.H.G. fior and G. vier, though the loss of 
 the medial dental is a riddle as yet unsolved. 
 ' 5. I.-E. *pefiqe, Sk. pdnca, Gr. Trefnre irivrt, L. qulnque, Go. 
 fimf, O.N./irara, O.E. /if and ~E.five, O.H.G. finf, then funf from 
 the intrusion of the vocalism of the ordinal number, M.H.G. 
 viinf with metaphonical vocalism imitated from viinf-zic (fifty), 
 and G. funf. 1 
 
 6. I.-E. *sweJcs *seks, Sk. SdS, Gr. €£=*ofe£, L. sex, Go. saihs r 
 O.N. seta, O.E. seox = seohs, and E. «M) (metaphony caused by 
 the h), O.H.G. sehs and G. seeks. 
 
 7. I.-E. *septm, Sk. saptd, Gr. «rra, L. septem, Go. si6«ra (the 
 final from sihun-da " 7th "), 2 O.E. seo/cm and E. seven, O.H.G. 
 sibun (u- metaphony) and G. sieben. 
 
 8. I.-E. *oktow, Sk. aStau, Gr. 6ktw, L. oc£o, Go. ahtdu, O.E. 
 eataa and E. eight, O.H.G. ahto and G. acht. 
 
 9. I.-E. *net0W, Sk. ndva, Gr. evvca = *c-w€fa, L. novem = 
 *nerem (cf . notms "new" = Gr. veos), Go. niun = *niwun, O.E. 
 nigon and E. wine, O.H.G. mttn. and G. net*«. 
 
 10. I.-E. *dek-m, Sk. dapa, Gr. Scko, L. decern, Go. taihun T 
 O.E. flen £en and E. £ew, O.H.G. zehan, M.H.G. z'ehen and G. 
 zehn, the last three being traceable to an I.-E. form *dehom(t) 
 with deflected last syllable. 
 
 11-12. These two numbers assume a peculiar shape only in 
 Germanic : they are formed by non-syntactical composition, the 
 last term being *-lihi-, a stem abstracted from the root which 
 produced Go. bi-leib-an = G. b-leib-en (to remain), thus *aina- 
 litri- " 1 left [above 10] " : Go. din-lif, O.N. ellifu, O.E. andleo- 
 fan 3 endleofan and E. eleven, O.H.G. einlif, M.H.G. eilif eilf and 
 G. eilf elf; Go. tva-lif, O.N. tolf, O.E. twelf (metaph.) and E. 
 twelve, O.H.G. zwelif (metaph.), M.H.G. zwelf, and G. zwolf for 
 zwelf.* 
 
 1 E. and G. fing-er seems to be a cognate word. 
 
 * Cf. supra 39, 1, and 40, 1. Of course the group pt is here reduced to 
 simple p. 
 
 8 A euphonic d is inserted between n and I. Final an is here probably a 
 superadded plural exponent, from which twelf has remained free. 
 
 4 Compare the spellings schopfen, schworen, instead of schepfen, schweren, 
 and a few others. 
 
206 ENGLISH AND GERMAN GRAMMAR. 
 
 13-19. As in every other I.-E. language (cf. Gr. Iv-Sexa, L. 
 un-decim), these numbers are compounds, the last term being 
 10 : O.H.G. dri-zehan, etc., and Gr. drei-zehn ; O.E. ftri-ten-e, 
 with a subjoined ending, owing to which the preceding vowel, 
 since it no longer stood in a close syllable, was not shortened 
 as it was in ten, whence E. -teen. 1 The initial unit undergoes 
 no change ; in English, however, the shortening in a close 
 syllable changed ftftene to fifteen, whereas ftritene, sounded M.E. 
 thritteen, became by metathesis thirteen. 
 
 § 2. Tens. 
 
 (122) The names for the I.-E. tens are compound words : 
 ■Sk. vin-^ati, Gr. Toid-Kovra, 2 L. quadrd-gintd, etc. The Germanic 
 corresponding terms are only apparent compounds. In origin 
 they are locutions, consisting of two words, the name for the 
 unit governing in the plural number the name for the ten, thus 
 *' two tens "=20. But this archaic formation is hidden from 
 the eye, both in German and English. 
 
 In order to understand it, we must consider some equivalent 
 Gr. form, e.g. rpeis 8eKa8es = 30. The word Se>cas (stem ScxaS-, 
 observe the accent) represents an I.-E. noun *dekmt- which is 
 derived from *dekm (ten) and has become Preg. *te%un, then 
 *te%u, whereupon, as it appeared to be a stem in -u-, it was 
 declined as such : Go. sg. *tigu-s, pi. tigjus, ace. tigu-ns, etc. 
 From this arose such juxtapositions as nomin. tvai tigjus and 
 ace. tvans tiguns = 20, which, by dropping their endings, reduced 
 the last term to O.E. -tig and E. -ty, O.H.G. -zug, s M.H.G. -zic 
 and G. -zig, which are invariable. 
 
 Hence grammatical exactness would require the noun 
 governed by a ten to stand in the gen. pi., and this rule in fact 
 often holds good in early O.H.G. , e.g. feorzuc wehhono, literally 
 
 1 This was afterwards abstracted from such compounds, and became an 
 independent word. 
 
 * The last term is a shortened derivative from number 10 : cf. infra 123, 
 I. 
 
 8 This strange and as yet unexplained vocalism is probably due to a 
 reduced grade of the radical syllable in *d€krn. — As to G. z for Preg. t, 
 observe that the t here is really an initial consonant, supra 49, III. 2. 
 
THE SYSTEM OF NUMERATION. 207 
 
 ""four tens of weeks," as in L. tria milia hominum. But, as 
 the unit-nnmbers governed the nominative, the same case was 
 used throughout. 
 
 When a unit is added to the ten, they form together a merely 
 syntactical compound : E. three and twenty, G. drei und zwanzig. 
 English, however, probably influenced by French, also admits 
 •of the simple juxtaposition : twenty-three. 
 
 20. Go. tvdi tigjus, O.E. twen-tig and E. twenty, O.H.G. zwein- 
 zug and G. zwanzig, cf. the nomin. msc. twegen zwene. 
 
 30. Go. ace. J>rins tiguns, O.E. (Srl-tig tSrittig and E. thirty 
 {like thirteen), O.H.G. drl-Tfiug (the t treated as it usually is 
 when medial, but also dri-zug) and G. dreiszig. 
 
 40. Gro.fidvor tigjus, O.E. femcer-tig and E. forty, O.H.G. fior- 
 zug and G. vierzig. 
 
 50. Go. ftmf tigjus, O.E. flf-tig and E. fifty, O.H.G. finf-zug 
 funf-zug, M.H.G. viinf-zic and G. funfzig. 
 
 60. Go. saihs tigjus, O.E. siex-tig and E. sixty, O.H.G. sehs-zug 
 .s'ehzug (by reducing the group) and G. sechzig. 
 
 70-90. The three following tens are now formed after the 
 pattern of the first five; but such was not always the case, for 
 Go. has no type *sibun tigjus, etc. Indeed, the process formerly 
 nsed was exactly the reverse : the term is sibunte-hund, wherein 
 ■sibunte- is a gen. pi. governed by -hund (this shortened from 
 *taihund), the whole meaning " the tithing of sevens " ; and so 
 also, ahtdute-hund, niunte-hund, and even taihunte-hund "tithing 
 of ten8" = 100. In O.E. the two terms changed places, whilst 
 analogy substituted the syllable -tig for the syllabic -te: hund- 
 seofontig, hundeahtig, hundnigontig ; then, the syllable hund, 
 which seemed meaningless, was dropped. In O.H.G., however 
 far we go back, this hund has already disappeared : sibun-zo 
 (instead of -50, owing to the analogy of the z in -zug), ahto-zo, 
 even zehan-zo (100) ; whereupon, analogy working its way 
 throughout, we have sibun-zug, now siebenzig, etc. 1 
 
 1 Except that zehanzug is replaced by hundert. The process, though it 
 has quite disappeared, is interesting inasmuch as it throws light upon the 
 pre-historical formation of number 100. — As to -zug and -zo, observe moreover 
 that the sibilant often followed a nasal, and cf. supra 49, III. 2 in fine. 
 
208 ENGLISH AND GERMAN GRAMMAR. 
 
 § 3. Hundreds and Thousands. 
 
 (123) I. The word for 100 was I.-E. *km-t6-m, neuter 
 noun : Sk. cu-ta-m nt., Gr. i-Ka-ro-v (one hundred), L. cen-tu-m > 
 Lith. szim-ta-s, Go. tva hunda " two hundreds," which literally 
 translated would be L. duo *centa. If we traced farther back 
 the origin of this form, an easy explanation could be suggested, 
 namely, a derivation from *dekm by addition of the suff. -to- ; 
 this syllable, being accented, suppressed the accent of the first 
 syllable, which then appeared in the reduced grade : an I.-E. 
 *dkm-to- or *tkm-to-, being barely pronounceable, might have 
 lost its initial consonant, and, finally, *km-to-m would very well 
 mean " the tithing to ten." 
 
 This is, without doubt, the meaning of Go. taihunte-hund > 
 O.E. hund-teontig, O.H.G. zehanzo, which are the only ancient 
 Germanic terms for 100. But the simple *hund=li. centum- 
 nevertheless also existed, since we find it in Gothic forming 
 every multiple of 100 : Jrrija hunda, fimf hunda, etc. Now,, 
 this word very early entered into composition with a noun 
 *raj>-, akin to Go. raj>-jan (to count) = G. reden (to speak) > 
 whence the modern forms, E. hund-red, G. hund-ert, are obviously 
 substantives literally signifying " a count of hundred." 
 
 II. The old multiples are O.E. tu hund, etc., O.H.G. zwei hunt > 
 etc. = Go. tva hunda, etc. The substitution of hundred and 
 hundert took place as a matter of course. But, although the 
 latter were originally and, in fact, still are substantives in both 
 languages, being as such capable of receiving the sign of the 
 plural (E. hundreds, G. hunderte), yet they never assume this 
 sign in numeration {two hundred, zwei-hunderi), 1 and moreover 
 have ceased to govern the following noun in the genitive. The 
 tens and units, if needed, are added by juxtaposition. 
 
 III. The word for 1000 is only found in Germanic and Letto- 
 Slavonic 8 : it is Go. Jjusundi, a fm. noun (once nt.), which has. 
 
 1 This is not at all an irregularity, but very likely a survival from a time 
 when a great many nouns kept the singular form in the plural without, 
 alteration : cf. G. vier-mal, zehn mann, sechs fusz, and popular E. three year,, 
 infra 139, HI. 1 ; 143, II., and 147, 4. 
 
 * In Sk. and Gr. (x^Xioi) we have quite a different term, and another still 
 in L. {nitlle, pi. milia). 
 
THE SYSTEM OF NUMERATION. 209 
 
 become nt. in O.E. Susend and E. thousand, O.H.G. dusunt tu- 
 sunt and G. tausend. As the word occurs, in Salic Frankish 
 and elsewhere, with an inward aspirate (thus-chunde, cf. O.N. 
 fius-hundraft " 1200 "), we are led to explain it as an old com- 
 pound meaning " the large hundred." x 
 
 IV. Though this term is a substantive, it does not vary in 
 the multiples (two thousand = zwei-tausend) and governs the 
 multiplied noun in the nominative. 
 
 V. Numbers beyond 999999 are, as elsewhere, learned terms 
 constructed on the basis of Latin stems. 
 
 Section II. 
 
 DERIVATIVES FROM CARDINAL NUMBERS. 
 
 (124) I. Ordinal numbers. — Among the derivatives, the 
 ordinals are the simplest and most regular : several suffixes 
 were used, but Germanic preserved but one, I.-E. suff. -to-, cf. 
 Gr. 7rc/A7r-To-s (fifth), L. quinc-tu-s quintus, etc. The accent of 
 such words, being variable in Sk. and even in Gr. (Sckci-to-s, but 
 cikckt-to-s), probably also varied in the primitive speech. Now, 
 in Germanic, after an unaccented syllable, as in Sk. catur-thd-s 
 (fourth), we should expect regularly E. *four-d and G. vier-te ; 
 after an accented syllable, as in Sk. saptd-tha-s (seventh), we 
 should expect, on the contrary, E. seven-th and G. sieben-de ; 
 lastly, after an / or an s, the I.-E. t here remaining unaltered, 
 the result would be E. six-t and G. sechs-te. 2 All these distinc- 
 tions hold good for the earliest state of both languages, but we 
 see at once what has become of them in later times : excepting 
 only number 3, analogy in E. extended the exponent -th 
 throughout, and the exponent -te in G., so that E. and G. now 
 disagree in the consonant of their ordinal suffix. 
 
 1. The ordinal is nowhere cognate with the cardinal: Go. 
 fru-ma (cf. Gr. irpo " before " and L. pri-mu-s), superlative 
 frum-ist-s = 0.1$. fyrm-est (cf. the E. comparative form-er), 
 
 1 For Preg. *>us, cf. Sk. tduas (strength) and tavds (strong, mighty, 
 great). 
 
 2 Of course the final G. e is a declensional ending, cf. infra 156. 
 
 P 
 
210 ENGLISH AND GERMAN GRAMMAR. 
 
 which a popular etymology finally corrupted to E. fore-most ; 
 another superlative is Go. *faur-ist-s, O.N. fyr-st-r, O.E. fyr-st 
 and E. first, O.H.G. fur-isto, M.H.G. viir-ste (first), and G. fiirst, 
 now confined to the sense of " prince " (cf . L. princeps " first ") ; 
 another still, from the same root that produced Gr. rjp-i (early), 
 E. ere and ear-ly, etc., is O.E. cer-est (lost in E.), O.H.G. er-ist 
 and G. er-st, cf. Go. air-is (sooner). 
 
 2. No relation with the cardinal : Go. an-}>ar, the compara- 
 tive of the root occurring in Sk. an-yd-s (other) 1 ; O.E. *on-<5er, 
 whence o<5er, but E. second borrowed ; O.H.G. an-der, 2 but G. 
 zwei-te, a late derivation from zwei. 
 
 3. Go. J>ri-dja=Sk. tr-tiya-s (for *tri-) = h. ter-tiu-s ; O.E. 
 tSri-dda (and Sir-da), E. thir-d ; O.H.G. dri-tto and G. dri-tte. 
 
 4. O.E./eower-6"a 3 and E. four-th, after the analogy of seven-th 
 and the like (cf. L. quar-tu-s) ; O.H.G. feor-do and G. vier-te. 
 
 5. O.E. flf-ta, M.E. fif-t? altered to E. fif-th ; O.H.G. finf-to 
 funf-to regular (the latter probably reproducing a form with 
 reduced radical syllable, I.-E. *priq-t6-s) and G. funf-te (the 
 metaphony as in funf). 
 
 6. Go. saihs-ta ; O.E. siex-ta, M.E. six-t,* altered to E. six-th ; 
 O.H.G. sehs-to and G. sechs-te, regular forms. 
 
 7-10. E. seven-th, eighth (syncopated from *eight-th), nin-th, 
 ten-th (the noun ti-the being an older form with O.E. compensa- 
 tory lengthening). G. sieben-te (corrupted from M.H.G. sieben- 
 de), achte (syncopated from O.H.G. ahto-do), neun-te, zehn-te. 
 
 11-19. The formation is the same throughout. 
 
 20-90. The same formation in English ; but O.E. g developed 
 before the suffixal 6" a euphonic vowel, which is kept in E. twen- 
 tieth, etc. German here changes its system : to the cardinal is 
 now added a superlative-suffix (O.H.G. zweinzug-osto), whence 
 the modern type zwanzig-ste. 
 
 1 It must be borne in mind that suff. -tero- involves an idea of choice or 
 alternation between two things only ; cf. Gr. Setf-repo-s " second " and 6ei/- 
 raro-s ,l last." 
 
 2 A meaning preserved in G. ander-t-halb (infra III.) and in the current 
 E. locution every other day. 
 
 3 Where the f might very well be still voiced. 
 * Still used in some provincial dialects. 
 
THE SYSTEM OF NUMEEATION. 
 
 211 
 
 100-1000. E. hundred-th, thousand-th ; but G. hundert-ste, 
 tausend-ste : as above. 
 
 Ordinal adverbs are derived : in E., by means of suff. -ly ; in 
 G., by adding an old genitive-ending, erst-ens. 
 
 II. Multiplicatives. — 1. The formation of multiplicative 
 adjectives, namely, E. -fold and G. -fait -fiiltia, has been given 
 above. But Mod. G. uses by preference a composition with the 
 word fetch (compartment, class), viz. zwie-fach or zwei-fach 
 (double), etc. 
 
 2. The E. multiplicative adverbs, once (O.E. an-es), twice 
 (M.E. twi-es) and thrice are old genitives used adverbially; next 
 comes a syntactical locution, four times. G. einst (once) = 
 O.H.G. ein-est seems to have been influenced by the ordinals • 
 for it looks very like an old genitive ein-es, the more so be- 
 cause the corresponding ander-s (otherwise) has also become 
 ander-st dialectally. Apart from these two, all are syntactical 
 (ein-mal, zwei-mal, etc.), the invariable nt. noun mal being per- 
 haps akin to the L. final in semel (once) and simul (at once) = 
 semol. The very late adverbs in -lei are likewise syntactical 
 forms, but in the genitive case, thus einer leie "of one sort," 1 
 now einer-lei, and analogically zweier-lei, etc. 
 
 III. Partitives. — The partitive *halha-, though common to 
 all the Germanic dialects, is not found in the rest of the I.-E. 
 family. Its original meaning was probably " side," so far as 
 we can judge from such curious compounds as G. meiner-halb- 
 (e»)=M.H.G. min-halp, as it were "on my side" (in my con- 
 cern, upon my account) 2 . Preg. *halha- is O.E. healf, E. half, 
 G. halb, and its derivative hdlf-te.' The compounded type 
 *J>ri&ja-halha-, literally "the half of a third," that is "two 
 and a half," is likewise Pregermanic : O.E. (Sridda-healf, feorfta- 
 healf, etc. ; G. dritt-halb, viert-halb : but it is quite lost in Mod. 
 English. 3 
 
 Further partitives are: in E. (as in F.), the ordinals, a third, 
 a fourth, etc. ; in G., non-syntactical compounds, the last term 
 
 1 M.H.G. leie (manner) is borrowed from O.F. ley, now loi (law). 
 
 2 Sk. drdha- likewise means " side " and "half." 
 
 8 Analogy transferred the medial t of dritt-halb, viert-halb, etc., into 
 ander-t-halb = OM. oSer-healf " one and a half." 
 
212 ENGLISH AND GERMAN GRAMMAR. 
 
 of which, as unaccented, is reduced to a dull syllable, drittel = 
 *dritt-theil (M.H.G. drit-teil), viertel, etc. 
 
 IV. Distributives. — The old distributive derivation by 
 means of the I.-E. suff. -no- (L. bi-nl tri-ni, Go. tvei-h-nai and 
 O.N. tven-ner " two by two ") is no longer represented except 
 by a few survivals : E. twi-n = I.-E. *dtoi-no-s, and to twi-ne ; G. 
 zwir-n (twist) = I.-E. *dwis-no-s, 1 and vb. zwir-n-en. The dis- 
 tributive locutions now used are E. two by two and G. je zivei. 
 
 1 For the medial s, compare E. twist and G. zwist (discord). 
 
THIRD PART. 
 
 DECLENSION. 
 
 (125) Under the name of Declension are comprised the 
 various modifications, either in stem or termination, under- 
 gone by nominal and pronominal stems, and corre- 
 sponding to the various relations of meaning of which 
 they are capable. 
 
 These relations are three in number, namely : gender, 
 number, and case. 
 
 Indo-European had three genders, masculine, feminine, 
 and neuter. All these are kept in the Germanic languages. 
 
 Indo-European had likewise three numbers, singular, dual, 
 and plural. But the dual, which is preserved in Sk. remark- 
 ably, and in Gr. tolerably well, has left in Latin but two 
 isolated representatives (ambo and duo = *duo), and is com- 
 pletely lost in the Germanic group, 1 being replaced everywhere 
 by the plural. 
 
 Lastly, the relation between an I.-E. noun or pronoun and 
 the other members of any given sentence comprised eight 
 cases, which are still exhibited by the Sk. and partly even by 
 the Mod. Russian declension. But this number is well known 
 to have been already reduced to six in Latin, and in Greek to 
 five. A further simplification took place in Germanic : the 
 vocative or calling-case was everywhere blended with the 
 nominative ; the locative (situation in) and the ablative 
 (coming from) were expressed by means of prepositions 2 ; the 
 
 1 The few cases in which it survives have been just mentioned above 
 (121, 2). 
 
 2 Respectively :—E. in or at, G. in or zu (with the dative), cf. Gr. iu, L. 
 in ; — E./rom or out, G. von or aus (with the dative). 
 
 213 
 
214 ENGLISH AND GEBMAN GRAMMAE. 
 
 instrumental (means or accompaniment) survived a little longer, 
 though Gothic has already lost it; O.E. and O.H.G. still show 
 it in use, 1 but it died out later and was also replaced by a 
 periphrastic locution. 2 In short, M.H.G., like Go., is confined 
 to four cases, still surviving in Mod. German, viz. : nomina- 
 tive, or subject-case ; accusative, or object-case (direct 
 complement) ; genitive, or case of possession and adnominal 
 determination ; and dative, or case of attribution to a 
 given subject ; with this restriction, however, that the accusa- 
 tive in nouns very seldom differs from the nominative. In 
 English, the same confusion is constant, and the dative is 
 denoted by a preposition 3 : so that the English declension is 
 reduced to two cases, nominative and genitive, the latter, 
 moreover, being often replaced by the concurrent use of the 
 particle of. 
 
 The words which admit of these three relations of gender, 
 number and case, are distinguished under four grammatical 
 categories, respectively called article, noun, adjective, and 
 pronoun. 4 
 
 1 The adverbial formation (supra 110, I. 4) still exhibits a number of (at 
 least probable) instrumentals. 
 
 2 E. with, by, or through; G. mit (dative) or durch (accusative). 
 
 8 Usually to = G. zu. The possibility, however, of construing occasionally 
 a so-called indirect complement without any preposition (the land yields the 
 owner two hundred a year, cf. G. es gilt dem besitzer . . .) is an impor- 
 tant survival from the old dative, which will further be found in the prono- 
 minal declension (infra 161 and 166). Hence, the dative may be said still 
 to be a living case, as far as function is concerned, though, in form, it is one 
 with the nominative-accusative. 
 
 4 The participle, of course, inasmuch as it is a declinable word, is a mere 
 adjective ; the adverb has been seen to be a case-form of the noun or adjec- 
 tive : so that six among the ten parts of speech fall under the above divisions. 
 The verb (with the participle as a form of it) is the subject of our Fourth 
 Part. The remaining three classes of invariable particles require but little 
 etymological explanation ; the reader should refer for these to the alpha- 
 betical indexes where very many will be found. Hence the work comprises 
 the whole of Anglo-German comparative Grammar. 
 
CHAPTER I. 
 AETICLES. 
 
 (126) Properly speaking, the article is simply a demon- 
 strative pronoun, a part it often plays in German. As snch, 
 it ought to fall nnder the head of chapter IV., and it might 
 appear to be an error in method to separate it. In a strictly 
 historical sense, it is an error, no donbt ; for Indo-European, 
 and even Pregermanic never had an article ; Slavonic still does 
 without it ; and this essential part of English and German 
 speech is but a late product of the improper and somewhat 
 blundering use of an old demonstrative stem in a weakened 
 and vague meaning. From a practical point of view, however, 
 which is the one here chiefly requiring consideration, it seems 
 advisable to isolate the article: being of frequent and often 
 obligatory use, it occupies a distinct place among the demon- 
 stratives ; having strongly influenced the other parts of speech, 
 it is well fitted to afford us a general paradigm of declension, 
 which it allows us to analyse once for all ; and lastly, at least 
 in German, where it has preserved its endings unaltered while 
 the noun has mostly dropped them, the article is the chief 
 exponent of gender, number, and case, so that the declension of 
 the noun necessarily supposes a knowledge of this prefixed and 
 declined particle. 
 
 Yet more accurately, the article is either an old demonstra- 
 tive or an old numeral ; for it seems, in every language, the 
 unavoidable destiny of the little word meaning "this," to be- 
 come, sooner or later, an insignificant aid to speech, as of the 
 numeral " one " to degenerate into the indefinite article. L. 
 caballus means equally " the horse " and " a horse " : so, when- 
 ever a speaker wanted to express more precisely one of the two 
 meanings, he naturally said either ille caballus or unus caballus, 
 
 216 
 
216 ENGLISH AND GEEMAN GEAMMAE. 
 
 whence O.F. li chevals and uns chevals ; the same thing exactly 
 happened in passing from Pregermanic to English and German. 
 
 Section I. 
 
 DEFINITE ARTICLE. 
 
 (127) The greater part of the declension of the definite 
 article proceeds from an I.-E. demonstrative *so *to, which, 
 though more or less altered, may be found in any member of 
 the I.-E. family. A few forms, however, are borrowed from 
 another demonstrative, nearly akin indeed, viz. *syo *tyo, which 
 scarcely appears anywhere but in Sk, though there with per- 
 fect clearness. 
 
 § 1. Origin and Primitive Declension. 
 
 (128) I. The demonstrative, I.-E. nomin. msc. sg. *so, fm. 
 sg. *sd, 1 never showed an initial s except in these two forms 
 only : Sk. msc. sd, fm. sd; Gr. 6 fj (article). Everywhere else, 
 that is to say, in the nomin. nt., in every other case of the 
 three genders, and in any case of the plural, the stem-form 
 became *to- : I.-E. nomin. nt. sg. *to-d, Sk. td-d, Gr. to, 2 L. 
 (is-)tu-d ; I.-E. ace. msc. sg. *to-m, Sk. td-m, Gr. to-v, L. 
 (is-)tu-m; I.-E. nomin. msc. pi. *toy, Sk. te, Gr. toi, 3 L. (is-)ti, 
 etc. Now, in Sk., Homeric Gr. and even in the L. compound, 
 the word has kept its fall demonstrative value ; but, in classical 
 Gr., it has already become an article, merely meaning " the " 4 
 and accompanying the noun throughout. In the history of the 
 Germanic languages we find a similar evolution : the same 
 stem, which is essentially a demonstrative pronoun, only with 
 somewhat weakened stress, in Go., O.E. and O.H.G., is an 
 article, that is to say, a kind of nominal prefix, almost constant 
 and obligatory in German speech, and likewise obligatory, in 
 the great majority of cases at least, in English. 6 How this 
 
 " ' Cf. supra 73 (HI.). 
 
 * For *t6-S, since any final explosive disappears in Greek. 
 8 A Doric form. Ordinary Gr. oi after the analogy of 6. 
 
 4 The meaning " this " was then confined to the juxtaposition 6-5e. 
 
 5 Except when the noun is conceived in a general and undefined meaning 
 (man delights me not), an exception which confirms the rule, as strictly 
 agreeing with the original demonstrative value of the article. 
 
ARTICLES. 
 
 217 
 
 change came to pass, is also easily seen : when we read, for 
 instance in Ulfilas * " manne sums dihta tvans sununs, jah qa}> sa 
 jilhiza ize du attin, of-men some-one had two sons, and quoth 
 this younger of-them to father," or in Old English 2 " soSlice ut 
 eode se scedere his seed to sdwenne, sooth ly out went this sower 
 his seed to sow," we translate at once without any difficulty 
 u the younger told . . . ," " the sower got out . . . " ; 
 and, in fact, the demonstrative here is already a mere article. 
 
 The Gr. declension of this demonstrative nearly reproduces 
 the I.-E. accidence ; again, the Gothic declension is truly re- 
 flected hy the German : nothing, therefore, would appear better 
 fitted to smooth the way from one to the other, than a Go. 
 paradigm with the similar Gr. cases. The agreement will at 
 once strike the eye. 
 
 Singular. 
 
 
 msc. 
 
 
 fm. 
 
 nt. 
 
 N. 
 
 sa = 6 
 
 
 so = T7 4 
 
 J>a-t-a z = r6 
 
 A. 
 
 }>a-n-a s = 
 
 TO-V 
 
 ]pb = rrj-v 
 
 Jja-t-a = To 
 
 G. 
 
 Ms 
 
 
 pi-zos 
 
 ]pi-s 
 
 D. 
 
 Jja-mma 
 
 
 . pi-zai 
 Plural. 
 
 Jja-mma 
 
 
 msc. 
 
 
 fm. 
 
 nt. 
 
 N. 
 
 ]>ai = tol 
 
 
 fids 
 
 >d = Ta 6 
 
 A. 
 
 fia-ns = to 
 
 ■vs 5 
 
 J>os 
 
 _/»o = ra 
 
 G. 
 
 fii-ze 
 
 
 pi-zo 
 
 Jfi-ze 
 
 D. 
 
 fiaim 
 
 
 Jjdim 
 
 Jjdim 
 
 In comparing, however, this old declension with that of Mod. 
 German, the reader must remember that the demonstrative, 
 after it had become confined to the function of an article, was 
 apt to become proclitic and, therefore, to undergo several 
 phonetic alterations. 7 
 
 1 Parable of the Prodigal Son, Luke xv. 11-12. 
 
 2 Parable of the Sower, Matt. xiii. 3. 
 
 3 The a is a subjoined particle (as in F. ce-ci), which maintained the 
 •consonantal ending by covering it, supra 39 and 46. 
 
 4 Remember that I.-E. 3 is always shifted to common Gr. i\. 
 
 5 Cretan. Lengthened to Gr. robs. 
 
 6 Originally *ri (Sk. td), but shortened in Greek, as also in Latin. 
 
 7 Cf. supra 34 and 66, II. 
 
218 ENGLISH AND GERMAN GRAMMAR. 
 
 II. The I.-E. demonstrative *syo likewise assumes this stem- 
 form in the nomin. sg. msc. and fm. only, Sk. syd syd. Every- 
 where else, the stem is *tyo- : Sk. nt. tyd-d, etc. The declension 
 was the same as the preceding ; bnt, since Germanic borrowed 
 from it only two forms, we need not lay any further stress 
 upon it. 
 
 § 2. Modern State. 
 
 (129) Of this elaborate declension English is well known 
 not to exhibit even a trace : it is all reduced to a single mono- 
 syllable. Hence German, as having to a large extent retained 
 it, claims here the foremost place. But, at the outset, a general 
 observation is indispensable : even German no longer knows 
 anything of a stem either *so or *syo in the nomin. sg. msc. 
 fm. 1 ; in other words, the analogy of the other cases, in which 
 the stem began with a t, transferred the t to the nominative 
 and extended to the whole declension a stem *to- 2 or *tyo-, the 
 Preg. form of which is naturally */>a- or *}>ja-. This being 
 understood, we have to consider successively, in the order of 
 the above paradigm : — the sg. forms which proceed from the 
 stem */><z- ; — the pi. forms from the same stem ; — the few sg. 
 or pi. forms which proceed from the stem *}>ja- ; — and the 
 indifferent English form. 
 
 (130) I. Singular. — 1. Nominative. — In the masculine, Sk., 
 Gr. and Go. agree in showing the bare stem without any 
 ending; but the demonstrative, nevertheless, could always 
 take the ordinary ending -s of the nomin. msc. sg, thus Sk. 
 also sd-s z ; and this is just what has happened in German. 
 Now, an I.-E. nomin. *to-s, next *te-s 4 would become in turn 
 Preg. *J>e-s, then *}>e-s proclitic, whence *}>e-z, West Germ. 
 *J?e-r, O.H.G. d'e-r and G. der. — The fm. is taken from *tyo-, 
 infra III. — The I.-E. nt. exponent is -d : Sk. td-d, Gr. to, L. 
 
 1 Both of them were kept in O.E., and we shall find the fm. *sya among 
 the personal pronouns, infra 166. 
 
 2 The same is the case with Latin: nomin. sg. msc. (is-)te, fm. (is-)ta. 
 
 3 Cf. L. (is-)tu-s, (ip-)su-s, and infra 137. 
 
 4 The vocalism of the oblique cases (infra 3) intruding into the nomina- 
 tive (?). Or else, simply der = *\>ar = I.-E. *t6s, owing to lack of accent (?). 
 
ARTICLES. 219 
 
 (is-)tu-d. An I.-E. *to-d was changed to Preg. *pa-t, which 
 should have become *J>a; but the added -a prevented final t 
 from dropping, Go. pa-t-a, so that, after this -a was again lost, 
 there remained O.E. ftce-t and O.H.G. cZa-3. E. tha-t is still 
 living, but only as a demonstrative or relative pronoun and a 
 conjunction. 1 As to O.H.G. da-%, no change is found in the 
 modern article das, the relative das and the conjunction dasz 2 
 being, of course, only particular varieties of it, the latter dis- 
 tinguished by a somewhat stronger accent which is rather 
 awkwardly represented by the usual spelling. 
 
 2. Accusative. — The ending of ace. msc. is -m : Sk. td-m, Gr. 
 to-v, L. (is-)tu-m. I.-E. *t6-m gives Preg. *}>a-n, which should 
 have become *J>a ; but, final n being covered by the particle -a y 
 we have Go. fia-n-a, O.E. db-n-e, O.H.G. d'e-n and G. dew.— The 
 fm. belongs to class III. — The nt., even in the I.-E. period, was 
 always identical with the nt. nominative. 
 
 3. Genitive. — In order to obtain a correct view of this and 
 some further formations, we must recall the fact that every 
 I.-E. stem in -0- is at the same time a stem in -e- 3 : hence, a 
 stem *to- may, under certain conditions, assume a form *te-, 
 which becomes Preg. *}>e- and Go. J>i- ; and such is the case 
 here. Now the I.-E. exponent for the gen. sg. msc.-nt. in the 
 demonstratives was a double ending : either *-syo, Sk. td-sya, 
 Gr. *to-ctio changed to toio and further contracted to tov ; or 
 *so, cf. Slav, ce-so gen. of the relative pronoun. The latter 
 type, let us say I.-E. *te-so, would give Preg. *}>e-s, whence Go. 
 ki-s, O.E. tfe-s, O.H.G. d'e-s and G. des. — The fm. ending corre- 
 sponding to msc.-nt. *-syd was *-syds : Sk. td-syds. But Ger- 
 manic, following the analogy of msc.-nt. *-so, likewise sup- 
 pressed the j in the fm. exponent, whence Preg. *-sds, gen. fm. 
 *}>e-zdz, Go. }>i-zds, where the final as well as the medial s is- 
 a 2, owing to the syllable being unaccented, as shown by O.N. 
 Jy&i-rar. Final z was lost in West Germanic, and there re- 
 mained O.E. frce-re, O.H.G. de-ra de-ru and G. der. 
 
 1 Though we have some traces of it in the article, supra 66, II. 1. 
 
 2 Cf. the use of quod as a conjunction both in Latin and the Romance 
 languages. ^^,*-*<^. 
 
 3 Supra 72, II., and cf. the L. nomin. msc. (is-)te. ^^%&$? " '/ »•),>*. 
 
 (TTN I VERS ITT 
 
220 ENGLISH AND GERMAN GRAMMAR. 
 
 4. Dative. — The Gr. form tu> is not primitive, but imitated 
 from the similar form in the dative of nouns. 1 The I.-E. 
 msc.-nt. dative inserted, between the stem *to- and the specific 
 ending *-oy, a group -sm- of uncertain origin, Sk. td-sm-di ; a 
 formation well preserved in Go. f>a-mm-a, 2 O.E. &ce-m, O.H.G. 
 d'e-mu and G. dem. — The fm. dative (Sk. td-sydi, Go. f>i-zd%) 
 was like the fm. genitive, except in its final diphthong, which 
 is still visible in Go., but could not help being lost in the 
 uniformity of the "West Germanic unaccented syllables : O.E. 
 $ce-re, O.H.G. d'e-ru and G. der. 
 
 (131) IT. Plural. — 1. Nominative: msc. I.-E. Hoy, Sk. te, 
 Gr. tol ol, L. (is-)ti, Preg. *J>ai, whence Go. J?di, O.E. ttd and 
 G. de, the latter afterwards diphthongized to dea, dia, die, 
 under which shape it became blended with the nomin. pi. fm. 
 and nt. belonging with their accusative to class III. 
 
 2. Accusative. — As the ace. pi. nt. was identical with the 
 nomin. even in I.-E. times, and the ace. pi. fm. had become so 
 at least as early as the Pregermanic period, it was only natural 
 that the ace. pi. msc. should also reproduce the corresponding 
 nomin., O.E. &d, G. de, die. 3 
 
 3. Genitive. — The Preg. stem is */>e-, as in the gen. sg. The 
 •ending, in the demonstratives, was I.-E. *-som, as shown by Sk. 
 fm. td-sdm, Homer. Gr. fm. rawv = *t5.-o-wv and L. fm. (is-)td-rum 
 
 — Hd-som. Now, the Preg. msc. and nt. being of course identi- 
 cal, the fm. could not differ very much from them, apart from 
 the vocalic force of the Go. ending (viz. J?i-ze J>i-zo) which is not 
 yet satisfactorily explained, but may well be neglected here, 
 ;since it disappeared in the later languages : hence, in the three 
 genders, O.E. &ce-ra, O.H.G. de-ro and G. der. 
 
 4. Dative. — This case has nothing to do with the Gr. and L. 
 so-called dative, tois, (is-)tis, which corresponds to a Sk. in- 
 strumental tdis. Here the ending is *-ww(s), 4 and this, as also 
 in the corresponding Sk. instrumental te-bhis, is not attached 
 directly to the stem H6-, but to its pi. msc. nomin. form, thus 
 
 1 Dat. twin? from nomin. iVwos (horse), etc., cf. infra 152. 
 
 8 Supra 61 (III.) and 34, § 1, III. 
 
 8 Neither O.E. nor O.H.G. show even a trace of Go. J>a-ns = Gr. t6-v%. 
 
 4 Which will appear again in the nominal declension, infra 152, II. 
 
ARTICLES. 221 
 
 I.-E. *toy-mi(s), O. Slav, te-mi (with them), Preg. *J?ai-rni r 
 whence, in the three genders, Go. J>ai-m, O.E. ftce-m, 1 O.H.G, 
 de-m, next de-n, 2 and G. den quite regularly. 
 
 (132) III. From the I.-E. declension of the stem *to-, we 
 have been able to deduce the whole G. accidence, with the 
 exception of three forms, viz. : nomin.-acc. sg. fm., nomin.-acc. 
 pi. fm., and nomin.-acc. pi. nt. 
 
 1. The stems in -0- form their fm. in -a : to msc. *tyo the 
 corresponding fm. is *tyd, Preg. *}>j6, whence O.H.G. diu and 
 G. die. The ace. is naturally *J)jd-m, which also became G. die? 
 
 2. The nomin.-acc. pi. fm. had the exponent -s, Sk. tyd-s, 
 whence Preg. *}>jo-z, O.H.G. deo dio, G. die. 
 
 3. The nomin.-acc. pi. nt. is always identical with the nomin. 
 sg. fm., cf. L. (is-)ta bona in both cases : I.-E. *tyd, Preg. *}>jd, 
 whence O.H.G. diu and G. die. 
 
 Hence, nothing could prevent the various forms which pro- 
 ceeded from stem *tyo- being blended with the nomin. pi. msc. 
 belonging to stem *to-, so that the four-and-twenty theoretical 
 forms of the G. article become finally reduced to six. 4 
 
 ( 1 33) IV. This process of assimilation was carried on much 
 farther in English, where the stem *tyo- plays no part at all. 
 O.E. possessed a regular nomin. msc. sg. se se = Go. sa, and a 
 corresponding fm. seo, which has become E. she. In later O.E., 
 under the influence of the nt. nomin. tce-t and the oblique 
 cases, there arose 6"e and 6"eo, and these new forms again became 
 blended together, owing to their being unaccented ; later still, 
 the neuter was confounded with the msc.-fm., and all three 
 resulted in the uniform unaccented M.E. the. On the other 
 hand, the nomin. pi. ftd, from the same cause, assumed a duller 
 
 1 Cf. the E. pronominal dative the-m, infra 160, 3, and 166, II. 
 
 2 Here the m was final. It was not final in the dat. sg., cf. supra 130, 4. 
 — In order to simplify the reader's task, I have reminded him, throughout 
 the section, of every phonetic law that found an application in it ; but 
 henceforth I shall suppose him to be familiar with them. 
 
 8 The O.H.G. ace, however, is not 'diu, as theory would require, but dia, 
 a somewhat puzzling form in its vocalism. But we may leave the point 
 unsettled : for, as dia would also become die, which was favoured moreover 
 by the analogy of the nomin., the confusion between the two forms was 
 unavoidable. 
 
 4 Namely : 8 die, 6 der, 4 den, 2 dem, 2 des, and 2 das. 
 
222 ENGLISH AND GERMAN GRAMMAR. 
 
 vowel ; the identity of the nomin. and ace., in the nt. and in 
 the pi., induced the speaker to confuse the cases everywhere 
 else ; the increasing use of prepositions, influenced by the 
 imitation of French syntax, gradually turned the other cases 
 into useless and obsolete forms : in short, ultimately, English 
 was confined to the single form the, a few others, however, 
 being retained in the pronominal declension. 
 
 Section II. 
 
 INDEFINITE ARTICLE. 
 
 (134) In the sg. E. uses an and a, the unaccented and now 
 invariable forms of owe = O.E. an. G. spells ein the word for 
 the indefinite article, as well as for the number, but the stress 
 in pronouncing the former is much weaker. The Or. negative 
 kein is O.H.G. dihh-ein, M.H.G. dech-ein, where, the e no longer 
 being sounded, the initial t has become unpronounceable * ; but 
 the element dihh- is quite unexplained. 2 The declension of 
 both is the ordinary declension for all determinatives. 3 
 
 G. has no indefinite article for the plural number : mensch-en 
 (some men), wort-e (some words). E. either uses an-y = G. 
 ein-ige (several, some), or some = Go. sum-s (a certain) 4 = Gr. 
 a/x,6-<s= Sk. sama- (whoever), both undeclined. 
 
 1 Cf. I.-E. *kmtdm=*tkmtdm, supra 123. 
 
 2 The O.H.G. doublet nihh-ein is much plainer. The E. negative is the 
 pronoun none=n-one, shortened to no (adjective and negative particle) = O.E. 
 na, cf . G. n-ein. 
 
 8 Cf. infra 159 and 169. 
 
 4 See the sentence quoted above (128) from Ulfilas. Likewise sg. in E. 
 some-body, some-thing, etc., cf. any-body, etc. The inserted -sm- in Go. 
 J>a-mm-a=I.-E. *t6-sm-0y (supra 130, 4) is probably a reduced grade of this 
 stem. 
 
CHAPTER II. 
 NOUNS. 
 
 (135) In the declension of E. and G. nouns, we have to deal 
 with three grammatical categories: gender, number, and case. 
 
 Section I. 
 
 GENDER. 
 
 (136) As a general rule, — excepting the class of stems in 
 -a, which are all feminine nouns, and the specific ending of 
 the nomin.-aoc. in neuter stems in -o-, 1 — I.-E. gender had no 
 influence on the accidence of nouns : thus, in Latin, soror is 
 declined like dolor, and fructus, like manus. On the contrary, 
 in the Germanic declension, the genders have become of much 
 importance, and are now heading the G. paradigms. The 
 reason for this change is obvious : when the final syllables had 
 been dropped, and the stem-class to which a noun originally 
 belonged was no longer known, several nouns of various genders 
 left the declension that suited their etymology, and were forced 
 into agreement with other nouns of the same gender, but of 
 different formation. Of this process, many instances will occur 
 below. Here, one will suffice: G.flut (flood), which ought to 
 be masculine, since it is an I.-E. stem in -tu- (Go. flo-du-s, 
 O.H.G./ttoi msc), became feminine, together with many like 
 it, in imitation of words which came from stems in -U-, thus 
 
 "fm. bucht (bent, gulf) 2 : whereupon both bucht and flut, and a 
 large number of other feminines, were declined after the pattern 
 olfrau (lady), which is a stem in -en-. 
 
 1 L. bonus bona bonu-m, supra 72-73. 
 
 2 Supra 78, 2-3. Cf. L. msc. fructus borrowed and changed to G. fm. 
 frucht. 
 
 2S3 
 
224 ENGLISH AND GERMAN GRAMMAR. 
 
 English and German, although they kept the three I.-E. 
 genders, distributed them in quite a different way. 
 
 It must be owned that the I.-E. distribution was very arbi- 
 trary and corresponded as little as possible to the meaning of 
 the nouns to which it applied. Generally speaking, a noun 
 denoting a male being or agent was masculine ; the names of 
 female beings or agents, feminine. But there was hardly any 
 further attempt made to lend any system or order to the chaos 
 of grammatical gender. We are not even able to assert that 
 the names of male or female animals were all, respectively, 
 masculine or feminine. 1 As for the names of inanimate beings 
 and the abstract nouns, there is no rule whatever : some are 
 masculine, some feminine, and some neuter ; thus, among the 
 nouns denoting instrument, though many end in -tro-m, there 
 are also several stems in -tro-s or in -trd. 2 
 
 In German this state of chaos still continues. This, how- 
 ever, must not be misunderstood as meaning that German has 
 retained unaltered the primitive distribution. Since the dis- 
 tinction of genders is arbitrary, it is apt to become unsteady, in- 
 asmuch as a word may easily change its gender if the mind does 
 not perceive any reason for assigning it to one class rather than 
 to another. Such changes even took place in the short interval 
 between Latin and French 3 ; and they would be much greater 
 in the course of forty centuries. Much more surprising, in 
 fact, than this occasional gender- shifting, is the marvellous 
 agreement with I.-E. gender, shown in a great number of cases: 
 thus, in a single word, I.-E. *yugo-m, Sk. yuga-m,. Gr. £vyo-v>. 
 L. jugu-m, Go. juk, G. joch, Slav, igo, all neuter ; in the whole 
 class of feminines in -w- = L. ion- (L. fm. leg-id, G. fm. hbh-e = 
 hoh-i) ; in the whole class of feminines in -ti-, G. -t ; in the 
 neuter formation of plurals in -er, etc., etc. 4 
 
 1 Indeed, most European languages, however logical they claim to be,, 
 have not passed this rudimentary stage : thus, in G. and F., maus and souris 
 (mouse) are exclusively feminine, schmetterling and papillon (butterfly) ex- 
 clusively masculine, so that the notion of their being male mice and female 
 butterflies actually requires a slight effort of thought. 
 
 2 L. aru-tru-m (plough), cul-ter (knife), mnlc-tra (milking-pail). 
 
 8 L. arbor fm., F. arbre msc. ; L. mOres msc, F. maurs (customs) fm. 
 4 See the formation of words, and especially 74 (2 C), 78, 80. 
 
nouns. 226 
 
 But the changes, none the less, embraced whole classes of 
 words. So it happened that the masculines in -tu-, after the u 
 had vanished, were classed with the feminines in -ti- whereof 
 the i had been lost, and thus became feminine ; a further 
 consequence was, that the suffix -heit and -heit, which was 
 a masculine noun (Sk. msc. Jce-tii-s), became an essentially 
 feminine exponent. So also, the Go. suff. -n-assu-, regularly 
 masculine, became the G. suff. -nisz, which forms only neuter 
 nouns and a very few feminines. From its origin, suff. -sal 
 ought to have been applied indifferently to masculine or neuter 
 nouns, whereas it is now exclusively neuter, with the exception 
 of three feminine words. Lastly, the diminutives, probably as 
 forming logically a class in themselves, were all made neuter. 1 
 
 The triumph of logic is seen in the English distribution, 
 where, apart from a few immaterial exceptions, the three 
 genders rigorously correspond to the male, female and neuter 
 sex (the last-named including animals in which the sex is 
 neglected, inanimate beings, and abstract nouns). Of course, 
 the language only arrived by degrees at this artificial levelling, 
 another example of which could hardly be found throughout 
 the whole of linguistic science: the O.E. genders still admit 
 the anomalies that now survive in Mod. German ; in M.E., the 
 remnants of such old finals as would still allow the recognition 
 of the original gender were dropped altogether, whereupon 
 the gender itself became slowly modified, mainly under the 
 influence of Anglo-French, and without exhibiting as yet any 
 decidedly logical tendency ; this, however, may be seen to 
 prevail from the beginning of the fourteenth century, and, at 
 the end of the sixteenth, arrived at the universal agreement 
 which is now found to exist. 
 
 Further details with regard to gender should be sought in 
 the ordinary grammars. 
 
 1 The typical diminutive, viz. the word kind (child), is neuter, though it 
 involves the idea of a sex : still more so such as do not involve it. 
 
226 ENGLISH AND GERMAN GRAMMAR. 
 
 Section II. 
 
 NUMBER. 
 
 (137) The nominative singular, whether in English or 
 German, has no ending, and resembles a bare stem. But 
 this is already known to be the combined effect of phonetic 
 reduction and analogy. In fact, Pregermanic possessed, for the 
 nominative singular, as for all other cases, even more than one 
 ending, bequeathed to it by I.-E. ; but from their very nature 
 these endings were doomed to disappear. 
 
 From the Greek declension, which is extremely well pre- 
 served, we are able to restore five primitive types of nomin. sg. 
 The ordinary masculine and feminine stems formed this case : 
 either by a mere lengthening of the vowel in the final syllable, 
 Gr. stems a.Kp.ov- (anvil, ace. a/cju,ov-a), <pepovT- (carrying, ace. 
 <pepovr-a), irarep- (father, ace. irarep-a), nominative sg. cik/awv 
 (cf. L. homtn-em and homo), <pepu>v, Trarrip, etc. ; or by adding a 
 final -$, Gr. «nro- (horse, ace. ltttto-v), vvkt- (night, ace. vvkt-o), 
 nomin. sg. ?7r7ro-s (cf. L. equo-s), vu£=*vvkt-s (cf. L. nox), etc. 
 But feminine stems in -a, -I, -u, even in the earliest period, 
 lacked any nominative-ending : Sk. deva (mare), Gr. x™P°- 
 (land), L. terra (earth), etc. Neuter stems in -0- added a 
 final -m in the nominative as well as in the accusative : Sk. 
 yugd-m (yoke), Gr. £,vyo-v, L. jugu-m. Every other neuter stem 
 remained unchanged: Sk. nt. svddu (sweet), Gr. nt. rjhv, cf. 
 msc. 178U-5; Sk. pdgu (cattle), L. pecu, Go. faihu, etc. 
 
 Hence, West Germanic inherited a large number of neuter 
 nominatives which lacked any case-sign. With these were 
 already confounded neuter nominatives ending in -m, since 
 final m was dropped in Pregermanic. Feminines ending in a 
 long vowel first shortened, and then occasionally lost the vowel, 
 which besides was not exclusively found in the nominative. 
 The lengthened nominatives, type aK/xwv or homo, likewise 
 shortened the vowel: thus, I.-E. *kano ( = L. *cano "the 
 singer") became Go. hana (cock), O.E. hona, O.H.G. hano; 
 and this short vowel again vanished later on, G. hahn. In 
 
nouns. 227 
 
 short, four of the five I.-E. types could not help becoming 
 nominatives without a case-sign in West Germanic. 
 
 There yet remained one type, and that the most important, 
 the nominative in -s. This -s ought to have survived, at least 
 after an accented vowel. Gothic, indeed, kept it, and even 
 shows it by analogy 1 after an unaccented syllable. On the 
 contrary, in Old Norse and West Germanic, after the -s of 
 nomin. sg. had become -z when following an unaccented vowel, 
 this -z analogically intruded everywhere else : whence the 
 regular O.N. -r, and the total loss of the termination in West 
 Germanic. A further consequence in the latter group is that 
 the plural-exponents and case-endings henceforth seem to be 
 adapted, — not to a declensional stem, as theory would require, 
 and Go. and O.N. grammar still show them to be, — but to 
 the nominative singular : which is as much as to say that this 
 case is no longer a case ; in the eyes of an English or German 
 speaker, it is the word itself. 
 
 The singular having attained this simplicity of form, the 
 plural necessarily followed : the various stems being no longer 
 distinguishable from one another by the presence or absence of 
 a nominative-ending, or by the character of their stem-vowels 
 which had vanished long before, their declensional varieties, 
 based on hidden differences in etymological structure, naturally 
 tended to become levelled under the influence of analogy, so 
 that a merely outward likeness, or occasionally, as noted above, 
 a logical association would substitute, here and there, some new 
 classification for the grammatical categories inherited from the 
 past. In what measure the German plural, 2 after being thus 
 shifted, either reproduces or alters the primitive formation, is a 
 point easily settled by glancing at the curious and, in the main, 
 regular forms exhibited by the Gothic declension. 
 
 1. Stems in -o- and -yo- : 
 msc. sg. dags, pi. dagos, cf. G. tag tag-e ; 
 
 nt. „ vaurd, „ vaurd-a, „ „ wort wort-e (but also wurt-er) ; 
 msc. „ harji-s, „ harjos, „ „ heer heer-e (nt.). 
 
 1 Cf. supra 63 and the instances below. 
 
 2 English, with its almost universal plural in -«, does not here concern us: 
 in it the analogical levelling is complete. 
 
228 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 2. Stems in -a and -yd : 
 fm. sg. airfia, pi. airpos, 
 
 „ ,, lidifci, ,, hdiJ?jos, 
 
 3. Stems in -i and -«- : 
 msc. sg. gast-s, pi. gasteis, 
 
 fm. ,, de}>-s, „ defieis, 
 msc. ,, sunu-s, „ sunjus, 
 fm. ,, handu-s, „ handjus, 
 
 4. Stems in -ew- and -?/ew- (so-called weak declension) : 
 msc. sg. bloma, pi. bloman-s, cf. G. blume blumen (fm.) ; 
 
 „ „ arfe/a, „ arbjan-s, „ 
 
 „ ,, garda, „ gardan-s, but 
 
 ,, ,, liana, ,, hanan-s, „ 
 
 fm. „ tuggd, ,, tuggon-s, cf. 
 
 ,, ,, hauhei, „ hduhein-s, „ 
 
 hairto, „ hairton-a, „ 
 
 but G. ertZe er<2en ; 
 „ „ fteide heiden (heaths). 
 
 cf . G. grosi ^ttste ; 
 but „ f/iai that-en ; 
 cf. ,, so^w so/iwe ; 
 ,, „ frawd hiinde. 
 
 erbe erben ; 
 garten giirten ; 
 AaTrn holme * ; 
 zunge zungen ; 
 Tw/ie hohen ) 
 herz herzen. 
 
 nt. 
 
 5. Stems in -es- (Gr. epe/Jos, gen. €pe'/3€(o-)-os) : 
 nt. sg. ro'^'s, pi. riqiz-a, cf. G. &aZ6 kalber. 2 
 
 6. Stems in any other consonant : 
 
 msc. sg. frijond-s, pi. frijond-s, but G. freund freund-e ; 
 fm. „ dauhtar, „ daiihtr-jus, cf. „ tochter tochter. 
 
 Now, let the reader reflect that O.E. and O.H.G. had exactly 
 the same declensional classes as Go. itself, and he will under- 
 stand how far the modern languages have since gone astray. 
 Our further task is to show the roads they took. 
 
 § 1. General Remarks. 
 
 (138) Nowhere do English and German more widely dis- 
 agree, than in the part of their grammar which deals with 
 plural-formation: the characteristic English type is unknown 
 to German ; to the manifold German types, English opposes 
 scarcely a dozen survivals. And yet nothing is more certain 
 
 1 Also regular hahnen, but less used. 
 
 2 In this and the preceding class, the outward agreement of Gothic and 
 German is striking at first sight ; but, if we scrutinize it closely, we perceive 
 an intrinsic difference, namely, that an integral element of the Go. noun 
 has become a G. plural-exponent. 
 
nouns. 229 
 
 than that both languages once started from a common declen- 
 sion, as illustrated by the Go. examples given above, and traced 
 by linguistic comparison up to I.-B. Hence, our best course 
 will be to define the starting-point and the present state of 
 these endings. 
 
 I. The I.-E. nouns formed their plural with great simplicity, 
 namely : the masculine and feminine, by adding a syllable -es ; 
 the neuter nouns, by adding a vowel -a. 
 
 1. When the stem ends in a consonant, the ending remains 
 pure: Sk. pdd-as (feet), Gr. 7ro8-es, L. ped-es ; Sk. pi. bhdrant-as 
 (carrying), Gr. <pepovT-es, L. ferent-es. 1 
 
 2. When the stem ends in -6- or -a (Gr. and L. so-called 
 second and first declension), the £ is contracted with the pre- 
 ceding vowel to a long vowel : I.-E. *ekwo-s (horse), pi. *ektvo-es, 
 whence *ekioos, Sk. dgvds, Preg. *ehtvoz; I.-E. *ekwd (mare), 
 pi. *ekwd-8s, whence *ekwds, Sk. dqvas, Preg. *ehivoz, etc. 2 
 
 3. When the stem ends in -i- or -u-, this vowel passes to 
 the normal grade -ey- or -ew- before the plural-ending -es : 
 I.-E. root *tri (three), declined *trey-es, Sk. trdy-as, Gr. *Tpiy-t% 
 *rpe€<> Tpets, L. *trey-es *trees tres; so also, Sk. gali-s (march), 
 pi. gatdy-as, Gr. fxdvn-s (prophet) pi. /u,avT«s, L. hosti-s (enemy), 
 pi. hostes; I.-E. *swddu-s (sweet) pi. *swadew-e's, Sk. svdddv-as, 
 Gr. *rj8iF-e<> ^Sces ^Sets, cf. L. manu-s (hand) pi. manus, etc. 3 
 
 4. If the neuter stem ends in -o-, the plural sign -a simply 
 replaces the o : I.-E. *yugo-m (yoke) pi. *yugd, Sk. yugd-m 
 yugd, Gr. £,vyo-v £vyd, L. jugu-m juga,* Go. juk juka, etc. 
 
 5. In any other case, the -a is added to the stem- form : I.-E. 
 nt. *gen-os, stem *gen-es-, pi. *gen-es-d; Gr. ycV-os, pi. *yev-ecr-a 
 •yevca ycvrj ; L. gen-us, pi. *gen-es-a genera ; Go. riq-is (darkness) 
 
 1 The L. lengthening is later and secondary, being borrowed from the 
 following class of words : see Henry, Gramm. of Gr. and Lat., no. 206. 
 
 2 The Gr. and L. forms are quite different : respectively Xiriroi and equi— 
 *equoi, *'{irirai. and equcc = *equdi. But, that is because here the plural of 
 these stems has followed the analogy of the similar demonstratives, cf. Gr. 
 ol oi (article) and L. (is-)ti (is-)tee, supra 131. 
 
 3 In other words, here as in the preceding case, final s follows a long 
 vowel, which is the product of contraction between the final vowel of the 
 stem and the e of the termination. 
 
 4 Shortened in Gr. and L., but still kept long in Preg., since it occurs as 
 a short vowel in Gothic : if it had been short in Preg., it must have vanished 
 in Gothic, supra 34. 
 
230 ENGLISH AND GERMAN GRAMMAR 
 
 = Gr. eoe/Jeo--, pi. riq-iz-a; O.H.G. chalb (calf) = Preg. *TcaVb-iz t 
 pi. chelb-ir(-u) = Preg. *MZb-iz-d. 
 
 II. On the other hand, the English and German plurals may 
 be divided into five main classes. 
 
 1. Plurals in -5: exclusively English ; without any dis- 
 tinction of gender; no equivalent in German. 
 
 2. Plurals in -en or -n : no metaphony ever occurring ; 
 masculines (bote boten, christ christen), feminines (zunge zungen, 
 frau frauen) and a few neuters (auge augen, herz herzen) ; so- 
 called weak declension. 
 
 3. Plurals in -e without metaphony : German masculines 
 and neuters (tag tage, arm arme, wort worte). 
 
 4. Metaphonical plurals in -e or without any ending : 
 masculines (sohn sohne, vater viiter) and feminines (hand hande, 
 mutter mutter) ; cf. E. goose geese. 
 
 5. Metaphonical plurals in -er : neuters (kalb Mlber, 
 lamm lammer, haus hduser) and a few masculines. 
 
 Let us now inquire how it could happen that the latter 
 system proceeded from the former. 
 
 § 2. Plurals in -s. 
 
 (i39) I* has been seen that every I.-E. masculine or 
 feminine noun formed a plural which ended with s. Now this 
 s, according as it was or was not immediately preceded by the 
 accent, ought either to survive, or to be shifted to z and then 
 disappear. History confirms this theory : O.E. has a minority 
 of plurals in -s, and a large majority of plurals without -s, that 
 is to say, in -z 1 ; but in German, however far we go back, the 
 latter had eliminated the former and are found throughout. 
 O.H.G. in consequence as completely lacks the -s in the 
 nominative plural, as West Germanic lacks it in the nomin. 
 sg. : a new way has been found for plural-formation, and thus, 
 English having taken to propagating by analogy 2 the 
 
 1 As to these being in a majority, the reader is referred to § 1, where it 
 appears that the 8 in the termination -es never had the I.-E. accent, so that 
 the * could never be immediately preceded by the accent, except when the 
 stem itself ended with an accented vowel which had coalesced with the 
 initial e~ of the ending. Now this case, of course, is by far the rarer. 
 
 8 Analogy here is not the only responsible factor: after the Norman 
 
NOUNS. 231 
 
 very exponent that had become lost by analogy in 
 German, it happened that the two sister- tongues diverged 
 from each other as widely as possible. 
 
 The main processes of the English evolution may be easily 
 and shortly retraced as follows. 
 
 1. Masculines. 1 — 1. The plural-exponent s could nowhere be 
 better preserved and nowhere more widely diffused, than in 
 the class of primitive masculine stems in -o-, because there the 
 words accented on their final syllable are rather common, so 
 that the final s very often followed an accented vowel. We 
 need but quote such Gr. words as oTpa/3o'-s, <f>opo-<;, 8eivo-s, 
 Avto'-s, A.a/«rpo-9, t7T7rtKo-s, etc., etc. Thus, if we consider an 
 I.-E. stem *dhoghos (day), 2 pi. *dhoghos = *dhogho-e's, the cor- 
 responding Preg. forms are sg. *b~ , a%as, pi. *Sagds, Go. dags 
 dagos, and nothing could be more regular than that the s should 
 be retained in the O.E. pi. dagas of sg. dceg, whence E. day 
 days. From this the O.E. plural-ending -as, which became E. 
 •es or s, was extended to all similar nouns, whatever might 
 have been their original accent: O.E. wulf (L. lupus) wulf-as, 
 weal (L. vallus) weall-as, paift (Gr. 7tcitos) pab"-as, ende (Sk. 
 dntyas) end-as, earm (L. armus) earm-as, etc. ; E. wolf wolv-es, 
 ivall walls, path paths, end ends, arm arms, and so indefinitely. 
 
 2. The same analogy, also as early as the O.E. period, largely 
 contaminated the old stems in -i-, since, their -i- having been 
 dropped, they were no longer distinguishable from stems in -o-. 
 O.E. lyge, for instance, is Preg. *lu%i-z, as shown by the early 
 metaphony which did not take place in G. lug : now, the old 
 plural is lyge = *lu^lz (G. luge) ; but another analogical plural 
 is lyg-as, whence E. lie lies. Words in which the i had left no 
 
 conquest, French doubtless played a great part in this expansion of E. s ; 
 for, since words borrowed from F. used to form their plural in this way, 
 native words could easily adopt the same exponent, the more so as some of 
 them already admitted of it. At any rate, it is a remarkable fact, that the 
 s-plural should have reached such a high degree of development in the two 
 Germanic languages which were brought into nearest contact with French, 
 that is, in English and Dutch, whereas it is lost everywhere else. A few 
 instances occur also in Low German (madcliens " lasses," frauleins "misses "), 
 but these are generally acknowledged to be due to a modern French influence. 
 
 1 Starting, as a matter of course, from the historical gender. 
 
 * Cf. supra 72, II. 
 
232 ENGLISH AND GERMAN GRAMMAR. 
 
 trace were naturally even more liable to be altered in this 
 way: thus O.E. giest — *%asti-z ( = Gr. gast) has no other plural 
 but giest-as, E. guest guests, and wyrm — *wurmi-z (cf. Sk. 
 kfmis, Gr. wurm) has only wyrm-as, E. worm worms, and so 
 forth. But the stems in -u-, as long as they keep the u, remain 
 unaltered (O.E. sunu, pi. suna) ; this having vanished, their 
 plural also gives way to the influence of analogy (E. son sons). 
 
 3. Again, and as early as O.E., the same s became attached 
 to one masculine noun denoting relationship : feeder, pi. fcedr-as ; 
 E. father fathers. But of bro<5or the pi. is still hrob~or or brotSru ; 
 later we have brother brothers. 
 
 4. No less intelligible is the analogical process which modified 
 the plural of the stems in primitive -en- and nearly effaced in 
 English the so-called O.E. weak declension. We have seen 
 and are yet to see further on 1 how it happened that the n, 
 though lost in the nomin. sg., re-appeared in the other cases. 
 Hence, a Preg. *yiman- (man), sg. *%umd = "L. homo, pi. 
 *£uman-iz = ~L. homin-es, becomes O.E. sg. guma, pi. guman; 
 and, in the same way, we have mona monan, flea ( = G. floh) 
 flean, creda credan, etc. As long as this characteristic a was 
 retained in common pronunciation, the weak declension re- 
 mained free from any external corruption ; but, the a being lost 
 in M.E., nothing was left which could prevent such words as 
 moon, flea, creed, etc., from forming such plui'als as moons, 
 fleas, creeds, etc. The pi. oxen from ox (O.E. oxa oxan) alone 
 survived from this class. 
 
 5. The same is the case with the old metaphonical plurals. 
 A Preg. *frijonds *frijond-iz had become O.E. sg. freond and 
 pi. friend (cf. Mod. E. foot and feet) ; but, the plural s being 
 added by analogy to the latter form, it gave way to a pleonastic 
 plural friends, from which analogy again easily transferred the 
 metaphony to the sg. friend 2 ; and so also with O.E. feond fiend, 
 now fiend fiends. 5 
 
 1 Supra 74, 87, 114, 117, and infra 140. 
 
 2 The more so, because there already existed in the sg. a metaphonical 
 form, namely the O.E. dat. friend =*freond-i (cf. L. d&t.ferent-i). 
 
 8 Many other analogical blunders might have taken place : from sg. foot 
 might have come a pi. 'foots (O.E. fot-as is occasionally met with), or a pi. 
 *feet-s, whence there might have proceeded a sg. 'feet, or, in another form 
 
nouns. 233 
 
 II. Feminines. — 1. O.E. brofror has been seen to keep its 
 plural unaltered. It stands to reason that the names of female 
 relationship should keep it even better : modor, dohtor, sweostor ; 
 pi. modru, dohtor, etc. But, in imitation of father and brother, 
 these nouns also took the ending s in M.E., mothers, daughters, 
 sisters. 
 
 2. The path being thus opened, no feminine could escape, 
 especially as the stem- vowels in all had become dull and in- 
 audible, while in a great many the gender had a tendency to 
 shift. Thus, the long specific a of the fm. sg., which had 
 become a dull short vowel in O.E., being no longer sounded in 
 M.E., we find such contrasts as: O.E. cearu, pi. ceara, but E. 
 care cares; O.E. ondswaru, pi. ondswara, E. answer answers, etc. 
 
 3. The same is the case with fm. stems in -i- and in -u- : 
 O.E. tid ( = Gr. zeit), pi. tld-e, E. tide tides; O.E. hond, pi. 
 hond-a, E. hand hands; OJ&.fl&r, ipl.fldr-a, E. floor floors, etc. 
 
 4. Further with stems in -en- (weak declension) : O.E. tunge, 
 pi. tungan (cf. Gro. tuggo tuggon-s), E. tongue tongues; O.E. 
 eortSe, pi. eortSan (passed to the weak declension like G. erde 
 erden), but E. earth earths, etc. 
 
 5. Lastly, with the metaphonical stems : O.E. boc, pi. bee, E. 
 book boohs 1 ; O.E. burg, pi. byrg, now borough boroughs, etc. 
 
 III. Neuters. — 1. Neuter nouns could never have an s-plural, 
 and therefore O.E. knows nothing of this formation. But the 
 analogy of masculine stems first introduced it into neuter stems 
 in -o-, a natural process ; for a great many masculine nouns had 
 become neuter through logical gender-shifting, and, these of 
 course keeping their old s-plural, there appeared no reason 
 why such a plural might not as well be added to any other 
 neuters ; on the other hand, the old neuter plural-ending -a (-6) 
 being now phonetically dropped, the plural in this class did not 
 differ from the singular, 2 although in most cases the speaker 
 
 of declension, the pi. *child-s *ox-es from child ox, etc. The few plurals 
 which are known to be formed without -» and which will occur below, are 
 interesting archaisms which succeeded in resisting the assaults of progressive 
 analogy. 
 
 1 The metaphony survives in beech, which was transferred to the sg., like 
 friend above. Cf. supra 33, 1. 
 
 2 Still sheep, pi. sheep, infra 143, II. 
 
234 ENGLISH AND GEEMAN GKAMMAE. 
 
 would -wish to distinguish them. Hence: O.E. wif, word, geoc y 
 deor ; pi. wif, word, geoc ( = Go. juka), deor; but E. wife wives, 1 
 word words, yoke yokes (deer still remaining deer in the pi.), 
 etc., etc. 
 
 2. Stems in -i- : O.E. spere, pi. speru, E. spear spears; O.E. 
 gewile, pi. gewilu, E. will wills, etc. 
 
 3. Stems in -en- (weak declension) : O.E. eage = Go. a«ad, pi. 
 eaaaw = Go. dugon-a, E. e^/e eyes; O.E. eare earan, E. ear ear-s, 
 etc. 
 
 4. Stems in -es- : here analogy was long kept at bay. Several 
 representatives of the G. type kalb kcilber occur in O.E., viz. 
 did, pi. cild-ru, cealf cealf-ru, lomb lomb-ru, etc. This ending 
 not only survived, but even spread by analogy at the beginning 
 of the M.E. period. It disappears in later times, slowly giving 
 way to calv-es, lambs, folks (cf. G. volk-er), which are directly 
 derived from the sg. calf, lamb, folk, so that child-r-en 2 remains 
 the only instance of the former formation. 
 
 § 3. Plurals in -en. 
 
 (140) The plural type which belongs to the Weak 
 Declension is extremely simple, if we only bear in mind that 
 I.-E. stems in -dn- and the like 3 lost the n in the nomin. sg. 
 and kept it everywhere else. So, an I.-E. stem *uks-6n- (ox) 
 was sg. *uks-6, pi. *uks-6n-^s, whence Sk. sg. uk§-d, pi. tiM-dn-as,. 
 consequently Go. auhs-a auhs-ans, E. ox ox-en, G. ochs ochs-en. 
 So also, I.-E. stem *kan-6n- (singer), sg. *kan-6, pi. *kan-6n-es,, 
 which would be L. *can-d *can-on-es : Go. has regularly han-a 
 (cock) Jian-ans, O.H.G. han-o han-on, and G. Tiahn hahn-en. 
 
 It is easy to restore the historical process : when the final s 
 in the plural, retained in Gothic, had ceased to be sounded in 
 West Germanic, the preceding nasal syllable, the only 
 remnant of the ending, was mistaken for the plural- 
 
 1 This, moreover, was shifted to the fm. gender at an early date. 
 
 2 This again corrupted from another cause, infra 140 and 147, 2. But 
 popular speech sometimes keeps it pure : thus we read childer in Jane Eyre r 
 as used by an old servant-maid. 
 
 8 Cf. supra 74 and 137. 
 
nouns. 235 
 
 exponent, for the sole reason that it was missing in the 
 singular. This gave birth to a class of plurals formed by an 
 exponent which Pregermanic knew nothing of. The observa- 
 tion, of course, applies not only to masculines as exemplified 
 above, but also to similar feminines, which had lost the final n 
 of the nomin. sg. only in the Preg. period (Go. tugg-0 tugg-on-s, 
 O.E. tung-e tung-an, O.H.G. zung-a zung-un), and even more to 
 neuter nouns, which lacked the plural -s and never had it (Go. 
 hairt-o hairt-on-a, O.E. heort-e heort-an fm., G. herz herz-en). 
 This formation, nearly forgotten in English, is a living and 
 active one in Mod. German. 
 
 The further point, why here it proscribes metaphony 
 altogether, is likewise easily settled by referring to its origin. 
 The word-classes to which it can apply are but five in number : 
 — (1) masculines in -on-, Go. hana hanans, bloma blomans, G. 
 hahn hahnen, blume (now fm.) blumen ; — (2) masculines in -yon-, 
 Go. arbja arbjans, G. erbe erben; — (3) feminines in -on-, Go. 
 tuggo tuggdns, Q.zunge zungen; — (4) feminines in -yon- and -%n- y 
 Go. managei manageins, hduhei hduheins, G. menge mengen, hohe 
 hohen; — (5) neuters in -on-, Go. dugo dugona, G. auge augen. — 
 Now, in these five classes, we see plainly that the first, third 
 and fifth can never have any metaphony in the plural, whereas 
 the second and fourth must have it already in the singular ; 
 which is as much as to say that the plural of this declension 
 can never differ from the singular as to the vowel of the stem- 
 syllable. 
 
 Such being the case, it may well seem somewhat strange 
 that, English having preserved three plurals of this kind, viz. 
 oxen, brethren and children, one of them at least should be a 
 metaphonical form. But brethren is not an English word (O.E. 
 pi. brdftor) : O.N. had for this noun a metaphonical plural, like 
 the G. pi. brilder; thus, M.E. brethre was borrowed from O.N. 
 and afterwards received an added n on the analogy of children. 1 
 Besides the pi. oxan, O.E. had also a metaphonical plural exen> 
 which probably proceeded from the other vocalic form of suff. 
 
 1 Because both denoted relationship. We need scarcely observe that 
 brethren now belongs exclusively to ecclesiastical language. 
 
 OF THE 
 i VfWT T7 T 13 o T T \ 
 
236 ENGLISH AND GEBMAN GKAMMAK. 
 
 -on- (-en-), found in L. hom-in-es. Lastly, child-r-en obviously 
 ■contains a pleonastically double plural-ending. Every other 
 weak plural has been replaced by the s-plural. 
 
 In German, the weak plural was also exposed to the influence, 
 though far less active, of other formations ; and, though it gave 
 way here and there, yet on the whole it has rather gained than 
 lost ground. The facts may be stated briefly, as follows. 
 
 (141) I. Masculines. — 1. The n-plural requires no ex- 
 planation in the masculines which still end in -e ; for this -e 
 represents, as a matter of course, a prehistorical long vowel, — 
 otherwise it would have been dropped, — and this again is as 
 a rule the representative of the final vowel of L. hom-6 = Go. 
 gum-a, pi. gum-an-s = archaic L. hom-on-es. Thus, we have 
 regularly : knabe hnabe-n, robe rabe-n, lowe lowe-n (cf. L. led 
 leones), etc.; a class which naturally comprises the names of 
 nations, Sachse Sachse-n (cf. L. Saxo Saxones), and further the 
 adjectives used with the same meaning, Deutsche Deutsche-n, 
 and, generally speaking, every adjective used substantially. 1 
 
 2. But this sg. -e was not likely to be kept everywhere ; for, 
 though the old law of Germanic finals would protect it, the 
 gradual weakening of modern finals, or even analogy alone 
 might easily prevail against it : thus, the sg. for " ox " is ochse, 
 but also ochs, so that, in ocJisen when contrasted with ochs, the 
 plural-sign seems to be -en. The same is the case in many 
 words: bauer (peasant) = O.H.G. (gi)buro; graf (earl) =M.H.G. 
 grave = O.H.G. grdvo grdvio ; O.H.G. heri-zogo "leader of an 
 army," wherein the element -zogo bears the same relation to 
 the vb. Ziehen — Go. tiuhan (to lead, to draw) that a substantive 
 *duco would bear to the L. vb. ducere ; the pi. of *duco would 
 be *ducones, hence O.H.G. pi. herizogon, M.H.G. herzoge herzogen, 
 G. herzog (duke) and pi. herzog-e (now corrupt, see below) ; G. 
 7ierr = O.H.G. herro (lord), pi. Jierr-en; G. mensch = O.H.G. 
 ■mennisco, an adjective used substantially, or a noun derived 
 from an adjective *mann-iskd-s, the latter being formed in the 
 same way as Deutsch = Gro. J>iud-isk-s, cf. Go. adjective mann- 
 inks (human) and observe the regular German metaphony, etc. 
 
 1 Cf. infra 156. 
 
nouns. 237 
 
 3. This plural formation in -en being now steadily fixed in the 
 words which had a legitimate claim to it, analogy caused it to 
 spread : — either, through a merely material association, to nouns 
 ending with an -e in the nomin. sg., though this -e had actually 
 no relation at all to that of the weak declension, as, for instance, 
 to primitive stems in -yo-, namely hirte (also hirt) = O.H.G. hirt-i 
 = Go. haird-ei-s, 1 cf. E. shepherd = O.E. sceap-hyrde, etc.; — or, 
 above all, through logical association, and because the nouns 
 mentioned above chiefly denoted male beings or agents (knabe, 
 bote, herr, graf, narr, etc.), to other nouns of similar meaning as 
 they were imported from foreign languages into German (christ, 
 soldat, philosoph, etc.). Hence it happened that this plural was 
 deemed the specific one for borrowed words, which still for the 
 most part follow the weak declension, including even names of 
 animals (elephant) and inanimate things (planet). 
 
 4. As opposed to this increase, a small apparent decrease 
 may be observed, owing to the fact that the ending -en occa- 
 sionally intruded into the singular. This process 2 is marked, 
 in its gradual development, by three main types, as follows. — 
 Gothic has a nt. namo, pi. namn-a, which is exactly reproduced, 
 except in gender, by G. msc. name, pi. name-n. But, as some 
 nouns are seen to have regularly the pi. like the sg. (ofen, 
 boden)? the pi. namen gives birth to the sg. namen, whereupon 
 both name and namen survive side by side. So also, friede and 
 frieden* f unite and funken, wille and willen (Go. vilja, pi. 
 viljans) ; further, the -e being dropped in the nomin. sg.. fleck 
 and flecken, fels and felsen, daum and daumen, etc. — Gothic has 
 brunna, pi. brunnans. In German, the pi. brunnen produces a 
 sg. brunnen, which replaces and supersedes the regular brunn = 
 M.H.G. brunne (source). — Gothic has garda (enclosure), pi. 
 gardans. The G. pi. *garten, produces a sg. garten, and this 
 
 1 As it were Preg. *herS-ja-z " belonging to the herd." 
 
 2 Similar to the extension, referred to above, of metaph. pi. friend to the 
 sg., tupra 139, I. 5. 
 
 s O.H.G. ovan, bodam. The metaphony in the plural is later, infra 146, 
 1.3. 
 
 * For this word,, the process is still more intricate, for the stem did not 
 legitimately belong to the weak declension, the nomin. sg. being Preg. *fri- 
 \>u-z. It became first shifted to the weak declension and then corrupted. 
 
238 ENGLISH AND GERMAN GRAMMAR. 
 
 again, after the analogy of vdter, cipfel, etc., introduces a new 
 metaphonical plural gdrten. So also : laden (shop), pi. laden 
 and laden ; rnagen (stomach), pi. magen and mdgen, etc. 
 
 5. Lastly, though seldom, a noun may be shifted from the 
 weak to the strong declension and assume in consequence a new 
 ready-made plural formation. Thus, the often quoted hahn still 
 keeps the old pi. hahnen = Qxo. hanans ; but htihne is far com- 
 moner, and the whole case- declension agrees with the latter 
 (gen. sg. hahn-s, no longer *hahn-en). The case-forms of herzog 
 likewise exclusively correspond to the strong declension ; in- 
 deed, the plural is now herzoge. 
 
 (142) II. Feminines. — 1. Most weak fm. nouns still ex- 
 hibit, in the shape of final -e, the ending -6 which characterized 
 their Preg. and Go. nomin. sg. : G. zunge, pi. zunge-n, cf. O.H.G. 
 zunga zungun and Go. tuggo tuggon-s ; so also, G. taube, sonne, 
 wittwe (O.H.G. tuba, sunna, wituwa) ; G. blume, pi. blumen, 
 shifted to the fm. gender (Go. msc. bldma blomans), etc. 
 
 2. The fm. nouns ending in -* in their Preg. and Go. nomin. 
 sg. belong to the same class in Gothic (managei manageins). 
 Not so, however, in O.H.G., where the plural does not differ 
 from the sg. : hdhi, pi. hdhi ; though the type hohln is found 
 occasionally. Later on, the I being shortened, the words menge, 
 hohe, lange, etc., become blended with the words zunge, taube, 
 sonne, etc., in the nomin sg., and consequently also in the nomin. 
 pi., viz. mengen, hbhen, Idngen, etc. 
 
 3. The weak plural-formation is no less regular in nouns 
 which no longer show an -e in the nomin. sg., but are known to 
 have formerly ended with that representative of the Preg. fm. 
 -0 of stems in -on- : G. /ra« = M.H.G vrouwe = O.H.G. frouwa, 
 which would be Go. *frdujo, fm. of frduja (lord), pi. frdujans; 
 hence, G. pi. frau-en = Go. *frdu-jon-s ; G. au (meadow) = aue = 
 M.H.G. ouwe; G. ader (vein) = O.H.G. ddara, etc. 
 
 4. After this termination, -en or -n, had become the apparent 
 plural-exponent of a large number of feminine nouns, it gradu- 
 ally invaded the whole feminine accidence, either owing to the 
 similar ending in the nomin sg., or merely from the influence of 
 the gender itself. A brief enumeration must suffice. — (a) 
 Stems in I.-E. -a : Go. airpa, pi. airfcos ; O.H.G. erda, pi. erdd ; 
 
NOUNS. 
 
 239 
 
 but G. erde erden. 1 — (b) Stems in I.-E. -I, which in O.H.G. had 
 yielded to the analogy of the preceding class : Go. bandi (tie), 
 pi. bandjos ; but O.H.Gr. sunta (sin), pi. suntd, and Gr. silnde 
 siinden; O.H.Gr. kuningin 2 kuninginnd, but Gr. kbnigin koniginn- 
 en, etc. — (c) Stems in -ti-, for the most part: Go. ndup-s 
 (need), pi. ndudeis, O.H.G. not noti, G. noth nothe regular ; but, 
 on the contrary, G. that that-en, schlacht schlacht-en, etc. ; so 
 also, tugend, arbeit, and all nouns ending in -schaft. — (d) Stems 
 in -tu-, wholly assimilated to the preceding class, viz. flut flut-en, 
 kost ( = L. gus-tu-s) kost-en, and further every substantive ending 
 in -heit or -keit? — (e) All abstract nouns in -ung, which, how- 
 ever far we go back, appear to have been distinguished from 
 masculines of similar formation by being shifted to the weak 
 declension (O.H.G. nomin.sg. -unga). — (/) One noun of female 
 relationship : Go. svistar, pi. svistrjus ; O.H.G. swester, pi. 
 swester ; but G. schwester schwester-n. — (g) Lastly, at any period, 
 any feminine noun borrowed from a foreign language, from 
 O.H.G. phlanza phlanzun, lira lirun, kiriKha kirihhun (G. 
 pflanze-n, leier-n, kirche-n), down to the latest and contemporary 
 words, as universitat-en, photographie-n, etc., etc. — In short, this 
 termination was used for every feminine, unless it originally 
 possessed and kept, or subsequently adopted a metaphonical 
 plural 4 ; and, since these two exceptional cases hardly apply to 
 more than forty nouns, the weak plural having spread here as 
 elsewhere, the termination -en or -n may be deemed the specific 
 and characteristic exponent for the plural of feminine German 
 nouns. 
 
 III. Neuters. — Neuter stems in I.-E. -on-, which naturally 
 could never have an ending -s in the plural, regularly dropped 
 their plural-exponent -a in West Germanic, and, like the mascu- 
 lines and feminines, kept the bare nasal as a plural-sign. They 
 are few in number and did not increase. We need but mention 
 the three Germanic words : Go. dugo dugon-a, O.H.G. ouga 
 
 1 M.H.G. already erden, though many nouns of this class are kept un- 
 altered in M.H.G. : gube (gift), pi. gdbe, not gaben. 
 
 2 Supra 87. Beware of confounding such nouns with stems in -i=-in-. 
 
 3 Cf. supra 109, I. and 136. 
 
 4 Indeed, it has even often replaced the metaph. pi., since every stem in 
 -i- ought to have a metaph. plural, infra 146, II. 
 
240 ENGLISH AND GERMAN GRAMMAR. 
 
 ougun, G. auge auge-n ; Go. duso duson-a, O.H.G. ora drun, G. 
 ohr ohr-en ; Go. hairt-o hairt-on-a, O.H.G. herza herzun, G. herz 
 herzen. 1 
 
 Summing up the results, we find that the plural in -en gained 
 no ground in the neuter, though it influenced the masculine to 
 a large extent, and was quite overwhelming in the feminine 
 gender. 
 
 § 4. Plurals in -e ivithout metaphony. 
 
 (143) This class was originally large, but became reduced 
 to scanty dimensions by the gradual influence of analogy : it 
 may be said to exclude all feminine nouns, 2 with the exception 
 of a few words in -sal and -nisz, which moreover were shifted 
 to the feminine gender and did not originally belong to it ; all 
 original feminines adopted the weak plural, whereas masculines 
 and neuters generally gave way to the analogy of the meta- 
 phonical plural. 
 
 I. Masculines. — Final -e in the G. pi. is the exact 
 equivalent of the E. -s (-es)-plural, 3 with but this differ- 
 ence, that English extended the s after the analogy of the cases 
 in which it had been kept, whereas German suppressed it 
 everywhere after the analogy of the cases in which it had 
 become voiced and been dropped: in other words, the E. pL 
 days being I.-E. *dhoghos and Preg. *ftay)s, the G. pi. tage no less 
 clearly suggests I.-E. *dhoghos unaccented, changed to Preg. 
 *&a%dz, whilst Go. dagos of coarse may represent either the one 
 or the other form. O.H.G. in consequence has sg. tag = Go. 
 dag-s, and pi. tagd=Go. dagos; but, as early as O.H.G., the 
 final vowel was shortened, thus taga, later on changed to the 
 duller sound of tag-e, where the -e was mistaken for the actual 
 plural-ending. So also : L. armus (flank), of which the regular 
 pi. is *armos, G. arm arm-e; Gr. 71-aTos (path), regular pi. 
 *7raT(os, G. pfad pfad-e ; L. calamus, G. halm halm-e ; Sk. caphds 
 
 1 Go. namu has been shifted to the msc, and O.H.G. wavga (G. wange), to 
 the fm. On the other hand, ende end is an old msc. in -yo- (Sk. dntya-s). 
 
 2 The regular plurals of erde, konigin, etc., have been seen to be *erde, 
 'koniginne, etc., like tage to tag. But they have disappeared. 
 
 3 And, therefore, of the so-called Gr. and L. second declension, supra 
 139, 1. 
 
NOUNS. 
 
 241 
 
 (hoof), pi. gaphds, G. huf 1 huf-e ; Sk. qrutds (heard), pi. grutds, 
 G. laut (sound) laut-e, etc. 2 
 
 Had this class retained all the nouns which would regularly 
 have been comprised in it, that is to say, all the German sub- 
 stantives which etymologically belong to the same suffix-forma- 
 tions as Gr. and L. nouns of the so-called second declension, 
 it would have been an extremely rich one. Bat it was quite 
 overwhelmed by the influence of the metaphonical class 3 and 
 is now reduced to about sixty nouns, whereof: — a half, roughly 
 speaking, may be traced baek, with certainty or at least great 
 probability, to I.-E. stems in -o- (pi. -os), in which therefore 
 this formation must be deemed primitive ; — a few others are of 
 undefined structure, as hund (dog), which may be a secondary 
 derivative, with suff. -to-, from the L-E. stem *kun- (I.-E. *kun- 
 to-s, Preg. *hun-6"a-z), or perhaps a mere corruption from the 
 originally consonantal stem *kun- (Sk. qvdn- gun-, Gr. kvwv 
 kw-6%), in which latter case the pi. in -e is analogical 4 ; — and, 
 lastly, words borrowed from L. (grad, punki), to which the 
 plural in -e was applied by analogy. 
 
 English has nothing resembling this formation : people is not 
 a plural, but a collective singular; and, besides, the word is not 
 a Germanic one. 
 
 II. Neuters. — In this class the neuters correspond like 
 the masculines to the second Gr. and L. declension, the more 
 so since here there was no possibility of an s-plural. The type 
 is I.-E. *yugo-m, pi. *yugd, Sk. yugd-m yugd, Gr. £vyo'-v £vya, L. 
 jugu-m juga, Preg. *jukd-m *jukd, whence Go. juk juka : final 
 Preg. -o ought to have become O.H.G. -ii, or to have been 
 dropped, according to a well-known distinction 5 ; but, in point 
 of fact, it was dropped everywhere, after the analogy of cases 
 where it disappeared regularly, and we find O.H.G. joh, pi. joh, 
 wort, pi. wort (cf. Go. vaurd vaurd-a = L. verbum verba), jar, pi. 
 
 1 O.H.G. huof, which presupposes a Preg. *hofaz = Sk. *$dphas- 
 
 2 It is clear that such a formation cannot, any more than the plural in -en, 
 admit of metaphony. 
 
 8 Infra 146, I. The Southern dialects have gone so far as to create the 
 plurals tdg(e), arm(e), and the like almost throughout. 
 
 4 Such it is certainly in herzog-e for herzog-en, supra 141, 2 and 5. 
 » Supra 34, § 1, II. 
 
 R 
 
242 ENGLISH AND GERMAN GRAMMAR. 
 
 jar, fuir, pi. fuir (G. feuer, pi. feuer), and a great man}' similar 
 plurals. 1 
 
 This equivocal plural type subsisted during the whole O.H.G. 
 and nearly the -whole M.H.G. period : a circumstance which 
 must have strongly favoured the analogical shifting of neuter 
 nouns to the metaphonical plural-formation in -er. 2 But, 
 though this shifting was carried out on a large scale, M.H.G. 
 in its later period also exhibits a tendency to add a final -e to 
 these plurals, under the obvious influence, either of the pre- 
 ceding masculine nouns, or of the demonstratives and adjectives 
 which occasionally accompany the neuter plural 3 ; thus, the 
 formation gives Mod. G. joch joch-e, wort wort-e, jahr jahr-e, bein 
 bein-e,* etc. Besides the greater part of neuter collectives with 
 pref. ge-, nouns in -nisz (formerly masculine), and nouns in -sal 
 = I.-E. *-tlo-, contemporary German comprises under this head 
 about fifty neuter nouns, which according to their etymology 
 may be divided as follows : — a few which immediately corre- 
 spond to neuter nouns of second Gr. and Latin declension, as 
 schwein = Jj. sulnum (adjective), werh = Gr. Zpyov, recht = L. 
 rectum; — a great many nouns of possibly the same origin, as 
 thor (door), which would be Gr. *6vpov, cf. the fm. dvpa, or rosz, 
 which can scarcely be anything but Preg. *hrossam (cf. E. 
 horse), pi. thor-e ross-e ; — some words which doubtless underwent 
 an analogical shifting, as nt. heer = O.H.G. heri= Go. msc. 
 harji-s, knie = Go. hniu = L. genu, salz = h. msc. sdl — *sald; — 
 lastly, a few boi'rowed nouns, pferd (horse), pfund, nt. kreuz 
 (cross) =L. fm. crux, f est (feast), etc. 
 
 1 Metaphony, of course, cannot occur here, and, in fact, appears nowhere, 
 but in words which already show it in the singular, that is to say, in words of 
 other classes which were analogically transferred into this ; namely stems in 
 -i- : meer =L. mare = *mari, pi. meer-e=li. mari-a ; and the neuter collectives, 
 yestirn= g< stirn-e, cf. supra 96, I. 
 
 2 Supra 80, 137, and infra 147 : thus wort had a pi. wort-e, but also a pi. 
 wort-er, both exactly synonymous, though now a slight difference in meaning 
 is observed between them. 
 
 3 In such a combination as diese wort (these words), the final -e of diese 
 naturally intruded into wort-e. 
 
 * The old plural, like the singular, survives in numeral locutions (vier 
 pfund, zehntausend pferd), because here there is no ambiguity. So also in 
 popular English (thirty year) : cf. Wright, Dialect of Windfall, 337. And 
 even currently, ten pounds, but a ten-pound note . 
 
NOUNS. 213 
 
 This very important class in German is represented in 
 English by only three plurals formed without -s : deer, pi. deer, 
 O.E. deor deor, O.H.G. tior tior, G. thier thier-e ; sheep, pi. sheep, 
 O.E. sceap sceap, G. schaf-e ; swine, pi. swine, O.E. swln swin, 
 O.H.G. swin swln, 1 G. schwein-e ; though the double (sg. and 
 pi.) sense of E. hair might well be due to its having been a 
 primitive neuter invariable in the plural number, cf. Mod. G. 
 haar haar-e. 
 
 § 5. Metaphonical plurals with or without an -e. 
 
 (144) Here, an apparent likeness, at first sight, between the 
 English and German plurals may be said to further emphasize 
 the curious and constant contrast which keeps the two lan- 
 guages apart. German has a great many metaphonical 
 plurals, Old English had some, and English has preserved six, 
 which coiTespond exactly to six German metaphonical plurals : 
 man men, foot feet, goose geese, tooth teeth % mouse mice, louse lice ; 
 cf. G. mann-er, 2 fusz-e, gdns-e, zahn-e, maus-e, laus-e. And, 
 notwithstanding all these coincidences, the English and German 
 metaphonical plurals must be traced back, respectively, to quite 
 a different origin, as having nothing in common but the form, 
 and corresponding only by aooident. This statement, strange 
 and paradoxical as it appears, will prove clear and natural to 
 any reader who bears in mind the chronological relation of the 
 two metaphonies, namely, that the English metaphony is much 
 older than the German, and therefore must have been occa- 
 sioned, in primitive O.E., by some suffix-syllables which in this 
 remote period could not yet produce it in O.H.G., whereas later, 
 that is, when German metaphony took place, these same sylla- 
 bles, being then completely dropped, could no longer produce it. 
 
 Hence, the E. metaphony is the proper characteristic of the 
 plurals of consonantal stems, while the German metaphony 
 takes its origin from the declension of stems in -*-. Both points 
 are easily verified. — A consonantal stem Preg. *fot- (foot) is 
 nomin. sg. *fot-s, pi. *fot-iz = l.-~Et. *pod-es. Now, in pi'imi'tive 
 
 1 Also, though late and rare, swin-ir, tior-ir, cf. infra 147. 
 * Only in this word the G. ending differs, cf. infra 147, 4. 
 
244 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 O.E., the i in the termination causes metaphony in the stem, 
 and we have pi. *jeti, whence historical jet. Nothing of the 
 sort can happen in O.H.G. : the 5 remains unaltered in its 
 O.H.G. form uo (dat. pi. fuoyum), so that, after the plural- 
 ending had vanished, the nomin. pi. must have been *juo$, 
 exactly like the nomin. sg. (cf. the pi. man of O.H.Gr. man). 
 Afterwards, the word analogically passes to the declension of 
 stems in -i-, and we have O.H.Gr. pi. juoyi. Finally, after 
 many more centuries, that is, in M.H.G., the new ending pro- 
 duces metaphony in the stem, whence Mod. Gf. fiisz-e. — Let us 
 now take a primitive stem in -%-. Why should it not exhibit 
 metaphony in the E. plural ? Plainly because it has it already 
 in the singular, so that the stem-vowel in the plural cannot 
 differ from the stem- vowel in the singular. A Preg. sg. *yistiz 
 is *yistlz in the plural : then, O.E. % in *yistiz occasions the 
 metaphony in the stem, and disappears, whence O.E. sg. giest, 
 E. guest ; now, whatever else may become of the plural, whether 
 O.E. regular *giest-e, or analogical giest-as, E. guests, the stem- 
 vowel appears the same in both numbers. But, on the con- 
 trary, in O.H.G., the final t-syllable had vanished long before 
 it could cause metaphony in the stem, the word thus being sg. 
 gast, pi. *gastl ; in the latter, final I is shortened, *gasti, where- 
 upon it produces metaphony, gesti; and thus comes the contrast 
 which is still seen in gast gdst-e. 
 
 (145) A. English metaphony : consonantal stems. 
 
 I. Masculines. — The I.-E. stem *pod- (foot) is declined Gr. 
 sg. 7rous 7rws = *7rw8-s, pi. 7to8-€9 ; on the other hand, we learn 
 from Sk. sg. pad and pi. pad-as that the long vowel in the 
 nomin. sg. may extend to the other cases, as, for instance, to 
 the nomin. pi., which might equally well be Gfr. *7ru>S-es : this 
 is what happened in Preg., where we have sg. *fot-s, pi. *jot-iz. 
 
 In consequence, such languages as admit of an early meta- 
 phony, viz. O.N. and O.E., agree in showing here a metaphony 
 of the plural stem : jot-r, foet-r (long ce) ; jot, jet, whence E. 
 joot, jeet. So also, we have an I.-E. stem *dont- *dent- *dy,t-, 
 meaning "tooth," Sk. ddnt-, Gr. oSovs = *o8ovt-s, L. den-s = 
 *dent-s, etc. : the form *dont-s produced Preg. *tdn)>-s, pi. 
 *tdnj>-iz, whence O.E. to$ te$, E. tooth teeth. The third plural 
 
NOUNS. 
 
 245 
 
 of this class is due to analogy ; for it is primitively a stem in 
 -o-, a secondary derivative from I.-E. *man-u- (man, Sk. mdn- 
 u-s), thus I.-E. *man-u-o-s *manwos (human, man), whence 
 Preg. *mannaz, O.E. man, mon, of which the plural is men = 
 *manniz, just as if the sg. were consonantal, E. man men. The 
 compound woman —-wif -man, pi. women = M.E. wim-men = unf- 
 men, afterwards follows without difficulty. 
 
 II. Feminines. 1 — I.-E. had a stem *mus- (mouse), nomin. sg. 
 *mus = *miis-s, nomin. pi. *mus-es, Gr. /xvs //.u-es, L. mils mur-es, 
 consequently Preg. *mus *mus-iz, exactly represented by O.E. 
 mus mys, whence E. mouse mice. The same formation holds 
 for louse lice, though the word has no acknowledged kin. 2 Gr. 
 6 S- XV V > pl- x*7 v "" c? = *X t " /(r " £S 5 similarly points to a consonantal 
 stem *ghdns- (goose), sg. *ghdns, pi. *ghdns-es, 3 Preg. *yins 
 *-&dns-iz, O.E. gos ( = *gons) ges, whence E. goose geese. And so 
 we have done with the English metaphonical plurals. 
 
 (146) B. German metaphony : stems in -i-. 
 
 1. Masculines. — 1. The T.-E. stems in -1- assume the normal 
 form -ey- before the plural-ending -es : I.-E. *ghos-1i-s *ghos- 
 tey-es, Sk. agni-s (fire) agndy-as, Gr. /xdVri-s (prophet) /AavT£is = 
 */AavT€€s = *yu.avT€j/-«s, L. hovti-s liostei = *hostees = *hostey-es, con- 
 sequently Preg. *yisti-z *yistiz, Go. gast-s gasteis, O.H.G. gast 
 gesti, G. gast gaste. To this class, in which the etymology 
 strictly requires the metaphonical process, belong such words 
 as balg (E. bellows) bdlg-e, sang sdng-e (Go. balg-s balgeis, 
 saggv-s saggveis), most probably also schlag schldg-e, wurf wiirf-e, 
 and a few others. 
 
 2. As early however as O.H.G. this plural begins to spread : 
 first, to stems in -«-, Go. sunus sunjus, but O.H.G. sunu sani, 
 
 1 Neuter nouns, since they never could have had an ending -iz, do not 
 enter into the question. 
 
 2 The c in mice and lice is an artificial spelling, intended to indicate a 
 voiceless sibilant sound. The two remaining plurals in -ce have nothing to 
 do with metaphony : dice is a mere graphic variety for 'dies (cf . twice = 
 twiet, supra 124, II., and the spelling ice=0 E. i» = O.H.G. is = G. ei*), the 
 ordinary pi. of the word die borrowed from French ; and pence is merely the 
 regular *penn(i)s, which lost its vowel when rapidly uttered. From this 
 again a new plural, some sixpences. 
 
 3 Which likewise occurs, but accompanied by a derivative suffix, in Sk. 
 hdns-as (swan, goose), and L. anser (goose) = *hans-er. 
 
246 ENGLISH AND GERMAN GRAMMAR. 
 
 now sohn sohn-e ; next, to some stems which formerly ended in 
 -0-, Go. ast-s ( = Gr. o£os = *ocr8o-9) astds, but O.H.G. ast esti, G. 
 ast aste ; to consonantal stems, O.H.G. zand zendi, fuoj fuoifii, 
 G. zahn zahne, fusz fiisze, etc. As the language grows older, 
 the tendency grows stronger : O.H.G. still keeps stuol stuola, 
 hals halsa, scalk scalka, slaf slafa, whereas Mod. G. has stuhl 
 stiihle, hals halse, schalk schalke, schlaf schldfe, ivolf wolfe, etc. 
 In short, the metaphonical plural has become the prevailing 
 formation for masculine monosyllables. 1 
 
 3. No less early it is occasionally found in some dissyllabic 
 nouns ending with a liquid : O.H.G. zahar (tear), pi. zahari, 
 now fm., with the metaphony transferred to the sg., and a new 
 weak plural, zahre zdhre-n; O.H.G. aphul (E. apple), pi. ephili, 
 now apftl dpfel. Here the plural-ending is dropped altogether, 
 and there remains no other plural-sign but the metaphony : in 
 consequence of which, metaphony was mistaken for a plural- 
 sign and assumed this function in many other words. Thus, 
 the pi. to nagel (nail) is still M.H.G. nagel-e = 'Preg. *na%ldz, but 
 now nagel; so also Preg. *fu%laz (bird), pi. *fu%ldz, has become 
 G. vogel vogel. Further, in nouns of male relationship : O.H.G. 
 bruoder bruoder, fater fater-a 2 ; but G. bruder briider, vater vater. 
 Further, hammer hammer, acker acker, etc. ; then, for ofen (Gr. 
 i7rvo-?), boden (E. bottom), hafen (a late borrowing), were formed 
 the metaphonical plurals ofen, boden, hafen, etc., whence, for 
 many words ending with a liquid or a nasal, metaphony alone 
 is now considered to be the proper plural-sign. 
 
 4. On the other hand, it may be seen from nagele above that 
 M.H.G. possessed a great many nouns with unaccented final 
 syllable, which formed their plural without metaphony, so that, 
 the plural-ending -e being once dropped, the pi. became exactly 
 like the sg. Several nouns afterwards swelled this class : 
 either such as were already metaphonical in the sg. (gdrtner) ; 
 or such as contained a vowel which was no longer capable of 
 
 1 It must be understood that, if the vowel is not capable of metaphony in 
 M.H.G. and Mod. G. (berg, weg), or if it never was {finch), then, of course, 
 both formations are outwardly identical, though historically distinguishable. 
 
 8 It need scarcely be observed that this again is an irregular plural bor- 
 rowed from stems in I.-E. -o-. 
 
nouns. 247 
 
 metaphony in M.H.G. or Mod. G. (degen) x ; or, lastly, such as 
 deserted a class in which no metaphony conld primitively take 
 place (laden). 2 Now, this type likewise, though partly influ- 
 enced by the metaphonical class, held its own and even in- 
 creased : whence the pi. always like the sg. in adler = O.JI.G. 
 adel-ar " noble eagle," sommer (cf. winter, where metaphony is 
 quite impossible), and generally in the nouns, denoting agent, 
 in -er, bohr-er, etc. 3 
 
 II. Feminines.— 1. The I.-E. plural, in stems ending in -i-, 
 is the same for feminine or masculine nouns : cf. Gr. pi. •n-o'Aei? 
 like fj.dvT€i<s, L. pi. vestes like hostes, etc. G. haut (skin) haute 
 may be said therefore to reproduce the L. cutis cutes. It is 
 chiefly in abstract nouns with suff. -ti-, that the process is 
 seen : O.H.G. maht mahti, G. macht machte, and so also kraft 
 krdfte, kunst kiinste, etc. A great many of them, however, have 
 been seen to have passed to the weak declension. 4 
 
 2. An early confusion, as often noted, took place between the 
 stems in -ti- and the stems in (primitive) -tu- : whence G. luft 
 (air) lilfte, lust liiste, etc. 
 
 3. Moreover, in O.H.G., the feminines, as well as the mascu- 
 lines, in -u- had adopted the -i-declension : O.H.G. hant ( = Go. 
 hand-u-s) hentif G. hand hdnde. 
 
 4. This formation was extended to some fm. monosyllables 
 ending either with a consonant or a long vowel : gans gdnse, magd 
 mdgde, maus mduse, laus lause, nacht ndchte, kuh kiihe, sau saue. 
 In a quite different class, the analogy of vater and briider gave 
 rise to mutter and tochter, the only feminines which show no 
 other plural-sign but metaphony. 
 
 5. In spite of this slight expansion, the weak plural, as 
 stated above, got the upper hand, so that the feminine meta- 
 phonical plural is now reduced to about forty representatives, 6 
 
 1 F. dague imported into late M.H.G. 
 
 2 Supra 141, 4 : pi. laden, but also the analogical Idden. 
 
 3 In the neuters with similar endings, feuer, fen*ter, and the diminutives 
 in -chen and -lein, the result is naturally the same, the cause being slightly 
 different, though indeed of the same kind, supra 143, II. 
 
 * Supra 142, II. 4. 
 
 * M.H.G. has also the weak declension, which still survives in the locution 
 vor-hand-en (ready, to be disposed of), literally " before the hands." 
 
 6 For the sake of completeness may be mentioned a few irregular though 
 
248 ENGLISH AND GERMAN GRAMMAR. 
 
 § 6. Metaphonical plurals in -er. 
 
 (147) Unless a neuter noun forms its plural in -e or -en 
 •without metaphony, its regular plural-sign must be metaphony 
 and an ending -er. Such is the case for the great majority of 
 neuter nouns. After all that has been said of this curious 
 word-class, 1 very few further details are here required. 
 
 1. Let us consider the Gr. nt. ($p£<p-o<s (child, young animal), 
 gen. fipecpovs = ftptfpeos = * ftp€(j>-€(r-os, pi. (3pe<prj = fipt.<pea = *f3pe<p- 
 €cr-a. From a nearly identical root 2 might have been derived 
 an I.-E. nt. *golbh-os, pi. *golbh-es-d 1 which would become 
 Preg. *kdlbaz, pi. *kdlh-iz-o, O.H.G. chalb, pi. chelb-ir(-u) chelb- 
 ir, G. Icalb halber. So also, most likely, in lamb lembir, hrind 
 hrindir, huon huonir, G. lamm lammer, rind rinder, huhn hiihner, 
 ei eier, etc., the metaphonical plural is attributable to an I.-E. 
 -es-sufnx. 
 
 2. O.E. has been seen to possess the equivalent forms, cealfru, 
 vombru. The pL of did, however, is did, and dldru is excep- 
 tional. The latter type, being, as it were, pluralized anew by 
 the addition of tbe weak plural-ending, is now child-r-en, the 
 only survival from this nt. class. 
 
 3. While English lost this characteristic formation, it became 
 largely diffused even in the earliest period of O.H.G. : thus, 
 from sg. grab, radi, liolz, kriit, etc., the pi. grebir, redir, holzir, 
 krutir, etc., now grdb-er, rad-er, holz-er, krdut-er, etc. This end- 
 ing gradually gained ground and was considered the regular 
 one, whence Mod. G. hand-er, wort-er, biich-er, hdus-er, 3 etc., etc. 
 
 easily explained forms : (1) the fm., with un-metaphonical pi., drang-sal and 
 triib-sul (formerly either nt. or msc), and the fm. in -nisz (as kennt-nisz), 
 formerly either msc. or nt. ; (2) on the contrary, four metaphonical neuters, 
 two of which, at least, are old masculines (chor chore, flosz flosze), whereas 
 the two others may well proceed from analogy (rohr rohre, kloster kloster 
 borrowed from L.). Further details belong to ordinary grammar. 
 
 1 Supra 80, 90 (VIII.) and 117 (I. 4). 
 
 2 It is only an approximation, the vocalic grade being different, and Ger- 
 manic showing an I instead of the Gr. r ; but it must be understood that 
 Gr. /3^0oy is quoted here only as a representative of an I.-E. neuter class 
 which is known to have comprised a great many primary stems. 
 
 8 O.H.G. still possesses side by side the plurals hus and hwir, bant and 
 bantir. In German, if preserved at all (bande and bander), the doublets 
 always assume different shades of meaning. 
 
NOUNS. 
 
 249 
 
 4. The few masculines which follow this declension are only 
 slightly irregular : irr-thum and reich-thum, the only nouns in 
 -thum which have kept the msc. gender, borrowed their plurals 
 from the other nouns in -thum which had been shifted to the 
 nt. gender (supra 109, II.); msc. wald is probably an old neuter, 
 G-r. nt. dAo-os (forest) ; msc. Gott is certainly a primitive neuter, 
 for Go. gup is nt. in the meaning "false god," and O.H.G. abgot 
 has already the pi. abgotir ; msc. mann, the regular pi. being 
 mann, 1 adopted an analogical plural to that of Gott, on account 
 of " G-od " and "men" being frequently opposed to each other; 
 and the same is probably the case with msc. geist (mind) — Sk. 
 nt. hedas (anger), to which msc. leib (body) was very often 
 opposed ; the other words of this class (msc. rand, ort, wurm, 
 dorn, bosewicht) are quite insignificant. 
 
 Section III. 
 
 CASES. 
 
 (148) Though German preserves twice as many cases as 
 English, it has only four, and these moreover are often repre- 
 sented by the same external form. The nominative was the 
 subject of our preceding section. Here we have to examine 
 the accusative, genitive, and dative, in both numbers. 
 
 § 1. Accusative. 
 
 (149) Either primitively, or more commonly owing to pho- 
 netic decay, the Preg. accusative was, almost everywhere, either- 
 like, or at least scarcely different from, the nominative : hence 
 it afterwards was quite blended with this case, in English 
 everywhere, and almost everywhere in German. 
 
 1. Singular. — 1. Even in I.-E., the accusative of neuter 
 nouns is always and everywhere like the nominative. 2 
 
 2. In msc. and fm. nouns, the accusative-exponent is an end- 
 ing -m, which remains -m if preceded by a vowel, 3 and becomes 
 
 1 Kept in numeral locutions : vierzehn mann. 
 
 2 Gr. rb {vylv, ytvos, Kp4as, etc. ; L. jugum, genus, caput. 
 8 Gr. Xmro-v, x^pi-". iriiki-v ; L. equo-m, terra-m, manu-m. 
 
250 ENGLISH AND GEBMAN GRAMMAR. 
 
 -m if following a consonant. 1 Since, in either case, the Ger- 
 manic laws require the final nasal to be dropped, the accusa- 
 tive, in most Gothic declensions, will be seen not to differ from 
 the nominative except by suppressing the nominative-sign -s 2 ; 
 and, since this -s again has been lost in later languages, no 
 English declension, and scarcely any German one, keeps a dis- 
 tinct form for the two cases. 
 
 3. In one class, however, the two cases are to be distin- 
 guished, not of course by the accusative-ending, which is com- 
 pletely lost everywhere, but by the form of the declensional 
 stem : namely, in the L. type hom-o hom-in-em, ley-id leg-ion- 
 em., where the n has been seen to disappear in the nomin. sg. 
 and to persist in every other case ; in other words, in the Ger- 
 manic so-called weak declension. Gothic answers the L. 
 forms with nomin. gum-a, ace. gum-an, han-a (cock) han-an ; 
 and, in fm. nouns, tugg-o tugg-on, hduh-ei hduh-ei?i, etc. O.E. 
 has likewise : gum-a gum-an, ox-a ox-an (Sk. uks-d uks-dn-am) ; 
 fm. tung-e tung-an, eor<S-a eortS-an. Lastly O.H.G. : han-o han- 
 un, has-o has-un, bot-o bot-un, Tierizog-o herizog-un, etc. ; zung-a 
 zung-un, bluom-a bluom-un, sunna sunnun, diorna (now dime 
 "lass ") diornun, etc. But this regular formation of the weak 
 accusative suffered from several encroachments. 
 
 (a) As the accusatives everywhere else were like the nomina- 
 tives, M.E. effaced the distinction in this class also : E. ox, 
 tongue, earth, etc. 
 
 (6) In the feminines in -i = -in, O.H.G. has already sup- 
 pressed it, that is to say, the form is optionally either hdhi or 
 hohin in the nomin. and the ace, whereas Mod. G. has but hohe 
 in either case, hohen being confined to the plural. In other 
 feminines M.H.G. still preserves the distinction : ace. sg. die 
 zungen, etc. But, either from the analogy of feminines belong- 
 ing to a different class (erde, gabe), where the two cases could 
 not differ, or because such a form in the sg. was liable to an 
 
 1 Gr. w6S-a, voi/iiv-a ; L. ped-em, homin-em. — For the first case, compare 
 Go. v«Z/=L. lupu-m (nomin. vul/-8 = li. lupus); for the second, Go. fot-u = 
 *fdt-um=li. ped-em =*ped-m. 
 
 8 And not to differ from it at all, when the nomin. has no case-sign (supra 
 137), as in Go. air\>a (earth), ace. a%r\>a (L. terra terra-m). 
 
NOUNS. 
 
 251 
 
 undesirable confusion with the plural (ace. pi. likewise die zun- 
 gen), Mod. G. confined the ending -n to the latter number, and 
 the sg. is die zunge, frau, sonne, taube, dime, nominative and 
 accusative alike, in all feminine nouns. 
 
 (c) A similar confusion being impossible in masculine nouns, 1 
 their declension remained by far the purer: most of them have 
 kept their regular accusative sg. (ochs-en, bote-n, hase-n, bauer-n), 
 indeed, it has even extended to some new words (christ-en, 
 philosoph-en, planet-en). 2 Yet a few are already known to have 
 been shifted to another declension, whence den hahn, den herzog, 
 etc. It happened even more frequently that a masculine noun 
 kept the weak declension in the plural, but assimilated its 
 accusative singular to the nominative ; then, of course, the 
 weak forms are as completely lost in the whole sg., as they are 
 in feminine nouns: thus we have pi. die pfauen — h. pdoones, 
 but no longer ace. sg. den *pfauen = lj. pdvonem, the identity, as 
 it were, in (der) vogel = (den) vogel having here produced a 
 similar identity in (der) pfau = (den) pfau, etc.; and further, 
 though a German correctly says die bauern and den bauern, and 
 also die nachbaren, he will say den nachbar like den mann. The 
 tendency to assimilation proved here stronger than the histori- 
 cal system of accidence, as shown even by the plural-forma- 
 tion. 3 
 
 II. Plural. — 1. The accusative of neuter nouns is even in 
 I.-E., and everywhere, exactly like the nominative. 
 
 2. In the msc. and fm. gender, the exponent for the ace. pi. 
 was everywhere an ending -ns (Sk. dg van = *d$va-ns, Gr. i7r7roi>s 
 = *?7r7rov-s, L. Iupds = *lup8-ns), which is still visible in the 
 Gothic forms, viz. nomin. vulfos, ace. vulfa-ns.* But, either 
 after the analogy of the neuter nouns, or because the nomin. 
 and ace. of several declensions had become alike, owing to a 
 peculiar phonetic treatment of the nasal when following a long 
 
 1 Because ace. sg. den boten was clearly distinct from ace. pi. die boten. 
 
 2 Cf. supra 141, 3. 
 
 8 The masculines of this class, that is to say, weak in the plural, but with- 
 out nasal ending in the singular, are about thirty : for a list, the reader may 
 refer to any grammar. 
 
 * Cf. supra 128, I. : Go. \>di = rol (61), and \>ans = rots. 
 
252 ENGLISH AND GERMAN GRAMMAR. 
 
 vowel, 1 Germanic speakers at an early period took to using the 
 former case for the latter, 2 so that they were blended together 
 as early as O.E. and O.H.G., and still more so of course ten 
 centuries later. 
 
 § 2. Genitive. 
 
 (150) I. Singular. — The formation of the gen. sg. was 
 rather complicated in I.-E., and probably even in Pregermanic ; 
 but it became analogically simplified in German and especially 
 in English. In order to understand it thoroughly, it will suffice 
 here to distinguish three main I.-E. terminations : — an expo- 
 nent -so, borrowed from the demonstratives, and used exclu- 
 sively with stems ending in -0- (second Gr. and L. declension), 
 as I.-E. *wlqd-s (wolf), gen. *wlqe-so (cf. Gr. Xvkoio = *\vko-(tio), 
 whence Preg. nomin. *wulfa-z, gen. *wulfi-za ; — an ending -os 
 or -es, which appears very clearly in consonantal stems, Gr. 
 7roS-os kw-os, L. ped-is nec-is ; — lastly, a simple ending -s, which 
 results from the contraction of the preceding exponent with 
 the final I.-E. -a of feminine stems, Gr. rjixepa ^/icpd-s, archaic 
 L. escd-s for classical escce, etc. — Observing at the outset that 
 every one of these exponents contained an s, let us enquire 
 what has become of them in Germanic. 
 
 1. German. — (a) In the last class, the $ was but seldom 
 preceded by the accent 3 : hence, West Germanic dropped it 
 and the genitive here differed but slightly from the nomina- 
 tive : thus Go. airpa, gen. airfros; but O.H.G. erda, gen. erda 
 erdu erdo. Since, on the other hand, feminines in -a formed a 
 majority of feminine nouns, analogy favoured this assimilation, 
 so that the genitive of fm. nouns, though still distinct from the 
 nominative in a great many O.H.G. words, 4 finally became 
 identical with it and the accusative. Weak feminines offered 
 a stronger and longer resistance, but gave way in their turn : 
 
 1 Go. nomin.-acc. pi. au\>bs (earths), tugguns (tongues). 
 
 * Cf. the Gr. ace. raj iroXetj, the L. nomin.-acc. manus, oves, etc. 
 
 3 The first Gr. declension, for instance, chiefly consists of nouns accented 
 on the penult ; nouns accented on the last syllable are here in a minority. 
 
 4 Namely, for instance, in stems in -ti-, the i here being kept in the gen. 
 and causing metaphony : Go. ansts (favour), gen. ansteis (cf. Sk. agni-s 
 " fire," gen. agne*) ; O.H.G. anst, gen. ensti. 
 
nouns. 253 
 
 O.H.G. d'ero zungun, M.H.G. dere zungen, but Mod. G. der zunge, 
 derfrau, etc. 1 
 
 (b) Whilst the weak feminines thus everywhere lost the n 
 of the oblique cases, the weak masculines preserved and even 
 propagated it. In fact, here, the genitival s was likewise pre- 
 ceded by an unaccented vowel ; and, in consequence, from the 
 whole genitival ending, Gr. iroifiev-os, L. homin-is, Go. hanins, 
 the n alone survived, the genitive being henceforth exactly like 
 the accusative, G. des ochsen, des boten, des herrn, 2 etc. Now, 
 we have seen that this class occasionally underwent a slight 
 decay : in the same way as ace. den hahn, den herzog, den pfau, 
 den nachbar, etc., there were created the genitives hahn-s, 
 herzog-s, pfau-es, nachbar-s, etc., with a genitival s imported 
 from the following class ; an analogy which likewise reached 
 and completely swept away the weak declension of neuter 
 stems, auge-s, ohr-(e)s, instead of O.H.G. oug-en, or-en. As to 
 the rare genitives in -ens, name-n-s, herz-en-s, we must beware 
 of classing them with the Go. genitives nam-in-s, hairt-ins ; for 
 O.H.G. and even M.H.G. drop the final s here as everywhere 
 else. Gen. namen-s proceeds from the nomin. namen, which is 
 known to be an analogical type 3 ; whereas herzens is an hybrid 
 product of two cumulative genitive-exponents (regular O.H.G. 
 and M.H.G. herz-en, and irregular Mod. G. herz-es), and schmerz- 
 ens, a later imitation of the hybrid herzens. 
 
 (c) In the ending -e-so, the s often followed an accented 
 vowel : I.-E. msc. *dhogh-eso, nt. *yug-eso, whence Go. dag-i-s, 
 juk-i-s, G. tags, jochs. The same was the case, very often, 
 with the ending in consonantal stems : Sk. pad-as, Gr. 7roS-ds, 
 
 1 A gen sg. frauen is still found as late as Goethe, Herm. und Doroth., IX. 
 123. 
 
 2 Supra 149, 3. More precisely, the O.H.G. gen.-dat. is hanin hanen, 
 while the ace. is hanon hanun ; but this difference has become lost in the 
 uniformity of unaccented syllables. — The reason must now have become ob- 
 vious, why weak masculines have all cases alike both in the sg. and pi. : it is 
 because, properly speaking, they have nowhere kept any case-sign or number- 
 exponent whatever. 
 
 * Supra 141, 4. The spoken language even went farther : then were once 
 created the genitives knabens, rabens, etc. ; but they could hardly survive, 
 since there was no nominative 'knaben, *raben (cf. E. raven) to support 
 them. ^^^■JT'^Z" '. '»~ — ~ — 
 
 S^ ^fe.* fc i. , PH4JS CNN 
 f fJt'THt r ~ v 
 
 \ UNIVERSITY 
 
254 ENGLISH AND GERMAN GRAMMAR. 
 
 O.E. fot-es. Hence, particularly in all the masculine and neuter 
 nouns which correspond to the 2nd Gr. and L. declension, — a 
 remarkably rich class, — West Germanic may be expected to 
 have, and actually has, an ending -es, O.E. dceg-es, word-es, 
 O.H.G. tag-es, wort-es ; and this case-sign, which appeared to 
 be a most convenient one, easily became diffused. Thus, 
 O.H.G. has already lamb-es, chalb-es, 1 gast-es, sun-es, and even 
 fater-es side by side with gen. fater. Next, this syllable -es, 
 often reduced to a mere -s, was gradually considered to be the 
 specific exponent of the gen. sg. for all neuter nouns without 
 distinction, and for all masculine nouns, except those which 
 either adopted or regularly retained the weak declension. 
 
 2. English. — The latter process proved still far more widely 
 energetic in English. As early as O.E., every msc. and nt. 
 nonn, unless of the weak declension, has a gen. sg. in -es 
 (dceg-es, word-es, giest-es, wyrm-es, sun-es, 2 fot-es, fcedr-es side by 
 side with gen. feeder, freond-es, lomb-es, etc.) 3 ; whereas the 
 feminine nouns are free from contamination, and preserve the 
 old genitive, which differs . but slightly from the nominative 
 (cearu " care," sorg " sorrow," lyft "air," gen. ceare, sorge, lyfte), 
 and the weak declension, in the three genders, retains its nasal 
 exponent (gen. ocean, tungan, eagan = O.H.G. ougen) : to sum up, 
 O.E. is still much purer than Mod. German. But all these 
 varieties soon disappeared : even Chaucer knows but little of 
 a genitive formed without -es, which element, now spelled 's, is 
 added to the nomin. sg. of any noun, including even feminine 
 nouns (mother's after the analogy of father s, queen's, wife's, 
 etc.), which theoretically ought not to have taken it, and never 
 have taken it in German. 4 The syntax and use of this genitive 
 belong to English grammar. 
 
 1 It need scarcely be observed that the only regular form would be *chelbir 
 (a dat. sg. chelbir-e occurs omce)=Preg. *kalh-iz-az, wherein final s, changed 
 to z, was regularly dropped, and medial s, likewise changed to z, had become 
 r ; and so also, 'lembir, etc. But lamb-es (now lamm-s) was derived imme- 
 diately from the nomin. lamb. 
 
 2 Barer however than the regular suna, and much later. 
 
 8 E. day's, word's, guest's, worm's, son's, father's, friend's, lamb's, with this 
 restriction however, that the genitive is now obsolete for nouns denoting 
 anything but a man or woman. 
 
 4 It must be noted, however, that in German also some feminine nouns 
 
nouns. 255 
 
 (151) II. Plural. — 1. Even without investigating the for- 
 mation of the I.-E. genitive plural, we are aware of the simple 
 fact, that it always substitutes, for the ending of the nomin.pl., 
 another ending with final nasal, either -dm or -dm : Gr. lttjtol 
 (horses) iiririov, £vya (yokes) £vy(ov, 7ro8-cs 7roS-u>v, L. ped-es 
 ped-iim, etc. This nasal having vanished, there remained a 
 vocalic ending, not quite the same, to be sure, as that sounded 
 in the final syllable of the nominative, 1 but not sufficiently dis- 
 similar to prevent the two being blended together when the 
 vowels were weakened ; and, since on the other hand the 
 O.H.G. nomin. pi. in -es had everywhere lost its consonant, the 
 two cases could no longer be distinguished from each other. 
 Hence arose the constant likeness of nominative, accusative 
 and genitive plural throughout the whole German declension. 
 
 2. It is a curious and singular exception, that English, poor 
 as it is in declensional forms, should appear in this single case 
 even richer than German ; for it has secured a distinct genitive 
 plural, merely by transferring to the plural the case-sign of 
 the genitive singular. German could do without it, because 
 the article was a sufficient case-exponent ; but English, having 
 ceased to decline the article, would have been confined to the 
 nse of a preposition. Hence, the genitive of child being child's, 
 it seemed a matter of course that that of children should 
 similarly be children's* Further, since in time almost every 
 E. plural became provided with an s, this ending was deemed 
 to represent at once both the number-exponent and the geni- 
 tival s, the actual form being fathers', mothers', sons', the only 
 distinction from the nomin. pi. being in the spelling. 
 
 § 3. Dative. 
 
 (152) I. Singular. — The Germanic so-called dative is in 
 reality an old locative, which was formed by adding to the 
 
 adopted it, but only proper names, Claras, Marias, concurrently with 
 Maria (Lat.), Marien and Marienx. 
 
 1 Go. dagos dage, vaiirda vaurde, gasteis gaste, etc. ; O.E. dagas daga, 
 word worda, giestas giesta, etc. ; O.H.G. taga tago, wort worto, gesti gestio, 
 etc. (now, in both cases, tage, worte, gaste). 
 
 2 We find the same process in Swedish, viz. nomin. sg. fisk, gen. fisks, 
 nomin. pi. Jisk-ar, gen. jisk-ars. 
 
256 ENGLISH AND GEEMAN GEAMMAR. 
 
 stem the I.-E. ending -%. When the stem ends in a consonant, 
 the case-sign is isolated and visible : Sk. pdd-i (at the foot), 
 Gr. xoS-t, L. ped-e l = *ped-i, etc. If the stem ends with a 
 vowel, either -o- or -a, contraction gives a final diphthong : 
 Sk. dgve (to the horse) = *aqva-i, Gr. olkol (at home), L. huml 
 (on the ground) = humoi, Go. vulfa (to the wolf) = Preg. *wulfai, 
 etc. The whole of the modern formation may be said to pro- 
 ceed from these simple premises. 
 
 1. In feminines in -a, the primitive diphthong still kept in 
 Gothic (airpa, dat. airfcai), became a simple vowel in O.E. 
 (cearu, dat. ceare) and in O.H.G. (erda, dat. erdu erdo), where- 
 upon the final vowels, both in the nominative and the dative, 
 were weakened to the same form : G. erde. Analogy completed 
 the assimilation: the dative, as well as the genitive, 2 of 
 feminine nouns, was everywhere blended with the nominative, 
 even in stems in I.-E. -ti-, which formerly had a characteristic 
 metaphonical dative, 3 and in the feminine weak declension, 
 which, in consequence, no longer shows any trace of its old 
 nasal element in any case of the singular. 4 
 
 2. On the contrary, masculine nouns, inasmuch as they re- 
 tained 5 or adopted the weak system, quite regularly exhibit 
 the nasal as the sole remaining case-sign (Go. han-in " to the 
 cock " = *han-in-i, cf. L. hom-in-e) : whence we may say, that 
 the dative, in the German weak masculine declension, differs 
 from the nominative, and is identical with the accusative and 
 genitive, thus dem ochsen, dem herm, dent bauern, etc. 
 
 3. In the case of the other masculine and all neuter nouns, 
 the influence of the dat. sg. in -o-i (2nd Gr. and L. declension) 
 prevailed and formed a new termination. To the Preg. type 
 *wulfai, Go. vulf-a (nomin. vulf-s) corresponded in O.E. and 
 
 1 Termed ablative ; but really locative in such locutions as in pede, etc. 
 
 2 Supra 150, 1 a. — The identity was regular in nouns denoting relation- 
 ship : nomin. muoter=I.-E. 'mater, dat. muoter=I.-TL. *mdtSr-l. 
 
 8 O.H.G. anst (favour), dat. ensti = Go. anstdi. 
 
 4 O.H.G. deru zungun, M.H.G. dere zungen, G. der zunge. The last 
 survivals of a fm. or nt. dat. in -en are found in such locutions as auf erden 
 and von herzen. 
 
 5 Cf. supra 150, 1 b. Of course, we have nt. dem auge, dem herze, msc. 
 dem hahn, etc. for the reason explained above. 
 
nouns. 257 
 
 O.H.G. a duller ending, respectively wulf-e wolf-e, which was 
 dropped in English, but spread in German. Now, after the 
 pattern of the regular datives, msc. tag-e = I.-T$. *dhogho-i, 
 fisk-e (also fisk-a), stuol-e, stein-e, nt. johhe = I.-~E. *yugo-i, wort-e 
 (also loort-a), jdr-e, lamb-e, 1 etc., O.H.G. has the same types : — 
 in masculine stems in -i-, gast-e, for *gesti ; — in those in -u-, 
 sun-e ; — in those ending with a consonant, fater-e, etc. 2 So this 
 -e.. was added everywhere ; but, being almost mute, it is now 
 suppressed in dissyllables in which the accent rests on the first 
 syllable (dem vater, dem adler), and may also be dropped in any 
 other noun (dem sohn, dem joch, dem lamm). The main result 
 is a dative which scarcely differs from the nominative in all 
 classes except the weak declension. 
 
 II. Plural. — Several I.-E. cases in the plural number 
 (instrumental, dative, ablative) are formed by adding to the 
 stem a peculiar element, in which, though the vowel and final 
 may vary, the initial is always a labial consonant : Sk. -bhis, 
 -bhyas ; Gr. (Homeric) -<f>i, -d>iv ; L. -bos, -bus. In Germanic, 
 as well as in Bal to- Slavonic, the initial labial is an m, and the 
 form of the endings seems to have been -ml or -mis, -mil or 
 -mils. 3 The initial m is the only sound we need lay stress 
 upon, since the rest of the syllable naturally disappeared in the 
 Germanic languages ; and we find that, however far we go 
 back, the case-sign for the dat. pi. is invariably an ending -m : 
 Go. daga-m, vaurda-m, air}>5-m, gasti-m, sunu-m, brofrr-um, etc. ; 
 O.E. dagum, wordum, cearum, giestum, sunum, fcedrum, etc. ; 
 O.H.G. tagum (tagom, whence also tagon, tagun), wortum, erddm, 
 gestim (also gestin and gesten), sunim (instead of sunum,, analogi- 
 cally assimilated to the preceding), faierum, muoterum, etc. 
 English has altogether lost this form. M.H.G. completed the 
 assimilation, by transferring to this case the plural-metaphony 
 exhibited by the other cases, and, final -m being further regu- 
 
 1 With the restriction mentioned in 150, 1 c ; for, if the declension were 
 quite pure, the dative ought to be *lembir=*lamb-iz-i. 
 
 2 By the way, observe the remarkable purity of the O.E. datives fet, 
 friend, where the final -I has caused metaphony (*fot-l, *frennd-i). 
 
 3 Still in Mod. Russian: slov-o (word), pi. slov-d, dat. pi. dov-d-m(u), 
 instrum. pi. slov-d-mi, etc. 
 
 S 
 
258 ENGLISH AND GERMAN GRAMMAR. 
 
 larly changed to -n, 1 the whole process may be summed up in 
 the simple rule : 
 
 The dative plural is formed by adding an ending -n to the 
 ordinary plural-form, unless this already ends in -n. 8 
 
 1 Supra 39, I. 1, in fine. 
 
 * It must be observed that a great many proper names, of feasts or places, 
 are historically traced back to old dative cases, which have become invari- 
 able in this shape, because they were more frequently used than the corre- 
 sponding nominatives ; for it stands to reason that we should have more 
 occasion for saying " about Easter-tide," or " he is in London," than for 
 saying " Easter is a great feast," or " London lies on the Thames," etc. : 
 whence, such German words as Ostern (E. merely Easter = O.E. eastre), 
 Pfingsten, Weihvachten, etc., geographical names as Meyrivgen (a very 
 common type), etc. (In Vnterwaldev, the second term is a dat. pi. governed 
 by the first term.) A parallel English type to Meyrwgsn is Canterbury, the 
 second term being tbe dat. sg. of borough, supra 65, 5. 
 
CHAPTER III. 
 ADJECTIVES. 
 
 (153) Every German adjective is capable of two de- 
 clensions, which, as is well known, are used in distinct 
 syntactical combinations. Further, the adjective, when used 
 as a predicate, is invariable, or incapable of receiving any 
 exponent, either of gender, or number, or case : msc. sg. der 
 mann ist blind ; fm. die frau ist blind ; rit. das kind ist blind ; 
 ace. er machte ihn blind ; pi. die manner sind blind, etc. 
 
 The English adjective is always invariable : a blind 
 man, a blind girl, a blind dog, the blind men, etc., like the man is 
 blind, etc. Now, in order to understand how this twofold, or 
 even threefold system could proceed from one and the same 
 primitive declension, we must first distinguish the declined 
 and invariable adjective. 
 
 Such a distinction, though already existing in Pregermanic, 
 was quite unknown to the I.-E. language. The Sk., Gr. and 
 L. adjective, for instance, agrees with the noun it qualifies, in 
 whatever grammatical construction they may be found : L. 
 pater est bonus, like bonus pater, etc. And, again, the declension 
 of the adjective in these languages has no peculiar character, 
 except that it is capable of the three genders ; that is to say, 
 an adjective is declined in the same way as a noun, provided 
 they be of the same gender : Gr. koAos Ka.\y koXov, just like 
 lim-os KecfxxXrj epyov ; L. bonus bona bonum, in the same way as 
 dominus terra jugum, etc. 1 
 
 1 The adjectives we are here dealing with are only those which belong to 
 the so-called (Gr. and L. ) 2nd and 1st declension ; for this declension is the 
 only one preserved for adjectives, in Germanic, and it has extended so far 
 beyond its limits as to overwhelm any other : thus, the comparatives (G. 
 
 859 
 
200 ENGLISH AND GERMAN GRAMMAR. 
 
 Hence we infer, at the outset, that Pregermanic has com- 
 pletely altered the primitive type, whereas English again shows 
 not even the slightest trace of the Pregermanic declension. 
 Let us now proceed to explain this simple evolution. 
 
 Section I. 
 
 DECLINED ADJECTIVE. 
 
 (154) The old I.-E. type bonus bona bonum is not dead in 
 Germanic : it still survives in the invariable adjective. But, 
 side by side with this ancient declension, analogy had created 
 two other systems, which may be traced back to Pregermanic, 
 and which survived into Mod. German, becoming in the latter 
 the double, so-called strong and weak, declension of the 
 adjective when used as an attribute to a following noun, viz. 
 blind-er mann, and der blind-e mann. 
 
 The general rule for the use of either declension, in Go., 
 O.E. and O.H.G., is nearly the same as in Mod. German, 
 namely: if the adjective is not preceded by a demonstrative 
 (article or the like), it supplies the place of the missing demon- 
 strative, and assumes the demonstrative endings, or, in other 
 words, the case-signs of the strong declension ; but, if the 
 demonstrative precedes, and sufficiently indicates the gender, 
 number, and case, then the adjective assumes uniform and 
 simplified endings which belong to the weak system. 
 
 § 1. Strong Declension. 
 
 ( 1 55) The fact is, that the so-called stroug adjectival 
 declension merely proceeds from an analogical adaptation, to 
 the adjectival stem, of the endings which were originally the 
 property of the definite article and similar demonstrative 
 words: an extension logically justified, since the adjective 
 resembles the demonstrative so far as both are accessories to 
 the noun. Hence, most of the Go. endings which have occurred 
 
 better), though actually consonantal stems, are declined with the case- 
 endings of their positives (G. put), whereas in Greek and Latin they regu- 
 larly retained their primitive (3rd) declension. 
 
ADJECTIVES. 
 
 261 
 
 
 msc. 
 
 fm. 
 
 N. 
 
 blint-er 
 
 blint-iu 
 
 A. 
 
 „ -an 
 
 „ -a 
 
 G. 
 
 ,, -es 
 
 ,, -era 
 
 D. 
 
 „ -emu 
 
 „ -eru 
 
 
 Plural. 
 
 
 msc. 
 
 fm. 
 
 Tit. 
 
 blint-e 
 
 blint-o 
 
 blint-iu 
 
 ii -e 
 
 >. -o 
 
 „ -in 
 
 n -e^o 
 
 „ -ero 
 
 n -ero 
 
 „ -em 
 
 „ -em 
 
 ,, -em 
 
 above in the declension of the article also became attached to 
 the Go. adjective, whilst in O.H.G. the likeness is even more 
 striking and almost complete. The reader need but compare 
 with the earliest declension of the article that of the O.H.G- 
 adjective, as given below : 
 
 Singular. 
 
 nt, 
 blint-aj 
 „ -a$ 
 ii -es 
 ,, -emu 
 
 Farther, remembering the duller sound of modern final syl- 
 lables, the change of final m to n, and the analogical processes 
 stated above for the article, we shall easily understand the con- 
 temporary type and may even dispense with transcribing it, 
 since it exactly corresponds to the article. 2 
 
 O.E. possessed a similar declension. 3 How it was lost in 
 English, will be explained when we come to the invariable 
 adjective. 
 
 § 2. Weak Declension. 
 
 (156) In order to trace back to its origin the adjectival 
 weak declension, we have but to note one historical fact, namely 
 that, side by side with a great many adjectives in -o-s -a -o-m 
 (L. type bonus bona bonum), there existed other adjectives, with 
 the same meanings, wherein the suffix-element was -en- or -on-, 
 the common and familiar formative syllable which loses its n 
 in the nomin. sg, and retains it everywhere else. Thus, in 
 Greek, side by side with <f>dyo<; (glutton), we have <f>ayu)v (glut- 
 ton), the plural of the latter being <f>ayu>v-e<;, and of the former 
 simply <f>ayoi. Thus also, in Latin, catus (wise, sensible), and 
 the doublet *cato, which may be assumed from the proper 
 name (nickname) Cato, gen. Catonis. Further instances are : 
 
 1 Sg. also -emo, -ero ; pi. also -en. 
 
 3 Cf. supra 128-131. The only difference is nomin. -ace. sg. nt. -es instead 
 of -as = O.H.G. -aj. 
 
 3 And even as late as Shakespeare we find an adjectival genitive plural, 
 alder-liefest " dearest of all," which corresponds to G. aller-liebst, and would 
 be O.E. ealra leofetta. 
 
262 ENGLISH AND GERMAN GRAMMAR. 
 
 Grr. cu#o? (soot- colour) aWtttv, ide\r)p.6<; (willing) i$e\rjp.wv, 
 ovpavios (heavenly) OupavtWcs (the gods) , o-Tpafios (squint-eyed) 
 a-Tpdfiuv Hrpdfitov (proper name), etc.; L. (multi-)bibus bibo 
 (drunkard), rebellis (from bellum "war") rebellio, scelerus 
 (felon) scelero, susurrus (murmuring) susurro, etc. Now, sup- 
 posing analogy to create other such doublets after these pat- 
 terns, and that any adjective may assume a form either without 
 or with a nasal suffix : Latin, for instance, along with bonus, 
 2/robus, prdvus, niger (pi. boni, probi, prdvl, nigri), would exhibit 
 the synonymous adjectives, *bono, *probd, *prdvo, *nigro (pi. 
 bonon-es, *probon-es, *prdvon-es, *nigron-es), and so forth, all 
 interchangeable. 
 
 This is precisely what happened in Pregermanic. Besides 
 the strong form, which is declined after the demonstrative, 
 every Gothic adjective has a weak form, which is declined, if 
 masculine, like hana, if feminine, like tugqo, if neuter, like dugo, 
 which nouns are well-known to be the Germanic representatives 
 of the I.-E. stem-formation with suffix -on- 1 ; and the same cor- 
 respondences are rigorously reproduced in O.E. and O.H.G. 
 
 English lost the weak declension in its adjectives as well as 
 in its nouns. But, on the contrary, the German adjectives 
 retained it even with more vigour and purity than the nouns 
 themselves. For the latter have been seen to keep the regular 
 ending -en only in the plural of the three genders and in the 
 masculine singular declension, whereas, in the adjectives, this 
 ending characterizes, quite regularly, any oblique case of any 
 gender, whether plural or singular: thus, we find gen. des auges 
 and dat. dem auge, for des augen and dem augen ; but always 
 gen. sg. nt. des guten, and dat. sg. nt. dem guten ; so also, gen. 
 dat. sg. der zunge, for der zungen ; but gen. dat. sg. fm. der 
 guten, etc. This strictly coherent system suffered encroachment 
 in one point only, that is, the ace. fm. sg. was assimilated to the 
 nominative, 2 owing to the preceding demonstrative having be- 
 come the same in both cases: O.H.G. nomin. sg. diu blinta, ace. 
 dea blintun ; M.H.G. nomin. diu blinde, ace. die blinden ; but 
 Mod. G. die blinde in either case. 
 
 1 Supra 74, 140-142, 150 and 152. 
 
 2 The ace. nt. sg., of course, can never differ from the nominative. 
 
ADJECTIVES. 263 
 
 Section II. 
 
 INVARIABLE ADJECTIVE. 
 
 (157) The English invariable adjective and the adjectival 
 form -which is commonly called " unfiektiert " in German 
 grammar, were both originally declensional forms, and primi- 
 tively no less variable than any that have occurred above. But 
 this flexion no longer appears, because phonetic laws effaced 
 most of their endings, and analogy afterwards disposed of the 
 remainder. 
 
 § 1. In German. 
 
 Let us return to the original type represented by L. bonus 
 bona ( = *bond) bonum. After the loss Of final s or m together 
 with the preceding vowel, both masculine and neuter as- 
 sumed an indifferent form, the phonetic equivalent of which 
 may be seen in F. bon, thus O.H.G. guot, Mint, etc. The primi- 
 tively long final vowel in the feminine, shortened in late Pre- 
 germanic (Go. blind-a), was destined to disappear in West 
 Germanic, at least when the preceding syllable was long, as is 
 the case with these two adjectives and with many others, thus 
 fm. guot, blint, etc. Now O.H.G. analogically lost this termin- 
 ation even when the preceding syllable was short, so that the 
 form of the adjective in the nominative singular became the 
 same for the three genders. 
 
 Since this form is historically a declined one, no less than 
 the so-called forms of strong or weak declension, it is but na- 
 tural that it should appear fit for the same use and be con- 
 strued in the same sentences, without any noticeable difference 
 in the earliest documents. Hence O.H.G. has either: blinter 
 man, or blint man ; blintiu magad, or blint magad ; blintaj kind, 
 or blint kind, etc. And, no less optionally, even : der man ist 
 blinter, or blint ; diu magad ist blintiu, or blint ; da% kind ist 
 blintaj, or blint, etc. As early, however, as this period, there 
 may be observed a slight tendency to decline the qualificative 
 adjective rather than the other, and to assign the invariable 
 form to the predicative in particular. 
 
264 ENGLISH AND GERMAN GRAMMAR. 
 
 Next, from this increasing tendency, arises a new conse- 
 quence: the speaker, being now accustomed to keep the adjective 
 invariable in gender, began to make it invariable also in num- 
 ber. The plural, indeed, is still required in the construction 
 blinte man, 1 but the agreement is optional in the sentence die 
 man sint blinte, or blint ; and similarly, in the three genders. 
 
 Lastly, that artificial regulation occurred, by which any 
 language, sooner or later, aims at making the most of its 
 resources. M.H.G., it is true, still often uses the invariable 
 adjective as an attribute, 2 but rarely shows the variable form 
 in a predicate. 3 And this adventitious distribution gradually 
 became the modern German rule : the qualificative adjective 
 agrees with its noun ; the predicative adjective is always 
 invariable. 
 
 § 2. In English. 
 
 It is somewhat surprising to find that the so-called invariable 
 adjective is less invariable in O.E. than in O.H.G. : of course, 
 since the phonetic law applies equally to the whole West Ger- 
 manic group, the type god, blind, etc., here also, is the same for 
 the three genders in the singular ; but, when the preceding syllable 
 is short, O.E. regularly keeps the feminine ending, and declines, 
 for instance, msc. blmc, fm. blac-u, nt. blwc, forms now replaced 
 by E. black, and even in O.H.G. by msc. fm. nt.flah (flat). 
 
 The later process is quite plain : the same analogy that had 
 corrupted the O.H.G. declension entered into action in M.E. 
 and thus an invariable adjective black was formed on the 
 pattern of the l-egularly invariable good. On the other hand, it 
 must be observed that the analogy of the demonstratives, either 
 in O.E. or in Go., had never affected the nominative of the de- 
 clined adjective : in other words the O.E. strong declension had 
 nothing like the G. nomin. blind-er blind-e blind-es, so that the 
 adjective, either as predicate or attribute, had no other form 
 but god in the three genders for the nomin. sg. The corre- 
 
 1 It must not be forgotten that the pi. to man is regularly man. 
 
 2 Rural dialects and popular speech still retain this use, which is likewise 
 illustrated by many geographical names, as Neubvrg, Neukirch, etc. 
 
 ' * A survival is found in the adj. roll, still declined, though a predicate, in 
 such sentences as der ort war voller leute. 
 
ADJECTIVES. 265 
 
 sponding plural forms were msc. god-e, fm. god-a, nt. god, the 
 latter again like the singular ; the primitive plural sign -s, 
 which as a rule was omitted in the feminine, had been lost in 
 the masculine, as it had also disappeared in the nominal G. 
 type tag-d (tag-e) ; and there was no reason for analogy to re- 
 store it in the adjectives as it had restored it in the nominal 
 declension : in short, all circumstances conspired to a complete 
 levelling of the adjective in the nominative of the three genders 
 and the two numbers, whence the indifferent form good. 
 
 Now there remained the other cases and the forms of the 
 weak declension. But in all these also the final syllables had 
 become obscured ; and, since the old distinction of cases had 
 perished in demonstratives and nouns, it could not be expected 
 to survive in adjectives, as there it was even less needed, the case 
 being always indicated either by the noun itself 1 or else by a 
 preposition. Hence the English adjective has kept no flexional 
 forms but an analogical plural in -s which appears when the 
 word is used substantially (the goods). 2 In any other position, 
 whether qualificative or predicative, the English adjective is 
 altogether invariable. 3 
 
 1 English, for instance, in the good wife's, expresses the case quite as 
 clearly, as it would by the *gooden wife's, or as German by its threefold 
 case-ending des guten weibs. Cf. however the archaic and still used locution 
 in the olden times. 
 
 * So also eatables, etc., but the wise = the wise men; in M.E. the other is 
 still plural, now the others • and the plural the . . . ones is even later. 
 
 3 For comparatives and superlatives, adverbs, and other points which are 
 usually dealt with in connection with the adjective, the reader must be re- 
 ferred to our Second Part, especially 80, 90, 111, etc. 
 
CHAPTER IV. 
 PBONOXJNS. 
 
 (158) The I.-E. Pronouns must be divided first into two 
 great classes : Demonstrative and Personal Pronouns. 
 
 The declensional system of the two classes is quite different. 
 But their main distinction lies in the following fact : the 
 demonstratives, theoretically, agree in gender with the person 
 or thing which they determine or represent, whereas the per- 
 sonal pronouns do not mark any difference in sex. Thus, in 
 Latin, msc. Me, fm. ilia, nt. Mud, but ego "I," tit "thou," in- 
 differently for the three genders. 
 
 The latter distinction holds good, in English ' as well as in 
 German, for the pronouns of the first and second person, and, 
 in German alone, for the reflexive pronoun; while the third 
 person is an exception in both languages, the so-called pronoun 
 of the third person being really an old demonstrative used for 
 this purpose. 2 
 
 Section I. 
 
 DEMONSTRATIVES. 
 
 (159) The general paradigm for the declension of every 
 I.-E. demonstrative is that of the stem *to-, which has become 
 the Germanic article. Since this system has already been 
 
 1 But here the contrast disappears, because the demonstrative has also be- 
 come invariable in gender. It will be seen to have varied in O.E. 
 
 2 Sanskrit, Greek, Latin, and even Mod. Bussian agree in showing no 
 other personal pronouns, except those of the 1st and 2nd person and the 
 reflexive, which primitively represented all three persons, as it Btill does in 
 Bussian. Cf. injra 166. 
 
TRONOUNS. 267 
 
 studied and has not suffered change, either in German, or in 
 the scanty relics of a demonstrative declension which English 
 has retained, we need no longer dwell upon it. It will be 
 sufficient to point out, as they occur, a few interesting pecu- 
 liarities which present themselves in the different classes of 
 demonstratives, namely: — demonstratives (either adjectives 
 or pronouns) properly so-called ; — interrogative and indefi- 
 nite pronouns ; — relative pronouns. 
 
 § 1. Demonstratives properly so-called. 
 
 (160) I.-E. was well furnished with demonstratives of 
 various forms and meanings ; but most of them were lost in 
 Pregermanic, and the whole of English and German demonstra- 
 tives might be said, with one exception, to proceed from a single 
 primitive pronoun, either shifted in its form, or altered in its 
 meaning. 
 
 1. O.E. o~e-s b"eo-s b"i~s, and E. this in the three genders; O.H.G. 
 d'ese desiu dij, and G. dies-er dies-e dies-es 1 : demonstrative for 
 things close at hand. — This word was originally a mere juxta- 
 position of the ordinary demonstrative and an indeclinable 
 particle -se, 2 as still clearly shown by the O.E. declension. In 
 German, this inward declension is but partly visible, having been 
 replaced by an outward declension imitated from the endings of 
 the declined article ; that is to say, the indeclinable particle 
 adopted the case-signs, whilst the declinable stem lost them. 3 
 Lastly, in English, every trace of a declension was swept away, 
 and nt. this serves for all genders. The O.E. pi. is Sd-s, and, 
 with a slight vocalic shifting, Sees, which respectively became 
 E. those and these: the latter is retained as the invariable plural 
 of this, whereas the former now belongs to another system. 
 
 2. Go. jain-s; O.E. geon and E. yon ; O.H.G. jen-er, etc., and 
 G. jen-er jen-e jen-es : demonstrative for distant objects. — This 
 pronoun is of doubtful origin and belongs exclusively to the 
 
 1 The regular form dies remains still in use as a neuter pronoun : dies mag 
 irohl sein = thu may well be. O.H.G. msc. sg. also deser, fm. sg. also disiu. 
 
 s As in F. celui-ci, celle-ci, ce-ci. 
 
 8 Such confusions are not at all rare in pronominal declension : cf. F. 
 pi. quelconques, which ought to be *queh-conque =L. quales-cunque. 
 
268 ENGLISH AND GERMAN GRAMMAR. 
 
 Germanic group. Besides it is nearly obsolete in English, 
 though, used in poetry ; compare however yon-der and be-yon-d. 
 
 3. E. that : usual demonstrative for distant objects. — It is 
 the neutral form of the ordinary demonstrative (article), 1 now 
 stiffened in an invariable shape which serves for all cases 
 and genders. The Mod. E. pi. is those, in reality a plural of this 
 and a mere phonetic doublet of these. Popular English still 
 shows a trace of the old dative : in them days. 
 
 4. G. der-selb-e, " this (pronoun), he," literally " the same," 
 die-selb-e, das-selb-e, pi. die-selb-en, etc., obviously, a purely 
 syntactical juxtaposition, wherein, as a matter of course, the 
 demonstrative showing the strong forms, the adjective assumes 
 the weak terminations. This adjective, Go. silba, O.N. sjalf-r, 
 E. self, is a stem of uncertain etymology, which, in the Ger- 
 manic family, plays the part of a pronoun denoting identity> 
 and is further combined with personal pronouns. 
 
 5. G. der-jenig-e, 2 etc., the demonstrative which precedes a 
 relative pronoun : a syntactical juxtaposition of the article 
 with an ad jective derived from the demonstrative jen- ; whereas 
 English in this construction merely uses he in the singular and 
 those in the plural. 
 
 § 2. Interrogative and Indefinite Pronouns. 
 
 (161) Both the interrogative and indefinite pronominal 
 categories are expressed by the same I.-E. stem, either a stem 
 in -o- or in -i-, characterized by an initial q, thus I.-E. *qo- and 
 *qi-, whence further: Sk. kd-s (who?), fm. hd, nt. hi-m, and 
 ci-d indefinite particle ; Gr. stem 7rd-, with its many forms, ttov 
 (where?), -rroOev (whence?), ttotc (when?), 7rws (how?), and 
 ti'-s (who?), fm. Ti-s, nt. Ti = *Ti-S 3 ; L. quo-, stem of the pro- 
 noun qui quai quo-d, and qui-, stem of the pronoun qui-s (who ?), 
 qui-d (what ?), etc. To these, the Preg. corresponding forms 
 naturally were *hwd- and *hwi-, the former surviving as an 
 
 1 Go. \>a-t-a, supra 128 and 130, 1. 
 8 Which replaced the simpler O.H.G. derjener. 
 
 8 The unaccented forms, ttov, vore, iru, t«, ti, are indefinite adverbs or 
 pronouns ; and so also L. qui* quid. 
 
PRONOUNS. 
 
 269 
 
 isolated pronoun, the latter as the first term of a compound, 
 both interrogative and indefinite. 
 
 1. Go. msc. hva-s, fm. hvo, nt. hva, declined like the demon- 
 strative, but without any plural- form; O.E. msc. hwd, nt. 
 hwoe-t, and E. who what, no feminine or plural ; O.H.G. msc. 
 hwe-r, nt. hwa-%, and G. wer was, no feminine or plural. — Here, 
 exceptionally, English retains almost the whole declension, 
 viz.: gen. sg. whose = O.T£. hwws; dat. sg. whom = O.E. hwd-m = 
 Go. hva-mma, shifted moreover to the accusative-meaning 
 instead of regular O.E. hwo-n-e = Go. hva-n-a. But the nt. is 
 invariable, and admits of no other form but what (like that = 
 Go. J>a-t-a) for all cases. 1 — German has the four cases : wer 
 was, wen was, wessen, wem. The regular gen. is O.H.G. hwe-s, 
 M.H.G. wes, kept in the juxtapositions wesz-wegen, wesz-halb, 
 etc., whereas wess-en is a recent and pleonastic form, created 
 by subjoining to the old genitive a new genitive-ending of weak 
 declension. The dat. wem, like the ace. wen, may only refer to 
 a person. 
 
 2. Go. hvi-leik-s, a compound of the stem *hvi- and the Preg. 
 adjective *lika- (like), 2 literally "to whom or what alike? of 
 what sort? which?"; O.E, hwyle and E. which invariable; 
 O.H.G. w'e-lich-, and G. welch-er welch-e welch-es. The Mod. 
 G. locution was fiir, which often replaces the interrogative 
 adjective, must be interpreted thus : was hast du fur ein kleid ? 
 " what hast thou for, instead of a garment ? " whence " what 
 garment do you wear ? " 
 
 Further, the indefinite meaning may be emphasized by adding 
 to these pronouns such adverbs or particles as E. ever, so ever 
 (whosoever), G. je, irgend. 3 But these syntactical constructions 
 have nothing to do with declension. 
 
 1 The analogical process here is quite plain : with whom (dat.) have you 
 been ? being regular, and the dative being next used for the ace. in such a 
 sentence as whom have you seen ?, the regular construction tohat have you 
 seen ? naturally produced the fourth term with what . . . ? For ace. 
 whom, see below 164, I. 4. 
 
 2 Cf. supra 110, I., and compare Go. sva-leik-s (<ro = E. so = G. so), E. 
 inch, G. solch (L. talis " such," quulis "which"). 
 
 8 O.H.G. to wer-gin: io, the particle which has become G. je; wer, 
 identical with E. where ; and -gin, the same as Go. -hun and L. -cum- in 
 qul-cun-que (whoever). So also, immer= O.H.G. io mer (now je mehr), and 
 ever = O.E. a-fre = *<e-mre. 
 
270 ENGLISH AND GEEMAN GRAMMAR. 
 
 § 3. Relative Pronouns. 
 
 (162) The so-called relative pronouns, which introduce a 
 subordinate sentence, cannot be properly said to have existed 
 in I.-E. times; in origin, they appear to be in all the I.-E. 
 branches either old demonstrative or old interrogative-inde- 
 finite pi'onouns, afterwards adapted to this new function. Thus, 
 the Gr. article is also the relative pronoun in Homer and 
 Herodotus, and even the ordinary relative, os ^ o, shows some 
 traces of its original demonstrative value. In Latin, on the 
 other hand, it is the interrogative-indefinite stem that plays 
 the part of the relative pronoun, qui quae quod. The E. and G. 
 relatives are taken from both sources concurrently. 
 
 1. Gr. der die das, identical with the article, demonstrative 
 used as relative. In the latter function, however, and not 
 earlier than Mod. German, the pleonastic process described 
 above has developed some new case-forms with redundant end- 
 ings : gen. msc. sg. dess-en, gen. fm. sg. and gen. pi. der-en der-er, 
 dat. pi. den-en, as if the regular monosyllables seemed too short 
 to express clearly what the speaker meant. 1 
 
 2. E. that, identical with the demonstrative (neuter article), 
 used as relative in all genders and numbers, though keeping 
 its etymological value, inasmuch as it stands only for a nomina- 
 tive or accusative and is never governed by a preposition. 2 
 
 3. E. who (gen. whose, dat.-acc. whom), primitively interro- 
 gative-indefinite, used as relative for both numbers and both 
 the msc. and fm. gender, never neuter. Its neuter what never 
 assumes the relative function. 
 
 1 Suyra 161, 1. Used also as a demonstrative pronoun (not adjective) : 
 dessen, " of him, his " ; es gibt derer zwei, " there are two of them." — It is 
 only use that gradually assigned to each of these forms a peculiar and 
 distinct function : as late even as Goethe we find der for the gen. fm. of the 
 relative pronoun, and on the other hand, as late as Klopstock, derer is used 
 even for the gen. pi. of the simple article. There existed besides a common 
 dat. sg. deren, which has become obsolete, dtren being exclusively assigned 
 to the genitival function, whereas der was used exclusively for expressing 
 the dative. The same and very late distinction has been drawn between 
 deren (now exclusively gen. fm. sg.) and derer (now exclusively gen. pi.), 
 which once stood indifferently for each other. 
 
 2 Further, that and dasz as conjunctions, supra 130, 1. 
 
PRONOUNS. 
 
 271 
 
 4. E. which invariable, and G. welch-er welch-e welch-es, usual 
 relative pronoun, the same as the interrogative-indefinite. 
 
 We need not deal here with the ellipse of the relative pro- 
 noun, which is common in English. 1 
 
 The various pronouns which have been purposely excluded 
 from this summary require but little etymological explanation. 
 Some have already appeared in the Study of Sounds 2 ; at any 
 rate, they belong rather to the province of the Dictionary than 
 to that of Grammar properly so-called. 
 
 Section II. 
 
 PERSONAL PRONOUNS. 
 
 (163) In this class, we have to distinguish the pronouns 
 of the three persons in both numbers, the reflexive 
 pronoun, which theoretically knows no distinction in number, 
 and lastly the possessive pronouns and adjectives, as 
 being, in both languages, peculiar forms or derivatives from 
 the corresponding personal pronouns. 
 
 Even in the earliest period, the primitive personal pronouns, 
 that is to say, those of the first and the second person, are re- 
 markable for the fact that they are not declined on a 
 single stem, like most demonstratives, and thus they often 
 assume different shapes, not only in passing from the sg. to the 
 pi., — as may be seen in L. ego and nos, tu and vos, — but from 
 one case to another in the same number, — from nomin. sg., for 
 instance, to ace. sg. in L. ego and me. — Germanic and, in con- 
 sequence, English and German reproduce all these peculiarities. 
 
 § 1. First Person. 
 
 (164) I. Singular. — 1. The probable I.-E. nominative was 
 *egom, Sk. aham, Gr. cyw, L. ego, Preg. *ik(a), whence : Go. ik, 
 
 1 The relative was likewise often dropped in O.H.G. : Fater unseer, thu 
 pist in himile (in the so-called "St. Galler Lord's Prayer"). It must be 
 borne in mind that the relative is but a slightly modified demonstrative ; 
 and a demonstrative never need be used when the sentence can do without it. 
 
 2 Refer to the Alphabetical Indexes of Words. 
 
272 ENGLISH AND GERMAN GRAMMAR. 
 
 O.E. ic, M.E. ich and E. I, O.H.G. ih and G. ich. These present 
 no difficulty. 
 
 2. Accusative. — Of the many accusative-forms of the 
 primitive stem, the simplest (I.-E. *me) may be inferred from 
 Gr. c/xe or fie, which is often accompanied by a strengthening or 
 emphatic particle, thus c/ic-ye, like nomin. eyw-ye, " as to me." 
 The parallel form I.-E. *me-ge has become the Germanic accu- 
 sative, Preg. *mi-k(i), Go. 7nik, O.E. me-c me me and E. me 
 (short because unaccented), O.H.G. mi-h and G. mi-ch. 
 
 3. The Go. genitive is meina, and really belongs, not to the 
 personal pronoun, but to the possessive, as given below. 1 O.E. 
 answers with min, lost in English. O.H.G. likewise has mln, 
 which as late as Mod. German 2 assumed a redundant case-sign 
 and became pleonastically mein-er, by adding to mein the 
 regular termination of the corresponding gen. pi. uns-er. 
 
 4. The puzzling Go. dat. mis cannot be traced back to any 
 primitive type, and seems to have been formed after the analogy 
 of a (lost) dat. pi. *nis = Sk. nas. According as the final s (z) 
 was merely dropped, or changed to r, we have now : O.E. me 
 and E. me, blended with the accusative 3 ; O.H.G. mir, G. mir. 
 
 II. Plural. 4 — 1. The curious nomin. pi. with initial to is 
 common to Sk. and Germanic (Sk. vaydm, Go. vein), whereas 
 it occurs nowhere else: it has become O.E. we and E. we, 
 O.H.G. wir and G. wir & ; compare me = mir. 
 
 2. Accusative. — The initial syllable of the oblique cases 
 clearly proceeds from an I.-E. nasal vowel, either n or m, 
 which again is simply the reduced grade of the stem exhibited 
 in its normal grade by Sk. nas "us" (cf. L. nos), or by Gr. fU 
 
 1 It is a case-form of it, but which case ? The question remains open ; 
 perhaps a weak msc. nominative, elsewhere obsolete, and surviving here in 
 this shape. At any rate, the use of the possessive in the function of genitive 
 to the personal pronoun (" my " instead of " of me ") is everywhere common : 
 thus, in L., nos has the gen. nostrum no«tri, and even mei, the gen. of ego, is 
 probably identical with mei the gen. of mew. 
 
 2 The pure form survives in the compound vergisz-mein-nicht (the flower). 
 8 This confusion further caused the two cases to be confounded in the 
 
 interrogative-indefinite declension, supra 161, 1. 
 
 * In all these declensions, Go. and O.E. had also retained some dual- 
 forms ; but O.H G. has lost the dual. 
 
 5 Dialectal (Alam.) mir imitated from the singular-stem. 
 
PRONOUNS. 273 
 
 (cf. L. me). The plural Sk. and Gr. forms moreover require a 
 particle 1 to be inserted between tbe stem and the endings : 
 I.-E. *7jb-sme- ; Sk. (ablative) a-sma-d; Homeric Gr. ace. afifj.e = 
 *a-a-fxi- ; Attic rjfxcLS for *d/xts (long a) = *d-<r/Ae-s. But Pre- 
 germanic has merely *uns = *ns, as attested by Go. uns. In 
 O.E. and O.H.G., the analogy of the ace. sg. caused a new 
 ending to be added to this simple form : O.E. usic = *uns-ic, and 
 O.H.G. uns-ih. O.E. us, however, which survived concurrently, 
 afterwards became the prevalent form : E. us ; and so likewise 
 G. uns. 
 
 3. Genitive. — Go. unsara is a case-form of the corresponding 
 possessive, like meina: O.E. user, syncopated to ure, lost in 
 English; O.H.G. unser and G. unser. 
 
 4. Dative. — On the primitive basis uns, Gothic had con- 
 structed, after the analogy of sg. mis and of the pi. pronoun of 
 2nd person, a new dative uns-is. But E. and G. have only the 
 pure form, us, uns, like the accusative. So in Sk. nas also 
 serves for both cases. 
 
 § 2. Second Person. 
 
 (165) I. Singular. — 1. The nomin. was I.-E. *tu and *tu, Sk. 
 tvdm = *tu-am, Gr. rv (tv) and a~v, L. tu, Preg. *}>u, whence: 
 Go. fru; O.E. ftu ftu, and E. tlnou — tSu; O.H.G. du du, and G. 
 du. These present no difficulty. 
 
 2. Accusative. — In the same way as in the 1st person, we 
 get Preg. *J>i-k. Go. has J>uk, which has evidently been in- 
 fluenced by the vocalism of the nominative. But O.E. and 
 O.H.G. retain the pure vowel : O.E. Sec d"e (Se, and E. thee = b~e 
 (the emphatic form 8 ) ; O.H.G. dih and G. dich. 
 
 3. Genitive. — Go. }>eina, O.E. (Sin, O.H.G. din and G. dein, 
 then dein-er, just as in the case of the 1st person. 
 
 4. Dative. — Go. Jjus instead of *]>is, like }>uk ; but O.E. 6"e <?e 
 and E. thee, O.H.G. dir and G. dir {supra 164, I., 4). 
 
 1 The same we have met with in Go. \>amma, supra 130, 4, and 134. 
 
 2 Hence it is obvious, that each personal pronoun ought to have two 
 accusative-forms, respectively, emphatic *mee and thee, unaccented me and 
 *the ; but the first person kept only the unaccented type, whereas the 
 emphatic type alone remained in the second person. 
 
 T 
 
274 ENGLISH AND GERMAN GRAMMAR. 
 
 II. Plural. — 1. The plural-stem of the 2nd person was I.E. 
 *yu-, which occurs, for instance, in Sk. yu-smd-d ablative, Gr. 
 vfi/jLc ace. Homeric, and vfxei<; = *yv-o-ix€-s. Gothic has the simple 
 nomin. jus. But West Germanic has altered the vowel under 
 the influence of the plural-stem of the 1st person : O.E. ge 
 (like we) and O.H.G. ir (like urir) ; E. ye and G. ihr. Usual 
 Mod. E. you is an accusative, while literary and dialectal E. 
 retains ye. 
 
 2. Accusative. — Go. izvis is not easy to explain and certainly 
 corrupt, though the last syllable seems identical with Sk. vas = 
 L. vos. West Germanic has a monosyllabic form, O.E. eow (cf . 
 pi. 1 us), and a dissyllabic form, O.E. eow-ic, O.H.G. iuw-ih, 
 the ending of which is imitated from the corresponding sg. 
 pronoun. Of these, eow became E. you, whilst iuwih survives 
 in G. euch. The difference in the initial diphthong is due to 
 a different accentuation : in English, the first component is a 
 semi- vowel, and the second a vowel ; and the reverse in German. 
 Further, ye and you, being nearly alike, were confused. 
 
 3. Genitive. — Go. izvara; O.E. eower and O.H.G. iuwer (like 
 user and unser) ; lost in English ; G. euer. 
 
 4. Dative. — Go. izvis, like ace. 1 ; O.E. eow and E. you (the 
 same) ; O.H.G. iu, also without ending, but Mod. G. euch, 
 assimilated to the ace. because in the 1st person uns is used for 
 both cases (M.H.G. dat. iu and ace. iuch). 
 
 § 3. Third Person. 
 
 (166) Indo-European, it has been seen, did not possess any 
 personal pronoun for the third person. Such languages as 
 secured one for their own use borrowed it from the demonstra- 
 tive class : thus F. il proceeds from L. ille. This being the 
 case, it is a remarkable fact that English and German, con- 
 sidering their strict correspondence to each other, should have 
 chosen to adopt a different stem to supply the want. Not a 
 single form, in the whole declension, is common to both lan- 
 guages, with the one exception of the nomin. fm. sg. : E. she = 
 
 1 It is doubtless from this dative that the analogy arose which created pi. 
 1 unsis and further sg. mis * \>is (G. mir dir). 
 
PRONOUNS. 275 
 
 O.E. seo, and G. ste = O.H.G. siu, both identical with Sk. syd — 
 I.-E. *sid, which is the fm. of the well-known demonstrative 
 *sio. 1 Everywhere else the comparison mnst be confined to the 
 mere endings. 
 
 I. Singular. — The E. pronoun proceeds from an I.-E. demon- 
 strative stem *kyo- and *ki-, from which some scanty remains 
 survive in other members of the family : Gr. o-rjficpov rrfixipov 
 (to-day) = *ky-o.[i£pov ; L. demonstrative hi-c hce-c ho-c 2 ; Go. 
 dat. hi-mma dag-a "in this day," and G. heute = *hio taga (to- 
 day) ; E. here, and G. hier, her, hin, etc., locative adverbs. — The 
 G. pronoun is the little demonstrative stem i-, which is chiefly 
 illustrated by the L. pronominal declension, msc. is, nt. i-d, 
 ace. msc. i-m (archaic), the latter identical with poetical Greek 
 l-v (fxiv vlv) " him." 
 
 1. Nominative. — O.E. msc. he (cf. the from the demonstrative 
 *to-) and E. he; O.E. nt. hi-t, with the characteristic pro- 
 nominal neuter-ending (cf. tha-t = I.-'E. *to-d), and E. it, the 
 initial aspirate being dropped. 3 — O.H.G. msc. er, instead of *i-r 
 = L. is, owing to the vocalism of der, and nt. i^ = Preg*i-t = lj. 
 i-d; M.H.G. msc. er, and nt. e'3 owing to the vocalism of er; 
 G. er and es. — The fm. she and sie borrowed from another stem. 
 
 2. Accusative. — E. nt. it ; in the two other genders the dative 
 is used as an accusative (O.E. acc.msc. hi-n-e). — O.H.G. msc. i-n 
 = L. i-m, also in-an with redundant termination, but G. merely 
 ihn = in lately lengthened; fm. siu, next sie, and nt. 13, next es, 
 like nominative. 
 
 3. Genitive. — O.E. msc. nt. his and fm. hi-re (hiere hyre) : 
 compare, on one side, the ordinary genitive-endings of the 
 demonstratives, and, on the other, the E. possessives. — O.H.G. 
 fm. i-ra, without difficulty ; but, in the msc. and (later) the nt. 
 
 1 Cf . supra 127 and 128, II. — Even this slight agreement disappears in the 
 oldest period of English and German, the latter always using siu, whereas 
 O.E. declines the whole of the pronoun of the 3rd person on the stem of he : 
 nomin. sg. fm. Mo (she), and even pi. heo (they). — The initial E. consonant 
 (* instead of s) is due to the influence of the following i, which was sounded 
 as a semi-vowel (group s + y, as in passion), whilst in G. the i kept its vocalic 
 value. 
 
 * The compared term is not the stem ho-, but the suffix -c, which repre- 
 sents a L. particle -ce — *ke (cf. L. gen. hujus-ce). 
 
 8 Cf. supra 66, II., 4, and further see the declension of the article. 
 
276 ENGLISH AND GERMAN GRAMMAR. 
 
 gender, the analogy of min and din in the other persons led 
 the language to adopt the correlative form of the reflexive 
 possessive sin, 1 which further became changed to a pleonastic 
 genitive sein-er imitated from mein-er and dein-er. 
 
 4. Dative. — O.E. msc. nt. hi-m = Gro. hi-mma (like J>a-mma), 
 and fm. hi-re hiere hyre ; E. msc. him, fm. her, but nt. it, the 
 accusative here serving for the dative. — O.H.G. msc. nt. i-mu, 
 fm. i-ru ; G. ihm and ihr. 
 
 II. Plural. — Here, deviating from the O.E. type, M.E. and 
 Mod. E. use the case-forms of the Germanic demonstrative *pd- 
 = I.-E. *to-, the same from which O.E. had already derived its 
 article: not the genuine O.E. case-forms however, but such as 
 were borrowed from O.N. at the beginning of the M.E. period. 
 German shows the same stem both in the singular and plural, 
 save that its nominative-accusative belongs to the demonstra- 
 tive I.-E. *sio. 
 
 1. Nominative. — E. they = O.N. }>ei-r. — O.H.G. msc. sie, fm. 
 sio, nt. siu, whence G. sie in the three genders. 
 
 2. Accusative. — E. them (original dative). — G. sie. 
 
 3. Genitive. — M.E. <7iei-r = O.N. )>ei-rra, to be found again 
 among the possessives. — O.H.G. i-ro regular, but Mod. G. 
 ihr-er, with pleonastic genitive-ending. 
 
 4. Dative. — E. the-m, with the usual and well-known dative- 
 ending. 2 — O.H.G. i-m, whence regularly in, and Mod. G. ihn-en 
 with a redundant plural case-sign. 
 
 § 4. Reflexive Pronoun. 
 
 (167) The reflexive pronoun is expressed in the two lan- 
 guages by a different process. The only primitive one is the 
 German. 
 
 I. In the first and second persons, the reflexive G. pronoun 
 is merely the personal pronoun : ich denke mir, " I am figuring 
 to myself " ; du befindest dich, " you find yourself," etc. But, in 
 the third person, German still possesses the representative of 
 
 1 Infra 168, 1., 3, and supra 164-165 (I., 3). 
 
 2 It is, of course, owing to a simple phonetic process (cf. supra 130, 4, and 
 131, 4), that the ending of this case has become identical in both numbers, 
 so that even O.E. has him in the plural as well as in the singular. 
 
PRONOUNS. 277 
 
 the general I.-E. reflexive pronoun, which, was characterized by 
 an initial sw or s, and appears clearly in the whole I.-E. family, 
 e.g. Sk. stem svd-, Gr. ace. c = *<rfe, L. se, etc. Of course, 
 this stem could have no nominative. It was, moreover, 
 indifferently singular and plural. All these characteristics 
 had been faithfully kept by Pregermanic; but the Saxonic 
 branch of West Germanic lost the pronoun altogether, as early 
 as O.E. 
 
 1. Accusative. — Go. si-Ze = I.-E. *se-ge, cf. mik a,nd}>uk; O.H.G. 
 sih and G. sich ; all both singular and plural. 
 
 2. Genitive. — Go. seina, O.H.G. sin, G. sein-er. 
 
 3. Dative. — Go. sis ; but G. sich, the accusative here being 
 used as dative (sg. and pi.). 
 
 In order to lay stress upon the reflexive meaning, the speaker 
 may add to the pronoun of any person, an invariable word 
 selb-st, which is apparently a superlative, but really a corrupt 
 form of M.H.G. selb-es, nomin.-acc. nt. and gen. msc. nt. sg. of 
 the adjective selb (self, same). This adjective formerly agreed 
 in gender, number and case with the governing pronoun ; but 
 it has now become invariable. 
 
 II. This stem, by means of a somewhat different, though 
 kindred construction, supplied English with a reflexive forma- 
 tion, which is now deemed obligatory. 1 Even in the O.E. period 
 are to be found such juxtapositions as se seolfa, literally " the 
 self" (the same, himself nomin., cf. G. derselb-e). From a 
 similar juxtaposition with the pronoun he, first proceeded the 
 accusatives-datives himself, herself, hitself, pi. themselv-es. 
 Next, since the form her could be equally referred to the posses- 
 sive adjective, and since, therefore, herself might easily be 
 mistaken for a substantival use of the word self, thus " the self 
 of her, her very person," the language at an early date also 
 created mlnself " my person," whence myself, thyself, our- 
 selv-es, yourselv-es, and (sg.) yourself. The equivocal sense is 
 the reason why the E. pronoun denoting identity, strangely 
 enough, in one case appears to be formed with the possessive 
 (lst-2nd person), and in another (3rd) with the pronoun itself ; 
 
 1 Authorised Version of the Bible, still ye clothe you (Haggai i. 6), now you 
 clothe yourselves, etc. 
 
278 ENGLISH AND GERMA.N GRAMMAR. 
 
 but rural dialects still show many traces of the old regular 
 types me-self, thee-self, ye-self, 1 etc. We now pass to the posses- 
 sive element. 
 
 § 5. Possessives. 
 
 (168) As a rule, the declension of the possessive adjectives 
 or pronouns comprises, — beside the gender, number, and case of 
 the possessed object, which is expressed in German by means 
 of the declensional endings, but not at all in English, — the 
 number and, only in the 3rd person, the gender of the 
 possessor, which must be inferred from the stem-form. 
 
 I. Singular of the possessor. — In every I.-E. language the 
 possessive adjectives and pronouns have been formed by some 
 method of derivation from the corresponding personal pronouns. 
 Now the derivative element adopted in Germanic in the singular 
 number seems to be the same we have already met with in E. 
 swine = G. schwein = h. su-lnu-m, literally "that which belongs to 
 hog." Thus, probably, Preg. *m-lna-z "which belongs to me," 
 etc. Go. rnein-s, J?ein-s and (nt.) sein, O.E. mln &ln sin, and 
 O.H.G. mln din sin. 
 
 In these words, when somewhat swiftly sounded and unac- 
 cented, either in English or in many German dialects, 2 the final 
 n often happened to be dropped, whence there arose such 
 doublets as E. mine and my. Further, in such monosyllables 
 the accent necessarily becomes weaker, when they are used as 
 adjectives, and consequently followed by an accented noun, 
 than it is when stress is laid upon them as pronouns and repre- 
 senting the noun itself ; and thus it is that English later made 
 the most of this accidental distinction by assigning its two 
 forms, respectively, to the adjectival and pronominal function, 
 my house is higher than thine. German, in either case, has but 
 the form with final n. Here, however, a process of secondary 
 derivation by means of suff. -ig gave rise to a periphrastic and 
 emphatic locution, concurrently used in the pronominal mean- 
 ing : der mein-ig-e, der dein-ig-e, further pi. der uns-r-ig-e, etc. 
 
 1 Cf. Wright, Dialect of Windhill, no. 353. — After the pattern of myself 
 was afterwards formed the late one's self, literally " somebody's person." 
 
 2 Alsatian and Swiss mi pfifli^mein pfeifiein (my little pipe). 
 
PRONOUNS. 279 
 
 1. O.E. mln, E. mine and my ; O.H.G. mln (nomin. msc. sg. 
 mln-er like blint-er, etc.) and G. mein. 
 
 2. O.E. Sin, E. thine and thy; O.H.G. din and G. dew. 
 
 3. The two languages are here at variance. — English, having 
 lost the reflexive, lacks the corresponding possessive, and 
 replaces it by the genitive of the pronoun of the 3rd person : 
 O.E. his "of him," if the possessor be of msc. or nt. gender; 
 hy-re "of her"; E. his, 1 her; whereupon, after the analogy of 
 the relation observed between he and his, a neuter possessive 
 was created, its. Again, the analogy of the relation observed 
 between dat. it and possessive its induced the speaker to derive 
 in the same way from dat. her a possessive her-s, which how- 
 ever, for the sake of distinctness, was confined to the pronominal 
 function. Hence : adjective 7m her its ; pronoun his hers its. — 
 O.H.Gr. sin and G. sein, which etymologically ought to serve for 
 the three genders of the possessor, since the reflexive, like any 
 other personal pronoun, knows no distinction of gender. 2 But, 
 for a possessor of feminine gender, German, like English, took 
 to using the genitive of the personal pronoun (G. stem i-), 
 whence O.H.G. i-ra " of her," G. ihr (her). 
 
 II. Plural of the possessor. — In the same way, and in both 
 languages, the plural of the possessor is denoted by the genitive 
 of the corresponding pronoun, as it were "of us, of you, of 
 them," but here the exponent of the gen. pi. is a syllable ending 
 in r, of rather obscure origin, which cannot be certainly identi- 
 fied with any I.-E. element. 3 
 
 1 E. his with a possessor of neuter gender is quite common as late as 
 Shakespeare, as also in the Authorized VersioD of the Bible : And it (the rock) 
 shall give forth his water (Exod. xvii. 6). 
 
 2 Supra 158 and 167 : G. sie sehnt sich exactly like es versteht sich. — Ob- 
 serve that G. sein is still clearly reflexive, though not so rigorously as L. 
 suus, since it is not allowed to stand in a sentence where it might be mis- 
 taken as referring to the subject. It would there be replaced by the genitive 
 of the demonstrative pronoun : er griff an sein schwert ; but Holland rilt 
 hinterm vater her, mit dessen (his father's) schwert und schilde (Uhland). 
 
 3 Perhaps it is the same r that appears in locative derivations, Go. he-r 
 = G. her hier^Fi. here, Go. >a-r = G. dar (also da) = E. there, Go. hva-r 
 ( = G. teo) = E. where, etc., so that Go. unsar and O.H.G. wiser would have 
 meant originally " among us," as in the Eussian sentence estt li u vasii 
 chlebu"? " is there with you bread ? " that is " have you any bread ? " — On 
 the other hand, Go. uns-ar might very well be a true possessive, a derivative 
 with a suffixal comparative element, like the Gr. Tposs^sixAJKUzrepo-s. 
 
 gg&Sg LJBTwrfjJg; 
 
 OP THE * 
 
 XT * ~7 T "O r- -r m V : 
 
280 ENGLISH AND GEEMAN GRAMMAR. 
 
 1. Go. uns-ar; O.E. user, and (syncopated) ur, whence B. our, 
 and (pronoun) ours, formed like hers ; O.H.G. uns-er (nomin. 
 sg. unser-er like min-er, etc.) and G. unser. 
 
 2. Go. izvar ; O.E. eow-er and E. you-r, pronoun yours ; 
 O.H.G. iuw-er and G. eu-er. 
 
 3. E. thei-r = 0J8. pei-rra (gen. pi.), pronoun theirs. O.H.G. 
 i-ro (gen. pi.) " of them," and G. ihr. 
 
 The use of the pronoun and possessive of pi. 2 (English and 
 German), and of sg. 3 or, chiefly, pi. 3 (exclusively German), 
 instead of sg. 2, is merely a matter of conventional usage. 
 
 (169) III. Declension. — English is well known to have 
 dropped every exponent of gender, number and case in its 
 possessives, even more constantly than in its demonstratives. 
 German has no less naturally transferred to its singular 
 possessives, inasmuch as they are genuine adjectives in the 
 nominative case (mein dein sein), the usual endings of its 
 demonstratives. In contemporary German, however, the nomin. 
 msc. sg. and the nomin. -ace. nt. sg. always assume the invari- 
 able form, and the terminations appear in the following 
 adjective ; mein vater, mein gut-er vater ; dein haus, dein schon-es 
 haus. 1 The possessive pronoun has the endings even in these 
 two cases : msc. mein-er and nt. meins (mine) = O.H.G. min-er 
 and min-ay. 
 
 The possessives which are really genitival forms (ihr fm., 
 unser, euer, ihr pi.), ought, of course, to be invariable : since 
 ihr nase means " of -her the-nose," the word ihr is theoretically 
 incapable of any further agreement with the word nase, and, in 
 fact, fm. and pi. ir, even as late as M.H.G., never appear with 
 any sign of gender, number or case. 2 But the reverse is the 
 case with the genitives unser and iuwer, which, even in O.H.G., 
 were declined like mm and din, as if they were genuine nomina- 
 tives. Their declension was somewhat arbitrary : thus, the 
 ending -er being mistaken for an ending of nomin. msc, the 
 
 1 In other words, the declension of mein dein sein has been completely 
 overruled by that of ein, a very plain and natural analogy. 
 
 2 Such is also the case with F. leur (their), which ought to be invariable, 
 as representing L. illorum (cf. Italian loro with either a singular or a plural 
 noun), but is now analogically sg. leur and pi. leur-s. 
 
PRONOUNS. 
 
 281 
 
 endings of the other genders, numbers and cases naturally 
 suppressed and replaced it with apparent regularity, msc. 
 uns-er, fm. uns-u, nt. uns-aj, etc. (Frankish) ; on the other 
 hand, unser being again mistaken for the invariable adjectival 
 form, the other terminations became added to it in the ordinary 
 way, msc. unser- er, fm. unser -iu, nt. unser-aj, etc. It is the 
 latter type that prevailed in classical German and was more- 
 over extended to the gen. ihr, msc. ihr-er, fm. ihr-e, nt. ihr-es, 
 though under the restriction stated above. The derivatives in 
 -ig require no further explanation. 
 
FOUBTH PART. 
 CONJUGATION. 
 
 (170) Under the name of Conjugation are comprised the 
 various modifications, either in stem or termination, under- 
 gone by the verbal stems, and corresponding to the 
 various relations of meaning of which they are capable. 
 
 These relations are five in number, namely: Tense, Mood, 
 Aspect (or Voice), Number, and Person. 
 
 The category of tense corresponds to the notion of time, 
 either present, past, or future. Indo-European possessed a 
 great many tenses, as we may gather from the extraordinary 
 richness of Sk. and Gr. conjugation. But Pregermanic kept 
 only two of them, Present and Historical Past. All the 
 other tenses are now supplied by various periphrastic locutions. 
 
 The category of mood states the manner in which the verbal 
 idea is thought of: either as simply affirmed or denied 
 (Indicative) ; or as eventually possible (Subjunctive), 
 desired (Optative), or subordinate to some uncertain event 
 (Conditional) ; lastly, as the object of command or prohibi- 
 tion (Imperative). The English and German conjugation 
 admits of but three moods : indicative, subjunctive, and impera- 
 tive. 1 All other shades of meaning, and occasionally even these 
 three, are expressed by means of a pei'iphrasis. 
 
 The aspect of the verb — what is termed Voice in Greek 
 Grammar — is either Active, or Reflexive (middle), or 
 
 1 The reader must not forget that neither the infinitive nor the participle 
 is a verbal mood. Both are mere nouns (supra 77-78), which however, for 
 the sake of completeness in the conjugational system, will be reviewed below 
 (229) at the head of the Chapter on Verbal Periphrases. 
 
 282 
 
CONJUGATION. 283 
 
 Passive, according as the agent (subject) performs the action, 
 or performs and at the same time suffers it, or simply suffers 
 it. Latin and, in particular, Greek denote the last two aspects 
 by various changes in the endings, which English and German 
 do not possess. A periphrasis must here again supply the loss. 
 
 The categories of number and person are the same for verbs 
 as for personal pronouns * ; indeed, as stated above in the 
 pronominal declension, they cannot be resolved into distinct 
 exponents, as it were, a number-sign and a person- sign. A 
 single element expresses both relations at once : whence a total 
 of six person -endings, three for the singular and three for 
 the plural. 
 
 In short, the Tenses, the Moods, the Person-Endings, 
 and lastly, the various Verbal Periphrases, which fill up the 
 gaps in English and German conjugation, are the four main 
 objects of study which fall under the head of Conjugation. 
 
 1 Supra 163 sq. The Gothic conjugation still keeps two persons (1 and 2) 
 in the dual number ; but nothing similar appears in the later languages. 
 
CHAPTER I. 
 
 TENSES. 
 
 (171) Logic, as well as grammar, distinguishes three essen- 
 tial classes of tenses, viz. Past, Present, and Future. In 
 grammar, however, the present-form is simply the form of the 
 verb itself, as analysed above in the study of word -formation. 1 
 Further, the I.-E. future was completely lost as early as the 
 the Preg. period, no Germanic tongue ever exhibiting the least 
 trace of it ; for all agree in replacing it by the present, or else, 
 when more precision is required, by using one of the periphras- 
 tic constructions considered below. Hence, of the whole tense- 
 system of English and German, there is nothing left for our 
 present study but the expression of the past, so far as it does 
 not itself consist of a periphrasis, but depends on some modifi- 
 cation of the verbal form. 
 
 (172) The latter mode of expressing the past tense is, again, 
 in Germanic, confined to one primitive grammatical form : of the 
 various original past tenses, namely imperfects, aorists, perfects 
 and pluperfects, best illustrated by the Sanskrit and Greek con- 
 jugations, Germanic has preserved but one, the perfect, which 
 plays the double part of an imperfect and an historical 
 past, E. he was and G. er war, both in stating a fact and in relat- 
 ing it. This somewhat awkward and ambiguous confusion took 
 place very early ; and, very early also, in order to remedy the 
 inconvenience, English and German were compelled to look for 
 a supply of periphrases. Indeed, it is a very remarkable fact, 
 that the form which in Germanic, as well as in Latin and 
 French, essentially assumes the function of an historical or 
 
 1 Supra 81 sq., 92-93 and 106-107. 
 
 284 
 
TENSES. 
 
 285 
 
 narrative past, has no such value at all in Greek, nor had it, 
 most likely, in I.-E. : a Greek narrator regularly uses the 
 aorists, whereas the Greek perfect involves the statement of an 
 achieved action, and therefore very often acquires the force of 
 a durative present * ; on the contrary, in Germanic, the his- 
 torical tense is the perfect, while the deficiency of a tense for 
 achieved action is made up, as in French, by the introduction 
 of auxiliary verbs. "We need but contrast I did with Fve done, 
 and er starb with er ist gestorben, to become aware of the rigorous 
 connection between the narrative meaning and the simpler and 
 primitive verb-form. 2 
 
 Section I. 
 
 PERFECT : GENERAL SURVEY. 
 
 (173) Thus we see that the study of the primitive English 
 and German tenses is confined to the Theory of the Perfect. 
 But this is worth a whole conjugation in itself, as the minutest 
 phonetic processes and the most various analogical resources 
 are here displayed, crossing and interacting upon one other in 
 every way. It is, in particular, impossible to study it apart 
 from the formation of the past participle, the first principles of 
 which have been already explained, 3 though many more particu- 
 
 1 Compare idave (he died) with rtdv^Ke (he is dead), etc., and observe that 
 this durative function is retained in a few Germanic perfects (termed preterito- 
 presents, infra 222). 
 
 2 An almost constant exponent in the I.-E. perfect is an initial reduplica- 
 tion (Sk. bi-bhed-a " he split," Gr. \4-\otir-e " he has left," L. ce-cid-it " he 
 fell"), which scarcely ever occurs in the aorist, and is likewise wanting in 
 most Germanic perfects. On the other hand, it is the I.-E. aorist that per- 
 forms the function of historical past, assigned to the perfect in the Germanic 
 family. From this double contrast, a very probable, and even certain infer- 
 ence may be drawn, namely, that the Germanic preterite — many scholars 
 preferably use this term in order to lay more stress upon the actual value of 
 the historical past — really consists of a nearly equal number of primitive 
 perfect-forms and aorist-forms, which were mixed together and blended into 
 one tense. This, moreover, is found in Latin. But the suggestion here 
 hinted at will not be repeated, since it would only cause needless compli- 
 cation; for, whether, for instance, Go. bit-um='E. (we) bit = G. (wir) bissen 
 was originally a perfect-form (Sk. bi-bhid-), or an aorist-form (Sk. bhid-), the 
 vocalism of the root, of course, must always become the same, so that the 
 distinction is of no interest, except as explaining the meaning of the tense or 
 the loss of the I.-E. reduplication. 
 
 * Supra 77-78 and 89-90. 
 
286 ENGLISH AND GERMAN GRAMMAR. 
 
 lars will naturally be required here. There exists a close relation 
 between these two grammatical categories : generally speaking, 
 if an English or German verb shows a so-called Strong 
 Perfect, — that is to say, with interior vocalic change and no 
 other tense-exponent, — it will also exhibit a so-called Strong 
 past Participle, — namely, a participle ending in -en, often 
 with either a similar or different change in the root- vowel, — 
 and reciprocally ; if a verb has a so-called Weak Perfect, — 
 without vocalic change, but formed by adding a suffix containing 
 a dental consonant, — the past Participle is likewise a Weak 
 one, — without vocalic change and with dental suffix, — and 
 reciprocally. Indeed, the vocalic shade in the strong perfect 
 must have often determined the vocalic shade in the strong 
 participle, or reciprocally : we have but to compare E. shine 
 shone shone and G. scheinen schien geschienen, wherein the 
 radical vowel, though the same in the present, differs in the 
 two languages in the other forms, but, in each, remains alike in 
 perfect and past participle. Hence we ought to group the two 
 forms side by side, as they are grouped in ordinary grammars, 
 and examine them together in every class of strong verbs in 
 which they were exposed to mutnal contamination. 
 
 (174) But, at the outset, a question must be solved: whence 
 arose, in the Germanic verbs, this fundamental and important 
 difference in conjugation ? or, in other words, why are 
 there Strong Verbs and Weak Verbs ? Language never 
 produces any arbitrary creations : there can be no doubt that 
 the distinction goes back to some ever-increasing difference in 
 one primitive and characteristic feature ; but this feature, 
 neither English nor German, nor their immediate ancestors, nor 
 even Gothic itself, would be able to reveal. In order to dis- 
 cover it we must trace the verb back to its earliest origin. 
 
 Such being the case, the rules of verbal derivation, as stated 
 above, are the necessary premises of our present search : conju- 
 gation, without them, is mere chaos ; with them, it becomes 
 perfectly regular. For the answer to the previous question is 
 found in a simple formula : there are strong and weak verbs, 
 because there are primary verbal stems,— that is to say, 
 verbs derived from the root merely by adding to it a single 
 
TENSES. 287 
 
 suffix, — and secondary (further tertiary, quaternary, etc.) 
 verbs, — that is to say, later verbal derivatives from a nominal 
 stem, the perfect of which must be a secondary and somewhat 
 complicated construction, since even their present is already an 
 aggregate of various elements added to the pure root-form. 1 
 
 Let us emphasize the contrast : E. (to) bear, for instance 
 (Go. bair-an), or G. nehm-en (Go. nim-an) has a pf. bare (Go. 
 bar) or nahm (Go. nam), and further apart, bor-n (Go. baur-an-s) 
 or ge-nomm-en (Go. num-an-s), because it corresponds to an I.-E. 
 verb derived directly from a root bher or nem, a simple verb 
 like Gr. <p€p-a> or vc/x-w, the perfect of which, if it existed in 
 that language, would be *Tre-<pop-a or *vc-vop.-a, a form likewise 
 constructed on the basis of the pure root by mere vocalic 
 gradation. Bat, still keeping the same examples, we find that 
 a Gr. causative like <pop-e-w, or a Gr. secondary derivative like 
 vo/a-cu-o), can no longer draw its perfect from the pure root, 
 since its present already contains other elements, or form it by 
 vocalic change, since the root-vowel has already undergone 
 such a change in the present : wherefore it forms a pf . ■m-<pop- 
 yj-K-a or ve-vo/x-ev-K-a, the perfect- exponent here no longer being 
 an internal change in a stem which has now become invariable, 
 but a new suffix -k- added to the prior derivative complex. 
 And this again is exactly the case with our weak perfect, except 
 that the Germanic suffix has a dental for the Greek guttural : 
 both elements, indeed, of equally obscure origin, but of equally 
 clear and similar meaning : E. ask and ask-ed-; G. lieb-en and 
 lieb-te. 
 
 Hence, — putting aside the cases in which a derivative verb 
 adopted a strong perfect and participle after the pattern of a 
 radical verb, and the other and more frequent cases in which a 
 primary stem has been analogically conjugated like the secon- 
 dary verbs, 2 — we may formulate the general rule as follows : 
 
 1 Cf. supra 70, 81 sq., 92-93, and 106-107. From this, however, it does 
 not follow that every primary verb is a strong one : of course, analogy has 
 interfered ; but the formula is convenient and simple, true in a general 
 sense, and quite satisfactory even for details, if restricted as shown below. 
 
 2 Since the latter were in a majority and daily increased in number, they 
 tended to ultimately prevail : a tendency, as will be seen, still far more ener- 
 getic in E. than it is in G. 
 
288 ENGLISH AND GEKMAN GRAMMAR. 
 
 strong verbs correspond to primary verbal stems ; weak verbs, 
 eitber to secondary stems, or else to such as were at an early 
 period assimilated to them. Still more accurately, strong 
 verbs are those which belong to classes II., III., V. 
 and VI. of our primary verbal derivation ; for class I., 
 being lost in Germanic, does not concern us ; and the causatives 
 (class IV.) even in Sanskrit have a periphrastic perfect 1 ; classes 
 VII.-VIIL, though really primary, are well-known to have 
 been treated as secondary in several parts of their conjugation, 
 and particularly in regard to their perfect-formation. 2 
 
 Of these so-called strong and often improperly termed irregular 
 verbs, English has preserved only about a hundred, counting 
 only simple verbs. 3 So great has been the power of analogy 
 that even some of these strong verbs only betray their primitive 
 character by forming a participle in -en, their perfect being 
 weak and ending in -ed. German is much purer and retains 
 twice as many, the complete enumeration of which belongs to 
 
 1 In some cases it happened, from a merely phonetic and regular cause, 
 that the present-vocalism of both the simple and the causative verb became 
 identical: thus E. sit = G. sitz-e ( = I.-E. *sed-yo) and E. *ei=G; setz-e 
 (=I.-E. *sod-Sy-o), are still contrasted forms ; but I.-E. *smeld-o (Gr. fj.4\d-w 
 " I become molten ") and I.-E. *smold-ey-5 (Germanic *smalt-ja " I cause to 
 melt ") became respectively O.H.G. smelzen and smelzen, which were further 
 blended into G. schmelzen. The same is the case with E. to melt, which 
 unites the two meanings. Next took place, analogically, either an extension 
 of the strong forms which swept away the weak ones (G. now only recognises 
 schmolz and geschmolzeri), or a confusion which led to the two being indiffer- 
 ently used (E. molt-en and melt-ed) : a process more or less illustrated by 
 such G. verbs as verderben " to become corrupted " and "to corrupt," loschen 
 "to become extinct" and " to extinguish," quellen, schwellen, etc., though 
 the primitive distinction is not forgotten. A correct speaker will say das 
 licht erlosch, but er loschte das licht aus. Likewise in E. a swollen leg, but 
 the rain has swelled the river, though the pf . is swelled in either sense. 
 
 2 Generally speaking, strong verbs correspond to the 3rd Latin conjugation,, 
 and weak verbs, to the three others, viz. : the type suchen (Go. sokjan — li. 
 sdgire), to the fourth ; the type salben (cf. Go. fiskon=li. piscarl), to the 1st ; 
 and the type haben (O.H.G. haben=Jj. habere), to the second conjugation. 
 Cf. the imperative-forms : suochi (seek) = sdgi ; salbo (salve) like amti ; habe 
 (have) = L. habe ; infra 196. 
 
 8 It need scarcely be observed that a compound verb follows in general the 
 conjugation of the simple verb : thus, kommen has kam and gekommen, in 
 an-kommen, zu-kommen, etc. But, of course, we must not be deceived by the 
 mere appearance of composition : the pf. and part, of bewillkommen are be- 
 willkommte and bewillkommt, because the vb. is not a compound of vb. 
 kommen, but a derivative from the adjective willkommen (cf. supra 118, 3). 
 And the same is the case with E. he overcame and he welcomed. 
 
TENSES. 289 
 
 ordinary grammar. Here our task will be to compare the 
 strong verbs in the two languages by classifying them accord- 
 ing to the regular vocalic changes which characterize their 
 tenses, and to ascertain the occasional influences which either 
 in English or German may have altered the original correspond- 
 ences in their conjugation. 
 
 Section II. 
 
 STRONG PERFECT AND PARTICIPLE. 
 
 (175) The so-called strong perfects of Modern Germanic 
 languages proceed from two distinct Pregermanic formations, 
 which are still clearly illustrated by the Gothic and, sometimes, 
 even by the later Germanic accidence, namely : perfects with 
 vowel-gradation and perfects with reduplication. The 
 former have been seen to change the root- vowel of the present- 
 stem : Go. bair-a (I carry), nim-a (I take), pf. bar and nam with 
 deflected root. The latter usually keep it unchanged, though 
 some may also alter it, and the leading characteristic of this 
 class, as of the Greek and Latin perfect, is a reduplication con- 
 taining a vowel e and prefixed to the root-syllable : Go. hdit-a 
 (I call), let-a (I leave), pf. hai-hait and lai-lot; cf. Gr. Ae-Aowr-a, 
 L. me-min-i, etc. 
 
 Of course, when taking these as two distinct classes, we are 
 considering exclusively the Germanic family ; for the fact is, 
 that the deflected and the reduplicated perfects formed origin- 
 ally but one class, or, in other words, that any I.-B. primary 
 perfect exhibited at the same time reduplication and vowel- 
 gradation. 1 But the adventitious distinction which took place 
 in Pregermanic may be easily accounted for : perfects in which, 
 owing to the phonetic laws of the Germanic group, the root- 
 vowel had become the same as in the present, kept their 
 reduplication, as the only means of distinction from the other 
 tenses ; while, in the perfects which kept their vocalism plainly 
 contrasted with that of the present, reduplication seemed a 
 
 1 Thus, the reduplicated Gr. pf. \4-\oiir-a, from a root Xe«7r (to leave), 
 shows exactly the same vowel-gradation that characterizes the primitive non- 
 reduplicated pf. ol5-a = fol8-a (I know) = Sk. ved-a, from a root fetS (to see). 
 
 U 
 
290 ENGLISH AND GERMAN GRAMMAR. 
 
 needless and rather cumbrous element which might well be 
 rejected. 1 
 
 § 1. Perfects with Vowel- Gradation. 
 
 (176) I. Let us recall the elements of I.-E. vowel-grada- 
 tion 2 : it is a vocalic alternation between three main grades 
 which have been termed respectively the normal, reduced and 
 deflected grade. To each of these grades correspond some well- 
 known I.-E. vowels. It remains to state the grade which 
 ought regularly to correspond to each tense- form of the conju- 
 gated root. This point too is easily settled : Sanskrit, and, in 
 particular, Greek, which are here in marvellous agreement 
 with Gothic, allow us to summarize it under four short and 
 invariable rules. 
 
 1. The present, — and consequently the Germanic infinitive, 
 which always resembles the present, — generally has the normal 
 grade (Gr. IS-cu <f>ep-o> AeiV-a) ^>evy-<o vcyn-w), sometimes — though 
 very seldom in Germanic — the reduced, never the deflected 
 grade ; — and this alone is sufficient to distinguish it from the 
 perfect. 
 
 2. The singular of the perfect always has the deflected 
 root: Gr. root f«S (to see), in cTS-os (image), cTS-o-v (I saw), but 
 pf . 018-a (I wot) = FolS-a ; Gr. XeiV-a), pf . Ac-Aowr-a ; L. fug-id and 
 -pi. fug -i, vtd-eo and pf. vid-i = *void-ei, etc. 
 
 3. In the plural of the perfect, the root is reduced 3 : Gr. 
 oTS-a, but pi. 18-fj.ev lo--fi€v (we know) . The fact is nowhere 
 better verified than in Sanskrit and Gothic : Sk. bi-bhed-a (I 
 split), pi. bi-bhid-imd ; Go. bait (I bit), pi. bit-urn. 
 
 4. The suffix -ono- of the participle (Sk. -and-, Go. -an-) is 
 accented and consequently always reduces the root : Go. bit- 
 an-s, cf . E. bitt-en and G. ge-biss-en ; Sk. root vart (to turn, L. 
 vert-ere), pf. part. vd-vrt-dnd-s = Go. vaurfr-an-s, cf. G. ge-word- en, 
 etc. 
 
 (177) II. Hence, if Pregermanic had kept the I.-E. tradi- 
 
 1 The more so, because, as noted above (172), some aoristic non-redupli- 
 cated forms intruded into the perfect and strongly influenced it. 
 
 9 Supra 43-45. 
 
 8 Because here the primitive accent was shifted to the personal-ending : cf . 
 the Sk. accentuation below, and supra 44, 45 (2), etc. 
 
TENSES. 291 
 
 tion unaltered, the theory of the strong verb would correspond 
 to the theory of vowel-gradation and would present a series of 
 phonetic equations of the utmost simplicity. But no language 
 of the European group is found in this state of archaic purity x : 
 at various times and in various circumstances, the four forms 
 stated above influenced one another, in such a way as to modify, 
 confound, and even differentiate one another by adopting new 
 characteristics, quite unlike their original features. Now these 
 various processes may again be summed up in four correspond- 
 ing formulas, the application of which will cross and limit the 
 effect of the preceding rules. 
 
 1. In imitation of presents, which primitively had the re- 
 duced root, — that is to say, the vowel-grade of the participle, — 
 it might well happen that other presents adopted, — chiefly as a 
 means of distinction from the perfect, with which the effect of 
 the Germanic phonetic laws tended to confound them, — the 
 vowel-grade of the participle, and thus opposed the reduced 
 root to the deflected root of the perfect. 2 
 
 2. In some verb-classes, Germanic phonetics had blended 
 both the reduced and the deflected grade to the same form ; 
 besides, in all of them, the vowel-grade was regularly the same 
 in the participle and the plural of the perfect : from this to a 
 closer assimilation and to the extension of the perfect-vocalism 
 to the participle itself, was but one step, and this step the 
 speaker often took. 3 
 
 3. Even within the perfect itself, such forms as bait bit-urn 
 might easily seem to be strange : hence there was a natural 
 tendency to adopt either one form or the other, and to con- 
 jugate the tense, either bait *bdit-um, or *bit bit-urn. 4 ' This 
 
 1 Up to a certain point, Gothic may be said, together with Sanskrit, to 
 abide by it most of all, since it still retains the curious alternation from the 
 sg. to the pi., which is entirely lost, apart from some immaterial survivals, in 
 Latin, and even in Greek. 
 
 2 Such is especially the case with the type slay = schlagen (Go. slah-a slah- 
 an-s), supra 45, 7, and infra 184. 
 
 8 A fact already stated in regard to the E. part, shone, contrasted with the 
 G. part, geschienen. Cf. infra 179 (2), 184 (2), 185 (IIL 2), and observe 
 moreover such popular blunders as I have took, etc. 
 
 4 Thus too, in Greek, the regular conjugation X^-Xo»7r-a *\e-\ur-fiei> was 
 soon forgotten, while a new pi. \e\oiiranev was formed from \£Koiira ; or else, 
 
292 ENGLISH AND GERMAN GRAMMAR. 
 
 analogy sooner or later worked its way throughout, and we are 
 able to follow its progress from the mother-tongues, which are 
 nearly free from it, down to the modern offspring, which pas- 
 sively obey its law. 
 
 4. Lastly, in imitation of the cases, in which the I.-E. reduced 
 grade did not differ from the normal vocalism, — the participle 
 and the present thus showing the same vowel, — it happened 
 that the participle adopted the present-vocalism in other cases 
 when it ought to have had the reduced syllable, — a corruption 
 which is found in all the Germanic dialects and is, therefore, 
 due to a vocalic peculiarity belonging to the common ancestor. 1 
 
 (178) III. These general principles being well understood, 
 we shall find that the English and German verbs with vowel- 
 gradation or deflected perfects fall under six distinct heads ; 
 the verbs with reduplicated perfects constitute by themselves 
 a seventh class. For the sake of brevity these seven types are 
 distinguished by letters. 2 
 
 (179) A. Type drive = treiben. 
 
 The type drive = treiben corresponds to our 2nd class of 
 vowel-gradation, that is to say, it comprises I.-E. verbs, the 
 root of which contains, in its normal grade, a diphthong ey 
 (Preg. i), which, deflected, becomes by (Preg. ai), and, reduced, 
 merely 1 (Preg. i) . Thus we shall have a triple alternation i ai 1, 
 which, when applied, for instance, to a Preg. vb. *orfb-ana-m 
 (to drive), and retained in the further evolution of the later 
 
 when the plural-vocalism prevailed, a sg. i\-f)\vda (for the regular Homeric 
 el\-^\ovd-a) was created in imitation of the regular pi. *4\-r]Xv6-fi€v (we are 
 come). Latin has many perfects with deflected root, like fug-i, and many 
 others with reduced root, like tul-i; but, whatever may be the actual vocalism, 
 it always holds good for the whole tense, singular and plural. Apart from a 
 few survivals, English and German have reached exactly the same point. — 
 Cf. also F. je treuve and nous trouvons (trdvotrovdmu$),faim and nous amons 
 (dmo amdmus), now levelled to trouve trouvons and, inversely, aime aimons. 
 
 1 The root kept without reduction in the participle is the leading character- 
 istic of the type see = sehen, supra 45, 1, and infra 183. 
 
 2 Hence our types A-G strictly correspond to the classes I.-VII. as distin- 
 guished in the historical grammars of the Germanic languages. The order 
 in which they are enumerated is rather arbitrary ; but it seemed convenient 
 to abide by it, since grammatical usage has adopted the terminology. 
 
TENSES. 
 
 293 
 
 languages, would produce the following historical, or theoretical, 
 forms : 
 
 Preg. inf. 
 
 "ffr'ih-ana-m 
 
 pf . sg. *ffrdib-a 
 
 pf. pi. 'tfrib-ume 
 
 part. *8rX£>-and-s ; 
 
 Go. „ 
 
 drcib-an 
 
 „ drdif 
 
 „ drib-um 
 
 „ drib-an-s ; 
 
 O.E. „ 
 
 drlf-an 
 
 , „ drdf 
 
 „ drif-on 
 
 „ ge-drif-en ; 
 
 E. „ 
 
 drive 
 
 , „ I drove 
 
 „ toe "drive 1 
 
 ,, driv-en ; 
 
 O.H.G.,, 
 
 trlb-an 
 
 „ treib 
 
 ,, trib-um 
 
 ,, gi-trib-an ; 
 
 G. 
 
 treib-en 
 
 „ ich * treib 
 
 „ wir trieb-en 
 
 „ ge-trieb-en. 
 
 We see at once the absolute purity of the Go., O.E. and 
 O.H.G. conjugation. By contrasting, on the other hand, the 
 historical and theoretical E. we *drive with the actual we drove, 
 and the historical and theoretical G. ich *treib with the actual 
 ich trieb, we easily become aware of the fact that, in the latter 
 language, the sg. of the pf. was accommodated to the pi., 
 whereas, in the former, the pi. borrowed its vocalism from 
 the sg. 
 
 Dealing now, in the same way, with a Preg. vb. *hit-ana-m 
 "to bite" (I.-E. root bheyd bhoyd bhid, cf. h.find-ere "to split," 
 pf.fid-i), we get the forms : 
 
 Preg. ir 
 
 if. *hit-ana-m 
 
 pf. sg. *hdit-a 
 
 pf . pi. *%lt-ume ; 
 
 part. *Vit-and-s 
 
 Go. 
 
 , beit-an 
 
 bdit 
 
 „ bit-um ; 
 
 „ bit-an-s ; 
 
 O.E. 
 
 , bit-an 
 
 „ bat 
 
 ,, bit-on ; 
 
 „ ge-bit-en ; 
 
 E. 
 
 , bite 
 
 „ I 'bote 
 
 „ we bit ; 
 
 „ bitt-en ; 
 
 O.H.G. 
 
 , bi^-an 
 
 n bev$ 
 
 ,, bijj-um; 
 
 ,, gibi^-an ; 
 
 G. 
 
 , beisz-en 
 
 ,, ich *beisz 
 
 „ wir biss-en ; 
 
 ,, ge-biss-en. 
 
 Here, both in English and German, the asterisk stands in 
 the same column, which is equivalent to saying that in both 
 languages the analogical process was carried on in the same 
 way, the sg. of the pf. being conjugated like the pi., thus E. 
 I bit like we bit, and G. ich bisz like wir bissen. 
 
 These two verbs represent the whole of this class, which 
 therefore requires but a few brief remarks. 
 
 1. English has about twenty and German about forty verbs 
 of this type ; though, excepting the constant alteration in the 
 pi. of the perfect, English here is by far the purer, having often 
 retained the vowel -gradation with three grades, e.g. in rise rose 
 
 1 Pronounced like give, not like drive in the 
 
294 ENGLISH AND GERMAN GRAMMAR. 
 
 risen, drive drove driven, shrive shrove shriven, smite smote smitten, 
 write wrote written, ride rode ridden, etc. 
 
 2. The last verb, however, also admits of a participle rode, 
 and, similarly, the vocalism of the pf. sg., after having invaded 
 the pi., further corrupts the part, also in abide abode abode, 
 shine shone shone, and strike struck struck,' 1 which have only two 
 vocalic grades ; though, of course, strick-en survives as an 
 isolated word. 
 
 3. The reversed process, as seen above, was no less possible ; 
 in that case the vowel-gradation was confined to the two grades 
 I and i, the vowel of the part, and pf. pi. being transferred to 
 the pf . sg., bite bit bitten (and bit 2 ), chide chid chidden, hide hid 
 hidden, even split 3 split split. 4 
 
 4. Now, in German, it is exclusively the latter process which 
 has spread and affected the whole of this class. German has 
 no more than two vocalic grades, namely, ei ( = *) for the 
 present, and, for the whole of the perfect as well as for the 
 participle, an i sounded either short or long 5 according as it 
 stands in a close or an open syllable : treiben trieb getrieben, 
 reiben, leihen, schreiben, etc. ; beiszen bisz gebissen, greifen griff 
 gegriffen, schneiden schnitt geschnitten, 6 leiden, weichen, streiten, 
 etc. And consequently, as opposed to the regular E. shone 
 wrote rode, we find the irregular G. schien risz ritt. 
 
 5. Having thus reduced its vowel -gradation, German applied 
 it even to some obviously secondary verbs which never had 
 anything to do with the primitive process : whence the strong 
 
 1 With vowel-shifting, owing to the shortening in a close syllable, supra 
 17, 4. The verbs stick stuck stuck and dig dug dug belonged originally to 
 type E (cf. G. steck-en stack ge-steck-t) and were transferred to type A by 
 analogy. 
 
 a On the accidental loss of the syllable -en, cf. supra 19, 2. 
 
 8 With shortening : cf . G. spleisz-en splisz ge-splUs-en. The' reader should 
 supply for himself any comparison that may have been omitted for the sake 
 of brevity. 
 
 4 The verbs rive and writhe still preserve their strong parts, riven and 
 writhen, but they have adopted a weak pf. rived and writhed. On the 
 contrary, strive (cf. G. streb-en) is a weak verb analogically adapted to the 
 strong conjugation. 
 
 5 Further, this long i is already known to be spelled ie, supra 12, 4. « 
 
 6 In this verb and the following we have a remarkable instance of " gram- 
 matical alternation" (supra 53 D and 55), the Preg. being inf. *sni\>-ana-m 
 and part. *sn\8-and-s, with accent-shifting leading to consonantal change. 
 
TENSES. 295 
 
 conjugation of gleichen, a derivative from the adjective gleich = 
 Go. ga-leik-s (cf. E. weak vb. like) ; and, even to preisen, a 
 derivative from preis (praise), which is a borrowed word (cf. 
 M.F. pris and F. prix). 
 
 (180) B. Type choose = kiesen. 
 
 This type corresponds to our 3rd class of vowel-gradation, 
 namely to the roots which respectively exhibited the three 
 vocalisms : normal ew, deflected ow, reduced u ; Preg. iu (eo) 
 au u. Thus, a root gews gows gus (to taste, cf. Gr. ycu-w and L. 
 gus-tu-s), treated according to our rule, assumes the following 
 forms, rigorously reproduced in the earlier Germanic languages : 
 
 Preg. inf. * keus-ana-m 
 Go. ,, kius-an 
 O.E. ,, ceos-an 
 O.H.G. „ kios-an 
 
 pf. sg. *kdus-a; pf. pi. *kuz-ume; part. *kuz-and-i 
 
 „ kdus ; ,, kus-um ; ,, kus-an-s : 
 
 ,, ceas ; ,, cur-on ; ,, ge-cor-en ; 
 
 ,, kos ; „ kur-um ; „ gi-kor-an. 
 
 Here, owing to the peculiar treatment of the I.-B. u in the 
 Germanic languages, 1 the threefold vowel-gradation has become 
 a fourfold one in O.E. and O.H.G. 2 But analogy has greatly 
 simplified this complicated conjugation. 
 
 1. English. — 1. First of all, the O.E. grade ea in the pf. sg. 
 vanished altogether: the O.E. type fleog-an ( = G . jiieg-en) fleag 
 flug-on ge-flog-en, although the purest in the later language, 
 is nevertheless confined to the three grades fly 3 flew flow-n ; 
 which is equivalent to saying, that, the infinitive and participle 
 respectively keeping their primitive vowel, the vocalism of the 
 pf. pi. was extended to the pf. sg., whence in the whole tense 
 flew = *flug, with the same change in spelling as in slew = *slog. A 
 
 2. No other verb has retained the three grades. The vowel 
 of the participle corrupted the perfect everywhere else, as 
 may be seen in : shoot ( = G. schiesz-en) shot shot ; cleave ( = O.E. 
 cleof-an) clove cloven ; even choose chose chosen, where the double 
 
 1 Cf. supra 28, I. 
 
 2 Observe also the " grammatical alternation " between s and z, which, 
 while entirely lost in Gothic, persists down to Mod. G. kies-en er-kor-en. 
 
 8 On y = eog, cf. supra 21. 
 * Supra 45, 7, and infra 184. 
 
296 ENGLISH AND GERMAN GRAMMAR. 
 
 vocalism is no longer betrayed except by the spelling, the a 
 being moreover extended throughout. 
 
 3. Further, two E. verbs show the gradation ee * o, viz. : freeze 
 froze frozen (instead of *fror-en, cf. G. frier-en fror ge-fror-en, 
 which, on the contrary, has extended the r), and seethe sod 
 sodd-en, which in both languages (G. sied-en sott ge-sott-en) has 
 kept the grammatical alternation. 
 
 4. These six verbs are the sole representatives of type B. 
 All the rest were shifted to the weak conjugation, so that 
 English, for instance, represents by verbs lose lost lost (formerly 
 loren, cf . for-lorn), lie lied lied, bow bowed bowed, etc., the G. verbs 
 ver-lier-en -lor -lor-en, liigen log gelogen, biegen bog gebogen, etc. 
 
 II. German. — 1. Here, on the contrary, there are no less 
 than thirty verbs of this class still existing ; but all of them 
 are confined to two vocalic grades. The old diphthong io in 
 the present is now, of course, spelled ie and sounded I. Now, 
 the diphthong in the pf. sg. is either O.H.G. ou or O.H.G. d, 
 according to the neighbouring sounds 2 ; but it has become 
 levelled to d, which again, in later times, was sounded either 
 long or short, according as the syllable was open or close. 
 This d, in consequence, became blended with o in the part., 
 which was itself either short or long under the same conditions, 
 and the it in the pf . pi. proved unable to resist this double 
 concurrence. In short, the modern paradigm is : ziehen zog 
 (Go. tduh) gezogen, fliegen, fliehen, wiegen (analogical), etc. ; 
 kriechen kroch gekrochen, gieszen, genieszen, triefen, sieden, etc. 
 
 2. In some as yet undefined circumstances, an I.-E. root 
 which had a u in its reduced grade could assume a normal 
 grade u instead of the regular diphthong ew. The fact is 
 verified as to two Germanic verbs, which, while exhibiting in 
 the pf. and part, the vocalism of the preceding verbs, merely 
 lengthen their root-vowel in the present : O.H.G. siif-an and G. 
 sauf-en (soff gesoffen) ; O.H.G. sug-an and G. saug-en (sog gesogen). 
 Their conjugation was afterwards applied to schnauben and 
 schrauben. 
 
 3. The metaphonical verbs er-kiir-en (cf . the primitive kies-en), 
 
 1 For the different vowel in the presents choose and freeze, see below 204. 
 
 2 Supra 32, the rule and the examples. 
 
TENSES. 
 
 297 
 
 trug-en (cf. the primitive (be-)trieg-en) and lug-en, which are 
 later derivatives from the nouns kur (choice), 1 trug and lug, 
 analogically adopted the conjugation of the primitive verbs 
 kios-an, triog-an and liogran. 
 
 4. With the exception of these seven presents, the leading 
 characteristic of type B is a vowel ie in the infinitive. 
 
 (181) C. Type drink — trinken and swell =schwellen. 
 
 Supposing the radical I.-E. e to be followed either by a 
 nasal or liquid, we reach the fourth and fifth classes of vowel- 
 gradation. But these two classes are not entirely comprised 
 in type C : they only belong to it so far as the radical nasal or 
 liquid is again followed by another consonant. When this is 
 the case, the e before a liquid remains e, but becomes i before 
 a nasal in West Germanic : hence, in West Germanic, type C 
 must be divided into two sub-classes, whereas in Gothic it is 
 the same throughout. 2 
 
 I. Let ns consider an I.-E. root dhrefig (to drink), deflected 
 dhroUg, reduced dhrng, and see what becomes of it in Preger- 
 manic and the modern languages. 
 
 Preg. inf. 
 
 *&rihk-ana-m 
 
 , pf. sg. *o r rdnk-a 
 
 pf. pi. *<frunk-ume 
 
 part. *ffruhk-and- 
 
 Go. 
 
 drigk-an 
 
 ,, dragk 
 
 „ drugk-um ; 
 
 „ drugk-an-s 
 
 O.E. „ 
 
 drinc-an 
 
 ,, drone 8 
 
 ,, drunc-on , 
 
 „ ge-drunc-en ; 
 
 E. 
 
 drink ' 
 
 ,, I drank 
 
 „ we * drunk ; 
 
 „ drunk ; 
 
 O.H.G. „ 
 
 trinch-an 
 
 „ tranch 
 
 ,, trunch-um 
 
 „ gi-trunch-an ; 
 
 G. ..„ 
 
 trink-en 
 
 „ ich trank 
 
 ,, wir * trunk-en * ; 
 
 „ ge-trunk-en. 
 
 1. The similar process in English and German is at once 
 obvious : the vowel of the pf. sg. was transferred to the plural, 
 we drank, wir tranken, etc. ; and consequently, even the purest 
 verbs of this class, namely such as keep unaltered the threefold 
 gradation i a u, never have any other grade but a in the 
 perfect-stem. 
 
 1 Also chur (election), whence chur-fiirst (high elector), and cf. the com- 
 pound will-kiir-lich (arbitrary). 
 
 * Because Gothic changes every e to i, supra 26. 
 
 8 O.E. o for E. a before a nasal, supra 39, 3. 
 
 4 Still M.H.G. (constantly) ich trank and wir trunken, ich fand and wir 
 funden, ich bran (I burnt) and wir brunnen, etc. 
 
298 ENGLISH AND GERMAN GRAMMAR. 
 
 2. Of the 25 E. verbs belonging to this type, eleven have pre- 
 served the regular alternation, e.g. : begin began begun, sing sang 
 sung, spring sprang sprung, sink sank sunk, swim swam swum, 
 etc. 1 ; cf. G. beginnen begann begonnen, singen sang gesungen, 
 springen, sinken, schwimmen schwamm geschwommen, etc. 
 
 3. But, even in the verbs which remain regular in English, 
 the perfect is 4iable to be assimilated to the participle : thus, 
 we may say either sank or sunk, span or spun, stank or stunk (G. 
 always sank spann stank). Next, the regular grade was com- 
 pletely forgotten, and the gradation was confined to two sounds 
 in cling clung clung, fling flung flung, string strung strung, win 
 won won (G. regularly gewinnen gewann gewonnen), where the 
 whole perfect is conjugated with the vowel of the plural-stem. 
 
 4. The latter has become the invariable rule for verbs in 
 which the primitive u has been lengthened to <m — u before the 
 group nd 2 : bind bound bound, find found found, wind, grind, con- 
 trasted with G. binden band gebunden, finden, winden, etc. 
 
 5. Though German, in this sub-class, is scarcely richer than 
 English, it is seen at once, from a glance at the previous in- 
 stances, to have preserved its original character far better. In 
 fact, the threefold gradation i a u (o) 3 here survives nearly in 
 every case. Occasionally, it is true, the vowel of the participle 
 appears in the perfect (ich begonn archaic) ; but this can hardly 
 be deemed an exception to the rule, as it is actually confined to 
 an optional hesitation in the metaphonical vowel of the sub- 
 junctive of this tense (rdnne or ronne, gewanne or gewonne, and 
 always begonne). There are but four verbs in which this cor- 
 ruption is consistently found, — and out of these four there are 
 two which once belonged to the weak conjugation, — viz. glim- 
 men (cf. the E. derivative glimm-er) glomm, klimmen ( = E. 
 climb, likewise a weak vb.) klomm (also klimm-te), schinden 
 schund geschunden, and dingen dung (bedingen bedung). 
 
 II. As the type of this sub-class we have chosen the verb 
 
 1 In run ran run (G. rinnen " to flow," not to be confounded with its caus- 
 ative rennen " to run ") the vowel of the part, has intruded into the present. 
 
 2 Supra 20, 4 B. 
 
 3 On the change of u to o (gesungen gefunden, but gewonnen geschwommen), 
 cf . supra 28, I. 
 
TENSES. 299 
 
 swell = schwellen, because it happens to be the only verb directly 
 comparable in the two languages. In the theoretical paradigm 
 however, it will prove advisable to replace it by another verb, 
 which has been preserved in Gothic, and to neglect the English 
 type, which, with many more, has passed to the weak conjuga- 
 tion. 
 
 Preg. inf. 
 
 *h£lp-ana-m 
 
 , pf. sg. *hdlp-a 
 
 pf. pi. *hulp-um6 
 
 part. *hulp-and-s; 
 
 Go. „ 
 
 hilp-an 
 
 ,, halp 
 
 ,, hulp-um ; 
 
 ,, hulp-an-s ; 
 
 O.E. „ 
 
 help-an 
 
 ,, healp 
 
 ,, hulp-on ; 
 
 „ ge-holp-en ; 
 
 O.H.G. „ 
 
 helf-an 
 
 ,< half 
 
 „ hulf-um ; 
 
 ,, gi-holf-an ; 
 
 G. 
 
 helf-en 
 
 ,, ich half 
 
 ,, loir half -en ; 
 
 ,, ge-holf-en. 
 
 1. Here, as above in type B, and owing to the same circum- 
 stances in Germanic phonetics, we find a fourfold gradation, 
 e a u o, which however in Mod. German has been reduced to 
 three grades, e a o, on account of the vocalism of the pf. sg. 
 having been extended throughout the whole tense. 1 
 
 2. But English scarcely shows a single representative of this 
 class, with the exception of a few participles : molt-en (cf . G. 
 ge-schmolz-en), as opposed to pf. melt-ed; holp-en, now replaced 
 by help-ed, etc. The whole set has been transferred to the 
 weak conjugation, and in particular the curious vb. starve (cf. 
 G. sterb-en " to die"), which, having adopted in the present the 
 vowel of the perfect (G. starb), 2 developed on this new stem a 
 new pf. starv-ed. Making these deductions, we find that only 
 three verbs fall under this head : swell swell-ed swoll-en, where 
 the pf . has become weak 3 ; burst burst burst (G. bersten barst ge- 
 borsten), where the vocalism is the same throughout; and, lastly, 
 fight fought fought. 
 
 3. German is much purer in every respect. Here, the type 
 schwellen comprises 22 verbs, namely : — 10 with consonant I 
 (helfen, melken, schmelzen, schwellen, quellen, bellen, etc.), including 
 also befehlen = O.H.(j. bi-felh-an, and erschallen (for *erschellen) 
 influenced by the vocalism of the substantive schall (sound) ; — 8 
 
 1 Still M.H.G. half hulfen, warf wurfen, etc., as above fand funden. 
 
 2 Possibly this is only an effect akin to the O.E. breaking, supra 21. — Ob- 
 serve the restricted meaning, and compare F. noyer (to drown) =L. necdre (to 
 kill). 
 
 8 The regular M.E. swal in Chaucer, e.g. Prioresses Tale, 72. 
 
300 ENGLISH AND GERMAN GRAMMAR. 
 
 with consonant r (werden, werfen, werben, sterben, bergen, verder- 
 ben, etc.), including wirren (verwirren) = O.H.Gr.werran, influenced 
 by the metaphonical vocalism of the adjective wirr (intricate) ; 
 — 2 in which the liquid precedes the vowel, the phonetic result 
 being naturally the same as far as the reduced grade is con- 
 cerned, 1 flecht-en = ~L. plect-o (cf. L. plied "I fold" and Grr. 
 ttAck-w), and dresch-en = ~E. thrash (weak verb) 2 ; — 2 analogi- 
 cally conjugated after the two preceding, fechten — 'E. fig ht, and 
 erldschen = O.H.&. ir-lesk-an (6 in spelling instead of e). 
 
 (a) The fourfold gradation persists but in one, namely 
 werden — Gro. vairj>-an = h. vert-l (to be turned, to be changed, to 
 become), which retains ich ward, wir wurden and geworden, 
 together with subj. pf. ich wurde ; though, even here, the 
 primitive conjugation is somewhat altered, inasmuch as the pi. 
 wir wurden gave birth to a sg. ich wurde, which is now oftener 
 used than ward. But there at least the grade u survives. 
 
 (fe) Such is no longer the case anywhere else, but the meta- 
 phonical form il appears occasionally in the subjunctive of the 
 perfect : sturbe, hiilfe, verdiirbe, wiirbe, wiirfe ; but also starbe, 
 warfe. 
 
 (c) Among the ten verbs that still retain the triple grada- 
 tion, several are already seen to admit concurrently either of 
 the perfect- vowel or the participle- vowel in the perfect-stem : 
 thus, side by side with barst, drasch, schalt, gait, befahl, we have 
 borst, drosch, scholt, golt, befohl, 3 with the vowel borrowed from 
 geborsten, gedroschen, etc. 
 
 (d) This anomaly becomes the rule in the remainder, whence 
 the merely double gradation e o in: schwellen schwoll geschwollen, 
 cf. E. swell ; and so also, quellen, melken, erschallen, erloschen, 
 etc. 
 
 (182) D. Type steal = stehlen. 
 
 This type comprises all the verbs of our fourth and fifth 
 
 1 An I.-E. root plek, for instance, and an I.-E. root pelk, if reduced, must 
 equally giveplk. 
 
 2 The present thrash =* thresh, like starve = *sterve and warp = *werp (G. 
 werfen). Besides, thresh as a rural form, is used in G. Eliot (Scenes of 
 Clerical Life) by a country-woman. 
 
 8 The last two are archaic, but subj. befbhle. 
 
TENSES. 301 
 
 classes of vowel-gradation, which do not belong to type C, or — 
 theoretically at least — those in which the radical e is followed 
 by a single nasal or liquid that is not followed by any other 
 consonant ; the nasal verbs, however, here are very few ; and, 
 on the other hand, analogy at an early date introduced into 
 this class some verbs which contained neither a nasal nor a 
 liquid, but were formerly, and still ought to be, conjugated 
 after the type E. 
 
 In fact, type D well marks a transition from the preceding 
 to the following verbs. It resembles the preceding, in that it 
 depends on the presence of a nasal or liquid in the verbal 
 stem: Preg. *nem-ana-m (to take), *fcer-ana-m (to carry), Go. 
 nim-an, bair-an. And, therefore, like the preceding, it strictly 
 requires the gradation e a u 1 : Go. nim-an nam num-an-s (G. 
 nehm-en nahm ge-nomm-eri), qim-an qam qum-an-s, bair-an bar 
 baur-an-s (G. ge-bar-en ge-bar ge-bor-en), etc. But, like the 
 following verbs, it assumes in the pf. pi. a peculiar vocalism, 
 namely a grade e (Go. nam nem-um, qam qem-um, bar ber-um), 
 which, being unknown in I.-E. vowel -gradation, has excited 
 the curiosity of many enquirers, without as yet betraying the 
 secret of its origin. 2 
 
 At any rate, the Go. alternation bar ber-um necessarily gives 
 O.E. beer bcer-on and O.H.G. bar bdr-um, that is to say, in each 
 language, a very similar sound for both numbers. Hence there 
 arose a tendency to unification, which took place everywhere, so 
 that English, which ought to answer O.E. with Ibare and we *bere 
 (cf. we were), has merely we bare. In German, the case is still 
 easier : for, as soon as the a was lengthened in an open syllable, 
 and shortened in a close one, there no longer remained any reason 
 why, — analogy moreover furthering the confusion, — the vowels 
 in ich gebar and wir gebaren, ich stdhl and wir stahlen, ich ham 
 and wir kamen, ich nahm and wir nahmen, etc., should be dis- 
 
 1 Indeed, apart of course from Gothic, the e here is never changed to t. 
 The u, naturally, is changed to o under the usual conditions. 
 
 2 To this e, have been compared the L. e in sedi, venl, etc., contrasted with 
 sedeo, venio, etc., and a Sk. e, which likewise occurs in the plural of certain 
 perfects ; but the connection between those various processes is quite doubt- 
 ful. We would rather adhere to the supposition of a so-called " lengthened 
 grade " of the root-syllable. 
 
302 ENGLISH AND GERMAN GRAMMAR. 
 
 tinguished from each other. Thus, the gradation admits of but 
 three grades at most : e a o. 
 
 German has 18 verbs of this type, namely: — 2 with consonant 
 m, nehmen = O.H.G. nem-an, and kommen (E. come) = O.H.G. 
 quem-an, where the present has adopted the vowel of the 
 participle; — 2 with consonant Z, stehlen (E. steal) = O.H.G. 
 stel-an, and verhehlen; — 4 with consonant r, g ebdr en = O.H.G. 
 ber-an, 1 scheren ga(h)ren and schwdren ; — 5 with a guttural, 
 though 4 of them have also an?* in the root, erschrecken, brechen, 2 
 sprechen, rdchen and stechen ; — 1 assimilated to the last type, 
 treffen; — and, lastly, 4 verbs which in O.H.G. belonged to the 
 type E and have passed to this, owing to the likeness of the 
 stem-vowel in the present, viz. pflegen, weben, bewegen and er- 
 wagen. On the contrary, the vb. verzehren ( = E. tear) has passed 
 to the weak conjugation ; but its old gradation still survives in 
 zorn (anger), a participle used substantially. 
 
 English has respectively : — come ; — steal ; — bear, shear, tear ; 
 — break, speak ; — weave, get, tread, three verbs from class E 3 ; — 
 heave, swear, from class F ; — wear, formerly a weak verb, now 
 assimilated to bear; — whereas play = O.E. pleg-ian = G. pflegen 
 has now become a weak verb, as the latter may also be in 
 German. 
 
 Of these 13 E. verbs, two only have kept the threefold grad- 
 ation : bear (in the meaning " to produce ") bare born, and come 
 came come. Two others still keep it in M.E., break brake (G. 
 brach) broken, and speak spake 4 spoken ; but the analogical broke 
 and spoke have decidedly prevailed. The vb. beget, in scriptural 
 style, has the pf. begat, whilst the only pf. of get is got, and so 
 with the rest : steal stole (cf . G. stahl) stolen, weave wove (cf . G. 
 wob) woven, bear (to carry) bore borne, shear, etc. 
 
 German is purer. Ten of its verbs preserve the three 
 grades : nehmen, kommen, stehlen, gebaren, erschrecken, brechen, 
 
 1 The a, here as in some other cases, is only a different spelling for etymo- 
 logical e. 
 
 8 Cf. supra 181, II. 3, and 42, note. 
 
 8 Compare forget forgot forgotten, tread trod trodden, with vergessen ver- 
 gasz vergexsen, treten trat getreten. 
 
 4 Usual in Shakespeare and his contemporaries. 
 
TENSES. 303 
 
 sprechen, stechen, treffen and pflegen (pflag or pflog). The re- 
 maining eight introduce the o of the participle into the perfect. 
 
 (183) E. Type see = sehen. 
 
 Excepting as regards the participle the two types D and E 
 are in full agreement : in the present, vowel e, Go. saihv-an • 
 in the pf. sg., vowel a, Go. sahv (cf. E. see saw and G. seh-en 
 sah) ; in the pf. pi., the same unexplained e, Go. sehv-um. The 
 only difference is in Go. saihv-an-s = 'E. see-n = G, ge-seh-en, which 
 has the infinitive- vowel. 
 
 The reason for this is that type E corresponds to our first 
 class of vowel-gradation, namely to the radical vowel e, standing 
 in any other position but in the four types A-D : such being 
 the case, the e, as noted above, when the syllable becomes 
 reduced, disappears in I.-E. without leaving any vocalic sound 
 to support the neighbouring consonants, so that the syllable 
 ought to disappear altogether ; but it has been seen that Pre- 
 germanic does not go so far as this. The reduced grade here 
 remains identical with the normal grade, that is, the stem of 
 the participle identical with the stem of the infinitive. Hence 
 there arises a threefold gradation, viz. e a e e (Go. gib-an gaf 
 geb-um gib-an-s, it-an at et-um it-an-s, etc.), which of course per- 
 sists in O.E. and O.H.G., but which still, by the process defined 
 above, is confined to two grades, e a a e, in the later languages : 
 E. give gave gave given, eat ate ate eaten ; G. geben gab gaben ge- 
 geben, essen asz aszen gegessen, etc. 
 
 Instead of an e, we may also find in the present, and ana- 
 logically even in the participle, 1 a metaphonical 1, that is, when 
 the verb is formed by means of the I.-E. suffix -yo- instead of 
 the simple -0- : thus, Preg. *et-ana-m (to eat), from I.-E. *ed-o 
 (I eat) = Gr. e8-a> = L. ed-o ; but Preg. *sit-jana-m (E. sit = G. 
 sitz-en), from I.-E. *sed-yo; and so also, E. lie=-Gc. liegen, E. bid 
 = G. bitten (to beg, to pray) = Go. bid-jan. 2 
 
 Briefly, the verbs of this class must have : in the infinitive, 
 
 1 In English only, in bidden part, of bid, whereas German has gebeten 
 part, of bitten. 
 
 2 In E. give = G. geben, the sound i is caused by the preceding g (O.E. 
 gief-an, cf. supra 50, II.). 
 
304 ENGLISH AND GERMAN GRAMMAR. 
 
 either e or i ; in the perfect, always a ; in the participle, theo- 
 retically at least, always e. 
 
 This main law holds good in English for six verbs of this 
 class, — which, when its losses are deducted (stick, tread, weave, 
 get), becomes reduced to nine in all, — namely : see saw seen, eat 
 ate eaten, 1 give gave given, bid bade (also bid) bid, spit spat (also 
 spit) spit, and lie lay lain. 2 In one of them, sit sat (sate) sat, 
 the perfect-vocalism has intruded into the participle. The 
 two remaining have neither a present nor a participle, viz. : 
 quoth = O.E. cwceb", 3 infinitive cwe&an = Go. qifran (to say) ; and 
 was, pi. were, the only one which, curiously enough, has kept 
 unaltered the regular gradation from the sg. to the pi., together 
 with the regular change of s to r, though, by being incorporated 
 with the vb. to be, it has lost every other tense. 4 
 
 The last verb is in German wesen (an infinitive used substan- 
 tially), pf. war (with the r imported from the pi. to the sg.), 
 part, ge-wes-en ; cf. O.H.G. wes-an was wdr-um gi-wes-an, and 
 Go. vis-an vas ves-um vis-an-s. And the same triple alternation 
 is reproduced in the fourteen verbs of type E which survive in 
 German : sehen sah gesehen, essen asz gegessen (a corrupt form 
 replacing the regular *gessen 5 ), geben, bitten, sitzen, liegen, etc. ; 
 lesen, genesen, messen, geschehen, lost in English ; treten, stecken, 
 vergessen, shifted in English to some other class. Here we can- 
 not but admire the marvellous preservation of the German 
 vowel-change. 
 
 (184) F. Type slay = schlagen. 
 
 With the five types A-E we have completed the five essential 
 classes of I.-E. vowel-gradation 6 ; the 6th class has no longer 
 
 1 Though a double analogy has produced the pf. eat and the part, ate, 
 both moreover unusual. 
 
 2 In lain = *lein<=d. gelegen, the vowel may as well be an e, as in way = 
 O.E. weg = Q. weg, etc., supra 50, II. 4. 
 
 3 E. for a owing to the influence of the preceding labial consonant. 
 
 4 Cf. infra 217, III. But illiterate people often say you was, thus effacing 
 the last trace of the traditional accidence. 
 
 5 O.H.G. gfi-e'35-an, syncopated to M.H.G. ge^en, and this again altered 
 to gegessen by receiving anew the half worn-out prefix ge-. 
 
 6 Let us state once more, in concluding, the general correspondence : 
 
TENSES. 305 
 
 any representative in Germanic accidence: there remains for 
 examination the 7th class, the characteristic of which has been 
 seen to be a mere alternation between ci,, in the infinitive and 
 participle, and 6, in the sg. and pi. of the perfect ; for, as early 
 even as the Pregermanic period, the vowel here had become the 
 same throughout the whole perfect-tense, Go. slah-an sloh sloh- 
 um slah-an-s, O.E. sle-an slog slog-on sleg-en, E. slay slew ( = *sZw) 
 slai-n, O.H.G. slah-an sluoh sluog-um gi-slag-an, 1 G. schlag-en 
 schlug schlug-en ge-schlag-en. 
 
 This type is usually kept without alteration in the verbs 
 which still survive, namely : E. shake shook shaken, take 
 (borrowed from Scandinavian), forsake, draw dreio draiv7i = 
 G. tragen trug getragen, G. fahrenfuhr gefahren, graben, tcachsen, 
 schaffen, laden, backen, etc. : in all, about ten English and a 
 dozen German verbs, besides the two metaphonical verbs 
 mentioned below. A very few remarks will suffice. 
 
 1. Several verbs have passed to some other system : E. fare 
 is now a weak verb ; so also E. shave and G. schaben, but the 
 former retains a part, shaven ; E. bake and lade are likewise 
 weak verbs, though the latter has a part, laden, cf. also the 
 vowel in the substantive load ; E. shape and wax preserve the 
 parts, shapen and waxen, but the pfs. are waxed and shaped, cf. 
 the G. forms schaffte geschafft from the vb. schaffen meaning " to 
 work " ; and, lastly, E. awake has awoke and awaked, whilst G. 
 wachen and E. wake are weak verbs. 
 
 2. The E. vb. stand = O.E. stond-an has several peculiarities. 
 The internal nasal, being a present-sign, does not pass to the 
 other tenses 2 : O.E. stod and E. stood. Next, the vocalism of 
 the perfect has been transferred to the part., now stood instead 
 
 Classes: 1st — 2nd — 3rd — 4th — 5th — 6th — 7th 
 
 Types: E— A— B — C — D— „_F. 
 
 1 Observe the "grammatical alternation," variously levelled in Go., O.E. 
 and Mod. German. 
 
 2 This should be the rule for any verb whatever assuming an interior 
 nasal, supra 82, 3 : cf. L. vinc-o vic-l,find-o fid-i, etc. But analogy every- 
 where transferred the nasal of the present into the perfect (as in L. jung-o 
 junx-i), and E. stand is the only one that remained free from such cor- 
 ruption. 
 
 X 
 
306 ENGLISH AND GERMAN GRAMMAR. 
 
 of O.E. regular gestonden. The same is the case "with awake 
 awoke awoke. The G. vb. *standen is lost, 1 but its pf. stand and 
 its part, gestanden are in use : the latter is regular (O.H.G. 
 gistantan) ; the former ought to be *stund (= O.H.G. stuont), 2 
 which is at any rate found in the pf. subj. stiinde, so that 
 German has certainly undergone the reverse analogy to that 
 of English. 
 
 3. Two Anglo- German verbs exhibit, in the present only, 
 instead of an a, the metaphonical vowel corresponding to a, 
 because they are derived by means of the Preg. suff. -ja- = I.-Tfi. 
 -yo-, namely: Go. haf-jan ( = L. cap-id), O.E. hebban and E. 
 heave, O.H.G. heffan and Gr. heben ; O.E. swer-ian and E. swear, 
 O.H.G. swer-ien and Gr. schworen. 3 Both being identical, in 
 their present- vocalism, -with the verbs of the type D, became 
 assimilated to this type, completely in English, and partly in 
 German : E. heave hove hoven (also weak), swear swore sworn ; 
 G. heben hub (also hob) gehoben ; schworen schivur (also schwor) 
 geschworen. 
 
 § 2. Reduplicated Perfects. 
 
 (185) G. The only Type being fall = fallen. 
 
 I. The last type we have to study now is merely differ- 
 entiated from the six preceding, — as the six preceding are 
 differentiated from one another, — by the nature of the alterna- 
 tion in its root- vocalism : in other words, this type appears, 
 no less than the types A-F, to proceed from vowel-gradation. 
 In reality it is nothing of the kind. For the history of the 
 Germanic languages and, in particular, comparison with Gothic 
 accidence show that verbs of this class, — those at least which 
 were primitive and served as patterns for English and German 
 conjugation, — kept the present- vowel both in the perfect and 
 the participle, and that the perfect took no other tense-sign but 
 
 1 Replaced by steJitn, infra 220. 
 
 2 German, less pure than English, has allowed the nasal of the infinitive 
 stantan to spread into the whole verb. Compare moreover Go. standan, pf. 
 std\>. 
 
 8 Spelled <5 = e, cf. supra 50, III. 2, and 121, 12; as also in loschen, supra 
 181, II. 3. 
 
TENSES. 307 
 
 the initial reduplication mentioned above : thus Gothic, for 
 instance, answers with hald-an hai-hald hald-an-s the verb 
 which is E. hold held hold-en (now held) and G. halt-en Melt 
 ge-halt-en ; and, if Gothic had a vb. *fall-an, it would likewise 
 oppose a pf. *fai-fal to O.E. feoll and E. /eZZ = O.H.G. fiat and 
 G.fiel. 
 
 Hence the perfects of type G were clearly reduplicated in 
 Pregermanic, and adopted an interchangeable vocalism only as 
 late as West Germanic. How the latter fact took place, will 
 be examined further on, though at the outset we may clear the 
 ground, by stating that their leading characteristic is a clearer 
 vowel (E. e and G. ie) as opposed to the obscurer vowel in the 
 present and participle. This obscurer vowel is not the same 
 throughout the whole class, as it is in the types A-F. Indeed, 
 it is a very capricious one, and here the unity of the class 
 comes by no means from the vocalism of the infinitive, but 
 only from the constant perfect-vocalism, as the whole category 
 admits of no less than nine sub-classes, fall = fallen, hold = 
 halten, blow — blasen} let = lassen, (hew = ) hauen, (leap = ) laufen, 
 stoszen, heiszen and rufen, all of which, in spite of the differences 
 in their infinitive-stern, assume the same vowel in their perfects 
 fell=fiel, held = hielt, blew = blies, let=liesz, hieb, Uef, 2 stiesz, 
 hiesz and rief : a remarkable likeness which at first sight 
 betrays a principle of formation common to all these various 
 stems. 
 
 And what could this common principle be but reduplication, 
 since the nine corresponding presents and perfects respectively 
 either are or would be Go. *falla *fai-fal, halda hai-hald, Mesa 
 *bai-bles, leta lai-lot, *hagva *hai-hagv, hldupa *hai-hldup, stduta 
 *stai-staut, hdita hai-hdit and hropja *hai-hrop ? The system 
 moreover is of the greatest simplicity : the vowel of the infini- 
 tive is always that of the participle, almost always that of the 
 perfect, and the latter always remains unchanged from the 
 singular to the plural. English and German reproduce all 
 
 1 More accurately speaking, E. blow would be G. blahen (weak vb.) ; but 
 the G. added s is reproduced in the E. derivative noun blas-t. 
 
 2 Unluckily, these two verbs, with many others, have passed in English 
 to the weak conjugation, cf. infra III. 2. 
 
SOS ENGLISH AND GERMAN GRAMMAR. 
 
 these characteristics with the single exception of the alteration 
 in the perfect-vowel. 
 
 II. The whole question, therefore, is how it happened that 
 West Germanic was led to substitute for the primitive re- 
 duplication a single syllable with clearer vowel ; and this 
 question answers itself; for, since the I.-E. or Preg. reduplica- 
 tion, being always characterized by a vowel e, necessarily con- 
 tained a clearer sound than that of the verbal root to which it 
 was prefixed, there can be hardly any doubt that the vocalic 
 change in the E. and G. root is due to the influence of this 
 reduplication. The only difficulty under which we labour is 
 to state the precise circumstances under which this process 
 took place ; for we have no historical documents to guide us, 
 and are left to choose between two different though equally 
 plausible and not altogether contradictory ways. Analogy may 
 have followed both at the same time, together with perhaps 
 some unknown paths. 
 
 1. Let us restore some Preg. perfects of the Go. type hai- 
 hald, that is to say, perfects of verbs beginning with an h, 
 *he-halfr-e (he held), *he-hd%w-e (he hewed), *he-hdit-e (he 
 called), *he-hdfii-e (he hung), etc. Nothing is more natural 
 than that, the medial h being sounded weakly and then 
 dropped, the two vowels thus brought together should coalesce 
 into one long vowel, whence *heWe, *he%we, *heite, *hei\^e, etc. 
 (G. Melt, hieb, hiesz, hieng). Now the visible relation between 
 such stems as hal& ha%w htiit hafi% and such as hel& he%w heit 
 heft&, respectively in the present and perfect, being analogically 
 transferred to other verbs, there arose, for instance, fel and 
 feft% (G. fiel and fing), created on the stems fdl and f&A^, and 
 so forth. Further, when the root contains a u, we find the 
 diphthongs O.E. eo and O.H.G. io in the perfect stem (O.H.G. 
 louf-an liof=G\ laufen lief) ; the simple e underwent the O.E. 
 breaking (feall-an feoll), while English still shows it unaltered 
 though shortened (held, fell) ; the same e became diphthongized 
 to H.G. ia (fial, hiang) ; and, lastly, O.H.G. io and ia both 
 equally gave G. ie or i (lief, fiel, hing) : these are matters of 
 simple inference for any one acquainted with phonetics. 
 
 2. We have seen that the root-vowel remained unchanged 
 
TENSES. 
 
 309 
 
 throughout the perfect. But it may be doubted whether this 
 was the case in Pregermanic, since it is a well-known fact that 
 in the I.-E. plural of the perfect the root-vowel tended to be- 
 come reduced, and even occasionally to disappear altogether. 1 
 Now, it is a curious coincidence that O.B. retains five perfects 
 which clearly show an initial reduplication, together with a 
 root which has lost its vowel, e.g. Tie-lit, from hdt-an (to call), 
 and reo-rd, from raid-an (to read) ; cf. Gr. hiesz, from heisz-en, 
 and rieth, from rathen (to guess). Hence, supposing Preger- 
 manic to have possessed such perfect-stems as *he-ht-, *re-rtS- y 
 etc., it was but natural, indeed, almost unavoidable that the 
 medial consonant in these forms should be eliminated, since the 
 speaker, when contrasting them with such present-stems as 
 *hait-, *re&"-, would obtain the misleading view that the interior 
 h or r (or other consonant) in each was a superfluous growth ; 
 and thus he would begin to suppress it, and to form the perfect- 
 stems *het-, *reft- (close e), 2 next analogically *fel- for *je-fl- y 
 and so on. 3 
 
 III. If however the origin of the phenomenon is lost in the 
 mists of the past, all its present applications are plain, simple 
 and absolutely constant. 
 
 1. German keeps fifteen perfects of the reduplicated type,* 
 all of them with vowel ie (t), and with the same vowel in the 
 infinitive and participle. To those quoted above we need but 
 add braten briet, schlafen schlief, fangen fing, and (gangeti) 5 ging. 
 
 2. English is far less pure. Besides the four verbs quoted 
 
 1 Cf. supra 44 and 45, 1. 
 
 2 We may even, in "red from reord, treat the transition as merely pho- 
 netic, inasmuch as we find E. meed and G. miethe (salary) =O.H.G. mlta = 
 O.E. mtcrdm Go. mizdo = Gr. fuad is, cf. supra 62. 
 
 3 On this difficult and intricate question, the most recent and complete 
 authority is: H. Lichtenberger, de Verbis qua in vetustissima Ger manor um 
 Lingua reduplicatum Perfection exhibebant (Nancy 1891), the work of an 
 acute and well-trained scholar. 
 
 4 O.H.G. had a few more, but some of them have passed to the weak 
 system : thus salzen (Go. salt-an, pf. sai-salt, but G. salz-te) still betrays 
 itself as a strong verb through its old part, ge.-salz-en (but ich habe ge-salz-t 
 in conjugating), whilst mahen, sden, etc., have lost every survival of the 
 strong type. 
 
 5 Infinitive and present obsolete in German, now replaced by gehen. In 
 English the whole verb is lost. Cf. infra 219. 
 
310 ENGLISH AND GERMAN GRAMMAR. 
 
 above may be mentioned grow grew, throw x threw, crow crew, 
 know knew and beat beat (O.E. beat-an beot). The vb. hold re- 
 places its part, hold-en by held after the analogy of the perfect. 
 The verbs hew ( = hauen), sow, mow, strow and strew, show and 
 shew (cf. G. sdev, mdhen, streuen, schauen, weak verbs) still keep 
 a part, in -n, but have a weak' perfect. Some others, as leap 
 ( = Gr. laufen) have become weak throughout their whole acci- 
 dence. Lastly, the vb. which is quite regularly O.E. *hong-an 2 
 heng ge-hong-en = G. hang-en hing ge-hang-en, is now E. hang 
 hung hung, owing to the intrusion of a vowel which has nothing 
 to do with this class and has probably been imported hither 
 from type C. 3 
 
 Section III. 
 
 WEAK PERFECT AND PARTICIPLE. 
 
 (186) Weak Verbs, as already stated, are those in which 
 the perfect and participle are formed, without any vocalic 
 change, by mere addition of a dental element, E. -d (-ed) or -t, 
 G. -te (-ete) and -t (-et). The great majority of English and 
 German weak verbs obey this fundamental rule ; but a con- 
 siderable minority, especially in English, are exceptions, whether 
 as apparently lacking the dental element, or as actually under- 
 going in their root some accidental vocalic shifting. Hence, 
 in order to fully understand the principle of formation, it 
 appears advisable to begin at once with stating all these 
 apparent anomalies and reducing them to the normal form 
 from which they are separated only by an insignificant detail. 
 
 § 1. Apparent Anomalies. 
 
 (187) In order to analyse the doubtful cases of weak 
 perfect and participle, we must first observe that the un- 
 accented vowel (now e mute) which preceded the dental suffix, 
 tended in early times, as it still does, to disappear in rapid 
 
 1 G. drehen " to turn," weak verb. 
 
 2 The actual O.E. infinitive is h0n = Go. hahan = *hanh-an, supra 24. 
 8 For the part, hang-ed, see above, p. 148, n. 2, and p. 288, n. 1. 
 
TENSES. 311 
 
 pronunciation. In fact, it vanished in any position, even after 
 a dental consonant, since German regularly has er sandte (he 
 sent), gesandt (sent), beredt (eloquent), so that the fuller forms 
 send-ete ge-red-et are ohviously refashioned, by a late analogical 
 process, on the basis of the infinitives send-en and red-en. 
 
 Next we must remember the undecided character of the 
 sound d as an English final ; for, although the consonant here 
 was a true Germanic o, it could nevertheless become shifted to 
 t, either before a voiceless initial consonant in the following 
 word, or when absolutely final at the end of a sentence, or 
 when immediately preceded by a voiceless consonant after the 
 intermediate vowel had been dropped. 1 
 
 No wonder, therefore, if we often find an English final t, or 
 if the d or t appears to be wanting altogether, that is, when 
 the verbal stem itself ends with a dental, so that the two 
 contiguous dentals coalesced after the syncope. Let us now 
 follow, in both languages, the further development of these 
 various processes. 
 
 1. Such a form as pas-t requires no explanation, being indeed 
 the true transliteration of the pronunciation of the word pass-ed 
 with which it alternates. And durs-t, instead of the analogical 
 dare-d, is hardly more puzzling ; for the Preg. root is not *$er-, 
 but *Sers- (I.-E. root dhers, cf. Gr. Oapcr-eiv " to cheer up," 
 6pacr-v-<s "bold"), and the present dare = O.E. dearr 2 merely 
 conceals the s by assimilating it to the preceding r, whereas it 
 reappears before the -t of the perfect. The difference in the 
 vocalism proceeds from dare regularly having the deflected 
 grade, while iiie remarkable archaic form durst no less regularly 
 exhibits the reduced grade (O.E. durs-te, from an I.-E. root- 
 form dhrs), which is likewise the case with the Greek words 
 compared above. 
 
 2. We can also easily understand the rather rare syncopes 
 which took place, usually in common words, between the final 
 consonant of the stem and the dental of the suffix, and led to a 
 complete absorption of the former: E. had = 0.~E. hmf-de, and 
 
 1 Remember moreover that the I.-E. suff. -to- of the participle is in some 
 cases represented quite regularly by a Germanic -t, supra 53 b and 78. 
 
 2 Exactly speaking, this present itself is an old perfect, cf. infra 222. 
 
312 ENGLISH AND GEEMAN GEAMMAR. 
 
 E. had=O.E. ge-hcef-d, cf. E. head = O.E. heafod; G. hatte = 
 O.H.G. hap-ta (also hab-e-ta), but part. ge-hab-t = O.H..Gc. 
 gi-hab-e-t; E. made = O.E. mac-o-de. 
 
 3. Now, whether these two consonants thus become assimil- 
 ated, or keep their original sound, in any case the loss of the 
 intermediate vowel produces a consonantal group, which again 
 may exert its usual influence over the preceding vowel, that is, 
 may either shorten it, or prevent it being lengthened : we have 
 but to compare the long vowel in G. haben — O.H.G. hab-e-n y 
 with the short one in hatte, gehabt. This is quite an elementary 
 distinction. But in English the phenomenon leads to further 
 consequences : the lengthening or shortening in its turn alters 
 the sound of the radical vowel, whence there arise such alterna- 
 tions as keep kept (O.E. cep-an cep-te), weep wept, feel felt, kneel 
 knelt, flee fled, 1 lean leant (also leaned) , mean meant, lose lost, 
 shoe shod, etc., all familiar to the reader. 
 
 4. Under the same conditions, if the syncope develops, after 
 a vowel capable of O.E. breaking, a consonantal group capable 
 of causing it (rd, Id), then, the breaking of the vowel produces 
 the alternations hear heard (O.E. her-an hier-de), sell sold (O.E. 
 sell-an seal-de), tell told, etc., so that the verb seems to vary in 
 its stem, whereas the history of the language explains this 
 apparent variation as the result of combined breaking and 
 metaphony, acting on an originally identical vocalism. 
 
 5. If the final consonant of the verbal stem happens to be a 
 dental explosive, d or t, like the suffix-consonant, the usual 
 syncope of the intermediate vowel, as a rule, takes place both 
 in O.E. and O.H.G., whereupon, the two dentals being blended 
 into one, the perfect and participle will seem as if they lacked 
 any exponent ; and, since English chiefly retains this primitive 
 formation, it inherits and keeps throughout such types as send 
 sent (O.E. send-an and sende = *send-de), rend rent, gild gilt, gird 
 girt, or cast cast, cut cut, put put, set set, hurt hurt, even cost cost 
 (borrowed from F.), indeed, even shed shed (O.E. scead-an 
 scead-de), wed wed, etc., etc. In the last two cases, the perfect 
 
 1 Supra 19 and 20. The vb. fleon = G. fliehen is strong in O.E. : it became 
 weak as a means of differentiation from E. X*/ = O.E. fleogan=G. fliegen y 
 and assumed a d which closes the syllable, whence the e is shortened. 
 
TENSES. 31$ 
 
 and participle have exactly the same relation to the infinitive 
 as in the types of strong verbs let let, split split, etc., and it is- 
 only the history of the language that allows us to distinguish 
 between these strong and weak forms. 1 Later on, analogy 
 caused the formerly syncopated final to be restored in many 
 verbs of this kind : from last, blot, wed, etc., were derived 
 last-ed, blott-ed, wedd-ed ; and this is almost constantly the case- 
 with literary German, rett-en rett-ete ge-rett-et, red-en red-ete 
 ge-red-et, etc. ; but the surviving forms, sandte gesandt, wandle 
 gewandt, beredt, etc., wherein the d is spelled but not sounded 
 and might be dropped altogether, 2 still shows in German an 
 early syncopated state, similar in regularity and constancy to* 
 the O.E. accidence. 
 
 6. Before the double dental group thus obtained in O.E., 
 English naturally undergoes the processes of shortening and 
 vowel-shifting, whence come the well-known alternations read 
 read (O.E. rved-an rced-de), lead led, meet met, light lit, etc., and 
 even clothe clad (O.E. clceft-an cloeft-de). The last shows us that 
 the change of ce to o is later than the shortening of a vowel in 
 a close syllable. Thus, here too the stem seems to' vary, 
 whereas the phenomenon is late and the result of accident. 
 
 7. Some variations appear more serious, merely because 
 they are older and carry us back farther into the past of 
 Germanic phonetics. Thus the present sometimes shows a 
 regular metaphony, which has no reason for appearing in the 
 perfect. 3 Taking, for instance, the Go. vb. vand-jan (to turn), 
 the pf. of which is vand-ida, we find the infinitive to be regu- 
 larly represented by O.H.G. wenten, and the perfect, no less 
 regularly, after syncope has taken place, by ivanta, after the 
 analogy of which German will replace the old part, gi-went-it 
 by a new one gewandt. These alternations were quite common 
 
 1 It is quite plain that, reciprocally, this identity must have contributed 
 on a large scale to further the loss of final -en in such strong participles as 
 bid (for bidden), and then, analogically, in a great many others. 
 
 2 The O.H.G. forms are infinitive went-en, pf. vrnnta, part, gi-went-it;. 
 and so also, rett-en (to save), pf. rat-ta (also ret-ita, which shows the work 
 of restoration to have begun very early), part, gi-ret-it, nomin. sg. gi-rat-t-er :. 
 a curiously varied, though now altogether uniform accidence, cf. infra 7. 
 
 3 This has already been seen above in the types sell and tell, where th& 
 present-metaphony actually combines with the perfect-breaking. 
 
314 ENGLISH AND GERMAN GRAMMAR. 
 
 in O.H.Gr., but the later language has levelled them almost 
 throughout, keeping merely a few survivals in the class of 
 verbs improperly termed " mixed verbs " — intermediate, as it 
 were, between strong and weak — in some grammars, namely : 
 ■nenn-en ( = Go. namn-jan) nann-te ge-nann-t, rennen, kennen, 
 brennen, senden and tcenden. English, more corrupt in this 
 case, has extended the vocalism of the present to the whole 
 conjugation: send sent, rend rent, burn burnt with metathesis. 
 
 8. English, notwithstanding, retains some traces of this effect 
 of present-metaphony, namely in verbs in which the present 
 and perfect had diverged too far to be again reconciled by 
 analogy. Nothing could be more unlike, at first sight, than 
 buy and bought ; but, on going back to O.E., we find bycg-an 
 and boh-te, and we at once recognise the y as the metaphony of 
 o ; in other words, we restore a Preg. *hu^-jana-m, wherein the 
 u has been mutated, while it was kept unaltered in the pf. 
 *huh-ta, no i here being inserted. So also, seek and beseech are 
 metaphonical forms of suchen and besuchen, whereas the perfects 
 ■sought 'and besought strictly reproduce the vocalism of suchte 
 a,nd besuchte. The relation is the same, though combined with 
 metathesis, in work and wrought, which are O.E. ivyrc-ean and 
 toorh-te, 1 and so with others. 
 
 9. Lastly, when to these various phenomena is superadded 
 the effect of the well-known and earlier law called Pregermanic 
 compensatory lengthening, 2 we have the curious and rare 
 alternations of the type bring = bring-en and brought = brachte, 
 think = denken (dunken) and thought = dachte, upon which we 
 need not dwell further. 
 
 10. There remain a couple of English verbs, nearly alike, 
 teach and taught, catch and caught. O.E. tcec-ean, contrasted 
 with pf. tdh-te, shows in the present a regular metaphony : the 
 sound of the two vowels, which at first scarcely differed, 
 -acquired in time the present striking dissimilarity. This ex- 
 planation, however, cannot hold good for M.E. caugh-te, since 
 the M.E. vb. cacch-en is not Germanic, but Romance, borrowed 
 from the Picardian cachier = M.F. chacier (to hunt, to catch, 
 
 1 For the interchangeable guttural, see above 53 C. 
 
 2 Supra 24, where the applications are enumerated. 
 
TENSES. 315 
 
 now chasser) : it has been suggested that the perfect was 
 imitated from that of another verb of Germanic origin and 
 similar meaninsr, which became obsolete after the former had 
 been introduced. 1 Everywhere else analogy worked in a con- 
 trary direction; we have, for instance, E. reached from reach = 
 O.E. rcec-an, 2 and many like it. 
 
 On summing up resnlts, we find, not only that the forms 
 with dental suffix Avhich appear most irregular agree with the 
 general rule of weak perfect and participle accidence, but even 
 that these are the most regular and primitive, inasmuch as 
 their changes are due to the phonetic laws of the two languages, 
 whereas the specious regularity of others really proceeds from 
 later and artificial restoration. 
 
 § 2. Principle of formation. 
 
 (188) I. It must now be well understood that the English 
 and German weak perfect and participle are always and every- 
 where identical, and consequently identical with the Gothic 
 types sok-ida (he sought) and sok-ip-s (sought). The next 
 question is : what is this dental element ? to which I.-E. 
 formation does it correspond ? This question has already been 
 asked and answered in the case of the participle : it has been 
 stated that the Preg. suff. -icTa- is the same, both in form and 
 function, that appears in Sk. dam-itd-s (tamed), Gr. Sa/A-aro-s, 
 L. dom-itu-s, Go. ga-tam-ida, O.E. *ge-taim-d and E. tame-d, 
 O.H.G. gi-zem-it and G. ge-zahm-t; and this identity is too 
 universal to admit of contradiction or even doubt. 3 
 
 II. But the case is quite different in the perfect, though 
 outwardly it looks so like the participle. The first thought, 
 indeed, that occurs to the mind, namely the suggestion of a 
 derivation from the participle itself, proves at once unable to 
 stand the test, since, in a transitive verb, the perfect has 
 always a transitive, and the participle an intransitive meaning, 
 and consequently no precise relation either in meaning or 
 derivation can be perceived between them. But, setting aside 
 
 1 Skeat, Principles, II., p. 187. 
 
 2 G. reichen and erreichen are likewise weak verbs. 
 
 3 Cf. supra 78 and 90. 
 
316 ENGLISH AND GERMAN GBAMMAR. 
 
 this hasty induction, we are compelled to acknowledge that the 
 Germanic weak perfect occupies a strangely isolated position 
 in comparative grammar; for there is not a single I.-E. tongue,, 
 even of the Letto- Slavonic group, which stands so near in 
 many other respects, that can show anything at all related ;: 
 and, unless we recur to the absurd theory of a creation from 
 nothing, we must suppose Germanic to have somehow de- 
 veloped a previous I.-E. form in which the dental element had 
 nothing to do with deriving the stem of the perfect-tense. 
 
 Many intricate processes may have had such an effect. Here 
 we may venture to sketch in rough outline the most plausible 
 of all. In order to understand it, the reader must remember 
 that every I.-E. verb possessed two conjugational systems, the 
 so-called voices, which were differentiated by the nature of 
 their person-endings, and that the so-called middle voice,, 
 though it theoretically involved an action returning to the 
 subject, in many cases hardly differed from the active, except 
 for a vague and almost immaterial shade of meaning: thus,. 
 Sk. bhdrati and bhdrate stand equally well for " he carries," 
 Gr. !So) and dxovco, actives, have the middle futures tSo-fxac and 
 aKovcro-fxai, and the L. so-called deponents, imitor, morior, etc., 
 though passively conjugated, do not differ in meaning from the 
 active forms imito, *morio, etc. Nothing, therefore, is more 
 common, and nothing will seem more natural than the use of a 
 middle form in the conjugation of an active verb. 1 
 
 Let us then consider any Gr. vb., for instance rpicp-u} (to 
 feed), with normal root in the present : with the regular grada- 
 tions and endings of the active and middle perfect, it forms a 
 pf. active sg. 3 Te-Tpo<p-e (deflected root) and a pf . middle sg. 3 
 t £-9 pair-rat (reduced root). Let us now take in the same way 
 an I.-E. root, for instance shab (reduced), the deflected grade 
 of which will be skob : it needs but slight reflection to see that 
 the Pregermanic forms strictly correlative to the Gr. forms- 
 above will be pf. act. sg. 3 *ske-skop-e and pf. m. sg. 3 *ske- 
 
 1 Particularly in reference to the present remarks, the L. pf. is certainly, 
 in the main, a perfect with middle person-endings, and the anomalies it 
 exhibits, especially when contrasted with the Gr. pf., could by no means be 
 otherwise explained. 
 
TENSES. 317 
 
 skap-tai. Now, carrying this twofold formation into German, 
 we shall find it to correspond sound for sound with two actual 
 •G. forms which have already occurred, viz. the two perfects (er) 
 schuf and (er) schaffte, the former strong, the latter weak. 
 Hence, speaking generally, and omitting many accessory 
 elements, 1 we may lay down this main formula : the strong 
 and the weak perfect look as if they were, respectively, the 
 active and the middle perfect of the I.-E. conjugation. 
 
 But, if this were the case, ought they not both to coexist 
 side by side in any verb, as they actually coexist in Greek, and 
 do here coexist accidentally 2 in the verb schaffen, which, owing 
 to this circumstance, has afforded us an appropriate example ? 
 Certainly, and it will be seen, moreover, 3 that in fact they do 
 coexist in a larger number of Germanic verbs than we should at 
 first sight expect. In general, however, one of them has dis- 
 appeared as early as the Pregermanic period ; but the loss is 
 attributable to a principle of economy, the effects of which 
 are obvious in every language : the memory of the speaking 
 subject has no room for useless words ; two exactly synonymous 
 forms, either survive by assuming peculiar shades of meaning, 
 or else, if they remain exactly synonymous, one or the other 
 soon becomes obsolete. So it happened with the Germanic 
 perfect, either strong or weak : wherever a strong perfect was 
 contrasted in its vocalism with the general verbal form, clearly 
 enough to be able to play without confusion the part of past- 
 tense, there was no reason for a weak perfect being retained ; 
 elsewhere, its dental element, denoting the 3rd person of the 
 singular, became associated in the speaker's mind with the 
 notion of past time, and thus the form remained indispensable 
 and was secured from decay. 
 
 1 Restricted to these terms, the theory must needs appear superficial, 
 inasmuch as it leaves out of account some phonetic and grammatical diffi- 
 culties, the details of which cannot conveniently find a place in this work. 
 Should the reader wish to proceed further, he should refer to Collitz' article 
 in either Amer. Journ. of Philology, ix., p. 42, or Bezzenbergtr's Beitriige, 
 xvii., p. 227. 
 
 2 Of course, schaffte is not primitive, but has been formed on schaffrv, 
 after the analogy of some other type ; but the twofold pf. of this verb proved 
 convenient for the purpose of the present demonstration. 
 
 3 Under the Preterito-presents, infra 222. 
 
318 ENGLISH AND GERMAN GRAMMAR. 
 
 III. Our task is not yet over ; for, now that we are dealing 
 with a new and complete past-tense, the cause no longer 
 appears adequate to the effect. The Greek exponent -tcu is 
 well-known to characterize only the third person singular : 
 now, if the final syllable in er schaff-te be the same, how does it 
 come that German says, quite as correctly, ich schaff-te, or, in 
 other words, that a whole tense is conjugated on a syllable 
 which really represents one of its person-endings P Here we 
 come to one of the most curious, though by no means rare, 
 phenomena due to grammatical analogy, namely, a single 
 and isolated form giving birth to a whole system of forms. 
 Let us briefly state the series of analogical facts. 
 
 1. Owing to grammatical and phonetic peculiarities which 
 will be explained below, 1 the two Germanic forms of sg. 1 and 
 sg. 3 of the strong perfect had come to be identical : Go. bar 
 "I carried" and "he carried," nam "I took" and "he took," 
 etc. ; cf. E. I bare and he bare, G. ich nahm and er nahm, and 
 so forth, everywhere. 
 
 2. On the other hand, as noted above, the form of sg. 3 
 in the weak perfect was Go. sok-ida, cf. E. he sough-t and G. 
 er such-te, from vb. sok-jan = seek = suchen. 
 
 3. The identity of sg. 1 and 3 in the strong perfect quite 
 naturally led the speaker to use similarly sg. 3 of the weak 
 perfect for sg. 1, and to say Go. soh-ida " I sought," E. J sought, 
 G. ich suchte, — a bold transposition, indeed, though so simple in 
 appearance. 
 
 4. After these two forms have become levelled, the others 
 must follow : hence, each Germanic language, according to its 
 own genius and to the conjugational resources it is able to 
 display, constructs on the basis of sg. 1 and 3 a new perfect- 
 accidence (Go. sokida sokides, etc.), which, even in O.E. and 
 O.H.G., and still more so of course in Mod. German and English, 
 is entirely modelled on the person-endings of the strong perfect. 
 
 5. In this marvellous growth of a single person-ending into 
 a whole tense, and in this new tense again extending through- 
 oat almost all the verbs, whether present or future, of the 
 
 1 Under the Person-endings, infra 208-209. 
 
TENSES. 
 
 319 
 
 Germanic family, we ought never to forget that the analogical 
 influence of the "weak participle must have played a large and 
 decisive part. It has been already explained that the weak 
 perfect cannot have arisen from it ; but these two grammatical 
 categories must have always gone hand in hand, and, by their 
 fortuitous and exact likeness, have protected each other from 
 the risks they ran, inasmuch as every verb which had a par- 
 ticiple with a dental suffix would naturally show a propensity 
 to develop a perfect with the same suffix, and vice versa. Hence 
 the actual correspondence, established long ago, with but a few 
 insignificant exceptions, between these two formative currents 
 which, starting from far distant sources, soon flowed together 
 and ended by uniting in a single channel. 
 
 § 3. Applications. 
 
 (189) I. The formation of the weak perfect being well 
 understood, its applications will appear of the utmost simplicity, 
 the more so because, in the later languages, the progress of 
 analogy, together with the dulness of unaccented finals, greatly 
 contributed to simplify them. Gothic distinguishes mainly 
 three classes of weak verbs, according as the Pregermanic 
 derivative syllable in each was -ja-, or -0-, or -e-, 1 viz. : 
 
 1. Infinitive s6k-ja-n, pf. sg. 3 sOk-i-da, part. sOk-i-\>-s ; 
 
 2. ,, salb-6-n, „ salb-O-da, „ salb-6-\>-s ; 
 
 3. ,, hab-a-n, ,, hab-di-da, ,, hab-di-\>-s. 2 
 
 II. Old English has the tolerably regular correspondences 
 to these three types, viz. : 
 
 1. Infinitive sec-ea-n, pf. sg. 3 toh-te, part. ge-sOh-t ; 
 
 2. ,, sealf-ia-n, 3 „ sealf-o-de, „ ge-sealf-a-d; 
 
 3. „ habb-a-n, „ hcef-de „ ge-hcef-d. 
 
 III. Old High German is either preserved or restored re- 
 markably well, viz. : 
 
 1 Besides a 4th class, comprising some intransitive verbs, which does not 
 occur in West Germanic. Cf. supra 83 (IV.), 85, 92 and 93. 
 
 2 The derivative vowel of the 3rd class is seen to disguise itself in Gothic. 
 But it appears clearly in O.H.G. 
 
 3 The present-form visibly influenced by the analogy of the 1st class. 
 
320 ENGLISH AND GERMAN GRAMMAR. 
 
 1. Infinitive suoch-e-n, pf. sg. 3 suoh-ta, part. gi-suoch-i-t ; 
 
 2. „ salb-6-n, „ salb-0-ta, „ gi-salb-6t. 
 
 3. ,, hab-e-n, ,, hab-e-ta, 1 „ gi-hab-et. 
 
 TV. Now, let the reader contrast the present states : 
 
 1. E. seek sough-t sough-t = G. such-cn such-te ge-such-t ; 
 
 2. ,, salve salve-d salve d = „ salb-en salb-te ge-salb-t ; 
 
 3. ,, have ha-d ha-d = ,, hab-en hat-te ge-hab-t. 
 
 Provided he has borne in mind our previous observations 
 (§ 1), he will easily perceive how it happened that, apart from 
 apparent exceptions, the three historical types were blended 
 into one, which may be reduced to the following rules : 
 
 (a) The English weak perfect and participle are formed by 
 adding to the infinitive a suffix -d, which, even when written 
 -ed, does not constitute a syllable (saved pr. sevd, pushed pr. 
 pusb, etc.), except when the infinitive itself ends with a dental 
 explosive (shade shaded, melt melted, etc.) ; 
 
 (&) The German weak perfect and participle are formed 
 by substituting for the -en of the infinitive a dental element, 
 respectively -te and -t, spelled and sounded -ete and -et when 
 preceded by a dental explosive (geredet, gestattet) . 
 
 V. These rules, without exception, apply: (1) to primary 
 verbs which, as early as the Pregermanic period, were shifted 
 to the weak accidence, especially to those in I.-E. -sko-, E. 
 wis7i = O.E. wyscean, G. forschen = OJI.G. forscon, etc.; (2) to 
 primary verbs, much more numerous in English than in German, 
 which in Pregermanic still belonged to the strong conjugation, 
 but later passed to the weak 2 ; (3) to all secondary, tertiary, 
 etc., I.-E. verbs ; (4) to all verbs Avhich in English or German 
 have arisen from later derivation ; (5) and, of course, to any 
 late created verbs; (6) lastly, to all verbs in particular which 
 English has borrowed from French, and, generally speaking, 
 to every foreign element imported into English or German ; — 
 excluding only the few cases in which a verb of regularly weak 
 conjugation has been compelled, through outward likeness to 
 a strong verb, to obey the laws of strong accidence. 3 
 
 1 Earlier hap-ta, but analogy restored the unsyncopated form, whereupon 
 the syncope again took place in Mod. German. 
 
 2 The most important instances have been quoted under 179-185. 
 » Cf. supra 179 (5), 180 (II. 3), etc., etc. 
 
CHAPTER II. 
 MOODS. 
 
 (190) The Indo-European conjugation distinguished four 
 moods, Indicative, Subjunctive, Optative and Imperative, all 
 well preserved in Greek and Sanskrit. But the subjunctive 
 and the optative had a kindred meaning, which was often 
 indistinct and soon became confused : thus, the so-called Latin 
 subjunctive, as a matter of fact, is a mixture of forms borrowed 
 from both the original subjunctive and optative. 1 In Pre- 
 germanic, the whole Subjunctive is simply the original 
 Optative, with perhaps vague traces of the l.-E. subjunctive. 2 
 Now, as the indicative, from its nature, did not need the use 
 of any exponent, being the stem of the verb itself, the study of 
 Anglo-German moods is confined to two verbal forms, Sub- 
 junctive and Imperative. 
 
 Section I. 
 
 SUBJUNCTIVE. 
 
 (191) The two tenses of the Indicative, Present and Per- 
 fect, are both capable of being conjugated in the Subjunctive, 
 the form of which is derived for each, by means of a peculiar 
 suffix, from the corresponding form of the Indicative. 
 
 § 1. Present Subjunctive. 
 
 (192) I. Referring to the Greek optative present, which 
 accurately reproduces the original I.-E. accidence, we find it, — 
 
 1 Cf. infra 211, the subjunctive of the vb. "to be." 
 
 2 Immaterial for our present subject : cf. infra 192, 1. 
 
 321 „ 
 
322 ENGLISH AND GERMAN GRAMMAR. 
 
 that is, in the so-called vei'bs in -to, as there are no others 
 left in the Germanic conjugation, — to be formed by inserting 
 an element -i- between the stem in -o- and the person-endings. 
 Thus, a verb, for instance, the stem of which was I.-E. *bher-o-, 
 formed an optative stem *bher-o-i-, whence sg. 1 *bher-oy-m 
 (may I carry), sg. 2*bher-oy-s, sg. 3 *bher-oy-t, etc., as obviously 
 represented by the conjugation of the optative of the Gr. vb. 
 ve'/x-w (I divide), and by the almost identical one of the sub- 
 junctive of the Go. vb. nim-a (I take). 
 
 Gr. 
 
 Go. 
 Gr. 
 Go. 
 
 Sg. 1 V€fX-OL-fXl, 2 VefJL-OLS, 3 v£fJ.-OL, 
 
 „ 1 nim-au, 2 nim-di-s, 3 nim-di, 
 
 pi. 1 vefx.-oi-fx.cv, 2 ve/x-ot-T€, 3 v£fx.-oi-ev, 
 
 ,, 1 nim-di-ma, 2 nim-di-}>, 3 nim-di-na. 
 
 Apart from a slight difference in pi. 3, the agreement is 
 exact ; though, only in sg. 1, Gothic has a different diphthong ; 
 but this difficulty, however it may be got rid of in Gothic, 1 
 no longer exists in the later languages, where all the endings 
 have become uniformly obscured. Here, as far as we go back, 
 the whole accidence rests on an identical stem. 
 
 O.E. 
 
 sg. 1 nim-e? 2 nim-e, 3 nim-e, 
 
 O.H.G. : „ 1 nem-e, 2 nem-e-s, 3 n'em-e, 
 O.E. : pi. 1 nim-e-v, 2 nim-e-n, 3 nim-e-n, 
 O.H.G. : ,, 1 niim-e-m, 2 nem-e-t, 3 n'em-e-n. 
 
 The uniformity of the O.E. person-endings being for the 
 present neglected, 3 this single paradigm explains the whole 
 formation of the subjunctive in the modern languages; for it 
 is the same throughout every class of verbs, strong or weak. 
 It is true that the O.H.G. weak verbs of 2nd class still exhibit 
 their characteristic vowel before the endings : sg. 1 salb-6 or 
 scdb-o-e, 2 salb-o-s or salb-o-e-s, etc. But it does not matter 
 for our purpose, since, the final syllables being everywhere 
 obscured, the types nemes, fares, suoches, salbos, habes, etc., 
 
 1 It is probably not a diphthong at all, but the transliteration of a vowel : 
 in other words, nimau ought to be accented nimaii and pronounced nimS = 
 Gr. subj. ve/xw, a Preg. remnant of the I.-E. subjunctive. 
 
 2 Infinitive niman, with i for e before a nasal. 
 » Cf. infra 212 and 214. 
 
moods. 323 
 
 necessarily became blended together in nehmest, fahrest, suchest, 
 salbest, hahest, etc. 
 
 II. These main points being settled, the analysis of the 
 modern subjunctive present is very simple. German keeps it 
 almost unaltered, and English hardly less pure. 
 
 1. The Preg. diphthong ai, preserved in Gothic, has become 
 G. e, more or less distinctly sounded : sg. nehm-e nehm-e-st 
 nehm-e, pi. nehm-e-n nehm-e-t nehm-e-n; and so everywhere. 
 Since this diphthong could, at no stage of its phonetic evolu- 
 tion, have caused metaphony, and since it was the same for 
 every person, the accidence of the subjunctive always remains 
 free from metaphony, even in strong verbs characterized by the 
 regular metaphony of sg. 2-3 in the indicative : ind. nehm-e 
 nimm-st nimm-t, 1 but subj. nehm-e nehm-e-st nehm-e; ind. fahr-e 
 fdhr-st fahr-t = O.H.G. far-u fer-i-st fer-i-t, but subj. fahr-e 
 fahr-e-st fahr-e = O.H.G. far-e far-e-s far-e, etc. Hence an im- 
 portant rule of practical grammar is accounted for in every 
 detail : verbs of this kind distinguish the subjunctive, in so-. 
 2 and 3, by the absence of the metaphony of the indicative ; 
 the remainder keep it distinct only in sg. 3 by the absence 
 of the person-ending ; everywhere else, the present tense is 
 identical in the two moods, except for a slight difference in 
 pronunciation. 
 
 2. English not only obscured the final syllables, but also 
 dropped all the person-endings, as will be explained below, 
 both in the singular and plural. But, since the indicative itself 
 has kept no endings except in sg. 2 and 3, the present tense 
 does not differ in the two moods save in these two forms. 
 
 § 2. Perfect Subjunctive. 
 
 (193) The subjunctive of the perfect, which is often called 
 the imperfect subjunctive, is derived from the perfect indicative 
 by means of a suffix very like the preceding. I.-E. verbal forms 
 which do not end with the conjugation-vowel -0-, when con- 
 jugated in the optative, add to the stem a syllable which is -ye- 
 in the normal, and -?- in the reduced grade: Gr. e.lrjv = *icr-i7)-v 
 
 1 Cf. infra 204 and 205. 
 
324 ENGLISH AND GERMAN GRAMMAR. 
 
 (may I be), Ti-Oe-ty-v (may I place), pi. n-Oe-l-fiev, middle sg. 1 
 Ti-de-i-fxrjv, etc. The two grades originally alternated according 
 to a regular law, namely, the normal in the sg. of the active, 
 the reduced in the plural of the active and throughout the 
 middle voice. But in time the conjugation became gradually 
 levelled, so that, for instance, the vocalism of the plural was 
 transferred to the singular : thus, in Latin, the contemporaries 
 of Plautus still conjugated s-ie-m (may I be. cf. Gr. elrjv) s-ie-s 
 s-ie-t s-i-mus s-l-tis, whereas a century later analogy had pro- 
 duced the sg. s-i-m s-i-s s-i-t by extending the plural- vowel 
 throughout the whole tense. This uniform exponent -i- is 
 likewise, however far we go back, the general mood-exponent 
 in Pregermanic, 1 so that a Preg. pf. *fg-fall-e (he fell) would 
 show a correlative subj. *fe-fall-i-J> (might he fall). In the 
 application of this rule, strong and weak perfects must naturally 
 be considered apart. 
 
 (194) I. Strong Perfect. — 1. Let us start from the double 
 conjugational stem of the perfect in the Go. vb. nim-an, thus 
 nam- and nem-, and observe, at the outset, that the subjunctive- 
 exponent -I- is always added to the reduced stem of the plural, 
 and never to the deflected stem of the singular. 2 The conju- 
 gation of the mood thus obtained will be as follows : 
 
 Go. sg. 1 nem-jau, 2 nem-ei-s, 3 nem-i, 
 ,, pi. 1 nem-ei-ma, 2 nem-ei-J>, 3 nem-ei-na. 
 
 In each of these forms, excluding the first, 3 we recognise at 
 once the suffix -1-, though regularly shortened as final in sg. 3. 
 
 2. It would be quite superfluous' here to give an O.E. 
 paradigm, 4 for it would lead to no knowledge of the modern 
 accidence, beyond the fact that, at this early period, and for 
 similar reasons, the subjunctive was as poor in endings in the 
 
 1 Cf. infra 211. 
 
 2 Because the I.-E. suffix -ye-, being accented, caused the stem to lose its 
 accent and consequently to be reduced. Compare, for instance, Sk. va-vdrt-a 
 (he turned), optative va-vrt-yd-t ; Gr. rt-Ovy-K-e (he is dead), opt. re-Ova-ir) : 
 L. root es (to be), subj. s-ie-t s-i-t, etc. Again, in the plural, the accent 
 was shifted to the person-ending : whence the suffix -ye- became in its turn 
 reduced to -i-. 
 
 8 Which of course is modelled on sg. 1 of the present subjunctive. 
 4 It will occur below (211). 
 
moods. 325 
 
 perfect as in the present : sg. -e throughout, pi. -en throughout. 
 The whole evolution of the B. subj. pf. is contained in the 
 statement of two very simple analogical processes. 
 
 (a) As shown by the Go. and even the Mod. G. conjugation, 
 the radical vowel of the perfect tended to suffer, and actually 
 suffered metaphony before the i of the following syllable. 
 Of this regular metaphony, however, O.E. itself exhibits but 
 scanty traces : it was already effaced almost everywhere by the 
 analogy of the indicative, so that the subjunctive of the perfect 
 had no other stem than that of the indicative plural, viz. : 
 infinitive bind-an, pf. sg. 3 bond, pf. pi. 3 bund-on, subj. sg. 3 
 bund-e (might he bind), pi. 3 bund-en, etc. 
 
 (b) Now, it happened, as seen above, that the E. perfect no 
 longer exhibited two distinct stems, but became entirely con- 
 jugated on one of them, thus either he bound like they bound, 
 or they drank like he drank. 1 Then, the uniform stem of the 
 indicative became analogically the stem of the subjunctive. 
 The only perfect which keeps the double stem in the indicative, 
 he was, they were, is likewise the only one which, according to 
 the historical rule, shows throughout the subjunctive, he were, 
 the plural-stem of the indicative. No other English subjunctive 
 perfect differs from the corresponding indicative, except in the 
 loss of the person-ending in the 2nd person singular. 
 
 3. German is far purer, and it will be worth while to quote 
 here the paradigm in which either Old or Modern German 
 agrees sound for sound with Gothic : 
 
 O.H.G. : sg. 1 ndm-i, 2 ndm-i-s, 3 ndm-i, 
 
 G. : „ 1 nahm-e, 2 nahm-e-st, 3 nah-me, 
 
 O.H.G. : pi. 1 ndm-i-m, 2 ndm-i-t, 3 ndm-i-n, 
 
 G. : ,, 1 niihm-e-n, 2 nahm-e-t, 3 ndhm-e-n. 
 
 Hence we should expect to find, in the G. subjunctive per- 
 fect, the radical vowel which regularly ought to appear in the 
 plural of the perfect indicative, though of course metaphonically 
 altered by the influence of the following i. Let us now carry 
 the application of the principle throughout the seven classes of 
 strong verbs. 
 
 1 Cf. snprn 177, 3, and the applications, 179 sq. 
 
326 ENGLISH AND GERMAN GBAMMAE. 
 
 A. The vowel of the pf. pi. is not capable of metaphony : so 
 we have, without any difficulty, trieb-e, biss-e, etc. 
 
 B. The vowel of the pf. pi. is u, and it is therefore this 
 vowel we find, for instance, in the pf. snbj. of the vb. kios-an, 
 in the common expressions O.H.G. ni curi and ni curit, translat- 
 ing respectively L. noil and nollte. 1 But, as the vowel-grade 
 o of the sg. finally prevailed in the whole indicative, it was 
 thence transported, with metaphony, to the subjunctive, and 
 consequently the modern formation is zbg-e (M.H.G. zilg-e), 
 fror-e, verlor-e, krbch-e, s'off-e, etc. 
 
 C. In the sub-type trinken, we should expect trilnk-e (M.H.G.) ; 
 but, here also, the stem having once become uniform in the 
 indicative formed the subjunctive with its ordinary vowel, 
 usually a (metaph. trdnk-e, fiinde, zwdnge), occasionally o 
 (metaph. klomm-e, and even begonn-e in spite of begann). The 
 u is either preserved or restored in schiind-e, bedung-e. — In the 
 sub-type helfen, the u is, as a rule, unaltered, and wiird-e, stiirbe, 
 verdiirbe, wiirbe, wiirfe are retained, in spite of starb, warf 
 (though wiirfe exists also). When corrupted, the vowel hesi- 
 tates between a and o, under the influence of the participle 
 (befdhle and befdhle, gdlte and golte). 2 
 
 D. The perfect-vowel being a, the types ndhm-e, kdm-e, 
 gebar-e, brach-e, etc., are quite regular; but verbs which have 
 adopted o in the indicative, naturally have 6 in the subjunctive, 
 schcren schor schbre. 
 
 E. Regular ware, gab-e, dsz-e, without difficulty. 
 
 F. Regular schliig-e, trilg-e, fiihr-e, without difficulty ; but, 
 owing to stand, the regular stilnd-e also admits of a doublet 
 stand-e. 
 
 G. The vowel being incapable of metaphony, we have fiel-e, 
 hielt-e, liesz-e, lief-e, rief-e, etc., without any modification. 
 
 (195) II. Weak Perfect. — The Gothic formation rests on 
 the same basis as the plural of the indicative, and both are 
 
 1 Literally " choose not to [do this or that] ," whence " do not," a general 
 formula of prohibition : cf. the actual use of E. choose. — The regular form 
 ■ni churls, which is also met with, had lost its final -s, under the influence of 
 sg. 2 of the imperative, as being used here in imperatival function. 
 
 * Cf. supra 181, II. 3 b. 
 
moods. 327 
 
 rather difficult to explain. 1 But, fortunately, it is a point we 
 need not deal with, because West Germanic has modelled its 
 subjunctive perfect, as also the plural of the indicative, 2 entirely 
 on the singular forms of the indicative; so that, upon the 
 whole, the result is the same as if the I of the subjunctive 
 merely replaced the a of sg. 2 and 3 of the indicative. 
 
 1. Consequently, owing to the general levelling of the person- 
 endings, which partly took place even in O.E., the English 
 subjunctive bears to the indicative exactly the same relation 
 in the weak perfect as in the strong perfect. 
 
 2. O.H.G. for the three classes of weak verbs has respectively 
 suoh-t-i, salb-d-t-i, and hab-e-t-i: in the last two, the radical 
 vowel, not being brought into contact with the i, could not 
 undergo metaphony, and in fact did not do so ; in the first, 
 however, it could undergo it, but the analogy of the other two 
 tended to impede or efface it. Hence Mod. German scarcely 
 shows it, 3 except in the so-called mixed verbs, where the out- 
 ward likeness to the strong verbs caused it to be kept : O.H.G. 
 ddhta and G. dachte (he thought), but subj. O.H.G. dahti and 
 G. dachte. So also : brachte brdchte, that thate, next kennte, 
 brennte, nennte, rennte, sendete, wendete, and the archaic dduchte ; 
 in the preterito-presents, 4 konnte, wiiszte, etc. ; lastly, hatte, 
 which certainly proceeds from the analogy of the parallel 
 auxiliary ware. Everywhere else, the subjunctive of the per- . 
 feet, both stem and endings, is identical with the indicative. 
 
 Section II. 
 
 IMPERATIVE. 
 
 (196) The Imperative Mood, in Germanic, appears ex- 
 clusively in the Present Tense; though, of course, owing 
 to its nature, the mood necessarily involves some idea of future 
 time, either deferred or immediate. 
 
 1 Go. suk-ided-um (we sought) and sok-ided-jau (might I seek). Cf. infra 
 233. 
 
 s Cf. supra 188, III. 4, and infra 210. 
 
 8 M.H.G. has many more instances of it : mechte, bedrechte, opposed to 
 indicative machte, bedrachte " betrachtete." 
 
 * See below 222 sq. 
 
328 ENGLISH AND GERMAN GRAMMAR. 
 
 The I.-E. imperative, generally speaking, is the simplest 
 possible form of the verb ; and, in particular, the 2nd 
 person sg. of this mood, which is the gennine and properly so- 
 called imperatival form, in such verbs at least as are charac- 
 ized by the interchangeable vowel -e- / -o-, is simply the pure 
 and bare verbal stem, always ending in -e-, and without any 
 personal termination : I.-E. *bher-e (carry), *nem-e (take), 
 *wegh-e (convey) ; Sk. bhdr-a, Gr. ve/x-e, L. veh-e, etc. Hence 
 the Germanic imperatives of strong verbs, quite regularly, are 
 Go. bair, nim, drigk, Huh (draw), and so everywhere. In the 
 weak verbs we have sok-ei ( = I.-E. *sag-ye), salb-5, hab-di, 
 wherein the final long vowel, being first shortened and then 
 obscured in the later languages, produced O.H.G. suoch-i salb-o 
 hab-e, G. such-e salb-e hab-e, and lastly the Mod. E. form. 
 
 The bare stem of the verb, altogether free from the meta- 
 phony we are to meet with in German, 1 and with no ending 
 whatever, is what we find without exception in the English 
 imperative of verbs either strong or weak : — drive, drink, bear, 
 give, eat, — seek, salve, have, — so that the mood requires no further 
 observation. 
 
 (197) German, on the contrary, as being less affected by 
 phonetic and analogical decay, here retains a primitive distinc- 
 tion, though not between properly strong and weak verbs, but 
 rather between two different classes of strong verbs. For its 
 imperatival formation is well-known to be twofold : in such 
 strong verbs as still retain a vowel e in the radical syllable of 
 the indicative, the imperative is usually formed by metaphony 
 and without adding any final vowel, nehm-en nimm, geben 
 gieb, essen isz ; whereas, in every other verb, strong or weak, 
 the imperative ends in -e and the stem is not metaphonized, 
 treib-e, flieg-e, trink-e, fahr-e, lauf-e, etc. 2 Let us inquire 
 whether the double process is traceable to an original variation. 
 
 "We must first remember that I.-E. final e had become Preg. 
 1 : thus, to the I.-E. imperatives quoted above, would cor- 
 
 1 The metaphony has been eliminated in the same way as the metaphony 
 of sg. 2 and 3 of the indicative, infra 204. 
 
 2 Of course, this short unaccented vowel may afterwards be dropped in 
 rapid pronunciation. 
 
MOODS. 
 
 329 
 
 respond such Preg. types as *nemi (take), *geb-2 (give), *fdr-i 
 (travel), *hlaup-i (run), etc. On the other hand, the meta- 
 phony of e has been seen to be far older, having taken place in 
 Pregermanic, whereas the metaphony of a is much later, and 
 still more so that of other vowels: hence the first two forms 
 had become Preg. *nim-i, *%ih-i, while *far-i and *hlaup-i re- 
 mained unaltered ; the result, after the loss of the final vowel, 
 was O.H.Gr. nim, gip, bat far, hlouf, etc. Shortly, no strong 
 verb had a final vowel in the imperative ; in those which had 
 an 8 in their stem, this mood was differentiated by a stem with 
 metaphonical % ; in the remainder, it had no sign at all : such 
 was the ancient and regular state. 
 
 Now, the weak verbs, on the contrary, as seen above, showed 
 in the imperative a final vowel, which in M.H.Gr. had become 
 a uniform and unaccented e : to the infinitives suchen, salben, 
 haben, were opposed the imperatives suche, salbe, habe, and the 
 speech only obeyed a logical and obvious tendency, when it 
 restored likewise fahre, laufe, as opposed to fahren, laufen, and 
 so also trage, trinke, treibe, ziehe, etc. ; the more so, because 
 there existed some strong verbs which seemed to exhibit no 
 metaphony in the imperative for the reason that their indica- 
 tive already had it, sitze, liege, hebe, etc. 1 Thence it followed 
 necessarily that final -e became added anew to the imperative, 
 either of verbs in which the imperative was originally like the 
 indicative, or of such as later and analogically had effaced the 
 difference : thus, schere, gebare, 2 rache (cf . the regular brich, 
 sprich), even dresche in spite of er drischt, and loerde in spite of 
 er wird. Here the final vowel is found throughout. 
 
 But verbs which had inherited from Pregermanic the meta- 
 phony of the imperative, naturally held aloof from the weak 
 verbs with invariable stem and remained free from their in- 
 fluence : wherefore we find them retaining the genuine forma- 
 tion in types C (2nd sub-type, hilf, wirf, stirb, erlisch), D 
 (nimm, brich, sprich, stiehl), and E (isz, gib, tritt, vergisz). The 
 duality in the German imperative may be deemed a very con- 
 sistent and important survival. 3 
 
 1 Cf. *upra 183 and 184 in fine. 
 
 2 Cf. infra the metaphony in the present (206 D). 
 
 8 The verbs that still keep it are thirty-five in number. 
 
CHAPTER III. 
 PEESON-ENDINGS. 
 
 (198) The Anglo-German Person-Endings constitute a 
 very consistent system, which comprises all verbs, strong or 
 weak, without distinction, though it does not apply, apparently 
 at least, to a few conjugations of anomalous appearance. The 
 latter must be reserved to be dealt with separately. We begin 
 by considering the ordinary conjugation of both languages. 
 
 Section I. 
 
 ORDINARY CONJUGATION. 
 
 (199) The I.-E. language distinguished four classes of 
 person-endings, which have become more or less altered and 
 blended together in its offspring, namely : — primary or pre- 
 sent-endings, which in Germanic are those of the indicative 
 present; — secondary endings, those of the so-called Gr. 
 augmented tenses, which are also those of the optative, and 
 therefore, in Germanic, characterize the subjunctive, either 
 present or perfect 1 ; — perfect-endings, in Germanic (indi- 
 cative) as everywhere else; — lastly, imperative-endings. 
 Though they are much mutilated in the later languages, a 
 thorough analysis will enable us to recognise them. 
 
 § 1. Present-Endings. 
 
 (200) I. Primitive Forms.— Let us conjugate, in the 
 present indicative, the I.-E. stem *bher-o- (to carry), together 
 
 1 Hence it will prove advisable to study them only after those of the 
 perfect indicative. 
 
 330 
 
PEBSON-ENDINGS. 
 
 331 
 
 with the corresponding forms in 
 
 the oldest 
 
 tongues of our 
 
 family. 
 
 
 
 
 I.-E. 
 
 8k. 
 
 Gr. 
 
 L. 
 
 Sg. 1 *bher-o 
 
 bhdr-d-mi 
 
 <p€p-(D 
 
 fer-o 
 
 2 *bher-e-si 
 
 bhdr-a-si 
 
 ^)€/OCtS 1 
 
 fer-(i-)s 3 
 
 3 *bher-e-ti 
 
 bhdr-a-ti 
 
 <f>epei 1 
 
 fer-(i-)t* 
 
 PL 1 *bher-o-mes 
 
 bhar-d-mas 
 
 <pip-0-[JL(V 
 
 fer-i-mus 
 
 2 *bher-e-te 
 
 bhdr-a-tha 
 
 cptp-z-re 
 
 fer-(i-)tis z 
 
 3 *bher-o-nti 
 
 bhdr-a-nti 
 
 <pep-o-VTL 2 
 
 fer-u-nt. 
 
 Two main facts are to be observed at the outset. 
 
 First, -we note the alternation, as stated above, 4 in the vowel 
 which precedes the termination, being o in sg. and pi. 1 and in 
 pi. 3, e in the three other persons. The I.-E. languages fully 
 agree as to this point, with the single exception of Latin in pi. 
 I, 5 and this curious variation, which probably proceeds from 
 the effect of primitive accentuation, will be seen to take place 
 in the Germanic group as everywhere else. 
 
 Next, we recognise the striking likeness of the person-end- 
 ings, with but insignificant differences between one language 
 and another. 6 That of sg. 1 is the only one which remains 
 obscure. It has been suggested that the stem *bher-o- could 
 not have become lengthened to *bher-o, unless a vocalic ending 
 had united with the final stem-vowel. But what was this 
 vocalic ending ? Perhaps an -a, as we shall find it in sg. 1 of 
 the perfect. 
 
 Further, it must be borne in mind that the I.-E. language 
 
 1 These two forms are corrupted, since they ought to be *<pep- e-<n and 
 *<pep-e-n. But Gr. shows elsewhere the ending -n of sg. 3 : Doric dl-Su-n 
 (he gives), Attic Si 8u-<ri (because Attic changes t to a before t). Cf. Henry, 
 Gramm. of Gr. and Lat., no. 249. 
 
 2 Doric ; but Attic (pkpovai = <pepov<ri. 
 
 8 Lest the unity of the L. paradigm should appear to be broken, we have 
 conjugated ferd like lego : the real forms are of course fers fert fertis ; but 
 this type is well known to be a unique exception. 
 
 4 Supra 82 (II.), and cf. Henry, Gramm. of Gr. and Lat., no. 269. 
 
 5 "We should expect to find *fer-u-mm. But we have s-il-mus, vol-u-mw>, 
 quas-u-mvs, and, in a totally different category of words (superlative opti- 
 mu8 = optumux), Latin shows a similar, probably phonetic, hesitation between 
 u and I before an m. Then, too, leg-i-mus was furthered by the analogy of 
 leg-i-tix. 
 
 6 All attributable to primitive doublets, infra 203, 1. 
 
332 ENGLISH AND GERMAN GRAMMAR. 
 
 had another termination for sg. 1, in verbs which did not form 
 their stem by means of the interchangeable vowel -e- / -o- (Gr. 
 so-called verbs in -/At). This exponent of sg. 1 was -mt, which 
 will occur below, 1 and even presently in O.H.G., though it 
 is best seen in Greek: ?o-7->7/ai (I place) = *ai-o-Ta.-fj.i, rt-drj-fu, 
 8t-8w-yu<., etc. But this ending, although important, only con- 
 cerns us here as completing the series of forms -si and -tt, and 
 in later comparison with the corresponding secondary termina- 
 tion -m. For the Germanic ordinary conjugation consists merely 
 of verbs which in Greek grammar would be called verbs in -w, 
 and here we have only to deal with the Germanic ordinary 
 conjugation. 
 
 (201) II. Pregermanic Forms. — Translating now, as it 
 were, into Pregermanic the T.-E. paradigm conjugated above, 
 we obtain a new series of forms, which, apart from unimportant 
 details such as may be reserved for further examination, is 
 rigorously verified in the accidence of the earliest and latest 
 Germanic languages. 
 
 Preg. Go. 
 
 Sg. 1 *ber-o bair-a 
 
 2 *t>er-e-si bair-i-s 
 
 3 *her-e-J?X bair-i-p 
 PI. 1 *ber-a-mes bair-a-m 
 
 2 *her-e-J?e bair-i-J> 
 
 3 *her-a-njji bair-a-nd 
 
 Let us set apart the obvious metaphony in the O.E. and 
 O.H.G. sg, and the O.E. uniform plural-ending, as both re- 
 quiring a paragraph for themselves. Considering only the 
 bare outline of the present-accidence, we find it as easy to go 
 back to Indo-European by starting from the earliest Germanic 
 tongues, as to descend from the latter to the contemporary 
 languages. 
 
 (202) III. Modern Forms. — A. Singular. — 1. The final 
 
 1 Infra 217, I. Cf. sg. 1 -m as a secondary ending, infra 211. 
 
 2 The dialects differ, but the principle of conjugation is the same through- 
 out : the type quoted here is the accidence in Wessex, except sg. 2 birest, as 
 also birist concurrently with biris in O.H.G. infra 202, 2. 
 
 O.E. 
 
 O.H.G. 
 
 ber-e 
 
 bir-ti 
 
 bir-e-s 
 
 bir-i-s 
 
 bir-e-ft 
 
 bir-i-t 
 
 ber-a-S 
 
 ber-a-vies 
 
 ber-a-ft 
 
 ber-e-t 
 
 ber-a-tS 2 
 
 ber-a-nt 
 
PEKSON-ENDINGS. 333 
 
 -5 has been shortened, and then variously obscured, down to 
 the dull vowel -e, which again is dropped in English, 1 as also 
 in German when swiftly sounded : gebdr-e, nehm-e, trink-e, 
 fahr-e, etc. In the weak verbs we have .likewise : Go. sok-ja, 
 salb-o, hab-a ; O.E. sec-e, sealf-ie, hcebb-e, whence E. seek, salve, 
 have; O.H.G. suoch-u, salb-o-m salb-o-n, hab-e-m, 2 whence G. 
 such-e, salb-e, hab-e : all the forms having become identical. 
 
 2. Final I being dropped, there remained -s, an ending well 
 preserved in Gothic, but already somewhat obsolete in the 
 sister-tongues, though it occasionally survives even in M.H.G. : 
 nim-i-s (thou takest), fer-i-s (thou travellest), etc. It will be 
 seen below that the regular Germanic ending of sg. 2 in the 
 perfect was -t, and that such perfects as retained it unaltered 
 possessed a decided sense of present-time 3 : from this arose the 
 analogical influence of the perfect on the present-accidence ; 
 for, since a West Germanic speaker would say, for instance, 
 *wais-t "thou knowest," *kan-t and even ultimately *kan-s-t 
 "thou canst," it was only natural that he should say likewise 
 *bir-i-st "thou carriest " instead of *bir-i-s. Yet another cir- 
 cumstance must be taken into account : the nominative of the 
 pronoun of sg. 2 *t>u was often added to the verbal form of sg. 
 2, whence proceeded such syntactical complexes as *biris tu,* 
 O.H.G. biris tu, which the speaker resolved unconsciously into 
 birist du, E. bearst (thou), G. gebierst (du). Thus the ending 
 -st prevailed, in both languages, either with or without syncope 
 of the preceding vowel : E. bear-e-st, drive-st, give-st, take-st, 
 drink-e-st ; G. gebier-st (gebdr-st), nimm-st, trink-st, fahr-st. 5 — 
 The weak verbs require no special remarks. 
 
 1 Even when written : there is no final e in drive, give, shake, take, any 
 more than in b*ar, drink, full, etc. 
 
 2 Hence O.H.G., purer in this respect even than Gothic, retains the I.-E. 
 ending -mi, which here is quite justified, the verbs of these two classes being 
 stems which did not end in -e- / -o-, supra 85 and 92-93. But M.H.G. effaced 
 the distinction and consequently lost the final m (n) of sg. 1. 
 
 8 Cf. infra 208, 209, and 222. 
 
 4 Do not forget that the s preceding a t impedes the consonantal shifting. 
 
 5 The syncope was once even more energetic than it is now, cf. supra 187 : 
 German, for instance, has restored trdgst trdgt after trage ; but M.H.G. 
 conjugated trage treist trcit, a syncope which is still kept in getreide (corn) = 
 O.H.G. gi-treg-idi (product). Compare moreover th e syncopa ted conjugation 
 in E. have hast hath (has) and G. habe hast haj^^^\^ (.IBK/^T"-^ /' 
 
 f OrTHF "r* 
 
 (UNIVERSITY 
 
 V nr 
 
334 ENGLISH AND GERMAN GRAMMAR. 
 
 3. The Preg. ending -J>X ought to remain -pi after an accented 
 syllable, and become Si after an unaccented syllable (Verner's 
 law). Now, some I.-E. verbs in -o- were accented on the root, 
 whereas others were accented on the suffix : so that the two 
 terminations -Si and -Jji must have coexisted in Pregermanic. 
 But, as already seen above in many other instances, 1 gram- 
 matical analogy led to a separate extension of each ending in 
 the two later languages : thus, the latter in O.E. became -6"; 
 in O.H.G. the former became -t : E. bear-e-th, give-th, etc. ; G. 
 gebier-t (gebar-t), nimm-t, fahr-t, irink-t, etc. German retains 
 this state ; not so English, though the ending -th is still current 
 in scriptural and poetical style. Ordinary speech replaces it 
 by an ending s, which however is quite modern : bears, gives, 
 drinlc-s, etc. The th and the s both being sibilant sounds, it 
 was easy to pass from one to the other; but this cannot be 
 deemed merely a question of phonetics, since a similar change 
 did not take place anywhere else. It must have arisen from 
 the s of sg. 2, which was extended by an unconscious desire to 
 systematise and, as it were, imposed its sound upon the th 
 of sg. 3 : in other words, because the speaker said givest, he 
 took to saying likewise gives instead of giveth. 2 
 
 (203) B. Plural. — 1. I.-E. seems to have had several ex- 
 ponents for pi. 1, viz. : -tries, retained only in O.H.G. ; -mes, 
 Doric -fits; -mos, L. -mils; simple -me, Gr. -/ac(v) and Go. -m, 
 an ending originally limited to secondary tenses, but often 
 ultimately extended to the present, as may be seen from Gr. 
 and Go. Thus the latter has everywhere a simple -m : bair-a-m, 
 nim-a-m, etc. O.H.G. hesitates between the two terminations, 
 that is to say, having primitively -mes for the present (ber-a- 
 mes) and -m for the secondary tenses, it extended -m into the 
 present (b'er-a-m), and -mes into the secondary tenses ; but the 
 latter form is antiquated and of no interest except from a his- 
 torical point of view. 3 Thus, O.H.G. still possesses nem-a-mes, 
 
 1 Supra 63, 124 (I.), 139, and infra 203 (2-3), 212. 
 
 2 Northumbrian at a rather early date had changed final th to s, so that 
 here a dialectal mixture is possible. 
 
 8 It has probably proceeded from the very rare cases in which the ending 
 of pi. 1 became attached to the root without any intermediate, thus Preg. 
 'xtd-mes (we stand), *&~>-m&s (we go), etc. ; for this position is the only one 
 
PERSON-ENDINGS. 335 
 
 nem-e-mes l and nem-e-m, zioh-e-mes and zioh-e-n, far-e-m and 
 far-e-n ; even M.H.G., nem-e-mes and nem-e-n 2 ; bat the latter 
 type ultimately prevails, and Mod. G. knows no other form but 
 gebar-e-n, nehm-e-n, zieh-e-n, fahr-e-n, etc. The same is the 
 case with the weak verbs. 
 
 2. The termination -fie was naturally treated like -fit, whence 
 -fie after an accented, and -6"£ after an unaccented syllable : the 
 Go. -fi, of course, corresponds to both at once ; the general ex- 
 ponent of the O.E. pi. comes from -fie, whereas the G. exponent 
 of pi. 2 represents Preg. -&"£, either with or without syncope : 
 gebar-e-t and gebdr-t, nehm-e-t 3 and nehm-t, trink-t, fahr-e-t, etc. 
 
 3. The ending -nfii likewise becomes either -nfii or -nfit : 
 Gothic has either ; O.E. has -nfii represented by -(«)&"; O.H.G. 
 has -nd"t shifted to -nt, O.H.G. ber-a-nt, n'em-a-nt, zioh-e-nt, 4 
 far-e-nt, M.H.G. n'em-a-nt and nem-e-nt, and so forth, without 
 exception. But it will be seen that in the perfect and subjunc- 
 tive the termination, quite regularly, was simply -n : wherefore 
 during the latest period of M.H.G., nement analogically became 
 riemen (on the pattern of namen " they took "), so that pi. 3 is 
 the same as pi. 1, G. gebar-e-n, nehm-e-n, fahr-e-n, etc. 5 
 
 § 2. Present-Metaphony. 
 (204) Our examination has been restricted as yet to the 
 endings, without any regard to the modification which might 
 
 that allows the ending to be accented in I.-E., cf. Sk. i-mds (we go) = Gr. 
 t-fiev ; and, if it had been unaccented, it could not have preserved its final s. 
 
 1 Owing to an encroachment of the vowel of pi. 2 nem-e-t. Cf. the similar 
 process in Latin, supra 200, note 5. 
 
 2 Observe moreover an optional apocope of final n when the pronoun sub- 
 ject is placed after the verb : M.H.G. neme wir (we take). 
 
 3 M.H.G. (Alamannic) nem-e-nt, pi. 2 having strangely enough adopted 
 the form of pi. 3 ; still in High Alsatian er name " you take " like si name 
 " they take." But O.H.G. and regular M.H.G. nem-e-t. 
 
 * The same vocalic corruption as in pi. 1. 
 
 5 Let us here, for the sake of completeness, give the paradigm of the three 
 weak verbs, which in the modern languages have reached the same dull 
 vocalic state : 
 
 Go. 
 
 sok-ja 
 
 sOk-ei-s 
 
 s6k-ei-\> 
 
 sok-ja-m 
 
 s6k-ei-\> 
 
 sOk-ja-nd = 
 
 O.H.G. 
 
 suoch-u 
 
 suoch-i-s 
 
 suoch-i-t 
 
 suoch-e-mes 
 
 suoch-e-t 
 
 suoch-ent ; 
 
 Go. 
 
 salb-0 
 
 salb-o-s 
 
 salb-0-\> 
 
 salb-O-m 
 
 salb-0-]> 
 
 salb-0-7id = 
 
 O.H.G. 
 
 salb-0-n 
 
 salb-0 s(t) 
 
 salb-O-t 
 
 salb-0-mes 
 
 salb-0-t 
 
 salb-O-nt ; 
 
 Go. 
 
 Jiab-a 
 
 hab-di-s 
 
 hab-di-\> 
 
 hab-a-m 
 
 hab-di-]> 
 
 hab-and = 
 
 O.H.G. 
 
 hab-e-n 
 
 hab-e-$(t) 
 
 hab-e-t 
 
 hab-e-mes 
 
 hab-e-t 
 
 hab-e-nt. 
 
336 ENGLISH AND GERMAN GRAMMAR. 
 
 affect the verbal stem itself. Now the fact is that, while all 
 the English verbs and the German weak verbs conjugate their 
 present indicative on the same invariable stem throughout, the 
 German strong verbs, in general, inasmuch as their radical 
 vowel is capable of metaphony, undergo it in the forms of 
 Sg. 2 and 3, all other persons remaining free from it : nehm-e 
 nimm-st nimm-t nehm-e-n, etc. ; fahr-e fdhr-st fahr-t fahr-e-n, etc. 
 This peculiarity of G. accidence we have now to explain. 
 
 Referring to Pregermanic, we find the existence of three 
 person-endings capable of producing metaphony, that is to say, 
 containing an i, namely, sg. 2 and 3, and pi. 3. 
 
 The last, of course, is out of question, for the vowel which 
 precedes it is I.-E. o, whence Preg. a, in short, a vowel which 
 could not undergo a Preg. metaphony. In other words, such a 
 type as *nem-a-nb"i *far-a-n}>i remained unaltered in Preger- 
 manic, since e was the only vowel capable of metaphony in 
 this early period ; whereas later, in O.H.G., when the a in its 
 turn became liable to similar mutation, the i which might pro- 
 duce it had long vanished, thus nem-a-n-t far-e-nt, in which 
 metaphony, for another reason, was equally impossible. 
 
 Secondly, English also may be neglected, though in its earlier 
 state it exhibits the metaphony precisely under the same con- 
 ditions as German 1 ; but no trace of it survives any more than 
 in the imperative, analogy having extended throughout every 
 verb a uniform vocalism : O.E. sg. 1 freos-e ceos-e, sg. 3 fries-S 
 cies-d~, but E. sg. 1 freeze choose, sg. 3 freeze-th choose-th, without 
 any mutation from one person to another. 2 
 
 Let us now analyse the forms of sg. 2 and 3 in German 
 strong verbs, and then try to account for the fact that the same 
 forms never undergo metaphony in the weak verbs. 
 
 (205) I. Strong Verbs. — 1. Considering in Pregermanic 
 
 1 See the O.E. paradigm, supra 201 ; many other similar instances could 
 be mentioned. 
 
 2 E. ee being the true representative of O.E. to, there can be no metaphony 
 in freeze as opposed to chooxe, though their vowels are different. Hence the 
 00 in choose (Chaucer has still cheese), shoot, etc., probably proceeds from the 
 influence of the preceding consonant. Cf. however E. £ose = O.E. leosan, 
 where such a reason cannot be given. — For the loss of the metaphony in 
 these types, cf. the same phenomenon in the similar G. kiesen, frieren, etc., 
 infra 206 B. 
 
TERSON-ENDINGS. 337 
 
 any strong verb, but preferably, for convenience of demonstra- 
 stration, a verb containing a radical a (type F), such as *far- 
 ana-m (to travel), we know it, as stated and proved above, to 
 have been conjugated, in the present indicative, *far-o *far-e'-s% 
 *far-e-J>Z *far-a-me *far-e-fie *far-a-nf>t. Now, such an acci- 
 dence could not persist ; for, even as early as Pregermanic, the 
 forms which contained an e followed by an i x namely sg. 2 and 
 3, metaphonized the e and became *fdr-i-sX *fdr-i-]n. Then, 
 in the dialects, final t vanishes, whence Go. fctr-i-s far-i-)> and 
 O.H.G. *far-i-s *far-i-t. But the two latter forms again be- 
 come impossible in O.H.G., as soon as in this tongue the a in 
 its turn undergoes metaphony, since, before the i which pre- 
 cedes the ending, the radical a must become e, whence finally 
 sg. 2 fer-i-s (fer-i-st) and sg. 3 fer-i-t, as opposed to sg. 1 far-u. 
 
 Let us now pass to the case when the Preg. root contaiued 
 an I.-E. e. Here the phenomenon will appear at once earlier 
 and more intricate. Though at first sight the two metaphonical 
 processes in far-an fer-i-s fer-i-t and n'em-an nim-i-s nim-it might 
 seem rigorously parallel, this view is incorrect, since the meta- 
 phony of e is known to be far earlier than that of a : hence, it 
 is not to O.H.G., but to Pregermanic itself, that we ought to 
 trace back the vocalic mutation of the latter type : in other 
 words, we have the series *nem-e-si *nem-e-(fi, then *nem-i-st 
 *nem-i-&i, and lastly *nim-i-si *nXm-X-8X, the whole already in 
 the Pregermanic period. So, on one side, the type nimis nimit 
 refers us to a much earlier Germanic state than the type feris- 
 ferit is able to do, though of course both rest on the same- 
 principle. 1 
 
 And, on the other hand, the former type shows a wider his- 
 torical extension than the latter, because in O.H.G. an e is 
 metaphonized under the influence, not only of an i, but also of 
 a following u, and the form of sg. 1 ends with a -« : so that, 
 O.H.G. far-u from far-an being kept unaltered, O.H.G. *nem-u 
 from n'em-an cannot but become nim-u, and so also bir-u, gib-u, 
 ijju, etc. 
 
 1 This is also the reason why the metaphony in the imperative nimm is 
 an isolated form, contrasted with the non-metaphonical imperative fahr-e, 
 supra 197. 
 
 Z 
 
33S ENGLISH AND GERMAN GRAMMAR. 
 
 Consequently, O.H.Gr. had a strong present-accidence with 
 metaphony in sg. 2 and 3, faru ferist ferit, which is correctly 
 reproduced in Mod. G. fahre fiihrst fiihrt, and another strong 
 present-accidence with metaphony throughout the whole sin- 
 gular, nimu nimist nimit, which is only partly reproduced in 
 Mod. G. nehme nimmst nimmt. How is it that the modern 
 speaker no longer says ich *nimme l ? The answer is obvious, 
 the analogical process here being as clear as possible : a general 
 levelling has effaced the metaphony of sg. 1. For, in the type 
 fahren, sg. 1 fahre being free from metaphony in contrast to sg. 
 2 and 3 fiihrst fiihrt, and similar in this respect to the corre- 
 sponding form pi. 1 fahren, it was but natural that, in the type 
 nehmen, there should be restored a sg. 1 nehme, contrasted with 
 sg. 2 and 3 nimmst nimmt and similar to pi. 1 nehmen. Even a 
 mathematical proportion could not have a more exact result. 
 
 Hence, through a kind of compromise between two types of 
 strong verbs, German has come to its present rule, namely : 
 apart from the rather rare cases in which the vocalism has been 
 entirely levelled by the power of analogy, 2 every strong verb, 
 provided its radical vowel be capable of metaphony, exhibits a 
 metaphony of the root in sg. 2 and 3 exclusively. 
 
 (206) 2. Now let us pi-oceed to the applications of the rule 
 throughout the various types of strong verbs. 
 
 A. The radical vowel (!) is incapable of metaphony. 
 
 B. In O.H.G. we find the regular forms, sg. ziuh-u ziuh-i-s 
 ziuh-i-t (metaph.), and pi. zioh-e-mes zioh-e-t zioh-e-nt (non- 
 metaph.) ; but Mod. G. has extended the non-metaphonical 
 form, and we have now zieh-e zieh-st zieh-t zieh-e-n, etc., 3 and so 
 in the whole class ; apart, however, from such slight traces 
 •of metaphony as may still be found, at least in archaic and 
 poetical language, in the verbs fliehen, fliegen, flieszen, gieszen, 
 schlieszen, kriechenf etc., but of course only in sg. 2 and 3. In 
 
 1 The present is still in M.H.G. nime nimes nimet ; and down to the pre- 
 sent day, in High Alsatian, the regular contrast exists between t nem (I take) 
 and mr ndviS (we take). 
 
 2 A*corruption often found in the dialects : Alamannic, for instance, con- 
 jugates er fahrt, er schlagt, er halt (however er treit " he carries"), etc. 
 
 8 The metaphonized stem would be *zeuh-, supra 29. 
 4 Was da kreucht xm&fleugt. Schiller, W. Tell, III. 12. 
 
PERSON-ENDINGS. 339 
 
 one of the four verbs the stem of which exhibits au = u, we find 
 an analogical metaphony, viz. sauf-e sduf-st silnf-t sauf-e-n. 
 
 C. The sub-type trinken does not here concern us. Bat the 
 snb-type helfen is quite consistent in metaphony : O.H.G. hilf-u 
 hilf-i-s hilf-i-t, stirb-u stirb-i-s stirb-i-t ; G. helf-e hilf-st hilf-t, 
 sterb-e stirb-st stirb-t; further, schwillt', quillt, befiehlt, erlischt, 
 and, the two dentals being blended together, after the syncope, 
 hirst for *birst-i-t, wird for *wirdit, 1 etc. This violent though 
 regular syncope has been replaced by an ending -t, restored 
 anew and preceded by an e, whenever sg. 3 happens not to bo 
 metaphonic : thus, binden, gestatten are conjugated er bindet, er 
 gestattet, because the syncopated forms er *bint, er *gestatt are 
 unable to show the exponent of sg. 3 hidden within them. But, 
 if the person is characterized clearly enough by the metaphony, 
 then the restoration is omitted as useless, and the syncope 
 retained: er gilt, er schilt, er ficht (from fecht- en), etc. 2 
 
 D. The metaphony is constant : O.H.G. biru biris birit, nimu 
 nimis nimit; G. gebdre gebierst gebiert (also gebdrst gebiirt), 3 
 nehme nimmst nimmt ; and likewise, bricht, spricht, trifft, even 
 kommt owing to a quite modern analogy. It is effaced, either 
 partly or entirely, in gebdrt, kommt, schert (scheert) and pflegt. 
 
 E. The metaphony is constant: O.H.G. sihu sihis sihit, gibu 
 gibis gibit; G. sehe siehst sieht, gebe gibst gibt ; and likewise, iszt, 
 miszt, liest ; putting aside, of course, such verbs as already have 
 the metaphony throughout their entire accidence (bitten, sitzen, 
 liegen) . 
 
 F. Here, as seen above, the metaphony is identical in O.H.G. 
 and Mod. G. : faru ferisferit, and fahre fdhrstfdhrt; and like- 
 wise, schlagt, trdgt, icdchst, etc. But laden, once a weak verb, 
 did not adopt it, and schaffen has lost it. It stands to reason 
 that heben and schwuren cannot show it at all, since their radical 
 vowel is metaphonical even in the infinitive. 
 
 1 Cf. supra 187, the syncopes in the weak perfect. 
 
 2 Even in sg. 2 in wirst m 'tcird-i-st. And also in the other classes of strong 
 verbs: er tritt = *trit-i-t, er halt = helt-i-t, etc. 
 
 3 The M.H.G. vb. w'egen = O.H.G. wegan = 'E. to weigh (weak vb., cf. also 
 G. wegen in be-wegen " to move "), being conjugated M.H.G. wige xcige*t 
 iciget wegen, thus gave birth to two synonymous verbs, respectively iccigen 
 (regular with a = e) and wiegen (metaph.). 
 
340 ENGLISH AND GERMAN GRAMMAR. 
 
 G. All the verbs of this type, excluding only hauen, which 
 have a radical a, require the metaphony : fdllt, halt, laszt, liiuft. 
 Indeed, it has been extended to stoszt as above to kommt. But 
 heiszen could not undergo it, and rufen remained free from it. 
 
 (207) II. Weak Verbs.— The first class of weak verbs 
 being characterized by an infinitive in -jan (Go. nas-jan " to 
 cure," 1 solc-jan " to seek), its radical vowel is, constantly and 
 in every form either of the sg. or the pi., followed by a syllable 
 which was able to cause metaphony : now, supposing this to 
 have taken place in earlier days, either in Pregermanic as in 
 sitzen, or at least in O.H.G. as in heben, — thus, for instance, 
 Mod. G. nahren (to feed) = Go. nas-jan, — then, of course, the 
 verb is metaphonized everywhere, and cannot appear particu- 
 larly metaphonized in sg. 2 and 3 ; the reverse being the case, 
 as in O.H.G. suochan, now suchen, 2 then the verb can be meta- 
 phonized nowhere, for it has no more reason for undergoing 
 metaphony in sg. 2 and 3 than in the rest of its accidence. A 
 type suochit (he seeks) could not undergo it in O.H.G., since uo 
 is not yet capable of it ; and, in M.H.G., when uo in its turn 
 became subject to it, the type was already altered to suochet and 
 no longer contained an i. Hence it is obvious that, in any case, 
 verbs of this class cannot exhibit in sg. 2 and 3 of their present 
 indicative a vocalism in any way different from that of their 
 whole accidence. 
 
 Still more so with the other classes ; for, in Preg. *saVo-6-b~l 
 *hab-e-d~i, the % is not yet brought into contact with a vowel 
 capable of metaphony, and in West Germanic the % is dropped 
 altogether. 
 
 Thus we have historically accounted for an important peculi- 
 arity which differentiates strong and weak verbs, no less than 
 the difference in the formation of their perfects and participles. 
 
 1 The causative of the. strong verb which is now Mod. G. ge-nes-en " 
 recover." 
 
 2 O.H.G. suochan is (regularly) a non-metaphonical form, whereas O.E. 
 secean (E. seek) has been seen to be (no less regularly) a metaphonical one. 
 
 8 With this main restriction, that, in M.H.G., as also in the dialects, the 
 metaphony has often, analogically, spread into the weak verb: there are 
 such conjugations as mache machst mcicht, sage sdgst sagt (the latter still 
 surviving in High Alsatian) ; but the literary language has proscribed these 
 forms, keeping only one of them, namely frfigst (askest) friigt (also fragt). 
 
PEKS0N-END1NGS. 341 
 
 § 9. Perfect-Endings. 
 
 (208) It must be understood that the perfect-endings have 
 no primitive character except in the strong perfect, which how- 
 ever has given its accidence to the weak one. Now, starting 
 from the Sanskrit conjugation, 1 e.g. the Sk. pf. ved-a (I wot), 
 
 sg. 1 ved-a, 2 vet-tha, 3 ved-a, pi. 1 vid-md, 2 vid-d, 3 vidur, 
 we may state the I.-E. endings of this tense as follows : 
 sg. 1 -a, 2 -tha, 3 -e, pi. 1 -me ; 
 
 the terminations of pi. 2 and 3 being of no interest in Germanic, 
 because there they were borrowed from another series. 
 
 (209) I. Strong Perfect.— Choosing, for the sake of 
 greater simplicity, such a perfect as, in the Pregermanic period, 
 had already completely lost the vowel-gradation from sg. to 
 pi., we see that its prehistorical and historical endings were as 
 follows : 
 
 Preg. Go. 
 
 Sg. 1 *sloh-a sloh 
 
 2 *sloh-ta sloh-t 
 
 3 *sloh-e sloh 
 PI. 1 *sld%-ume 2 sloh-um 
 
 2 *slo%-ud"e sloh-uf> 
 
 3 *slo£-iinJ> sloh-un 
 
 Sg. 1 and 3, after their final vowel had regularly disappeared, 
 became identical : O.E. slog and E. (J, he) slew drove chose drank 
 bound burst bare spoke gave was fell, etc. ; O.H.G. sluog and G. 
 (ich, er) schlug trieb erkor trank band barst gebar sprach gab 
 warfiel, etc. This is the reason why sg. 3 of the perfect never 
 has any ending, whereas sg. 3 in the present always shows E. 
 -th (-*) and G. -t. 
 
 Sg. 2 in West Germanic is quite a different form from what 
 it was in I.-E. and even in Gothic : properly speaking, it does 
 not belong to the perfect, but to the primitive aorist, 4 though 
 
 1 Controlled and corrected by the testimony of Greek, cf. the accidence of 
 oT5-a (I wot), viz. old-a olo-Oa. olS-e IS-fiev la-re ta-avrt. 
 
 2 With the regular " grammatical alternation," which everywhere else has 
 become variously levelled. 
 
 8 Also sluog-umes, supra 203, 1. 
 
 * Thus, for instance, sg. 1 zuh (I drew) and sg. 2 zug-i (thou drewest) bear 
 
 O.E. 
 
 O.H.G. 
 
 slog 
 
 sluog 
 
 sldg-e 
 
 sluog-i 
 
 slog 
 
 sluog 
 
 slog-on 
 
 sluog -urn 3 
 
 slog-on 
 
 sluog-ut 
 
 slog-on 
 
 sluog-un 
 
342 ENGLISH AND GERMAN GRAMMAR. 
 
 here confounded with the perfect. From this however it does 
 not follow that West Germanic knew nothing of a perfect-end- 
 ing sg. 2 -t, which in Preg. and Go. represents the I.-E. -tha ; 
 for we have already met with it and shall find it again in the 
 preterito-presents. 1 Next, the analogy of this eventual survival 
 of the ending -£,. together with the development of a new termi- 
 nation -st in the present, actually. influenced the perfect, and 
 introduced into it, separately though similarly in both the later 
 tongues, the same ending sg. 2 -st: while M.H.G. still conju- 
 gates nam ncem-e ( = O.H.G. ndm-i), Mod. G. has now nahm 
 nahm-st nahm, and so forth everywhere, E. slew-est like slay-est, 
 G. schlug-st like schlag-st, etc. 2 
 
 In pi. 1 the original ending is -me, in which, after a conso- 
 nant, the initial m easily develops its proper vowel, thus -mme, 
 whence Preg. -time and Go. -um. Later on, final m becomes n, 
 as may already be seen in the O.E. and even the O.H.G. acci- 
 dence (also sluog-uri) : further E. slexo, and G. schlug-en, so that 
 pi. 1 and 3 are as completely blended together as is the case 
 with sg. 1 and 3. 
 
 Like Greek (l<t-tc, Ae-Aour-a-Tc), Pregermanic borrows the 
 ending of pi. 2 from the series of the primary or secondary 
 endings, I.-E. -te, Preg. -}>e. The u developed by a phonetic 
 process in pi. 1 and 3 was analogically transferred to pi. 2, 
 whence a Gothic final syllable -uj> (-«&"), O.H.G. -ut, G. (ihr) 
 schlug-et, etc. 3 
 
 Greek borrows its ending of pi. 3 from the primary series, 
 whence -ntl ; Pregermanic borrows it from the secondary series, 
 whence merely -n}>, as in the subjunctive, and consequently, 
 with the vowel-sound n, an ending -un in the later languages 
 
 to one another exactly the same relation as Gr. (pf.) irt-favy-a (I have fled) 
 to (aor.) t-<pvy-e-s (thou fledst). Cf. moreover supra 172 and note. 
 
 1 Supra 53 A and infra 222 sq. 
 
 2 But the ancient aoristic form did not disappear without leaving some 
 traces of its existence : the complete assimilation, either in German or — 
 even more — in English, of the perfect subjunctive to the perfect indicative 
 certainly started from this form of sg. 2, in which both moods were charac- 
 terized at once by a termination -i and the reduced grade of the root. 
 
 3 The uniformity of the E. pi. endings has been reserved for further 
 examination : infra 214. 
 
PEBSON-ENDINGS. 
 
 343 
 
 which drop the final consonant. Nothing, therefore, is more 
 regular than the O.E. and O.H.G. contrast between slog-on 
 = sluog-un (never *sluog-unt) " they slew," and sle-a-ft = 
 slah-e-nt " they slay." But the modern present-accidence has 
 been seen to efface this original difference. 1 
 
 (210) II. "Weak Perfect. — The weak perfect, as stated 
 above, 2 probably proceeds from a single form of sg. 3 : to this 
 form again, West Germanic applied the endings of the strong 
 perfect, so exactly indeed, that it will prove sufficient to 
 quote the modern accidence without entering into any his- 
 torical details : 
 
 E. sought sought-est sought sought sought sought 
 G. suchte suchte-st suchte suchte-n suchte-t suchte-n 
 
 And so, likewise, E. salved salved-st, had had-st, G. salbte salbte-st, 
 hatte hatte-st, without either exceptions or difficulty. 3 
 
 § 4. Subjunctive-Endings. 
 
 (211) The so-called secondary I.-E. endings, which are 
 those of the I.-E. optative and, consequently, of the Preg. sub- 
 junctive, merely differ from the primary endings by dropping 
 the final % wherever it occurs, that is in sg. 1, 2, 3, and pi. 3. 4 
 Hence we may conceive the two series to run parallel as 
 follows : 
 
 Primary : sg. 1 -ml, 2 -si, 3 -ti ; pi. 1 -mes, 2 -te, 3 -nti ; 
 Secondary : „ 1 -m, 2 -s, 3 -t ; „ 1 -me, 2 -te, 3 -nt. 
 
 As Latin has not only kept, but even transferred to its 
 primary tenses, the secondary endings, it affords us the best 
 means of restoring them. Let us compare, in form and acci- 
 dence, the subjunctive (optative) of the root es (to be) : 
 
 1 Cf. supra 203, 3, and ivfra 211. 
 
 2 Supra 188, II. and III. 
 
 3 Gothic conjugates : sukida sdkides sOkida sdkidedum sukidedu\> sukidedun ; 
 salbdda talbodes, etc. ; habdida, etc. 
 
 4 Compare, for instance, Gr. pres. \6ov<ti = \ij-o-vti with impf. f\vov = 
 *t\v-o-vr. The relation is exactly the same in Germanic. 
 
344 ENGLISH AND GERMAN GRAMMAR. 
 
 Lat. : s-l-m s-i-s s-i-t s-i-mus s-l-tis 1 s-i-nt ; 
 O.H.G. : s-l s-i-s s-i s-i-m z s-i-t s-i-n; 
 
 Mod. G. : set sei-est sei sei-en sei-et sex-en. 
 
 It is indeed a marvellous thing to find so striking an agree- 
 ment, at a distance of so many centuries, between Italic and 
 Germanic. 
 
 From a glance at this table, the reader will derive more 
 profit than from any commentary, provided he has borne in 
 mind the essential laws of Germanic phonetics. For, seeing 
 that final m and t were equally dropped in Pregermanic, he will 
 find that sg. 1 and 3 have become identical even in this period, 
 as they are now* The same loss of final t in pi. 3 caused pi. 1 and 
 3 to be blended together, as soon as O.H.G. pi. 1 slm had been 
 changed to sin. Lastly, the substitution of sg. 2 slst for sis is 
 a simple and well-known process. Now let us turn to the 
 paradigms, ancient or modern, of both tenses in the subjunctive. 
 
 Present. Perfect. 
 
 
 Go. 
 
 O.E. 
 
 O.H.G. 
 
 Go. 
 
 O.E. O.H.G. 
 
 Sg 
 
 ;. 1. slih-au 
 
 bind-e 4 
 
 slah-e 
 
 s J 6h-jau 
 
 slOg-e sluog-i 
 
 
 2. slah-di-s 
 
 bind-e 
 
 slah-e-s 
 
 yloh-ei-s 
 
 $lOg-e sluog-i-s 
 
 
 3. slah-di 
 
 bind-e 
 
 slah-e 
 
 sloh-i 
 
 slOg-e sluog-i 
 
 r: 
 
 . 1. slah-di-ma 3 
 
 bind-e-n 
 
 slah-e-n 
 
 sloh-ei-ma 8 
 
 slog-e-n sluog-i-n 
 
 
 2. slah-di-\> 
 
 bind-e-n 
 
 slah-e-t 
 
 sloh-ei-\> 
 
 slog-e-n tluog-i t 
 
 
 3. slah-di-na 3 
 
 bind-e-n 
 
 slah-e-n 
 
 sloh-ei-na 8 
 
 slog-e-n sluog-i-n 
 
 
 Whence : 
 
 
 
 
 
 E. 
 
 slay slay 
 
 slay 
 
 slay 
 
 slay 
 
 slay, 
 
 G. 
 
 ichlxj-e schlag 
 
 ■e-st schlag-e schlag-e-n schlag-e-t schlag-e-n ; 
 
 E. 
 
 slew slew 
 
 slew 
 
 slew 
 
 slew 
 
 sleio, 
 
 G. 
 
 schlii^-i schliig- 
 
 c- t schliig-e schliig-e-n schliig 
 
 •e-t schliig-en. 
 
 (212) There is but one point left for examination : namely, 
 besides the uniformity in the O.E. pi., a question as yet re- 
 
 1 This is the only point of divergence in L., for we should expect *si-te ; 
 but -tis is borrowed from the indicative, and L. has lost -te everywhere but 
 in the imperative. 
 
 2 Also s-i-mes, which resembles still more the L. form. 
 
 8 These endings with an added vowel de not concern us, since West 
 Germanic knows nothing of them. 
 
 4 It proved convenient to replace here the vb. sle-an by the vb. bind-an, 
 because the former conceals its person-endings, its subjunctive being con- 
 tracted to sle. 
 
PERSON-ENDINGS. 34B 
 
 served, we find here a similar uniformity in the O.E. sg. ; in 
 other words, E. sg. 2 lacks the ending like all the other 
 persons. Bat this last difficulty must appear small indeed to 
 any one who remembers that a primitive final s, according to 
 the place of the accent, ought either to remain s, or to become 
 a and then disappear : it is the latter that took place in O.E., 
 with but this additional circumstance, that the cases in which 
 the ending had become -z analogically encroached upon those 
 in which the -s had been kept, so that the ending vanished 
 altogether, whereas German retained the -s (st) everywhere 
 after the analogy of the forms which preserved it quite regu- 
 larly. 1 Hence, the E. subjunctive is a new instance of a 
 general treatment applicable to every form with final -s, as 
 often stated above, and particularly in the contrast between 
 the German and English plural of declined nouns. 2 And thus 
 Modern English has acquired a subjunctive invariable for plural 
 or singular, present or perfect. 
 
 § 5. Imperative- Endings. 
 
 (213) English and German have lost all the original forms 
 of the imperative, with the exception of sg. 2 and pi. 2. The 
 remainder is simply borrowed from the subjunctive (G. nehme 
 er, nehmen sie), or else replaced by a periphrastic locution. 
 
 The form of sg. 2 being the stem of the mood, 3 our study is 
 confined to pi. 2, in which the ending is I.-E. -te, Preg. -pe or 
 -6"e: I.-E. *bher-e-te (carry), Sk. bhdr-a-ta, Gr. <f»ep-€-T€ vi/x-e-re, 
 L. leg-i-te dic-i-te, etc. ; Go. bair-i-J> (carry), nim-i-J? (take) ; 
 O.E. bind-a-5 (bind), help-a-ft (help), bidd-a-S (pray), etc. ; 
 O.H.G. nem-e-t (take), zioh-e-t (draw), far-e-t (travel), and G. 
 nehmet, ziehet, fahret, etc., etc., with the occasional syncope of 
 the unaccented e. 
 
 No remark is here needed, except that, since the syllables 
 
 1 As a matter of fact, the -s generally follows an unaccented syllable in 
 the optative of the present (Sk. vdrt-e-s), and an accented syllable in the 
 optative of the perfect (Sk. va-vrt-yd-s), so that in English the present may 
 be said to have influenced the perfect, and vice versa in German. 
 
 2 Cf. supra 63, 139 and 143, and also 202, 3, etc. 
 8 Supra 196 and 197. 
 
346 ENGLISH AND GEEMAN GKAMMAB. 
 
 which follow the stem are never of such a kind as to require it 
 to undergo metaphony, the well-known contrast between nirnm- 
 and nehmet is in strict accordance with the history of the 
 German language from its beginning. 
 
 In the weak verbs we find in the same way : Go. sok-ei-p- 
 salb-o-Jj hab-ai-Jj ; O.H.G. suoch-e-t salb-o-t hab-e-t ; G. suchet 
 salbet habet. 
 
 But in English there is no longer any difference between pL 
 2-and sg. 2, that is to say, here, as everywhere else, the plural 
 has altogether lost its endings. We must now direct special 
 attention to this strange peculiarity. 
 
 § 6. The English Verbal Plural. 
 
 (214) I. Having followed up to this point the inferences 
 which may be drawn, either from the original form of the 
 person-endings, or from their actual state in Gothic and 
 German, the reader would expect to find, in the O.E. plural,, 
 five series of regular endings, thus : 
 
 (1) Present indicative : *bind-a-n 1 *bind-e-ft bind-a-b" 
 ( = *bind-a-n& 2 ) ; -whereas we find only bind-a-b" in the three 
 persons, and in English, also in the three persons, the curtailed 
 bind : 
 
 (2) Perfect indicative : bund-on *bund-otS bund-on ; whereas 
 we have bund-on everywhere, and in English bound ; 
 
 (3) Present subjunctive: bind-e-n *bind-e-b" bind-e-n; but 
 actually bind-e-n everywhere, and in English bind ; 
 
 (4) Perfect subjunctive : bund-e-n *bund-e-& bund-e-n ; but 
 actually bund-e-n everywhere, and E. bound. 
 
 (5) Imperative *bind-e-b~, the real O.E. form being nearly 
 alike (bind-a-6~), but in E. merely bind. 
 
 II. How did such a curious levelling take place ? and how 
 could it happen that an accidence comprising various endings 
 became uniform in O.E., and completely lost these endings in 
 English ? Of course, phonetic change by itself is unable to 
 
 1 More accurately *bind-a-m, and so likewise in the other tenses ; but 
 remember that the final nasal is of an undecided character, supra 29. 
 
 2 With O.E. compensatory lengthening (*vpra 20, 4 A) and later shorten- 
 ing in an unaccented syllable (supra 65 and 66). 
 
PERSON-ENDINGS. 347 
 
 account for the fact : it must at least have been furthered by- 
 several analogical processes, which, plain as they are, would 
 nevertheless require a rather tedious explanation, were we to- 
 descend to minute particulars. But they may be briefly 
 summed up as follows. 
 
 1. The starting-point of the change is to be found in the two- 
 tenses of the subjunctive, where, as explained above, the three 
 forms of the sg. had become the same : since the sg. was in- 
 differently bind-e in the present and bund-e in the perfect for 
 all three persons, and since, moreover, the plural was indiffer- 
 ently bind-e-n and bimd-e-n in the same tenses for two of their- 
 three persons (1st and 3rd), it would seem only natural to say 
 likewise bind-e-n and bund-e-n in pi. 2. Here, the process is so 
 clear, that it may even be shown in a kind of mathematical 
 proportion, thus: (ge) binden: (we) binden=(d'u) binde : (ic) 
 binde, and (ge) bunden: (we) bunden=(tSu) bunde : (ic) bunde*. 
 There can be hardly any scientific problem which admits of a 
 more satisfactory solution. 
 
 2. It must be observed, however, that — as our formula clearly 
 suggests — the process is necessarily in strict connection with a 
 new phenomenon of so-called grammatical analytism, namely, 
 the concurrent use of the personal pronoun as a subject to the- 
 verb, a pleonastic locution still unfamiliar to Gothic. Thus r 
 for instance, as long as Old French conjugated its present 
 indicative aim aim-e-s aim-e-t ( = L. dm-o cim-a-s dm-a-t), 
 distinctly sounding all the vowels and consonants, there was no- 
 reason for prefixing to the verb the pronouns je tu il ; but, as- 
 soon as the speaker began to prefix them, there was no longer 
 any reason for distinctly sounding the terminations. Hence,, 
 the two facts of uniformity in the endings and analytism in. 
 conjugation are seen to go hand in hand and to further one- 
 another. If, quite early indeed, Old English was thus allowed 
 to level its accidence, it is because, also quite early, the use of 
 a pronoun subject had secured it from ambiguity. 
 
 3. The latter point being well understood, nothing could 
 prevent the uniform ending in -n, when fixed in the whole 
 subjunctive, from spreading also in the perfect indicative. It 
 is true that here the three forms of the singular wore not 
 
348 ENGLISH AND GEBMAN GRAMMAR. 
 
 identical ; but two of them were alike, and, above all, the 
 perfect indicative and the perfect subjunctive l'esembled each 
 other very nearly; and, since moreover two persons of the pi. 
 indicative (1st and 3rd) had phonetically become the same, the 
 identical form bunden in the pi. subjunctive logically led to a 
 similar identical form bundon in the pi. indicative. 
 
 4. In the pi. of the present indicative, though the three 
 forms were different, two of them at least, *binded" and bindaft, 
 could easily be confounded : both became bindab", which passed 
 also to the imperative. This being done, an analogical formula, 
 such as (we) bindab": (ge) bindab— (we) bundon: (ge) bundon, 
 was suggested, the more so as O.E. still preserved a pi. 1 im- 
 perative bindan, and it seemed convenient to distinguish " we 
 bind" from "let us bind." Consequently, here, the levelling 
 took place at the expense of the 1st person, which became 
 assimilated to the two others. 
 
 5. Let us now return to the endings in n. Final E. n, it is 
 well known, easily disappears. Even as early as O.E. it may 
 disappear from the terminations of the verbal plural when 
 followed by the pronoun subject : in other words, the speaker 
 ought to say we bundon and ge bundon, but he may say either 
 bundon we or bunde we, either bundon ge or bunde ge, etc. 1 From 
 this to the uniform E. type, pf. ind. and subj. bound, and pres. 
 subj. bind, the transition is merely phonetic. 
 
 6. This apocope, being regular as to final n, was extended, by 
 an almost uncontrollable analogy, to the final 6" of the present 
 indicative, whence, likewise optionally, binde we and binde ge. 
 Next, M.E. extended the use of this accidence. 
 
 7. The curtailed forms having become decidedly prevalent, 
 Mod. E. naturally knows no other form but bind in the plural 
 of the present, as also in pi. 2 of the imperative. 
 
 8. The same process having run parallel throughout the 
 whole, strong or weak, accidence, the final result is, that the E. 
 plural, in any tense or mood, has lost every characteristic 
 except its invariable and amorphous state. 
 
 1 A similar phenomenon has been stated in M.H.G., supra 203, 1. — 
 Observe, moreover, that the analogical ending -en everywhere, even in the 
 present indicative, spread out very early in the Mercian dialect ; further, 
 that English is an offspring of Mercian, not of Saxonic. 
 
PERSON-ENDINGS. 349 
 
 Section II. 
 
 ANOMALOUS CONJUGATIONS. 
 
 (215) The apparent anomalies of the Germanic conjugation 
 comprise: (1) the verb " to be," which has retained, almost 
 unaltered, its ancient accidence as a root-stem, that is to say, 
 a conjugation both more primitive and simpler than that of the 
 stems in -g-j-o-; (2) a few other verbs, which exhibit at 
 least some traces of the same accidence ; (3) the false pre- 
 sents, which in reality are genuine perfects and, for this 
 reason, are termed preterito-presents. 
 
 § 1. The Verb " to be." 
 
 (216) It has been already stated 1 that a great many I. -E. 
 verbs could be conjugated without inserting the interchange- 
 able vowel -e- 1 -0- between the root and the endings, so that 
 the latter were immediately added to the bare root. This is 
 the well-known class of verbs which in Greek Grammar are 
 properly called verbs in -/u, thus, el- pi (I am) =*ia--fii, el-pi (I 
 am going), rl-O^-pi (1 set), etc. It is still tolerably numerous 
 in Greek as well as in Sanskrit. But, in Latin, it was confined 
 to a single verb, namely sum (I am), and the same is nearly 
 the case with the Germanic languages. 
 
 In one point, however, the Germanic accidence is more 
 intricate: it is not a single one, but confounds several 
 verbs in one, in such a way that the verb " to be " is conju- 
 gated concurrently on three verbal roots, which moreover 
 are not the same even in the same persons and tenses of such 
 cognate languages as English and German, though all three 
 are found, with similar or different functions and various shades 
 of meaning, in every other group of the I.-E. family • namely : 
 — &, the proper root for the general sense " to be," cf. Sk. 
 ds-ti (he is), Gr. cV-ti, L. es-t, Russian jes-ti, etc. ; — bhew " to 
 become," cf. Sk. bhdv-a-ti (he becomes, he is), Gr. (f>v-e-Tai (he 
 sjrows), 7rc-<£v-K€ (he is grown, he is), L. fu-i-t (he was), Russian 
 
 1 Supra 81, and cf. 92. 
 
350 ENGLISH AND GERMAN GRAMMAR. 
 
 hu-du (I stall be), etc. ;— toes " to dwell," cf. Sk. vds-a-ti (he 
 dwells, he finds himself), Gr. eWa = *F ccr-Tt-a (hearth), L. Ves-ta 
 (hearth-goddess), Go. vis-a (I sojourn, I remain), etc. 
 
 As early even as Gothic, these three roots possess the common 
 meaning " to be," the third, however, with an accessory shade 
 of durative and permanent state. In Anglo-German they have 
 become entirely synonymous : English shows a decided prefer- 
 ence for the second, and German, for the first ; but both borrow 
 their perfect- tense from the third. 
 
 (217) I. Present indicative. — A. English. — Sg. 1 : O.E. eom 
 earn am = Go. i-m 1 = I.-E. *es-mt (Sk. ds-mi, Gr. ei/u = *eo--/u, 
 Russian jes-mi, etc.); E. am. 2: O.E. eart (broken) =*er-t, 
 which represents Go. w=I.-E. *esi *es-sX (Sk. asi, Gr. Homer. 
 icr-cri and class. eT=*co-i, Russian jesi, etc.), with the -later and 
 analogical addition of the ending -t of sg. 2, a process we have 
 already met with; E. art. 3 : O.E. is = h. es-t, with secondary 
 termination, whereas Go. is-t is equivalent to Gr. kr-n, with 
 primary termination 2 ; E. is. — PI. 1, 2, 3 : O.E. s-ind, owing to 
 a general extension of the form of pi. 3, as already seen in every 
 other present indicative, this form here being I.-E. *s-enti, 3 cf. 
 Sk. sdnti, L. sunt ( = *s-6nti), Go. sind, etc. But Mod. E. has 
 simply are, with the form of sg. 2 curtailed of its ending as 
 extended to the plural. 4 
 
 B. German. — Sg. 1: O.H.G. bi-m, replacing *6itt-m = I.-E. 
 *bhew-mi : whence bin, M.H.G. bin, G. bin. 2 : O.H.G. bi-st, 
 replacing *bi-s, the termination -st, as explained above, having 
 made its way here as everywhere; G. bist. 3: O.H.G. is-t = 
 Go. is-* = I.-E. *es-tX; G. ist.— PI. 1: O.H.G. birum birun, 5 and 
 
 1 The O.E. vocalism is clearly corrupt: it has been influenced by the 
 vocalism of the parallel verb beo-m (G. bi-n), which was extant in O.E., but 
 has become obsolete in this tense. German, on the contrary, though retain- 
 ing only *biu-m, introduced into it the vocalism of i-m. — The E. dialects 
 often keep true to beoin : thus, we read I binna for I am not in the first pages 
 of Adam Dede. 
 
 2 On this capital distinction, cf. supra 199 and 211. 
 
 8 The root es is reduced, because the accent rests on the ending. 
 
 4 The approximate formula may be are : art — were : wast. As early as 
 O.E. we have earun (arori), which is obviously refashioned on the double 
 pattern eart and wabron. 
 
 5 Also birumes, and pirum pirumes, etc. These strange and obsolete forms 
 are certainly attributable to the analogy of the perfect, since their w-vocal- 
 
PERSON-ENDINGS. 351 
 
 still M.H.G. birn, but also sin borrowed from the subjunctive, 
 and lastly sint transferred from pi. 3 ; for, since pi. 1 and pi. 3 
 had come to be identical in any verb, it seemed only natural to 
 say wir sint like sie sint, the accidence elsewhere being sie nemen 
 like wir nemen 1 ; G. sind. 2: O.H.G. birut ; M.H.G. birt, but 
 also sit borrowed from the subjunctive ; G. seid. 3 : O.H.G. 
 ■sint = Go. sind = I.-~E. *s-ent%; G. sind. 
 
 II. Present subjunctive. — O.E. sie, etc., as in O.H.G., and 
 beo, pi. beo-n; the former lost in E., the latter changed to 
 invariable be.— O.H.G. si, etc. ( = Go. sijau, etc.), and G. set, 
 etc. 2 
 
 III. Perfect indicative. — The root wes forms naturally an 
 I.-E. pf. sg. 3 *we-wos-e, whence Preg. *was-e, conjugated Go. 
 vas vas-t vas ves-um ves-uj> ves-un. O.E., still purer than Go., 
 shows in the plural the rhotacism, as proceeding from the 
 change of s to s, after an unaccented syllable : sg. 1 and 3 wees, 
 pi. 1, 2, and 3 wwr-on ; to which E. corresponds with sg. 1 and 
 3 was (2 was-t), and pi. were. Likewise : O.H.G. was and 
 wdr-un; M.H.G. was and war-en ; but G. war (2 war-st) and 
 waren, the rhotacism and the long vowel having been extended 
 from the pi. to the singular. 
 
 IV. Perfect subjunctive : O.E. wcere and E. were (sg. 2 also 
 wer-t by analogy), O.H.G. wdr-i and G. wdr-e; without difficulty. 
 
 V. Imperative: E. 6e = O.E. beo; M.H.G. wis regular, and 
 also bis (instead of *bi, after the analogy of wis) ; G. sex, bor- 
 rowed from the subjunctive. 
 
 VI. Infinitive: E. 6e = O.E. beo-n; M.H.G. wes-en — Go. vis-an, 
 kept only as neuter noun (das Wesen, " being, substance," and 
 ein Wesen, "a being"), and also sin, G. sein, likewise imitated 
 from the subjunctive. 
 
 VII. Participles. — Present : E. be-ing ; G. wes-e-nd regular 
 (in such compounds as abwesend, etc.), but usually sei-end after 
 the analogy of the subjunctive. — Past: E. 6een = O.E. ge-beo-n ; 
 G. ge-wes-en = O.H.G. gi-wes-an = Go. vis-an-s. 
 
 ism and non-etymological r cannot come from any other source. They must 
 be imitated from wartimes and wurum (G. ivaren). 
 
 1 Cf. supra 203, 1 and 3, and 211. 
 
 2 Cf. supra 211. 
 
352 ENGLISH. AND GERMAN GRAMMAR. 
 
 § 2. Other Boot-Verbs. 
 (218) I. E. do=G. thun.— The I.-E. root dhe "to place, to 
 do" (cf. Sk. da-dhd-ti = Gr. rl-Orj-a-i "he places" and L. fa-c-i-t 
 "he does" "with an inserted c), though still conjugated by 
 merely adding the person-endings to the bare root, shows some 
 unusual and even perplexing vowel-gradations, which Gothic, 
 moreover, proves quite unable to account for, since here the 
 verb has the form tau-jan and belongs throughout to the weak 
 accidence. 
 
 1. The present indicative, by a rare exception, seems to 
 exhibit the deflected grade of the root, thus I.-E. *dhd-mi, 
 whence O.E. do-m = O.H.G- tuo-n (I do). The metaphonical 
 conjugation O.E. dd-m dce-st (long or) dce-cf, pi. do-S, gives us 
 the reason for the present pronunciation of the English forms 
 dost and does ( = doth) in contrast with sg. 1 and pi. do. But 
 the ending -m of sg. 1 is seen to have been dropped in imi- 
 tation of the corresponding person in the ordinary conjuga- 
 tion. The same is the case with German : O.H.G. tuo-n ; 
 M.H.G. tuo-n, then tuo ; G. thue. The remainder, namely tuo-st 
 tuo-t tuo-en tuo-(n)t tuo-nt, are rigorously reproduced in the 
 modern accideuce ihust thut thun thut thun. 
 
 2. The present subjunctive is regularly derived from the in- 
 dicative : O.E. do, etc., and E. do ; O.H.G. iuo-e, etc., and G. 
 thu-e. 
 
 3. The perfect indicative is an intricate mixture of two dif- 
 ferent forms at the least, viz. : a strong perfect-stem with sur- 
 viving reduplication, thus I.-E. *dhe-dha-, cf. O.H.G. t'eta " I, he 
 did"; and an originally weak perfect-stem, with normal root and 
 reduplication dropped, thus I.-E. sg. 3 *(dhe-)dhe- tdy, whence 
 a Preg. *&e<5a, G. that. 1 But the latter is the only one kept in 
 Mod. German. O.E. again shows a third vocalism, perhaps that 
 of the present, though strangely shortened and undergoing the 
 metaphony which would be regular only in the subjunctive, as 
 though it were a stem of weak pf. subj. Su-ft-i-, whence O.E. 
 dyde dydes, etc., E. did did-st, etc. 
 
 1 For the vocalism is here the same as in the noun denoting action, G. 
 t/ui« = E. deed = Go. deJ>-s = Preg. *fe-ffi-8 — I.-E. *dhe-ti-t, cf. supra 26 (II.) 
 and 78, 2. 
 
PERSON-ENDINGS. 353 
 
 4. Perfect subjunctive : O.E. dyd-e dyd-e-n and E. did ; 
 O.H.G. tdt-i tdt-l-s, etc., and G. metaph. that-e. 
 
 5. Imperative: O.E. do, etc., and E. do; O.H.G. tuo and G. 
 ihue. 
 
 6. Infinitive: O.E. do-n and E. do; O.H.G. tuo-n, and G. 
 thun. 
 
 7. Participles : present, O.E. do-nde (E. do-ing) and G. thu- 
 end ; past, O.E. ge-do-n, and E. do-we with the vocalic shade of 
 ge-doe-n; O.H.G. gi-td-n and G. ge-iha-n, with the vocalism 
 borrowed from the perfect. 
 
 (219) II. E. go = G. gehen. 1 — This verb lacks any precise 
 equivalent in the I.-E. family ; but there is a general agreement 
 between scholars in considering it a Pregermanic juxtaposition 
 of two I.-E. elements, prefix ga-, the origin of which has been 
 stated above, and root ey 1 (to go) of Sk. e-ti (he goes) = Gr. 
 ei-cri = L. i-t. The vocalism here is still more puzzling than in 
 the preceding type: O.H.G. shows in the present two con- 
 current vocalisms, gd- and ge- : O.E. has only the former, gd-, 
 changed to E. go ; Mod. G., only the latter, ge-, now spelled 
 geh-. 2 
 
 1-2. Present. — Indicative : O.H.G. sg. 1 gd-m gd-n or ge-m 
 ge-n, and the other persons in keeping. M.H.G. has still gd-n 
 and ge-n; but Mod. G. has only gehe = *ge with the analogical 
 loss of the ending. The same corruption appears far earlier in 
 English : even O.E. has only sg. 1 gd, whence E. go, and the 
 rest in keeping with it. — The subjunctive requires no explana- 
 tion. 
 
 3-4. Perfect. — The G. pf. is derived from another root, 
 which, in spite of an outward likeness, has originally nothing 
 to do with gehen, namely, from a verb gang-an = Go. gagg-an (to 
 go) of unknown etymology 3 : it is a strong perfect belonging 
 to type G, thus gieng ging. Of this root English has no longer 
 
 1 Here, of course, as in stehen and similar words, the h has no etymologi- 
 cal value. It is simply a sign to denote a long vowel. 
 
 2 Without entering into particulars, we may observe that the same double 
 vocalism equally characterizes the following verb (stan sten), and that these 
 two verbs, owing to their obviously contrasted meaning, mighc very easily 
 have influenced each other. 
 
 3 A tempting comparison is Sk.jdiighu (leg)=I.-E. *ghe"Jtgh-a. 
 
 A A 
 
354 ENGLISH AND GERMAN GRAMMAR. 
 
 any trace in the conjugation of the verb, though it is kept in 
 the nouns gang (cf. G. gang "marching") and gang-way (cf. 
 Go. gagg-s "street"). The pf. of vb. go is borrowed from vb. 
 wend = Gc. (sich) wenden "to turn, apply one's self " x = Go. vand- 
 jan, causative of Go. vind-an=G. winden = ~E. (to) wind. — The 
 subjunctives went and gienge offer no difficulty. 
 
 5-6. Imperative and infinitive like the present. 
 
 7. Participle (pres. go-ing geh-e-nd) past : on the stem gd-, 
 O.E. ge-gd-n and E. go-ne; on the stem gaUg-, O.H.G. gi-gang-an 
 and G. ge-gang-en, according to type G. 
 
 (220) III. G. stehen, cf. E. (to) stand.— Here also two con- 
 jugations are blended together. English, however, no longer 
 possesses the corresponding verb to G. stehen, since (to) stay is 
 known to be of Romance origin : the whole E. vb. is conjugated 
 on the stem stand-, cf. Go. stand-an, pf. stop, and O.H.G. stant-an 
 now obsolete. German, on the contrary, retains in the present 
 the bare root-accidence, corresponding to Sk. d-sthd-t (he stood) 
 = Gr. t-o-r>7 = L. sta-t, etc.: O.H.G. sg. 1 std-m std-n or ste-m 
 ste-n, and the rest in keeping; M.H.G. std-n or std, ste-n or ste; 
 G. only stehe. The perfect and participle have occurred under 
 the strong verbs, 2 and the other forms present no peculiarities. 
 
 (221) IV. E. will = G. wollen. — Ordinary grammars usually 
 place this verb in the next following class, with which, in fact, 
 it exhibits a striking though merely outward likeness. The 
 common feature in both, that is to say, the lack of an ending in 
 sg. 3 of the present, is here due to an isolated peculiarity of 
 accidence, i.e., as shown by the corresponding Gothic conjuga- 
 tion, the present of will = wollen is not an indicative, but a 
 subjunctive, or, in other words, E. I will = Gr. ich will = Go. vil- 
 jau really means not " I wish," but "may I wish " (optative), 
 though the indicative reappears in the G. plural. The root is 
 I.-E. wil (cf. L. vel-le vol-o vul-t vel-i-m), which assumes the 
 following regular Preg. vocalisms : normal wel, whence wU if 
 metaphony is required ; reduced wul wol ( = I.-E. w]) ; deflected 
 
 1 E. went bears to G. wandte the same relation assent to sandte, supra 187, 
 7. The verb is now obsolete, except in such locutions as I wend my own way ; 
 but it is still met with twice in Shakespeare. 
 
 » Type F, supra 184, 2. 
 
PERSON-ENDINGS. 
 
 355 
 
 wal, whence wel with later metaphony. A correct application 
 of these few principles will prove sufficient for the analysis of 
 this verb which is irregular and difficult only in appearance. 
 
 1. Present indicative (formerly subjunctive). — Sg. 1 : Go. 
 vil-jau, O.E. will-e and E. will, O.H.G. will-u and G. will. 1 2 : 
 Go. vil-ei-s — Jj. vel-i-s ; O.E. already wil-t, with the intrusive 
 termination -t, and E. wil-t ; O.H.Gr. wil-i = *wil-i-z, 2 but G. 
 analogically will-st. 3: Go. vil-i = li. vel-i-t ; O.E. wil-e wil, 
 and E. will; O.H.G. wil-i and G. will. — PI. Go. vil-ei-ma vil-ei-fi 
 vil-ei-na, exclusively subjunctive. But the vowel of the O.E. 
 final syllable will-aS (E. will) already suggests an intrusion of 
 the indicative-form, which appears even clearer in the root- 
 vocalism of Northumbr. wall-aft, The same form becomes pre- 
 valent in O.H.G. well-e-mes well-e-t well-e-nt, which has become 
 G. woll-e-n woll-e-t ivoll-en, under the influence of the vocalism 
 of the perfect woll-te. z 
 
 2. Since the present subjunctive had passed to the function 
 of present indicative, a new subjunctive was analogically derived 
 from it: O.E. will-e and E. will; O.H.G. well-e (cf. Northumbr. 
 wcell-a) and G. woll-e. 
 
 3-4. The perfect is a regular weak one, with reduced root, 
 thus I.-E. sg. 3 *(we-)wl-tdy : Go. vil-da has borrowed its 
 vowel from the present; but the vocalism remains pure in West 
 Germanic, O.E. wol-de and E. woul-d, O.H.G. wol-ta and G. 
 woll-te. The subjunctive, being formed here without meta- 
 phony, is of course identical. 
 
 5-7. Imperative not used. Infinitive : O.E. will-an and E. 
 will, but O.H.G. well-an woll-an, and Gr. woll-en. Participle ge- 
 woll-t. 
 
 § 3. Preterito- Presents. 
 (222) We have seen that the I.-E. perfect was not a 
 historical past, but the tense denoting an achieved action 
 
 1 Indicative, or else (rather) subjunctive with analogical indicative-ending; 
 the pure subjunctive form would be *wil-i, supra 193, 194 and 211. 
 
 2 Cf. the regular loss of this ending in O.E., supra 212. 
 
 8 The e of the root is due to metaphony, because the indicative is conju- 
 gated on a stem in -no- with deflected root, thus well-e-nt = Preg. *wal-ja-nd r i. 
 — Even at the present time, in High Alsatian, mr vele (e open, £half mute), 
 " we will." 
 
356 ENGLISH AND GERMAN GRAMMAR. 
 
 and therefore able to play the part of a durative present, 
 and we have already emphasized the difference of meaning be- 
 tween Gr. e-6ave "he died" and Tc-$vr)-Ke "he is dead." Simi- 
 larly, Ae-Acwr-Tai means not " it was left," but " it has been left," 
 whence " it remains," and such is the case with every Gr. 
 perfect. We have also insisted upon the fact, that the I.-E. 
 pf. *woyd-e, literally " he has seen," means " he knows " in 
 almost every language of the I.-E. family 1 : Sk. ved-a, Gr. oT8-e 
 =FotS-c, Go. vait, etc. Now, the Germanic languages, though 
 currently shifting their perfect to the function of historical 
 past, preserved a trace of its original meaning in about a 
 dozen perfects, which down to the present time unite perfect- 
 accidence with present-meaning. 
 
 The isolated position of this class protected it, as a rule, 
 from the causes which tended to alter the conjugation of 
 ordinary perfects : thus, we shall find in it that vowel-degra- 
 dation from the singular to the plural, which has been so 
 constantly levelled in the later languages, and even, at least in 
 part, the old ending of sg. 2 -t ( = I.-E. -tha) which O.E. and 
 O.H.G. had already forgotten 2 ; in regard to the necessary 
 identity between sg. i and sg. 3, we need scarcely say that 
 it is the leading, specific and regular characteristic of all these 
 verbal forms. 3 
 
 Though every preterito-present is really a perfect, each 
 again is provided with a historical past (can could, weisz wuszte, 
 etc.), and the latter always belongs to the type of weak per- 
 fects ; in other words, the preterito-presents are Germanic 
 verbs which use their active perfect as a present, and their 
 middle perfect as a preterite. 4 
 
 Including will = wollen, ordinary grammars as a rule distin- 
 guish nine of them, viz. two exclusively English, one exclusively 
 German, and six common to both languages. There were once 
 
 1 Cf. L. novi " I have taken notice of " whence " I know," me-min-i " I 
 remember," vixit " he is dead," etc. 
 
 8 Cf. supra 209. 
 
 8 At least in the literary language ; for elsewhere analogy may easily have 
 exerted its influence ; er weiszt is a current form in several German dialects, 
 and he cans would not be unnatural from an illiterate speaker. 
 
 4 Cf. supra 188, II. in fine. 
 
PERSON-ENDINGS. 357 
 
 a few more ; but some have become obsolete ; some, as for 
 instance E. dare and G. taugen, 1 were mistaken for real presents 
 and passed to the ordinary conjugation. The best way to study 
 them is to classify them under the familiar heads of our types 
 of strong perfects. 2 
 
 (223) Type A. 
 
 I. Go. vdit 3 (he knows), E. wot obsolete, G. weisz. 
 
 1. Present indicative. — The sg. has regularly the deflected 
 root. 1 : Go. vdit = Sk. ved-a = Gr. F otS-a, cf . L. vld-l (I saw) ; 
 O.E. wdt and E. wot ; O.H.G. wei% and G. weisz. 2 : Go. vdis-t 
 = Sk. vet-tha = Gr. Foicr-Oa; O.H.G. weis-t and G. weisz-t (O.E. 
 wds-i). 3: Go. vdit = Sk. ved-a = Gr. foi8-e; O.H.G. wei'3 and 
 G. weisz. — The pi. has regularly the reduced root. 1 : Go. 
 vit-um = Sk. vid-md = Gr. FlS-/xcv icr/iev ; O.E. wit-on, but E. toot 
 with levelled apophony ; O.H.G. wvzfi-um and G. wiss-en. 2 : 
 Go. vit-up and G. wiss-et. 3 : Go. vit-un and G. wiss-en. 
 
 2. Present subjunctive : the root regularly reduced ; Go. 
 vit-jau vit-ei-s, etc. ; O.E. wit-e ; O.H.G. unSK-i and G. wiss-e, 
 etc. 
 
 3—4. Perfect (weak) : Go. vissa (regularly replacing *vis-ta), 
 etc. ; O.H.G. wissa, etc. ; M.H.G. wisse, next wis-te with ana- 
 logical restoration of the ending -te, lastly (late) woste wuste 4 ; 
 G. wuszte and subj. wiiszte. 
 
 5. The infinitive, O.H.G. wi^-an, M.H.G. wiff-en and G. 
 wiss-en, is formed anew after the analogy of the past participle 
 
 1 On dare, cf. supra 187, 1. As a perfect, taugen regularly shows the de- 
 flected grade of the same root that appears reduced in tugend, cf. Gr. t&x.-V- 
 
 * The following is an alphabetical list : E. can, may, must, need, ought, 
 shall, will, wot ; G. dilrfen, k'onnen, mogen, miissen, sollen, xoissen, wollen. 
 
 3 This vdit is clearly the pf. of a vb., the normal infinitive of which would 
 be *veit-an; cf. also the words, E. wit^G. witz, E. wise (sensible) = G.weise, 
 E. wit-ness, etc. 
 
 4 The vocalism is corrupted by several influences : the regular subj. wiste 
 was partly mistaken for wuste (woste), and on these forms, through reversed 
 metaphony, were remade wuste and woste ; but the leading factor, here, was 
 the analogy of such similar forms as muoste, kunde konde, dorfte, because the 
 speaker could not but perceive a connection between verbs with both cognate 
 terminations and meanings. This is the reason why in all the vocalism ap- 
 pears rather capricious. — Archaic English has the regular form he wist " he 
 knew," e.g. Act. Apost. xii. 8. 
 
358 ENGLISH AND GERMAN GRAMMAR. 
 
 and present plural, just like the preceding (wellen wollen) and 
 following verbs. 1 
 
 6. The past participle is very regular : O.H.G. gi-wvzfi-an — 
 Go. *vit-an-s = I.- 1 E. *wid-onos; M.H.G. ge-wi^-en (cf. G. 
 gewisz " certain ") and ge-wis-t (weak) ; bat G. ge-wusz-t like 
 the perfect. 
 
 (224) Type C. 
 
 II. Go. kann (he knows), E. can, G. Jcann. — The literal mean- 
 ing is "he has learnt," whence "he knows," further "he is 
 able, he can," the apparent present being in reality the perfect 
 of the I.-E. root gen (to know, Sk. jdnd-ti " he knows "), which 
 in Greek and Latin appears chiefly in the shape gno, Gr. yi-yvco- 
 otcu), pf. £-yva)-Ka, L. no-sco = gno-sco, pf. novl, part, no-tus, etc. 
 
 1-2. Present. — Indicative sg. 1 : Go. kann, O.E. con can and 
 E. can, O.H.G. kan and G. kann. 2 : Go. kan-t regular, but 
 already O.E. const and E. canst, O.H.G. kanst and G. kannst. 
 3 : like sg. 1. PI. 1 : Go. kunn-um, O.E. cunn-on, O.H.G. 
 kunn-un and M.H.G. kunn-en, regular. But English extends 
 the form can, whereas Mod. German, though keeping the 
 apophony, obscures it by borrowing the metaphony from the 
 subjunctive. — For this mood is, quite regularly, M.H.G. kiinne 
 = O.H.G. kunn-i. Since however, in the present of the ordinary 
 conjugation, indicative and subjunctive always show the same 
 root-vocalism, a similar correspondence took place here, whence 
 M.H.G. kiinnen and G. konnen in the plural of the indicative as 
 well as in the subjunctive. — So too pi. 2 konn-t and 3 kbnn-en. 
 
 3-4. Perfect. — Indicative : Go. kun-pa = I.-E. (ge-)gri-tay ; 
 O.E. cub~ , e = *cun-&'e, M.E. coud and E. could 2 ; O.H.G. kon-da, 
 M.H.G. kun-de kon-de, and G. konn-te. — Subjunctive : E. could, 
 G. konn-te. 
 
 5-6. The vocalism in the infinitive, E. can, G. konn-en, is 
 respectively that of the sg. and pi. of the present. The Mod. 
 G. part, ge-konn-t is formed from the perfect. 3 
 
 III. Go. }>arf (he wants), G. darf (is allowed, able). — The 
 
 1 Though it might also be deemed a primitive form : supra 82, 2. 
 
 2 The true spelling changed under the influence of would and should. 
 8 The regular one is the adjective kuud (known) = Go. kun\>-s. 
 
PEHSON-ENDINGS. 
 
 359 
 
 same as Sk. ta-tdrp-a, pf. of the root tarp = I.-~E. terp (to satisfy, 
 or be satisfied), cf. Gr. ripTr-t-Tai (he is delighted). Thus, the 
 literal meaning -would be " to find one's satisfaction in," whence 
 " to be in want of " 1 and "to have the faculty of," the latter 
 sense now being the only one surviving. 
 
 1-2. Present. — Indicative sg. 1 and 3 : Go. ]>arf, O.E. $earf, 
 O.H.G. darf, G. darf. 2 : Go. ]>arf-t, O.H.G. darf-t, G. darf-st. 
 PI. 1 : Go. J>aurb-um, with vowel-degradation and consonantal 
 change; O.E. Surf-on and O.H.G. durf-un, with degradation, 
 but the / extended ; M.H.G. durf-en, and dilrf-en, like kiinnen, 
 with the metaphony imported from the subjunctive ; G. durf-en. 
 — All the rest in keeping with pi. 1, and similarly G. subj. 
 diirf-e. 
 
 3-4. Perfect. — Indicative: Go. ]>aurf-ta, O.H.G. dorf-ta, 
 M.H.G. dorf-te, but G. durf-te influenced by the vocalism of the 
 present.— Subjunctive, likewise, M.H.G. dorf-te and G. durf-te. 
 
 5-6. Infinitive : durf-an, now dilrf-en. Participle : *ge-dorf-t 
 (in the compound be-dorf-t), now ge-durf-t. 
 
 (225) Type D. 
 
 IV. Go. skal, E. shall, G. soil, the pf. of a root which would 
 be I.-E. skel, but cannot be found outside the Germanic group. 
 The vowel o in German is due to the analogy of the plural ; but 
 the loss of the k, which takes place from O.H.G. to M.H.G., 
 remains as yet unexplained. 2 
 
 1-2. Sg. 1 and 3: Go. skal; O.E. sceal and E. shall; O.H.G. 
 seal sol, M.H.G. sal sol, and G. soil. 2 : Go. skal-t ; O.E. sceal-t 
 and E. shal-t, unaltered; O.H.G. scal-t sol-t, and occasionally 
 even sol-st ; M.H.G. sal-t sol-t, but G. soll-st. PI. 1 : Go. 
 skul-um ; O.E. scul-on, but E. shall ; O.H.G. scul-un, M.H.G. 
 sul-n or (metaph.) siil-n ; G. soll-en, with the perfect- vocalism, 
 or even (metaph.) soll-en. But the non-metaphonical form has 
 
 1 The O.E., O.H.G. and still M.H.G. meaning "to want" survives in 
 Mod. G. be-durf-en, whereas the active meaning " to satisfy " and conse- 
 quently "to thrive" may be seen in the lost primitive *derben, which, by 
 adding the pejorative prefix (ver-derb-en), has acquired the sense " to be 
 corrupted, to corrupt." 
 
 3 Still M.H.G. er schol (Bavarian). ^_^____ . 
 
 / --FTHE r 
 
360 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 decidedly prevailed and even invaded the subjunctive : O.H.G. 
 scul-i, M.H.G-. siill-e, G. solle. 
 
 3—4. Go. shul-da ; O.E. sceol-de and E. shoul-d ; O.H.G. scol-te 
 sol-te, M.H.G. sol-te and G. soll-te. The subjunctive is M.H.G. 
 sol-te and sol-te, but class. G. soll-te, like E. shoul-d. 
 
 5-6. Infinitive shall and soll-en. — Part, ge-soll-t. 
 
 (226) Type E. 
 
 V. Go. mag, E. may, G. mag, the pf. of an I.-E. root megh 
 (also meg), which expresses an idea of greatness and might, 
 and is found in Sk. mah-dnt- (great), mdh-as (greatness), Gr. 
 /xcy-a-s (great), cf. O.E. meah-t (power) and E. migh-t, O.H.G. 
 mah-t and G. mach-t. " r 
 
 1-2. Sg. 1 and 3 : Go. mag, O.E. maeg and E. may, O.H.G. 
 mag and G. mag. 2 : Go. mag-t ; O.E. meah-t mih-t, but E. 
 may-est ; O.H.G. and M.H.G. mah-t, but G. mag-st. As early 
 as Pregermanic the plural adopts the singular-vocalism : Go. 
 mag-um, O.E. mag-on and E. may, O.H.G. mag-un. But, owing 
 to the analogy of Tcunn-un, durf-un, we have also mug-un, 
 whence : M.H.G. mug-en mug-en (and meg-en) ; G. mog-en. — 
 Subjunctive : E. may, G. mbg-e. 
 
 3-4. Go. mah-ta, O.E. meah-te mih-te and E. migh-t, O.H.G. 
 mah-ta and M.H.G. mah-te, quite regular; but also O.H.G. 
 moh-ta (like kon-da), M.H.G. moh-te and G. moch-te (ver-moch-te) . 
 — Subjunctive : M.H.G. mehte mohte and G. mochte ; E. might. 
 — The rest present no difficulty. 
 
 (227) Type F. 
 
 VI. Go. ga-mbt (he finds an opportunity), E. must, G. musz, 
 pf. of an unknown root. 1 English has lost the strong perfect, 
 and its weak perfect plays at once the part of present and past 
 tense : he must = er muszte. 2 
 
 1-2. Sg. 1 and 3 : Go. ga-mot, O.E. mot (he can), O.H.G. 
 muoTf (he can) and G. musz (must). 2 : Go. *mos-t, O.E. mos-t 
 and E. mus-t, O.H.G. muos-t and G. musz-t. PI. 1 : Go. *mbt- 
 
 1 Cf. G. musze (leisure) = O.H.G. muo^a (possibility, opportunity), lost in 
 ,E. 
 
 2 Excepting, of course, sg. 2 of both tenses, since there is no form *must-est. 
 
 O.E 
 
PERSON-ENDINGS. 
 
 361 
 
 um, O.E. mot-on, O.H.G. muoyum, but M.H.G. (metaph.) mue^- 
 en and G. miiszen, etc. 
 
 3-4. Go. ga-mos-ta, O.E. mos-te and E. mus-t, O.H.G. muosa 
 (instead of *muos-sa, like wis-sa) and muos-ta (restored), M.H.G. 
 muos-te, G. musz-te and snbj. milsz-te, etc. 
 
 (228) Unclassed. 
 
 VII. E. ought. — A weak perfect: Go. dih-ta (he had), from 
 vb. dih-an, 1 cf. G. eig-en = ~E. ow-n; O.E. dh-te ; lost in German. 
 According to this etymology, a locution he ought to do, either 
 means " he had," or " he should have (subjunctive) to do " : 
 whence both the constant construction of ought with to, and its 
 absolute indifference in use for present and past time. 
 
 VIII. E. need. — A secondary verb, derived from the substan- 
 tive need = O.E. nead = G. noth — Go. ndup-s. But the analogy 
 of he must not, he shall not, etc., induced the speaker to say 
 likewise he need not : the only point in which this verb belongs 
 to the preterito-presents. 
 
 IX. E. will, G. will : is not a preterito-present, supra 221. 
 
 1 See the Go. sentence quoted above (128). 
 
CHAPTER IV. 
 VERBAL PERIPHRASES. 
 
 (229) Conjugation being reduced, as seen above, to the 
 utmost simplicity, — a single past tense and no future, — had 
 become unable to express the minute shades of the verbal 
 concept : incomplete as it was, however, it sufficed for Preger- 
 manic, even for Gothic and for the earliest state of English and 
 German ; but, as the speech grew more precise and refined, it 
 was led, no less than French and the other Romance languages, 
 to use certain periphrases in order to denote clearly such moods 
 and tenses as could no longer find adequate expression in the 
 forms of the simple verb. The Verbal Periphrasis, though 
 it may vary as to particulars, in its main and original structure 
 always and everywhere appears one and the same, namely, 
 a kind of compounded locution, in which a so-called 
 auxiliary verb governs a verbal noun, whether infinitive, 
 gerund or participle, derived from the principal verb. 
 Hence, it is the latter that contains the verbal concept, whereas 
 the auxiliary merely indicates the tense, mood, or aspect, which 
 modifies its general meaning. 
 
 I. Our first task will be, therefore, to recall the formation 
 and function of such Verbal Nouns as may be met with in a 
 verbal periphrasis. 
 
 1. Infinitive. — The primitive formation of the infinitive, as 
 originally a neuter noun and, besides, still capable of being 
 used and declined in German in this quality, and further the 
 extension of this grammatical category to all verbs, even non- 
 radical, from the earliest to the latest, and, lastly, the uniform- 
 ity of its exponent, reduced to -en in German, and dropped 
 
 362 
 
VERBAL PERIPHRASES. 363 
 
 altogether in English, have been illustrated in their proper 
 place, 1 and no further details are here required. 
 
 2. Gerund. — The gerund, though now completely con- 
 founded with the infinitive, was originally distinct, not only in 
 that it may be declined, while the infinitive verbal noun never 
 appears but in the accusative neuter, but chiefly because it 
 proceeds from a derivative and amplified stem : for, the infini- 
 tive-suffix being I.-E. *-ono-, the suffix of the gerund is I.-E. 
 -onyO', % as shown by the double nn ( = Preg. nj), which every- 
 where characterizes this form, O.E. dat. to sdwenne (in order 
 to sow), to drincenne, 3 O.H.G. gen. nem-anne-s (of taking), dat. 
 riem-anne (to take), etc. In the later languages, the genitival 
 gerund is extinct, — though, of course, German may decline its 
 infinitive in the gen. sg. (des nehm-en-s), like any other neuter 
 noun, — and the datival gerund, owing to the loss of the final 
 syllable, has been blended with the infinitive, E. to drink, G. 
 zu nehmen, etc. 4 
 
 3. Participles. — A. Present Active. The reader need 
 scarcely be reminded of the formation of the I.-E. active present 
 participle, 5 retained in Sanskrit, Greek, Latin, Gothic, Old 
 English, and German, whether Old or Modern, as well as of 
 the artificial and somewhat blundering way in which Middle 
 English has secured a new one 6 : G. trink-e-nd, E. drink-ing ; 
 
 1 Supra 77 (1), 89 (2), 106 and 107. 
 
 2 In other words, they bear to one another the same relation as Gr. 
 otp-wb-% (heaven) and ovp-dvio-t (celestial), and the gerund is the adjective 
 derived from the noun that plays the part of infinitive. 
 
 8 Cf. supra 128, and such sentences as ic to drincenne hcebbe " I have to 
 drink," ffa com hit to witenne Sam eorlum " then came it to knowing to the 
 earls," which clearly emphasize both the nature of the gerund, as a noun, 
 and the nature of the form which depends on a verb have, ought, am (he i» 
 to do), and generally on the preposition to, as a gerund, and not an infinitive. 
 Hence it is through a rather awkward confusion, and merely for the sake of 
 convenience, that practical grammar uses to as the specific exponent of the 
 infinitive. 
 
 * Already O.H.G. occasionally ze wesan, za galaupian (zu glauben). — 
 Observe however that, after any other preposition but to, E. uses not the 
 infinitive, but the noun of action in -ing (sxipra 103, II., 2), which therefore 
 serves for the gerund. 
 
 s Supra 79 (XI.) and 90 (VI.). 
 
 6 Supra 103, II., 2. Even passive ; for the clear and terse expression, 
 the tea is a-making has been replaced only as late as the end of the last 
 century by is being made. The passive present participle and the active 
 
364 ENGLISH AND GERMAN GRAMMAR. 
 
 and similarly in every verb, strong or weak, primitive or deri- 
 vative, without exception. 
 
 B. Past Passive. — This formation has been analysed in its 
 inner structure, the root-form and suffix being different as the 
 verb is strong or weak. 1 The origin of its prefix, which is 
 common, though not obligatory, in O.E., 2 and of constant use 
 in German, 3 has also been explained, and we have seen how 
 English got rid of it. 4 German, on the other hand, extended 
 it to all its verbs, excluding only such as were already provided 
 with some other inseparable prefix, 6 and even to the latter 
 when the speaker was no longer sensible of the existence of a 
 prefix : ge-fr-essen, ge-b-lieben, ge-g-laubt, etc. 6 In fact, no verb 
 remained free from it, but the modern and hybrid formations 
 in -ieren (-ier-i)? 
 
 C. Future Passive. — By blending together, as it were, 
 strangely enough, its gerund and present participle, German 
 has obtained a passive future participle : das zu lesende buch 
 " the book which is to be read " (F. shorter le livre a lire), der 
 zu schreibende brief, etc. ; a locution which belongs to the latest 
 period of speech. 
 
 II. The verbal periphrases have been said to denote the 
 modifications of the verb, either in Tense, or Mood, or 
 Aspect. Hence the threefold division of this Chapter. 
 
 past participle are now, in both languages, denoted by peculiar periphrases, 
 respectively : E. being done, G. gethan werdend (unusual) ; E. having done, 
 G. gethan habend. The active future participle does not exist in either. 
 
 i Supra 77 (1), 78 (1), 89 (1), 90 (1), 175 sq., 186 sq. 
 
 8 But unusual in Mercian, a further reason why it should be lost in English. 
 
 8 The current form worden is the only surviving example of the faculty of 
 deriving a root-participle without the prefix ge-. But compare the use of the 
 infinitive, instead of the participle, in such sentences as er haVs nicht thun 
 wollen, etc. 
 
 4 Supra 96 (I.) and 97. 
 
 5 And this for a good reason : such forms as * ge-er-funden, *ge-ver-ziert, 
 were quite impossible, since there never existed a vb. *ge-er-finden, etc. 
 
 6 Supra 19 (4), 65 (4) and 183 (twice in ge-g-essen). 
 
 7 Supra 107 (VI.). 
 
VERBAL PERIPHRASES. 365 
 
 Section I. 
 
 PERIPHRASTIC TENSES. 
 
 (230) Various tenses, either Present, Past, or (in par- 
 ticular) Future, either may or must be expressed by a peri- 
 phrasis. 
 
 § 1. Present. 
 
 Neither German nor French has a special term for the 
 Momentary Present. Thus, a German says, ich rauche sehr 
 wenig " I smoke very little " (customary present), like ich 
 rauche eine gute pfeife " I am smoking a good pipe " ; or, at any 
 rate, it would require a complete change of expression to mark 
 the distinction; whereas English simply says, respectively, I 
 smoke and 1 am smoking ; and it has been seen that from this 
 very locution English has abstracted its present participle. 
 
 The further development of this periphrasis gave birth to a 
 whole conjugation, wherein the principal verb, in the shape of 
 the invariable present participle, serves as a predicate to the 
 verb to fee, which latter may assume the form of any mood, — 
 though the imperative is unusual, — and, subsidiarily, every 
 tense whatever, viz. : he was working as I came in ; I have been 
 working while you were away, etc. 
 
 § 2. Past. 
 
 (231) I. We have already seen, that, in giving the sense 
 of a historical past to the I.-E. perfect which was originally 
 the tense of the achieved action, Germanic deprived itself of 
 any expression for the latter temporal relation x : in other 
 words, having lost the aorist, and using I spoke and ich sprach, 
 he died and er starb, etc., with a narrative meaning, English and 
 German, no less than French, missed the simple term correspond- 
 ing to the precise force of " I have spoken (I have done speak- 
 ing), he is dead (no longer alive)," etc. Both have filled the 
 gap by means of a very simple device, the same that Low 
 Latin bequeathed to the modern Romance languages. 
 1 Supra 172 and 222. 
 
366 ENGLISH AND GERMAN GRAMMAR. 
 
 1. It is a commonplace fact that an achieved action very 
 often results in either a real or, at least, a metaphorical pos- 
 session : thus, for instance, the dnrative consequence of the 
 act of acquiring is the act of possessing, 1 and such locutions as 
 E. I have bought a house, G. ich habe ein haus gekauft, F. fai 
 achete une maison, etc., may be translated word for word " I 
 have (I possess) a house [inasmuch as it is] bought." And 
 so likewise I have plucked this flower, I have filled this vessel ; 
 further, metaphorically, I have learnt this science, I have com- 
 prehended the fact (" I hold the cause of it"), etc. ; and, lastly, 
 such an abusive extension of the use of the verb " to have " 
 as a mere auxiliary, altogether deprived of its genuine and 
 primitive acceptation, that the speaker finds it quite natural 
 to say, in spite of common sense, E. I have sold my house, 
 G. ich habe mein haus verkauft, P. fai vendu ma maison. 
 
 2. On the other hand, the result of an achieved action may 
 be also very often a durative state, which is therefore con- 
 veniently denoted by the use of the verb " to be " followed by a 
 past participle, the latter, like any other adjective, playing in 
 the complex the part of a predicate : thus, such sentences as 
 E. he is dead, G. er ist todt, F. il est moi't, are, in their inner and 
 primitive structure, to be put exactly on a level with such as 
 E. she is pretty, G. wir sind stark, F. il est grand, etc. ; but the 
 function of the verb " to be " becomes gradually effaced and 
 confined to the meaning of a mere auxiliary and past-exponent ; 
 and, the angle of vision, as it were, having slightly changed, 
 such a sentence as G. er ist gestiegen, literally " he is mounted 
 (come up)," becomes equivalent to " he has ascended." 
 
 3. Of course, it is a matter of pure usage, in French and 
 German as well as English, to decide whether a given verb 
 actually expresses an action or a state, — for the dividing line 
 is often a vague one, — and, consequently, whether it ought 
 to be conjugated with the auxiliary " to have " or " to be." 
 English however, more complete and logical in this respect, 
 may use both at once with the same verb, and, by saying, for 
 instance, either he has come or he is come, may distinguish two- 
 
 1 Consequently, Gr. ict-tcrn-fiou "I have accfuired" signifies "I possess." 
 Cf. L. nCt-vi " I have learnt," whence " I know." 
 
VERBAL PERIPHRASES. 367 
 
 delicate shades of meaning, for which both French and German 
 have but one expression, il est venu, er ist gekommen. 
 
 II. Hence, the tense of achieved action, so-called Indefinite 
 Past, in contrast with the historical past (perfect), is ex- 
 pressed, in the indicative or subjunctive, by the present 
 indicative or subjunctive, either of the verb have — haben, or 
 be = sein, accompanied by the past participle of the principal 
 verb. 1 A further consequence is, that the perfect of these 
 auxiliaries appears well fitted for expressing a new shade of 
 past time, namely the Pluperfect, or tense denoting an action 
 achieved as prior even to another action equally achieved : I 
 had read (" done reading ") when he came in, G. ich hatte gelesen, 
 F. favais lu ; E. he had or was come, G. er war gekommen, etc. ; 
 and so likewise in the subjunctive (E. had or were, G. hatte or 
 ware) . 
 
 § 3. Future. 
 
 (232) Pregermanic bad completely lost the I.-E. Future. 
 Gothic translates the future by a mere present, occasionally, 
 though rarely, where more precision is required, using an 
 auxiliary skal or haba (I have to) with the infinitive. Such 
 being the case, O.E. likewise uses the auxiliary sceal, and 
 O.H.G. the auxiliary sculan or wellen, both however with a 
 decided preference for the simple present indicative as avoiding 
 a tedious periphrasis. Indeed, even at the present day, if the 
 meaning is unmistakable, the present is often used for the 
 future, chiefly in German, as also in French : G. morgen fahre 
 ich ab, F. je pars demain. 
 
 The values involved in the future-meaning being of a some- 
 what variable character, it is natural that we should find the 
 tense expressed by several periphrases. 
 
 1. The most usual value is that of a simple intention, thus 
 
 1 It is a curious fact — so great is the tendency, in all languages, to use 
 the well-known paths — that this past of achieved action has often become, 
 in its turn, a historical past : thus, spoken French may be said to have lost 
 the so-called definite past, and a Frenchman will say " j'ai vu" whether he 
 means " I saw " (relating) or " I have seen " (stating) ; the same is the case 
 with South German (Alamannic), which now knows nothing of ich sah and 
 uses only ich liabe cjesehen. 
 
368 ENGLISH AND GERMAN GRAMMAR. 
 
 "I will do." 1 Hence a verb denoting "will" may be used to 
 indicate a future event in regard to a subject upon whom no 
 outward constraint either is or seems to be exerted. This is 
 the case with English: I will go ("because I choose to go "), 
 you will go (" but there is nothing that compels you, or at any 
 rate I don't"). In German, though a speaker may very well 
 say, with the same meaning, ich will and wir wollen gehen, the 
 other persons are not used in a similar function, and the verb 
 wollen therefore cannot be properly termed an auxiliary or 
 future-exponent. 
 
 2. Now, it might also happen that the future involves an 
 accessory idea of obligation or necessity : in which case, an 
 auxiliary denoting duty equally answers the purpose, E. 1 shall 
 go. In this locution, however, the idea of duty is attenuated 
 and nearly effaced, — as it is also in F. je dois aller, — though it 
 reappears with great force in the other persons of the tense. 2 
 German likewise, but with attenuated force in all persons, may 
 say ich soil and du sollst ; but an actual command will require 
 du muszt. s 
 
 3. Exclusively English is another periphrasis, which also 
 contains a sense of obligation, though far less marked, and 
 consists of the verb to be combined with the gerund : he is to 
 go. It has certainly been furthered by the fact that the 
 imperfect of will and shall was shifted to the usual meaning of 
 a conditional 4 ; for it supplied the speech with an imperfect 
 of the future, which, if expressed otherwise, would have been 
 ambiguous : the member who was to read the report, cf . F. . . . 
 qui devait lire . . . and G. . . . der . . . vorlesen sollte. 
 
 4. The ordinary German future, although exclusively 
 German, is very like the preceding, only replacing the vb. " to 
 be " by the vb. " to become " and the gerund by the infinitive. 
 Besides, it is a rather late expression, the more so because, the 
 further we trace it back, — and it does not appear before the 
 
 1 Remember that Germanic " I will " was not originally affirmative and 
 peremptory : supra 221. 
 
 8 You shall do it being, in fact, such an authoritative command that it is 
 rarely used. 
 
 8 Cf. supra 225 and 227. 
 
 * Infra 235. 
 
VERBAL PERIPHRASES. 369 
 
 end of the thirteenth century, — the less it shows of a real 
 future meaning, merely denoting, as might be inferred at once 
 from its components, the beginning of an action, thus: sie 
 werdent ej$en, literally " they become [to] eat " ; er wart weinen, 
 "he began to cry," etc. 1 It was not until Modern German 
 that the speaker took to using this inchoative present in a 
 future sense. 
 
 5. By combining each of these locutions with (hose which 
 denote the past tense, English and German created a future of 
 the achieved action (so-called anterior or past future) : I will 
 or shall have done; ich werde gethan haben, etc. Further, in 
 English, a future of the momentary action : I will be doing, etc. 
 
 Section II. 
 
 PERIPHRASTIC MOODS. 
 
 ( 2 33) In the moods, the verbal periphrasis may be applied 
 to the Indicative, the Subjunctive, the Conditional, and 
 the Imperative. 
 
 § 1. Indicative. 
 
 Germanic, at any time, might replace the indicative of a 
 given verb by the indicative of the vb. do — thun governing the 
 infinitive of the principal verb, 2 in order to emphasize more 
 strongly either the reality of the action or the intensity of the 
 affirmation. Indeed, in the opinion of some scholars, it is such 
 a verbal agglutination that is to be recognised as the origin of 
 the Go. formation sdh'-ded-um 3 = we did seek. M.H.G. still 
 possesses this locution, though using it rather rarely : da^ sie 
 
 1 Such a sentence would seem to require the gerund rather than the 
 infinitive ; but, as early even as M.H.G., the use of either form is a matter 
 of choice and often of caprice. Even now, the use of zu, though regulated, 
 is known to be very arbitrary : ich weisz, for instance, cannot do without it, 
 whereas ich harm, which originally had the same meaning, altogether ex- 
 cludes it, as well as ich toll and many others. 
 
 2 The infinitive originally being only the accusative of a neuter noun 
 denoting action, it is clear that a Preg. locution *Sdmi Mnffanam (E. I do 
 bind, G. Kh time binden) must be literally translated " I perform binding " 
 whence " I b-nd." 
 
 3 Supra 195 and 210. 
 
 B B 
 
370 ENGLISH AND GEKMAN GBAMMAR. 
 
 uns tuon bewam, " that they are aware of us." In Mod. German 
 it is now nearly obsolete, though of course a speaker may still 
 say : ich thu' es leugnen, " I deny it expressly." But in English 
 it has been considerably developed, and its use has become an 
 important grammatical principle. 
 
 From this a double benefit was derived, namely, the possibility 
 of insisting strongly upon a statement (I do ask you), and that 
 of expressing either an interrogation or a negation without 
 being obliged to repeat the verb involved in it (you mean to go, 
 do you? I don't). But, since the speaker was accustomed to 
 cast in this mould an abbreviated interrogation or negation, he 
 was naturally led to mistake the words do you and I do not for 
 the specific exponent of the interrogative or negative force 
 itself ; and thus there became gradually consecrated the use, 
 which is now almost obligatory, of the auxiliary do did to 
 express the indicative in an interrogative or negative 
 sentence, 1 unless the verb is already provided with some 
 other auxiliary : do you see t but have you seen I 
 
 § 2. Subjunctive. 
 
 (234) The English potential subjunctive, or subjunctive 
 in a final sentence, 2 is really the indicative of the vb. may 3 
 followed by the infinitive of the principal verb : tell him so, that 
 he may know it ; I told him so, that he might know it, etc. The 
 same is the case, if a vague and indefinite eventuality is to be 
 expressed : I did not believe him, whatever he might say (G. was 
 er auch sagen mochte). Still more so, in the expression of a 
 wish 4 : God grant, he may succeed t (G. es moge ihm gelingen !). 
 As seen in these instances, German in such cases is likewise 
 always allowed to use the similar mogen ; but the use of it is 
 never obligatory. 
 
 § 3. Conditional. 
 
 ( 2 35) The Indo-European language had no special mood 
 
 1 Of course only when the negative is not. 
 
 2 That is to say, expressing a purpose, " in order that." 
 * Supra 226. 
 
 4 Here, of course, the auxiliary itself stands in the subjunctive. 
 
VERBAL PERIPHRASES. 371 
 
 for expressing the Conditional : in other words, the con- 
 ditional eventuality, no less than any other eventuality, was 
 denoted by the subjunctive or by the optative (Sanskrit, Greek, 
 Latin), two moods which in Pregermanic become the same. 
 Even contemporary German, at least in strong verbs, briefly 
 expresses the conditional by the perfect subjunctive: ich thdte 
 (I would do), ich hdtte (I would have), nahme er (if he took) ; 
 cf. E. were I, were you, etc. But, in English, and also in German 
 in the great majority of weak verbs and even a small number 
 of strong ones, the perfect subjunctive is scarcely or not at all 
 distinguishable from the perfect indicative. Such a gross 
 ambiguity could not be tolerated in any refined language, and 
 the inconvenience was remedied by recurring to a periphrasis, 
 the main principle of which is, that the conditional is conceived 
 as a past-tense of the future. 1 
 
 1. Consequently, English, together with the infinitive of the 
 principal verb, uses the perfect of its two auxiliaries will and 
 shall, thus would and should, respectively with the same shades 
 of meaning as their present may assume in the future-forma- 
 tion : I would go ; we should answer, etc. By means of the 
 auxiliaries have and be, is further derived a past conditional : 
 they would have done it. In the same sense, German may use 
 its verb sollte ; but this is not its ordinary conditional peri- 
 phrasis. 
 
 2. For, its ordinary future-auxiliary being the present of the 
 vb. werden, German naturally uses, as the auxiliary for its 
 conditional, the perfect subjunctive of the same verb : ich 
 wurde gehen, ihr wiirdet sehen, etc. ; though, of course, in order 
 to avoid such diffuse circumlocutions as ich wurde gekommen 
 sein and ihr wiirdet gesehen haben, the past conditional is better 
 formed by the simple perfect subjunctive of the auxiliary sein 
 or haben, thus, ich ware gekommen, and ihr hdttet gesehen. 
 
 § 4. Imperative. 
 (236) The periphrastic forms of the imperative, E. sg. 3 
 
 1 In French also (and in the other Romance languages) we find faurais : 
 j'aurai=j'avais: j'ai. 
 
372 ENGLISH AND GEBMAN GRAMMAR. 
 
 let him go, pi. 1 let us go = G. laszt uns gehen, pi. 3 let them go, 
 etc., are too plain to require explanation. 
 
 Like the negative form of the indicative, the prohibitive 
 form of the imperative, when the negation is not, usually intro- 
 duces in English the auxiliary do, thus : do not steal, but never 
 break your word. 
 
 Section III. 
 
 PERIPHRASTIC ASPECTS. 
 
 (237) It has been seen that every verb is capable of three 
 essential aspects, namely, Active, Reflexive, and Passive. 
 The active aspect having naturally no further expression but 
 the general form of the verb itself, we have only to deal with 
 the two others. 
 
 § I. Reflexive. 
 
 (238) The form denoting the reflexive aspect was in Indo- 
 European the so-called middle voice of the verb. But, since 
 the middle voice could also correspond, either to the active, 1 or 
 even, very often, to the passive aspect, there was likely to arise 
 a most inconvenient ambiguity, which is much felt in Sanskrit, 
 Greek and Latin, 2 and would have become quite intolerable in 
 the later and more analytical languages. As early as Pre- 
 germanic, the middle voice was no longer applicable either to 
 the active or even reflexive aspect, but was strictly confined to 
 the passive. Hence, a periphrasis had become necessary for 
 the reflexive function, and nothing could appear better adapted 
 for the purpose than the use of a pronominal complement to 
 the verb, denoting that the action returned to the subject (G- 
 ich besinne mich, E. he asked himself, etc.), a process which has 
 been already described in the study of the Pronouns. 3 
 
 1 Cf. supra 188, II. 
 
 2 In a single conjugation, for instance, imitor means " I imitate," lavor 
 " I wash myself," and amor " I am loved." 
 
 3 Supra 167. The reflexive verb ought not to be confounded with the 
 impersonal verb, the neuter subject of which may be either expressed or 
 merely understood: E. me-thinks = G. es diinkt mich (supra 24) "it seems 
 to me." 
 
VERBAL PEEI PHRASES. 373 
 
 § 2. Passive. 
 
 (239) Gothic, like Latin, still possesses a passive voice, 
 which, as stated above, is no other bat the former middle voice 
 of the Indo-European; that is to say, Gothic expresses in one 
 word what English, German and French are unable to express 
 except by two distinct words : nimada x " I am taken," hditada 
 " I am called (my name is) " 2 j subjunctive nimdidau hditdidau, 
 etc. This passive verb, however, has no tenses but the present, 
 and even Gothic, when a past tense was required, was com- 
 pelled to supply it by a periphrasis, a past participle joined as 
 a predicate to the verb visan (to be) or vair}>an (to become). 
 Still more so, indeed, in Old English and Old High German, 
 since neither shows a trace of a passive voice, even in the 
 present. The auxiliaries here are the same as in Gothic ; but 
 O.E. almost exclusively uses beon and wesan, which soon become 
 blended into one verb, whereas weortSan early becomes obsolete ; 
 while, on the contrary, German, from the ninth century, shows 
 a decided preference for werdan, which in fact, as appropriately 
 expressing a passage from one state to another, cor- 
 responds more exactly than wesan to the concept involved in 
 the passive verb properly so-called. 
 
 From this it follows that German accurately marks a dis- 
 tinction which is unknown in English as well as in French. 
 These two languages, outwardly at least, confound the true 
 passive verb (" he is wounded," that is " he becomes wounded 
 [at this instant], they are wounding him") with the locution 
 formed by the verb " to be " and a past participle serving as a pre- 
 dicate to it, as any other adjective would do ("he is wounded," 
 that is "he has been wounded [and still is so]," thus, in either 
 case, E. he is wounded, F. il est blesse. 3 In the past tense, the 
 
 1 Conjugated: sg. 1 mm-a-da, 2 nim-a-za (Gr. *i^i-e-er<u), 3 nim-a-da (Gr. 
 yi/i-e-rai), pi. nim-a-nda (Gr. vifi-o-vrai). 
 
 8 A last survival of this lost passive conjugation may be seen in the fact 
 that the G. vb. ich heinze means both " I call " and " I am called " (L. vocu 
 and vocor). 
 
 8 If however the present be not a narrative one, English may adopt the 
 momentary present (he is being wounded), whereas French, lacking this 
 form, but possessing an indefinite pronoun on (they), preferably uses the 
 active and says on le blesse. 
 
374 ENGLISH AND GERMAN GRAMMAR. 
 
 inconvenience appears still greater even in English, than it is 
 in French, since he was wounded may occasionally mean, either 
 " they wounded him," or " he was hurt." German, on the 
 contrary, expressly distinguishes ich werde and ich ward ver- 
 wundet, literally " I become " and " I became wounded," from 
 ich bin and ich war verwnndet, " I am " or " was hurt," and thus 
 possesses, for every tense or mood, whether simple or peri- 
 phrastic, a passive conjugation which, in meaning, exactly 
 answers the same purpose as the Greek and Latin passive verb. 1 
 
 1 Cf . L. volnerdtur, and volnerdtus est or fuit. 
 
CONCLUSION". 
 
 (240) We have now come to the end of our task. Aftei 
 having first established, by means of nnmerous and convincing 
 illustrations, the essential correspondence between English and 
 German sounds, we have studied in detail the grammar of the 
 two languages in regard to word-formation, declension and 
 conjugation, and in all these points we have found them 
 identical. Wherever any violent contrasts have appeared, we 
 have found a reason for them in the fact that each language 
 has extended the use of one of two forms which both originally 
 possessed in common. Further, the analogies we have dis- 
 covered are not such as strike the eye at the first glance and 
 afterwards prove misleading ; they were drawn, as it were, 
 from the heart of the language and have been traced back, step 
 by step, from their present state to the source from which they 
 originally sprang. We have gone beyond Germanic, and have 
 verified them in the classical languages and even in the older 
 eastern tongues. 1 For pi'actical purposes these scientific results 
 may have but little value. A knowledge of them will perhaps 
 encumber rather than smooth the path of those who merely 
 desire the conversational fluency of the courier, and the master- 
 pieces of literature can be equally well understood and enjoyed 
 by students who have not dissected the elements of which these 
 masterpieces are composed, just as a knowledge of botany is 
 not required for the Aesthetic appreciation of beautiful flowers. 
 
 1 The subject we have just left is most instructive and suggestive on this 
 point. Nowhere is the likeness clearer and more striking than in the com- 
 plex structure of the Anglo-German and Indo-European verb ; nowhere do 
 the different paths lead us more certainly to the centre from which they 
 diverge. 
 
 875 
 
376 ENGLISH AND GERMAN GRAMMAR. 
 
 But the scientific mind, which likes to penetrate into the reason 
 of things, will derive both profit and pleasure from being able 
 to trace in the history and evolution of a language the reasons 
 for its most minute and often apparently contradictory rules. 
 The enthusiastic student — and what is science without en- 
 thusiasm ? — will find no more fascinating field for his enquiries 
 than among the scientific methods and accurate documents of 
 language. He cannot but admire the unbroken tradition by 
 which our lips still retain the words of ancestors who have left 
 no other trace behind ; and he will equally marvel at the genius 
 of those who have re-forged, one by one, the links of a chain 
 which has its beginning in the mists of the past, and which 
 passes on to the future inextricably mingled with the destiny 
 of mankind. 
 
INDEX OF WORDS. 
 
 N.B. — This index does not include, as a general rule : (1) 
 secondary and farther derivatives, which must be sought under 
 their respective endings in the index of terminations below ; 
 (2) such compounds as may be resolved at first sight, though 
 quoted in the Chapter of Composition (112-118) ; (3) nominal 
 forms other than the nominative singular, and verbal forms 
 other than the infinitive, except in cases which have, a special 
 interest. 
 
 The references are to the sections (1-240). 
 
 
 I. English. 
 
 , 
 
 A... 19,66,121,134 
 
 angle ... 
 
 89 
 
 awkward Ill 
 
 a- 50, 96, 98, 108, 229 
 
 Angle ... 
 
 23 
 
 
 abide 179 
 
 another 
 
 114 
 
 Bag 17 
 
 above 54 
 
 answer 
 
 .. 96,139 
 
 bake 50, 57, 82, 184 
 
 accrue 17 
 
 ant 
 
 23 
 
 balk 17 
 
 ache (to) 107 
 
 any ... 
 
 .. 31,134 
 
 band 45 
 
 acre 30, 34, 50, 57, 76 
 
 apple ... 
 
 146 
 
 bare 17 
 
 adder 49 
 
 apron ... 
 
 49 
 
 batch 50 
 
 ado 96 
 
 arbalist 
 
 65 
 
 bath 17 
 
 afford 97 
 
 arch- ... 
 
 96 
 
 bathe 17 
 
 after 117 
 
 arise ... 
 
 98 
 
 be 217,230,231,232, 
 
 again 96 
 
 arm 
 
 .. 39,139 
 
 239 
 
 against 96 
 
 as 
 
 66 
 
 be- ...17,19,65,99 
 
 akin 96 
 
 ash 
 
 74 
 
 bear (to) 26, 82, 174, 
 
 alderliefest ... 155 
 
 ask 
 
 .. 50,84 
 
 182 
 
 alike 96 
 
 asunder 
 
 40 
 
 beard 21,48 
 
 all 17,117 
 
 at 
 
 96 
 
 beat 185 
 
 alone 117 
 
 aught... 
 
 96 
 
 become 99 
 
 along 96 
 
 await ... 
 
 103 
 
 beech 33, 48, 50, 56, 
 
 also 66 
 
 awake... 
 
 30, 106, 184 
 
 139 
 
 am (I) 217 
 
 awaken 
 
 106 
 
 beget 182 
 
 an 19, 66, 121, 134 
 
 aware ... 
 
 96 
 
 begin 181 
 
 and 66 
 
 away ... 
 
 96 
 
 behead 99 
 
 angel 39 
 
 awful ... 
 
 110 
 
 belief 32 
 
 377 
 
378 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 believe 
 
 
 19,32 
 
 bellows 
 
 
 ... 146 
 
 bend . . . 
 
 
 ... 45 
 
 beseech 33, 50, 99, 187 
 
 best ... 
 
 
 ... 80 
 
 better 49, 
 
 61, 
 
 80,106 
 
 better (to) 
 
 
 ... 106 
 
 beyond 
 
 
 ... 160 
 
 bid ... 
 
 
 83,183 
 
 bind 26, 28, 45 
 
 4b, 49, 
 
 
 56, 
 
 82, 181 
 
 birch ... 
 
 
 ... 17 
 
 birth ... 
 
 
 ... 17 
 
 bishop. . . 
 
 
 ... 17 
 
 bit 17,48 
 
 49 
 
 56, 57, 
 72, 179 
 
 bite 27, 48 
 
 49 
 
 56, 57, 
 82,179 
 
 blast ... 
 
 #- 
 
 ... 185 
 
 bleed ... 
 
 
 ... 33 
 
 blind ... 
 
 
 17,49 
 
 blood ... 
 
 
 ... 33 
 
 bloom . . . 
 
 
 33,75 
 
 blossom 
 
 
 ... 83 
 
 blot ... 
 
 
 ... 187 
 
 blow . . . 
 
 48, 5C 
 
 bolt ... 
 
 
 ... 17 
 
 bond ... 
 
 
 ... 45 
 
 bone ... 
 
 
 17,48 
 
 book 33 
 
 48 
 
 , 50, 56, 
 139 
 
 born ... 
 
 t< 
 
 ... 182 
 
 borough 
 
 50, 
 
 65, 139 
 
 both ... 
 
 
 ... 81 
 
 bottom 
 
 
 39,75 
 
 bough... 
 
 
 ... 110 
 
 bow ... 
 
 
 ... 50 
 
 bow (to) 
 
 "i, 
 
 50, 110, 
 180 
 
 boy ... 
 
 
 ... 33 
 
 boycott 
 
 
 ... 107 
 
 brake (pf .) 
 
 .. 
 
 100,182 
 
 breadth 
 
 
 ... 90 
 
 break ... 
 
 
 42, 182 
 
 breakfast 
 
 
 ... 117 
 
 breed ... 
 
 
 17,23 
 
 brethren 
 
 
 ... 140 
 
 bride .. 
 
 
 ... 48 
 
 bridegroon 
 
 L 
 
 ... 65 
 
 bring ... 
 
 
 24, 187 
 
 broad ... 
 
 
 ... 90 
 
 brood ... 
 
 "l7 
 
 ,18,23 
 
 brother 18 
 
 33 
 
 41, 48, 
 
 49, 51, 
 
 53, 
 
 54, 139 
 
 brow . . . 
 brown . . . 
 buck . . . 
 buckmast 
 bull ... 
 burial . . . 
 burn . . . 
 burst ... 
 bush . . . 
 bustle ... 
 busy . . . 
 but ... 
 butcher 
 butter... 
 buxom 
 buy 
 by y ... 
 
 Cab ... 
 calf ... 
 call ... 
 can 
 
 care . . . 
 cast 
 cat 
 
 catch ... 
 caterwaul 
 cave . . . 
 chalk ... 
 chew . . . 
 chide ... 
 
 ... 117 
 
 ... 18 
 
 ... 48 
 
 ... 83 
 
 17,48 
 
 ... 104 
 
 ... 187 
 
 17, 181 
 
 ... 17 
 
 ... 23 
 
 17,23 
 
 12,66 
 
 ... 17 
 
 ... 5 
 
 10, 19, 110 
 
 ,.. 83,187 
 
 17, 66, 99 
 
 , 17 
 
 48, 56, 139 
 
 54 
 
 39, 84. 224 
 .. 13,139 
 
 187 
 
 102 
 
 187 
 
 102 
 
 17 
 
 50 
 
 36 
 
 179 
 
 child 18, 139, 140, 147 
 
 chill 50 
 
 chin 50 
 
 choice 29 
 
 choose 29, 50, 57, 61, 
 82,180, 194,204 
 
 circle 
 clad . 
 clasp . 
 clean . 
 cleanse 
 cleave (split) 
 cleave (adhere) 
 climb . 
 cling . 
 clothe . 
 co- 
 
 17 
 
 .. 187 
 
 .. 106 
 
 17,77 
 
 .. 106 
 
 .. 180 
 
 82 
 
 82, 181 
 
 181 
 
 187 
 
 67 
 
 cold 17; 21, 50, 57, 78 
 
 colour ... 5 
 
 comb 39,48 
 
 come 89, 45, 50, 57, 
 
 182 
 
 cool 50 
 
 corn .., 
 cost .., 
 could ... 
 cow ...14 
 coxcomb 
 creed ... 
 cripple 
 crow (to) 
 cut (to) 
 
 Daisy ... 
 dale ... 
 danger 
 dare . . . 
 darling 
 
 18, 
 
 50,57 
 187 
 224 
 
 50,57 
 114 
 139 
 88 
 185 
 187 
 
 114 
 
 14 
 65 
 1»7, 222 
 ... 103 
 
 daughter 49, 56, 139 
 day 47, 50, 56, 72, 139, 
 143 
 dead 32,78 
 
 deaf . . 
 deal .. 
 death .. 
 deed 26, 
 deem . . 
 deep . . 
 deer . . 
 defile .. 
 depth .. 
 dew 
 die (a).. 
 dig .. 
 dip 
 dis- 
 
 dish .. 
 dive . . 
 dizzy 
 
 doom 
 
 door 
 
 dough 
 
 dove 
 
 down 
 
 drag 
 
 draw 
 
 dream 
 
 drink 
 
 32 
 .. 96 
 32,47 
 45, 47, 78, 218 
 ... 23,107 
 29, 57, 90 
 49, 189, 143 
 ... 23 
 ... 90 
 ... 17 
 ... 145 
 ... 179 
 29 
 100, 107 
 ... 50 
 ... 29 
 17,23 
 45, 56, 218, 233, 
 236 
 23,75 
 17,28 
 50, 56, 65 
 ... 82 
 ... 18 
 ... 50 
 17, 50, 184 
 49 
 117 
 
 drink (to) 39, 40, 181 
 drive (to) 45, 49, 179 
 due 17 
 
 dumb . 
 dung . 
 durst . 
 Dutch. 
 
 17,28 
 
 ... 17 
 
 ... 187 
 
 2.29 
 
INDEX OF ENGLISH WORDS. 
 
 379 
 
 Ear 32, 
 
 61, 74, 139 
 
 fight (to) 
 
 181 
 
 early ... 
 
 124 
 
 file (to) 
 
 23 
 
 earth 13 
 
 17, 20, 21, 
 
 fill ... 
 
 17, 23, 107 
 
 
 73, 139 
 
 filth .. 
 
 23, 70, 90 
 
 Easter... 
 
 152 
 
 find 19 
 
 20, 28, 181 
 
 eat 
 
 .. 26,183 
 
 fine 
 
 12 
 
 eatable(s) 
 
 .. 69,157 
 
 finger 26, 
 
 39, 50, 121 
 
 edge . . . 
 
 50 
 
 fire 
 
 29 
 
 egg ... 
 
 50 
 
 first 17, 
 
 23, 49, 124 
 
 eight ... 
 
 30, 53, 121 
 
 fish 27,48,50,54,72 
 
 either ... 
 
 79 
 
 fish (to) 
 
 ... 19,174 
 
 eke 
 
 32 
 
 fisher ... 
 
 102 
 
 elbow ... 
 
 .. 10,114 
 
 fist 
 
 49 
 
 elder ... 
 
 80 
 
 five 
 
 20. 54, 121 
 
 eldest ... 
 
 80 
 
 flea ... 
 
 ... ... 139 
 
 eleven 
 
 121 
 
 flee ... 
 
 187 
 
 else 
 
 .. 23,111 
 
 fleet ... 
 
 84 
 
 'em 
 
 66 
 
 flesh ... 
 
 ... 48,50 
 
 en- 
 
 107 
 
 flight ... 
 
 ... 49,78 
 
 end 
 
 .. 22, 139 
 
 fling ... 
 
 181 
 
 English 
 
 .. 23,91 
 
 flood ... 
 
 ... 33,78 
 
 enough 2C 
 
 1, 50, 55, 96 
 
 floor . . . 
 
 17, 76, 139 
 
 enow ... 
 
 50 
 
 flower., t 
 
 33 
 
 ere 
 
 124 
 
 flutter 
 
 106 
 
 even . . . 
 
 10, 17, 48 
 
 fly (a)... 
 
 50 
 
 ever . . . 
 
 161 
 
 fly (to) 
 
 ... 50,180 
 
 eye 50, 
 
 54, 74, 139 
 
 fodder... 
 
 ... 5, 49, 79 
 
 
 
 fold ... 
 
 110 
 
 Fair ... 
 
 13, 50, 76 
 
 folk ... 
 
 139 
 
 falcon . . . 
 
 13 
 
 follow . . . 
 
 50 
 
 fall (to) 17 
 
 21, 83, 84, 
 
 food . . . 
 
 23 
 
 
 185 
 
 foot 18, 
 
 33, 72, 139, 
 
 false ... 
 
 ... 13,17 
 
 
 144, 145 
 
 falsehood 
 
 109 
 
 for 
 
 99 
 
 far 
 
 ... 17,21 
 
 for- 
 
 ... 65,99 
 
 fare (to)17, 
 
 82, 83, 184 
 
 f ore 
 
 23 
 
 farther 
 
 17 
 
 fore- ... 
 
 99 
 
 farthing 
 
 103 
 
 foremost 
 
 124 
 
 fasten . . . 
 
 93 
 
 forget . . . 
 
 182 
 
 father 17 
 
 , 25, 48, 49, 
 
 forlorn 
 
 19, 45, 180 
 
 51-, 
 
 53, 54, 139 
 
 form . . . 
 
 5 
 
 fatherless 
 
 10 
 
 former 
 
 124 
 
 fathom 
 
 39 
 
 forsake 
 
 184 
 
 feather 
 
 23, 49, 54 
 
 forth ... 
 
 49 
 
 fee 
 
 34, 51, 54 
 
 foster ... 
 
 79 
 
 feed ... 
 
 23 
 
 foul 23, 2£ 
 
 , 50, 70, 76, 
 
 feel ... 
 
 187 
 
 
 90 
 
 feet ... • 
 
 .. 144,145 
 
 four ... 
 
 121 
 
 fell ... 
 
 26 
 
 fowl ... 
 
 .... 41,48 
 
 fell (to) 
 
 83 
 
 fox 
 
 23 
 
 fiend 22, 
 
 54, 90, 139 
 
 France 
 
 23 
 
 fifteen . . . 
 
 48 
 
 freeze ... 
 
 61, 180, 204 
 
 fifth ... 
 
 20, 49, 124 
 
 French 
 
 23 
 
 fifty ... 
 
 20, 48, 122 
 
 fresh ... 
 
 50 
 
 S* 
 
 fret 99 
 
 friend 22, 41, 54, 90, 
 139 
 
 frost 61 
 
 fulfil 114 
 
 full ...17,23,42,77 
 
 fulness 5 
 
 funk 17 
 
 furrow 42 
 
 further 79 
 
 Gander 102 
 
 gang 219 
 
 gangway 219 
 
 garden 30 
 
 gardener 102 
 
 Gardiner 102 
 
 garlic 32 
 
 geese 144, 145 
 
 get ... 50,182,183 
 
 ghost 50,62 
 
 gift ... 48, 50, 78 
 gild 23, 50, 107, 187 
 
 gird 106,187 
 
 girdle 106 
 
 give 45, 48, 50, 183 
 
 glass 20,61 
 
 glazier 14, 50, 61 
 
 glimmer ... 106,181 
 
 glitter 106 
 
 go 219 
 
 God 28,50 
 
 godfather 114 
 
 gold 17, 20, 23, 49, 50 
 
 good 50 
 
 goods 157 
 
 goose 18, 56, 102, 144, 
 
 145 
 
 gosling ... 18,103 
 
 Gospel 115 
 
 grasp 106 
 
 great 82 
 
 grind 181 
 
 groom 65 
 
 grow 185 
 
 guest 30, 50, 51, 56, 
 
 139, 144 
 
 guilt 50 
 
 Hail 50,54 
 
 hair 50,143 
 
 half ... 21, 48, 124 
 halm 75 
 
 haM-39, 49, 50,72, 139 
 *P8B r UD84^ 
 
380 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 handicraft ... 96 
 handiwork ... 96 
 
 handsome 110 
 
 hang ... 39, 83, 185 
 
 hard 17, 20, 50, 54, 72, 
 
 110 
 
 hare . . 
 hark . . 
 harvest 
 haste .. 
 hate .. 
 hater .. 
 hatred 
 have 
 
 ... 61 
 ... 20 
 ... 41 
 ... 17 
 50,93 
 ... 102 
 109 
 
 17, 19, 85, 187, 
 189, 231 
 he ... 17, 66, 166 
 head ... 19,32,187 
 
 headlong Ill 
 
 heal 90,93 
 
 hear 187 
 
 heart . 20, 21, 49, 74, 
 
 140 
 
 hearth 21 
 
 heath ... 
 heathen 
 heave 54, 
 hedge ... 
 height 
 heir ... 
 help (to) 
 her 
 
 herd ... 
 here . . . 
 hers . . . 
 hew ... 
 hide . . . 
 high ... 
 hill ... 
 hireling 
 his 
 
 . ... 137 
 . ... 19 
 J, 182, 184 
 . ... 50 
 . ... 90 
 . ... 74 
 . 45,181 
 . 17,168 
 . ... 21 
 . 166,168 
 . ... 168 
 . ... 185 
 . ... 179 
 50, 54, 90 
 . 50,88 
 . ... 103 
 . ... 168 
 
 hold 17, 21, 50, 54, 84, 
 185 
 ... 50 
 ... 65 
 
 holy ... 
 home ... 
 hoof . . . 
 horn . . . 
 horse ... 
 hot ... 
 hound . . . 
 house ... 
 hue 
 
 hundred 
 hunger 
 hunt ... 
 
 18, 
 
 . ... 17 
 
 . 42, 77 
 . ... 143 
 . ... 31 
 . 20,28 
 20, 28, 32 
 . 5, 17, 37 
 . 50,123 
 . ... 50 
 
 hurt . . . 
 husband 
 hussy .. 
 hut .. 
 
 I 
 
 ice 
 idle .. 
 
 if 
 
 in 
 in- 
 ink 
 
 ink (to) 
 is (he) . . 
 
 it 
 
 its 
 
 Keep .. 
 ken 
 kernel 
 kin 
 
 kindred 
 king .. 
 kiss 
 knave . . 
 knee . . 
 kneel .. 
 knight 
 know ... 
 knowledge 
 
 Lade (to) 
 lady ... 
 lamb ... 
 lame ... 
 land ... 
 landscape 
 last 
 
 last (to) 
 late ... 
 laugh ... 
 laughter 
 lay (to) 
 lay (he) 
 lead . . . 
 leaf ... 
 lean ... 
 leap . . . 
 learn 20, 
 
 lease ... . 
 
 leave ... . 
 
 leek ... . 
 
 lend ... . 
 
 ,. ... 187 
 5, 18, 114 
 65 
 17 
 
 164 
 145 
 110 
 66 
 66 
 96 
 13 
 107 
 .. 26,217 
 .. 66,166 
 168 
 
 .. 20,187 
 .. 39,84 
 .. ... 88 
 .. 50,96 
 .. 5,109 
 .. 50,103 
 
 50 
 
 17 
 
 .. 50,57 
 .. 106,187 
 
 21 
 
 ..1,84,185 
 .. 50,109 
 
 .. 50,184 
 
 65 
 
 .. 39,139 
 
 10 
 
 .. 89,49 
 
 109 
 
 80 
 
 187 
 
 80 
 
 .. 36,50 
 .. 50,79 
 .. 50,83 
 
 183 
 
 19, 20, 187 
 
 32 
 
 187 
 
 32, 50, 185 
 
 21, 31, 61, 
 
 106 
 
 61 
 
 20 
 
 32 
 
 82 
 
 length 
 let 
 
 lick .. 
 lid .. 
 lie (a) .. 
 lie (to) .. 
 lie (be laid) 
 
 90 
 
 ... 45,236 
 ... 27,50 
 
 117 
 
 17, 50, 139 
 
 180 
 
 50,83, 
 183 
 
 lief 48,56 
 
 life 48 
 
 light ... 29,41,50 
 
 light (to) 187 
 
 light (adj.) 50 
 
 like 
 
 lit 
 
 little 
 
 live 
 
 liver 
 
 load 
 
 loaf 
 
 loan 
 
 lone 
 
 long 
 
 loose 
 
 lord 
 
 lord (to) 
 
 lording 
 
 96 
 
 187 
 
 88 
 
 27.' 48, 85 
 
 27 
 
 184 
 
 ,. 17,65 
 
 82 
 
 117 
 
 39, 50, 90 
 .. 17,110 
 
 65 
 
 .. ... 118 
 103 
 
 lose 20, 45, 61, 180, 187 
 
 204 
 
 loss 45 
 
 loud "41, 50, 78, 110 
 
 louse 144, 145 
 
 love ... 48, 56, 107 
 low 50 
 
 Maid ... 17,19,50 
 maiden ... 19,114 
 make ... 20, 50, 187 
 man 37, 39, 48, 144, 
 145 
 
 many 74 
 
 mare 10,65 
 
 marshal 65 
 
 may ... 78,226,234 
 
 me 17,164 
 
 mead 39 
 
 mean (adj.) ... 96 
 
 mean (to) 187 
 
 meat 65 
 
 meed 62,185 
 
 meet 187 
 
 melt ...174,181,189 
 mere ... 17,19,72 
 mermaid 17 
 
INDEX OF ENGLISH WOEDS. 
 
 881 
 
 merry . . . 
 methinks 
 mice ... 
 mid ...26, 
 middle 
 midwife 
 might... 
 might (he), 
 mild ... 
 mile ... 
 mill ... 
 mine . . . 
 minster 
 mint ... 
 mirth ... 
 mis- ... 
 mischief 
 miss (to) 
 mist ... 
 molten 
 Monday 
 month... 
 mood ... 
 moon ... 
 more . . . 
 morn ... 
 morning 
 most ... 
 
 mount... 
 mouse 23, 
 mouth... 
 move ... 
 mow . . . 
 murder 
 murther 
 music . . . 
 must ... 
 mutton 
 my ... 
 mylady 
 mylord 
 
 Nail ... 
 name ... 
 napkin 
 nation 
 near . . . 
 need ... 
 need (to) 
 needle... 
 needs ... 
 neigh ... 
 
 90 
 
 neighbour 
 
 .. 50,65 
 
 pinafore 117 
 
 238 
 
 neither 
 
 .. 79,96 
 
 pink 13 
 
 ... 144,145 
 
 nephew 
 
 54 
 
 plant 48 
 
 », 36, 39, 56 
 
 nest 
 
 .. 27,62 
 
 play (to) 182 
 
 .. 10,49 
 
 net 
 
 17 
 
 plenty 5 
 
 117 
 
 never ... 
 
 96 
 
 poison 48 
 
 ... 50,78 
 
 next ... 
 
 80 
 
 pound 18, 48, 143 
 
 226 
 
 nigh ... 50,80,110 
 
 pretty 110 
 
 20 
 
 night 21, 39, 50, 54, 79 
 
 push 189 
 
 17 
 
 nightingale ... 114 
 
 put 17,187 
 
 42 
 
 nine 121 
 
 
 168 
 
 no 134 
 
 Queen 72 
 
 65 
 
 none 96,134 
 
 queer 17 
 
 49 
 
 nor 96 
 
 quick 27,50 
 
 90 
 
 nose 61 
 
 quicksilver ... 27 
 
 96 
 
 not 96 
 
 quite 117 
 
 96 
 
 note 10 
 
 quoth 183 
 
 96 
 
 nought 96 
 
 
 27 
 
 nut 50 
 
 Eadish 23 
 
 181 
 
 
 rain 50 
 
 18 
 
 Of 54,66 
 
 raise 83 
 
 90 
 
 off 54,66 
 
 rare 10 
 
 49 
 
 oft 49 
 
 rash 50 
 
 ... 18,139 
 
 old ... 17, 20, 21, 80 
 
 rather 17 
 
 ... 61,80 
 
 on 96,103 
 
 raven ... 13, 50, 54 
 
 ... 50,54 
 
 once 110, 124 
 
 re- 67 
 
 ... 50,54 
 
 one 31, 39, 66, 121, 157 
 
 reach 187 
 
 80 
 
 ordeal 96 
 
 read 17, 20, 41, 185, 
 
 3, 19, 33. 49 
 
 other ... 49, 124, 157 
 
 187 
 
 54, 139 
 
 ought 96 
 
 ready 109 
 
 26 
 
 ought (he)... 228,229 
 
 reason 65 
 
 31, 144, 145 
 
 our 168 
 
 red 41 
 
 40 
 
 out 66,96 
 
 reeve 65 
 
 17 
 
 oven 48 
 
 relations 5 
 
 ... 50,185 
 
 over 54 
 
 rend 187 
 
 42 
 
 own ... 50,54,228 
 
 rest 73 
 
 42 
 
 ox 17, 74. 139, 140 
 
 rich 50 
 
 17 
 
 Oxford 114 
 
 riddle 104 
 
 227 
 
 oxlip 117 
 
 ride 41,179 
 
 5,65 
 
 
 Riding 68 
 
 168 
 
 Pale (noun) ... 48 
 
 right 53,78 
 
 66 
 
 palfrey ... ... 48 
 
 ring (a) 50 
 
 66 
 
 pan 
 
 ... 23,48 
 
 ring (to) ... 13,181 
 
 
 pass . . . 
 
 187 
 
 rinse 106 
 
 50 
 
 past . . . 
 
 187 
 
 ripe (to) 107 
 
 75 
 
 path ... 
 
 .. 49,139 
 
 ripen ... 106 
 
 103 
 
 peacock 
 
 .. 48,115 
 
 rise 83,179 
 
 65 
 
 pen (to) 
 
 107 
 
 rive 179 
 
 50 
 
 pence ... 
 
 145 
 
 road 17 
 
 228 
 
 penny 
 
 .. 23, 145 
 
 room ...17,28,39,75 
 
 228 
 
 people... 
 
 143 
 
 rough 50, 54 
 
 26, 39, 79 
 
 person... 
 
 17 
 
 roundabout ... 117 
 
 Ill 
 
 pilgrim 
 
 23 
 
 rule 17 
 
 50 
 
 pillow 
 
 17 
 
 rummage 17 
 
382 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 run 181 
 
 Sad ... 
 saddle... 
 sail 
 
 sailor ... 
 sale 
 
 59 
 49 
 50 
 102 
 21 
 
 salt ... 30,49,59 
 salve ... 92,93, 189 
 
 same 110 
 
 sand 39 
 
 save 189 
 
 say 50 
 
 screw 17 
 
 second 124 
 
 see 17,45,50,54,183 
 
 seed ... 26,49,59 
 
 seek 33, 50, 187, 189, 
 
 207 
 
 seethe 53,180 
 
 seldom 23 
 
 self ... 21, 160, 167 
 
 sell 21,187 
 
 send 187 
 
 serve 17 
 
 set 17," 26, 36, 83, 187 
 seven ... 26,40,121 
 
 sew 59 
 
 shade 189 
 
 shaft 48 
 
 shake 184 
 
 Shakespeare 114, 117 
 
 shall 225,232 
 
 shamefaced ... 110 
 
 shape (to) 48, 50, 57, 
 184 
 
 sharp 20 
 
 sharpen 106 
 
 shave 30, 48, 56, 57, 
 82,184 
 
 she 133,166 
 
 she- 115 
 
 shear 182 
 
 shed 187 
 
 sheen .. ... 77 
 
 sheep "5, 50, 139, 143 
 
 shepherd 87 
 
 sheriff 65 
 
 shew 185 
 
 shilling 103 
 
 shine 18, 53, 84, 173, 
 
 179 
 
 ship 50 
 
 ship (to) 
 shire ... 
 shoe . . . 
 shoe (to) 
 shoot ... 
 
 107 
 65 
 50 
 
 187 
 180, 204 
 
 should 225,235 
 
 shove 50 
 
 shovel... 48,50,88 
 
 show 59,185 
 
 shrine 59 
 
 shrive 179 
 
 sight 50,78 
 
 silly 23 
 
 silver 21,59 
 
 sin 17 
 
 sing ... 181 
 
 sink ... 39,82,181 
 sister 10, 13, 59, 61, 
 
 139 
 sit 26, 59, 83, 183 
 
 six 21,121 
 
 slaughter ... 50, 79 
 slay ... 45,59,184 
 sleep ... 20,48,57 
 
 slide 17 
 
 slight 78 
 
 slip 59 
 
 slow 31 
 
 small 59 
 
 smile 59 
 
 smite 179 
 
 snow 37, 39, 50, 56, 59 
 
 so 66 
 
 sob 106 
 
 sodden 53 
 
 soft 48 
 
 some 40,134 
 
 son 20, 34 r 39, 59, 77, 
 
 139 
 
 soothly 110 
 
 sore 31 
 
 sorrow 93 
 
 soul 81 
 
 sound (adj.) ... 20 
 
 sour 32,76 
 
 sow 27 
 
 sow (to) ... 59,185 
 
 speak 17,182 
 
 spear 48,139 
 
 spew 48, 53 
 
 spin 48,181 
 
 spindle 88 
 
 spinster 90 
 
 spit 183 
 
 split ... 179 
 
 spouse 5 
 
 spring (to)... 82,181 
 
 stand (a) 49 
 
 stand (to) 53, 184, 220 
 
 star 21 
 
 starch 50 
 
 stark 50 
 
 starve 181 
 
 stay 220 
 
 steadfast 110 
 
 steal 182 
 
 steed 33 
 
 stick (to) ... 179,183 
 
 stiff 48 
 
 stink 181 
 
 stitch (to) 83 
 
 stone 105 
 
 stood (he) 184 
 
 stool 33 
 
 stove 20 
 
 stream . . . 59, 75 
 
 strength 90 
 
 strew 185 
 
 stride 59 
 
 strike 50,179 
 
 string 181 
 
 strive ... 179 
 
 strong 90 
 
 strow 185 
 
 stud 33 
 
 sty ... 26, 53, 82 
 
 such 50,161 
 
 suit 17 
 
 Sunday 117 
 
 swamp 59 
 
 swart 59 
 
 swear 182 
 
 sweat ... 49, 57, 59 
 sweet ... 49, 57, 59 
 swell ... 28,174,181 
 swim 17, 39, 59, 181 
 swine 27, 59, 89, 143 
 sword 21,59 
 
 Take ... 
 tale 
 
 tallow... 
 tame ... 
 taught 
 teach ... 
 tear (a) 
 tear (to) 
 teens ... 
 
 ... 32,184 
 ... 23,49 
 
 50 
 
 49 
 
 187 
 
 187 
 
 ... 30,76 
 41, 82, 182 
 121 
 
INDEX OF ENGLISH WORDS. 
 
 383 
 
 tell ... '23, 49, 187 
 ten 40, 47, 54, 57, 121 
 
 thank 50 
 
 that 46, 47,49, 66, 130, 
 
 160, 162 
 thatch 36, 57, 59 
 
 the 47, 49, 66, 133 
 
 thee ... 165 
 
 their 49,168 
 
 them 66,166 
 
 there 168 
 
 these 160 
 
 they 4), 166 
 
 thigh ... , 50 
 
 thimble 88 
 
 thin 47,54 
 
 thine 168 
 
 thing 24 
 
 think 24, 47, 50, 83, 
 
 187, 238 
 
 third 124 
 
 thirst 42 
 
 thirteen 121 
 
 thirty 122 
 
 this 160 
 
 thorn... 47,54,77 
 
 those 160 
 
 thou ... 47,49,165 
 
 though 14 
 
 thousand 19, 28, 49, 
 123 
 
 thrash 181 
 
 three 121 
 
 thrice 124 
 
 through ... 49, 50 
 
 throw 185 
 
 thumb 18, 47, 54 
 
 Thursday 114 
 
 thy 168 
 
 tide 49,139 
 
 tight 78 
 
 tile 50 
 
 timber 47, 57, 76 
 
 tin 49 
 
 tithe 124 
 
 to 66,100 
 
 toe 50 
 
 token 77 
 
 to-morrow 50 
 
 tong 39 
 
 tongue 13,39,47,50, 
 54, 74, 139 
 
 too 66 
 
 tooth ... 79,144,145 
 
 tother... 
 
 66 
 
 town ... 
 
 18 
 
 tread ... 
 
 ... 182,183 
 
 tree 
 
 37 
 
 trow . . . 
 
 37 
 
 true . . . 
 
 37 
 
 tub ... 
 
 17 
 
 Tuesday 
 
 17, 37, 114 
 
 tune . . . 
 
 17 
 
 twelve 
 
 121 
 
 twenty 
 
 ... 121,122 
 
 twice ... 
 
 .. 110,124 
 
 twig ... 
 
 50 
 
 twilight 
 
 121 
 
 twin ... 
 
 ... 121,124 
 
 twist ... 
 
 124 
 
 two ... 
 
 37, 49, 121 
 
 Udder 28 
 
 un- (negative) 17, 40, 
 
 96 
 
 un- (inversive) 98, 100 
 
 uncouth 78 
 
 under 40, 49 
 
 undo 10,98 
 
 up 66 
 
 us 20,164 
 
 Vat ... . 
 
 23 
 
 very ... . 
 
 110 
 
 view ... . 
 
 17 
 
 virgin 
 
 17 
 
 vixen ... . 
 
 .. 23, 87 
 
 vouchsafe . 
 
 118 
 
 Wait ... . 
 
 103 
 
 wake ... 
 
 36, 50, 184 
 
 Wales... . 
 
 .. 23,67 
 
 wall ... . 
 
 139 
 
 walnut 
 
 115 
 
 warm 21 
 
 , 50, 56, 75 
 
 warm (to) . 
 
 106 
 
 warp . . . 
 
 48, 57, 181 
 
 warrant 
 
 97 
 
 was (he) 
 
 17, 20, 61, 
 
 
 183, 217 
 
 wash ... . 
 
 .. 17,84 
 
 watch (to) 
 
 50 
 
 water ... 5 
 
 , 17, 37, 49 
 
 wax ... . 
 
 50 
 
 wax (to) . 
 
 184 
 
 way ... . 
 
 50 
 
 we 
 
 17. 63, 164 
 
 wear ... . 
 
 .. ' ... 182 
 
 weasel 88 
 
 weather ... 17, 49 
 
 weave 182,183 
 
 wed 187 
 
 wedlock 109 
 
 weep 187 
 
 weigh 206 
 
 Welsh... 23,67,91 
 
 wend 219 
 
 went (he) 219 
 
 were (he) ... 194,217 
 were (they)61,183,217 
 werwolf 27, 115, 117 
 what ... 50,54,161 
 
 wheat 50 
 
 when 66 
 
 where 168 
 
 whet 50, 54 
 
 which 161,162 
 
 while 17 
 
 white... 37,50,54 
 who ... 17,161,162 
 
 whole 90,93 
 
 widow 27 
 
 width 90 
 
 wife 48,139 
 
 wild 17,49 
 
 will 36,139 
 
 will (he) 17,221,228, 
 
 232 
 
 win 97,181 
 
 wind ... 17,26,37 
 wind (to) 17, 181, 219 
 
 wine 17,37 
 
 wise (adj.) 223 
 
 wise (manner) ... Ill 
 
 wish 84,189 
 
 wit 223 
 
 with 79 
 
 witness 223 
 
 wolf 34, 37, 42, 48, 54, 
 
 139 
 woman ... 115, 145 
 
 womb 48 
 
 wood 41 
 
 word 20,139 
 
 word (to) 107 
 
 work 21, 50, 53, 83, 187 
 world... 21,27,65 
 
 worm 139 
 
 worship 109 
 
 worship (to) ... 118 
 
 worth 21 
 
 wot (he) 81, 37, 45, 223 
 
384 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 would (he).. 
 
 . 221,235 
 
 wound 
 
 . ... 78 
 
 write 
 
 . 37,179 
 
 writhe 
 
 . ... 179 
 
 wrought . . 
 
 . 53, 187 
 
 Y- 
 
 . ... 97 
 
 yard 
 
 . 30, 74 
 
 Ab 54 
 
 abenteucr 6 
 
 aber 117 
 
 aberglaube 117 
 
 abgott 147 
 
 achsel 76 
 
 acht (eight) 30, 53, 121 
 
 acht (care) 93 
 
 achten 93 
 
 acker 30, 34, 50, 57, 
 7b, 146 
 
 ader 142 
 
 adler 146 
 
 afterwort 117 
 
 all 17 
 
 allein 117 
 
 allod 115 
 
 als 66 
 
 also 66 
 
 alt ... 17,20,21 
 
 am 66 
 
 ameise 23 
 
 ammer 65 
 
 an 96 
 
 andacht 78 
 
 ander 49,124 
 
 anders 124 
 
 anderthalb 124 
 
 angel 39 
 
 ant- 96 
 
 apfel 146 
 
 arbeit 142 
 
 arm 39,143 
 
 armbrust 65 
 
 armut 90 
 
 aache 74 
 
 ast 30,146 
 
 au ... 142 
 
 audi 82 
 
 aue 142 
 
 auf 66 
 
 auge 50, 54, 74, 140, 142 
 
 ye ... 165 
 
 yon 
 
 ... 160 
 
 year ... 36,76,143 
 
 yonder 
 
 ... 160 
 
 yearn 50 
 
 York 
 
 ... 54 
 
 yellow... 37, 50, 56 
 
 you 
 
 36, 165 
 
 yellowhammer ... 65 
 
 young 
 
 28,36 
 
 yesterday 41, 50, 56 
 
 youngster ... 
 
 ... 90 
 
 yield 50,125 
 
 your 
 
 ... 168 
 
 yoke 28, 36, 72, 139 
 
 youth 
 
 ... 90 
 
 II. German. 
 
 
 
 aus 66, 96 
 
 besuchen 
 
 ... 187 
 
 ausdruck 6 
 
 beten 
 
 ... 117 
 
 
 betriegen 
 
 ... 180 
 
 Baar 17 
 
 beugen 
 
 ... 29 
 
 bachen 50.57 
 
 beuegen 
 
 182,206 
 
 backen 50, 57, 82, 184 
 
 biegen ... 2£ 
 
 , 50, 180 
 
 bad 17 
 
 biegsam 
 
 19, 110 
 
 bahre 26 
 
 bin 
 
 ... 217 
 
 balg 146 
 
 binden 26, 28, 45, 49. 
 
 balken 17 
 
 56 
 
 , 82, 181 
 
 band 45,147 
 
 birke 
 
 ... 17 
 
 bart ... 20,21,48 
 
 bis 
 
 ... 96 
 
 bauen 18,32 
 
 bischof. 
 
 ... 17 
 
 bauer (peasant). . . 141 
 
 bisthum 
 
 ... 109 
 
 baum 10 
 
 bisz 17, 48,49, 
 
 >6, 57,72 
 
 be- 17, 19, 65, 94, 99 
 
 bitten 
 
 83,183 
 
 beben 81 
 
 bldhen ... 48 
 
 , 56, 185 
 
 bedingen ... 24,181 
 
 blasen 
 
 ... 185 
 
 bedurfen ... 55, 224 
 
 blech 
 
 ... 14 
 
 befehlen ... 181,206 
 
 bleiben 19, 65, 85, 1' 
 
 begehren 23 
 
 
 229 
 
 begierde 23 
 
 blind ... 17 
 
 , 49, 155 
 
 beginnen 181 
 
 blilhen 
 
 33,36 
 
 begleiten 19 
 
 blume 33, 75, 
 
 137, 140, 
 
 behaflet 110 
 
 
 142 
 
 behende ... 22,99 
 
 blut 
 
 ... 33 
 
 bei 17,99 
 
 bluten 
 
 ... 33 
 
 beide 31 
 
 bock 
 
 ... 48 
 
 beil 79 
 
 boden 39, 75, 
 
 141, 146 
 
 bein ... 17,48,143 
 
 bogen 
 
 ... 50 
 
 beiszen 27, 48, 49, 56, 
 
 bolz 
 
 ... 17 
 
 57, 82, 179 
 
 bdse 
 
 ... 109 
 
 bekommen 99 
 
 bbsewicht . . . 
 
 ... 147 
 
 bellen 181,206 
 
 boshaft 
 
 ... 110 
 
 bequem 65 
 
 bosJieit 
 
 ... 109 
 
 beredt ... 17, 19, 187 
 
 bote 
 
 ... 141 
 
 bereit 109 
 
 braten 
 
 ... 185 
 
 berg 146 
 
 braue 
 
 ... 117 
 
 bergen... 104, 181,206 
 
 braun 
 
 ... 18 
 
 bersten... 17,181,206 
 
 braut 
 
 ... 48 
 
 besser ... 49,61,80 
 
 br&utigam ... 
 
 ... 65 
 
 best 80 
 
 brechen 42, 
 
 182,206 
 
INDEX OF GERMAN WORDS. 
 
 385 
 
 brennen 
 brenzeln 
 bringen 
 bruder 
 49, 
 brunnen 
 brut 
 
 brilten .. 
 bube 
 buch 33, 
 buche . . 
 buchstab 
 bucht . . 
 bulle . . 
 burg .. 
 busch . . 
 butter .. 
 
 Chor .. 
 christ .. 
 chur . . 
 
 187 
 
 106 
 
 ... 24,187 
 18, 33, 41, 48, 
 51, 53, 54, 146 
 
 141 
 
 . 17, 18, 23, 48 
 . 17, 23, 107 
 . ... 18,33 
 48, 50, 56, 147 
 33, 48, 56 
 . ... 33 
 . ... 136 
 . 17,48 
 . 50,65 
 . ... 17 
 . ... 5 
 
 . ... 146 
 . 141, 149 
 . ... 180 
 
 Da 168 
 
 dach ... 36, 50, 57, 59 
 
 dank 50 
 
 dar 168 
 
 das 130,162 
 
 dasz ... 46, 130, 162 
 dduchte ... 24,195 
 
 d&uchten 24 
 
 daumen 18, 47, 54, 141 
 decken 36, 50, 57, 59 
 
 degen 146 
 
 dehnen 36 
 
 dein 168 
 
 denken... 24, 50, 187 
 der ... 66,130,162 
 deuchte ... 24,28 
 Deutsch 23, 29, 91, 141 
 
 dicht 78 
 
 die ... 66,132,162 
 
 die (pi.) 131 
 
 Dienstag ... 17,114 
 
 dies ... 160 
 
 dieser 160 
 
 Dietrich 29 
 
 ding 24 
 
 dingen 24,181 
 
 dime 149 
 
 dock 14 
 
 Donnerstag ... 114 
 darn 47, 54, 77, 147 
 
 drehen 185 
 
 drei 121 
 
 dreschen ... 181,206 
 
 dritte 124 
 
 du 47,165 
 
 dumm... 17,28,32 
 
 dung 17 
 
 dilnken 24, 47, 88, 187, 
 238 
 
 dilnn 47,54 
 
 durch 49 
 
 dilrfen 55,224 
 
 dttrr 42 
 
 durst 42 
 
 dusel 23 
 
 dutzen 106 
 
 Eben 
 eber 
 eck 
 ecke 
 
 48 
 54 
 50 
 50 
 
 ei 50,117,147 
 
 eiergelb 117 
 
 eigen ... 50,54,228 
 
 eilf 121 
 
 ein 31, 39, 66, 121, 134 
 
 ein- 66 
 
 einander 114 
 
 eindruck 6 
 
 einfalt 110 
 
 einige 31, 134 
 
 einOde 90 
 
 einst 124 
 
 eis 145 
 
 eitel 
 
 . ... 110 
 
 elend 
 
 . ... 23 
 
 elephant 
 
 . ... 141 
 
 elf 
 
 . ... 121 
 
 ellenbogen . . 
 
 . ... 114 
 
 emp- 
 
 . ... 98 
 
 emsig 
 
 . ... 23 
 
 ende 
 
 . 22,142 
 
 engel 
 
 . ... 39 
 
 ent- ... 95, 96, 98 
 
 entgegen 96 
 
 er 166 
 
 er- ... ... 98,99 
 
 erbe ... 74,137,140 
 
 erde 17, 20, 21,73,137, 
 
 142, 152 
 
 erktiren 180 
 
 erlauben 98 
 
 erlOschen ... 174,181 
 erquicken 27, 50, 107 
 
 erschallen 181 
 
 erschrecken... 182,206 
 
 erst 
 
 ... 124 
 
 ersuchen 
 
 ... 99 
 
 erwdgen 
 
 . ... 182 
 
 erz- 
 
 ... 96 
 
 erzdhlen 
 
 ... 23 
 
 es 
 
 ... 166 
 
 essen 26,82, 
 
 183, 197, 
 
 
 206, 229 
 
 euch 
 
 . 36,165 
 
 euer 
 
 ... 168 
 
 euter 
 
 28,49 
 
 ewig 
 
 . ... 81 
 
 Fach 82,124 
 
 faden 89 
 
 fahl 37 
 
 fahne 102 
 
 fdhnrich 102 
 
 fahren 17, 82, 83, 184, 
 204, 205 
 
 falb 37 
 
 falke 13 
 
 fallen 17, 21, 83, 84, 185 
 
 fallen 83 
 
 falls Ill 
 
 falsch 13,17 
 
 falte 110 
 
 fang en 82, 185 
 
 faul ...23,28,32,76 
 
 fdule 23 
 
 faust 49 
 
 fechten 181,206 
 
 feder ... 23,49,54 
 
 fegefeuer 50 
 
 fegen 50, 76 
 
 feind ... 22, 54, 90 
 
 fell 26 
 
 fels,felsen 141 
 
 fern 10,21 
 
 fernsprecher ... 6 
 
 fest (fast) 93 
 
 fest (feast) 143 
 
 feuer 29,143 
 
 finden 19, 20, 28, 181 
 
 finger 26, 39, 50, 121 
 
 fisch 27, 48, 50, 54, 72, 
 
 146 
 
 fischen 19,174 
 
 fischer 102 
 
 fittich 2$ 
 
 flaum 48 
 
 flechten 181 
 
 fleck, flecken ... 141 
 
 fleisch 48,50 
 
 C C 
 
386 
 
 ENGLISH AND GEKMAN GRAMMAR. 
 
 fliege 50 
 
 fliegen 29, 50, 180, 206 
 
 fiiehen 180,187 
 
 flieszen ... 84,206 
 
 floh 139 
 
 flosz 146 
 
 flucht 49,78 
 
 flur 17,76 
 
 fiut 33, 78, 136, 142 
 
 folgen 50 
 
 folgern 106 
 
 /order n 97 
 
 forschen ... 84,189 
 
 fort 49 
 
 fragen 207 
 
 frau 32, 74, 136, 142 
 frauenzimmer ... 114 
 
 freilich 110 
 
 fressen "99,229 
 
 fireund 22, 41, 54, 90, 
 
 137 
 
 friede, frieden ... 141 
 
 frieren ... 61,180 
 
 frisch 50 
 
 frohndienst 74 
 
 frost 61 
 
 frucht 136 
 
 fuchs 23 
 
 ftichsin ... 23,87 
 
 fuhren 83 
 
 fallen ... 17,23,107 
 filnf 17, 20, 39, 54,121 
 ftinfte ... 49, 53, 124 
 
 fUnfzehn 48 
 
 fiinfzig 48 
 
 funke, funken 17,141 
 
 filr 99 
 
 furche 42 
 
 filrst 17, 23, 49, 124 
 fusz 18, 33, 72, 144, 146 
 futter 5, 49, 79 
 
 Gabe 142 
 
 gtihren 182 
 
 gang 219 
 
 gans 56, 102, 144, 146 
 
 gdnserich. 102 
 
 giiren 182 
 
 garUn 30, 74, 187, 141 
 
 gdrtnetr 102 
 
 gast 30, 50, 51, 56, 187, 
 
 144, 146 
 
 ge- 19, 23, 50, 65, 69, 
 
 96, 97, 143, 229 
 
 0*- (part.) ... 97,229 
 ye- (nouns) 96, 143 
 gebdren 26, 82, 182, 206 
 geben 45, 48, 50, 183, 
 197, 206 
 
 gebirge 96 
 
 gehurt 17 
 
 gedeihen 24 
 
 gefdhrte 96 
 
 gef alien 97 
 
 gejilde 96 
 
 gegen 96 
 
 gehen 185,219 
 
 geld 37 
 
 geist ... 50,62,147 
 geldchter ... ... 79 
 
 gelb ... 37, 50, 56 
 
 geleiten 19 
 
 gelten 50, 125, 181, 206 
 
 gemald 96 
 
 gemein 96 
 
 genesen 97, 183, 207 
 
 genie 6 
 
 genieszen 180 
 
 genug 20, 50, 55, 96 
 
 gem 23.50 
 
 geschehen ... 183,206 
 
 geschweige 97 
 
 geselle 96 
 
 gesinde 96 
 
 gestehen 97 
 
 gestern ... 41,56 
 
 gestirn 23 
 
 gesund 20 
 
 getreide 202 
 
 gevatter 96 
 
 gewahr 96 
 
 gewinnen 181 
 
 gewisz 223 
 
 gieng 185,219 
 
 gier 23 
 
 gieszen 45, 84, 180, 206 
 gift ... 48,50,78 
 
 ging 185,219 
 
 glas 20,61 
 
 glauben 19, 32, 65, 229 
 
 gleich 96 
 
 gleichm 179 
 
 gleichung 6 
 
 glimmen 181 
 
 glitzern 106 
 
 gold 17, 20, 49, 50 
 golden... 22,23,89 
 Goit ... 28, 50, 147 
 
 grab . . . 
 
 147 
 
 graben... 
 
 184 
 
 grad ... 
 
 143 
 
 graf ... 
 
 ... 65,141 
 
 greifen 
 
 179 
 
 grosz . . . 
 
 32 
 
 gulden... 
 
 ... 22,23 
 
 gusz ... 
 
 45 
 
 gut 
 
 50 
 
 Haar 50,143 
 
 haben 19, 85, 187, 189, 
 231 
 
 hafen 146 
 
 haft (tie) 54 
 
 haft (capture) ... 54 
 
 hagel 50, 54 
 
 hahn 74, 137, 140, 141 
 halb ... 21,48,124 
 
 hdlfte 48,124 
 
 halm 75,143 
 
 hals 146 
 
 lialten 17, 21, 50, 54, 
 84, 185, 206 
 
 hammer 146 
 
 hand 48, 49, 50, 72 
 137, 146 
 
 handeln 106 
 
 handhabe 118 
 
 handhaben ... 114,118 
 hangen 39, 83, 185 
 
 hdngen 83 
 
 hart ...20,50,54,72 
 
 base 61 
 
 hassen 50,93 
 
 hast 17 
 
 hauen 37, 185 
 
 haupt 19,32 
 
 liaus 18, 20, 28, 32, 147 
 
 haut 146 
 
 heben 54, 55, 83, 85, 184 
 
 hecke 50 
 
 heer ... 41, 137, 143 
 
 heft 54 
 
 Jieide (heath) ... 187 
 heide (heathen)... 19 
 
 heil 90,93 
 
 Heiland 90 
 
 heilen 90,98 
 
 heilig 50 
 
 heifii 65 
 
 heimat 90 
 
 heischen ... 60,84 
 heisz 31 
 
INDEX OF GERMAN WORDS. 
 
 387 
 
 Jteiszen 185,239 
 
 helfen 45, 181 
 
 henker 13,83 
 
 her 166,168 
 
 herbst 41 
 
 herd 21 
 
 herde 21,87 
 
 herr 141 
 
 herz 20, 21, 49, 74, 137, 
 
 140, 142, 150, 152 
 
 herzog... 87, 141, 143 
 
 heute 166 
 
 hieb 37 
 
 hier 166,168 
 
 hin 166 
 
 hirte 87,141 
 
 hoch ... 50,54,74 
 
 hof 53 
 
 hShe 50, 74, 136, 137, 
 140, 142 
 
 hoheit 110 
 
 hdher 50,80 
 
 holz 147 
 
 horchen 20 
 
 horn 77 
 
 hilbsch 53 
 
 huf 17,143 
 
 htigel ... 50, 54, 88 
 
 huhn 147 
 
 hund ... 20, 28, 143 
 hundert ... 50,123 
 
 hunger 50 
 
 hiitte 17 
 
 Ich 14,164 
 
 ihr (you) 165 
 
 Mr (her) 168 
 
 ihr (their) 168 
 
 immer 96 
 
 in 66 
 
 irden 89 
 
 irdinch 91 
 
 irgend 161 
 
 ist 26,217 
 
 Jahr ... 36, 76, 143 
 
 jauchzen 106 
 
 je 96,124,161 
 
 jedenfalls Ill 
 
 jeder 79 
 
 jener 16') 
 
 joch 28,36, 72, 136, 143 
 jngend 90 
 
 jung .. 
 
 ... 28,36 
 
 jungfer 
 
 117 
 
 7TaZ648,56,80,137,147 
 
 kalk . . 
 
 50 
 
 kalt 17, 
 
 21, 50, 57, 78 
 
 kamm .. 
 
 ... 39,48 
 
 kann . . 
 
 ... 39,224 
 
 kater . . 
 
 102 
 
 katze . . 
 
 102 
 
 kauen .. 
 
 36 
 
 kaufen... 
 
 99 
 
 keck 
 
 ... ... 50 
 
 kehren . . 
 
 105 
 
 kehrieht 
 
 105 
 
 k-ein 
 
 134 
 
 kennen . . . 
 
 39, 84, 187 
 
 kiesen 
 
 29, 50, 57, 61, 
 
 
 82, 180 
 
 kind . . . 
 
 ... 136,147 
 
 kinn ... 
 
 50 
 
 kirche ... 
 
 142 
 
 klauben 
 
 32 
 
 kleben ... 
 
 82 
 
 klein . . . 
 
 ... 17,77 
 
 kleinod 
 
 17 
 
 klimmen 
 
 ... 82,181 
 
 Hosier . . . 
 
 146 
 
 knabe . . . 
 
 17, 36, 141 
 
 knappe 
 
 36 
 
 knecht ... 
 
 21 
 
 knie 
 
 50, 57, 143 
 
 knoblauch ..v ... 32 
 
 knobloch 
 
 82 
 
 kommen 
 
 39, 40, 45, 50, 
 
 
 57, 182 
 
 kdnig ... 
 
 ... 50,103 
 
 kdnnen... 
 
 ... 84,224 
 
 korn ... 
 
 ... 50,57 
 
 kost 
 
 142 
 
 kosten ... 
 
 29 
 
 krachzen 
 
 106 
 
 kraft . . . 
 
 146 
 
 kraut ... 
 
 147 
 
 kreuz ... 
 
 143 
 
 kriechen 
 
 ... 180,206 
 
 krilppel 
 
 88 
 
 kuh 
 
 14, 50, 57, 146 
 
 MM ... 
 
 50 
 
 kund ... 
 
 ... 78,224 
 
 kunfl ... 
 
 ... 40,78 
 
 kumt . . . 
 
 ... 78,146 
 
 kur 
 
 180 
 
 kiirze ... 
 
 74 
 
 Lachen . . . 36, 50 
 
 laden 141,146 
 
 laden (vb.) 17,50,184 
 
 lahm 10 
 
 laib 17 
 
 lamm 80, 147 
 
 land 39,49 
 
 Jang 39,50 
 
 lange 110 
 
 lange 74,142 
 
 lassen ... 45,185,236 
 
 lasz 45 
 
 laub 32 
 
 lauch 32 
 
 laufen 32,50,185,206 
 
 laugnen 106 
 
 laus 144,146 
 
 laut (sound) 41, 143 
 laut (loud) 41,50,78 
 
 lauter 110 
 
 leben 27,48,85 
 
 leber 27 
 
 lechzen 50 
 
 leclcen 27, 50 
 
 legen 20,50,83 
 
 lehren ...31,61,83,106 
 
 leib 147 
 
 leiche 110 
 
 leicht 50 
 
 leiden 179 
 
 lexer 142 
 
 leihen 82,179 
 
 foiten 19 
 
 lernen 20, 21, 31, 61, 
 63,106 
 
 lesen 61,183 
 
 letzt 80 
 
 leuchten 29 
 
 leugnen 106 
 
 leumund 75 
 
 leute 29 
 
 leutnant 65 
 
 licht 29,41,50 
 
 lid 117 
 
 lieb 29,48,56 
 
 liebe 48,56 
 
 lieben 107 
 
 liegen ... 50,83,188 
 los ... 17,45,110 
 
 Idnchen 174,181 
 
 losen 41 
 
 Idwe 141 
 
 lufl 78,146 
 
 lug 139,180 
 
388 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 Itige 
 liigen 
 lust 
 liitzel 
 
 ...14,17,50 
 50, 106, 180 
 146 
 
 Machen ... 20,50 
 macht 20, 50, 78, 146, 
 
 226 
 
 mag 226.234 
 
 magd ... 19,50,146 
 
 magen 141 
 
 mahen 50,185 
 
 mahnen 83 
 
 mdhre 65 
 
 maid 17,50 
 
 manch 74 
 
 mann 37, 39, 67, 144, 
 
 147 
 
 marsch 6 
 
 marschall 65 
 
 marschiren 6 
 
 maus 23, 61, 144, 146 
 meer ... 17,72,143 
 
 mehr 61,80 
 
 meiden 96 
 
 meile 17 
 
 mein 10,168 
 
 meist 80 
 
 meiszel 88 
 
 melken 181 
 
 menge ... 74,140,142 
 
 mensch 67,141 
 
 messen 183,206 
 
 messer 65 
 
 meth 39 
 
 miethe 62, 185 
 
 mild 20 
 
 mis- 96 
 
 missen 96 
 
 mist 27 
 
 misz- 96 
 
 mii 73,117 
 
 mitte 26,36,89,56,73 
 
 mittel 10,49 
 
 mode 6 
 
 mdgen... 20,226,234 
 
 monat 90 
 
 mond 18 
 
 Montag 18 
 
 mord 42 
 
 morgen ... 50, 54 
 
 miihle 42 
 
 mund 40 
 
 milnster 65 
 
 milnze .. 
 
 49 
 
 Pabst, papst 14, 48 
 
 mtissen.. 
 
 227 
 
 parlieren 
 
 6 
 
 musze .. 
 
 227 
 
 pfad ... . 
 
 .. 49, 143 
 
 muth .. 
 
 49 
 
 pfaffe ... 
 
 48 
 
 mutter 
 
 10, 18, 19, 33, 
 
 pfalil ... 
 
 48 
 
 
 49, 54, 146 
 
 pfalz ... 
 
 65 
 
 
 
 pfanne... 
 
 .. 23,48 
 
 Nach .. 
 
 50 
 
 pfau ... 
 
 .. 48,149 
 
 nachbar 
 
 50, 65, 149 
 
 pfennig 
 
 23 
 
 Nachbaur 65 
 
 pferd ... 
 
 48, 65, 143 
 
 ndchst .. 
 
 80 
 
 Pfingsten 
 
 48, 65, 152 
 
 nacht 
 
 14, 21, 39, 50, 
 
 pflanze. . . 
 
 .. 48,142 
 
 
 54, 79, 146 
 
 pfiaum... 
 
 48 
 
 nachtigall 114 
 
 pflegen... 
 
 182 
 
 nadel .. 
 
 ...26,39,79 
 
 pfund ... 
 
 18, 48, 143 
 
 nagel . . 
 
 ... 50,146 
 
 philosoph 
 
 141 
 
 ndhe . . 
 
 50 
 
 photograph' 
 
 ^ ... 142 
 
 ntiher .. 
 
 50 
 
 pilgrim 
 
 23 
 
 ndhren. . 
 
 207 
 
 planet ... 
 
 141 
 
 name .. 
 
 75, 141, 142 
 
 preisen 
 
 179 
 
 namen . . 
 
 141 
 
 punkt ... 
 
 143 
 
 namhaft 
 
 117 
 
 
 
 narr .. 
 
 141 
 
 Quecksilber. 
 
 .. 27,50 
 
 nase 
 
 61 
 
 quellen... 174,181,206 
 
 natter .. 
 
 49 
 
 quer ... . 
 
 17 
 
 neffe . . 
 nenmen 
 
 54 
 
 45, 82, 174, 
 
 Rabe ...36, 
 
 50, 54, 141 
 
 182 
 
 197, 204, 205 
 
 r&chen . . . 
 
 182 
 
 nein 
 
 10, 96, 134 
 
 rad ... . 
 
 10 
 
 nennex... 
 
 187 
 
 rand ... . 
 
 147 
 
 nest 
 
 ... 27,62 
 
 rappe ... . 
 
 36 
 
 netz 
 
 17 
 
 rasch ... 
 
 50 
 
 neun .. 
 
 121 
 
 r&sonnieren 
 
 ... 6 
 
 nicht . . 
 
 96 
 
 rast ... . 
 
 73 
 
 nichte .. 
 
 ... 53,87 
 
 rathen... 
 
 17, 41, 185 
 
 nie 
 
 96 
 
 rathsel. . . . 
 
 104 
 
 nimmer 
 
 96 
 
 rauchwerk 
 
 .. 50,54 
 
 noch ... 
 
 96 
 
 rauh ... . 
 
 .. 50,54 
 
 noth 
 
 ... 142,228 
 
 raum . . . 
 
 ..28,39,75 
 
 nur 
 
 96 
 
 recht ... . 
 
 .. 53,78 
 
 nusz 
 
 50 
 
 reckt (das) . 
 
 143 
 
 nutzen . . 
 
 107 
 
 recken . . . 
 
 99 
 
 niltzen . . 
 
 107 
 
 reden 17,20, 
 
 
 
 regen ... . 
 
 50 
 
 Oben . . 
 
 54 
 
 reiben ... . 
 
 179 
 
 ocean .. 
 
 65 
 
 reich ... . 
 
 50 
 
 ochs 
 
 17, 74, 140 
 
 reich (das) . 
 
 109 
 
 ochse . . . 
 
 141 
 
 reichen 
 
 187 
 
 6de 
 
 90 
 
 reifen ... . 
 
 107 
 
 ofen 
 
 48, 141, 146 
 
 rein ... . 
 
 77 
 
 oft ... 
 
 49 
 
 reiszen... . 
 
 .. 87,179 
 
 ohr 
 
 32,61,74,142 
 
 reiten ... 
 
 17,41,179 
 
 ort 
 
 147 
 
 reiter ... . 
 
 102 
 
 Ostern . . . 
 
 152 
 
 rennen ... . 
 
 .. 181,187 
 
INDEX TO GERMAN WORDS. 
 
 339 
 
 retten ... 
 
 187 
 
 schieszen 
 
 180 
 
 sein (his") 168 
 
 rettich . . . 
 
 23 
 
 schiff ... 
 
 50 
 
 sein (tc 
 
 be)... 217, 231, 
 
 richtig... 
 
 91 
 
 schiffen 
 
 107 
 
 
 239 
 
 rind . . . 
 
 147 
 
 schilling 
 
 103 
 
 selb 
 
 160 
 
 ring . . . 
 
 50 
 
 schinden 
 
 181 
 
 selbst 
 
 167 
 
 rinnen... 
 
 181 
 
 scMacht 
 
 142 
 
 selten 
 
 23 
 
 ritter ... 
 
 102 
 
 schlaf ... 
 
 .. 48,146 
 
 seltsam 
 
 23 
 
 ritz 
 
 37 
 
 schlafen 
 
 48, 57, 185 
 
 senden 
 
 187 
 
 rohr . . . 
 
 146 
 
 schlaff . . . 
 
 .. 48,57 
 
 setzen 17, 26, 36, 83,174 
 
 rosz 
 
 143 
 
 schlag ... 
 
 .. 72,146 
 
 seufzen 
 
 106 
 
 roth 
 
 ... 32,41 
 
 schlagen 
 
 45, 59, 184 
 
 sich 
 
 167 
 
 rufen . . . 
 
 185 
 
 schlank 
 
 13 
 
 sicht 
 
 50,78 
 
 ruhe ... 
 
 73 
 
 schlecht 
 
 78 
 
 sie 
 
 166 
 
 
 
 schlitten 
 
 17 
 
 sieben 
 
 .. 26,40,121 
 
 Saat ... '. 
 
 2, 26, 49, 59 
 
 schlilpfen 
 
 59 
 
 sieden 
 
 53,180 
 
 Sachse . . . 
 
 141 
 
 schmal... 
 
 59 
 
 sieg 
 
 117 
 
 scien 
 
 ... 59,185 
 
 schmeicheln 
 
 59 
 
 silber 
 
 21,59 
 
 sagen . . . 
 
 ... 20,50 
 
 schmelzen 
 
 .. 174,181 
 
 singen 
 
 181 
 
 salbe ... 
 
 92 
 
 schmerz 
 
 150 
 
 singer 
 
 102 
 
 salben ... 
 
 92, 93, 189 
 
 schnauben 
 
 180 
 
 sinken 
 
 .. 39,82,181 
 
 salbuch 
 
 21 
 
 schnee 37,39,? 
 
 sitte 
 
 59 
 
 salz 10, 3C 
 
 1, 49, 59, 143 
 
 schneiden 
 
 179 
 
 sitz 
 
 26 
 
 salzen ... 
 
 185 
 
 Schneider 
 
 102 
 
 sitzen 
 
 26, 59, 83, 174, 
 
 same ... 
 
 ... ... 75 
 
 schneien 
 
 50 
 
 
 183 
 
 Samstag 
 
 65 
 
 schnitter 
 
 102 
 
 so ... 
 
 66 
 
 sand ... 
 
 39 
 
 schon . . . 
 
 77 
 
 sohn 12, 2( 
 
 sanft ... 
 
 48 
 
 schdn . . . 
 
 77 
 
 
 77, 137, 146 
 
 sang ... 
 
 ... 39,146 
 
 schop/en 
 
 48 
 
 solch 
 
 50,161 
 
 stinger , . . 
 
 102 
 
 schraube 
 
 17 
 
 soldat 
 
 141 
 
 satt 
 
 59 
 
 schrauben 
 
 180 
 
 sollen 
 
 .. 225,232,235 
 
 sattel ... 
 
 49 
 
 schreiben 
 
 179 
 
 sommer 
 
 146 
 
 sau 
 
 27, 32, 146 
 
 schrein 
 
 59 
 
 sonder 
 
 40 
 
 sauer ... 
 
 ... 32,76 
 
 schreiten 
 
 .. 59,179 
 
 sonne 
 
 142 
 
 saufen . . . 
 
 ... 32,180 
 
 schuh ... . 
 
 50 
 
 Somitai 
 
 / 117 
 
 saugen. . . 
 
 180 
 
 schwtiren 
 
 182 
 
 sorge 
 
 93 
 
 sdule . . . 
 
 ... 59,76 
 
 schwarz 
 
 59 
 
 sorgen 
 
 93 
 
 schaben 3 
 
 0, 48, 56, 57, 
 
 schweigen 
 
 59 
 
 spcthen 
 
 53 
 
 
 82,184 
 
 schwein 27 
 
 59, 89, 143 
 
 spazier 
 
 en 107 
 
 schaf ... 
 
 ... 50,143 
 
 schweisz 
 
 ..49,57,59 
 
 speer 
 
 48 
 
 schaffen 
 
 48, 50, 57, 
 
 schwellen 28, 
 
 speien 
 
 48,53 
 
 
 184,188 
 
 schwert 
 
 .. 21,59 
 
 spielen 
 
 17 
 
 schaft ... 
 
 48 
 
 schwester 
 
 59, 61, 142 
 
 spindel 
 
 88 
 
 schalk ... 
 
 ... ... 146 
 
 schwimmen 
 
 17, 39, 59, 
 
 spinner 
 
 17, 48, 181 
 
 schall ... 
 
 181 
 
 
 181 
 
 spleisze 
 
 n 179 
 
 scharf ... 
 
 20 
 
 schwingung 
 
 ... 13 
 
 spreche 
 
 n ... 17,182 
 
 scli&rfen 
 
 106 
 
 schwitzen . 
 
 ..49,57,83 
 
 springi 
 
 m ... 82,181 
 
 schauen 
 
 ... 59,77 
 
 schworen 
 
 .. 96,184 
 
 stand 
 
 49 
 
 schaufel 
 
 ...48,50,88 
 
 sechs ... 
 
 .. 21,121 
 
 stand ( 
 
 pf.) 184 
 
 scheiden 
 
 ... 31,79 
 
 seele 
 
 31 
 
 stark 
 
 50 
 
 scheinen 
 
 31, 53, 84, 
 
 segel ... 
 
 50 
 
 stechen 
 
 182 
 
 
 173, 179 
 
 sehen . . .17 
 
 36, 45, 50, 
 
 stecken 
 
 179,183 
 
 schelten 
 
 181 
 
 54, 183 
 
 stehen 
 
 .. 53,184,220 
 
 scheren 
 
 182 
 
 sehr 
 
 31 
 
 stehlen 
 
 182,206 
 
 schieben 
 
 ... 50,180 
 
 sei 
 
 211 
 
 steif 
 
 48 
 
390 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 steigen . . . 
 stein ... 
 sterben... 
 stern ... 
 sticken... 
 stinken... 
 stoszen . . . 
 streben... 
 streich . . . 
 streiten 
 strom ... 
 stube . . . 
 studieren 
 stuhl ... 
 stute ... 
 suchen 33, 
 sump/ ... 
 stinde ... 
 siisz 
 
 Tag 
 
 26, 82, 179 
 
 105 
 
 ... 181,206 
 
 21 
 
 83 
 
 181 
 
 185 
 
 179 
 
 50 
 
 179 
 
 ... 59,75 
 
 20 
 
 107 
 
 ... 33,146 
 
 33 
 
 187,189,207 
 
 59 
 
 ... 17,142 
 ...49,57,59 
 
 , 47, 50, 56, 72, 
 137, 143 
 
 talg 50 
 
 taub 32 
 
 taube 32,142 
 
 taufen 29 
 
 taugen 222 
 
 lausend 19,28,49,123 
 
 teig 50,56 
 
 Teutsch 29 
 
 thai 14 
 
 that 14,26,45,47,78, 
 137, 142, 218 
 
 thau 17 
 
 theil 96 
 
 thier 49,143 
 
 thiergarten 49 
 
 thor ... 17,28,143 
 
 thrdne 22 
 
 thun 14,45,56,218,233 
 
 thUr 28 
 
 thiirel 88 
 
 tie/ 29,57 
 
 tie/e 74 
 
 tisch 50 
 
 toben 32 
 
 tochter 49,56,137,146 
 
 tod 32,47,48 
 
 todt 32; 78 
 
 trachi 20 
 
 tragen 17,20,50,184, 
 205 
 
 trauen 37 
 
 traum 49 
 
 treffen . . . 
 treiben. . . 
 treten ... 
 
 tretoir . . . 
 
 182 
 
 45,49,179 
 
 6, 182, 183 
 
 6 
 
 treu 
 
 37 
 
 trie/en... 
 trinken 
 
 180 
 
 39, 40. 181 
 
 trittoir. . . 
 
 ' 6 
 
 trilgen ... 
 
 tugend . . . 
 
 180 
 
 ... 142,222 
 
 Uber ... 
 
 54 
 
 urn 
 
 40 
 
 un- 
 
 ..17,40,96 
 
 und 
 
 66 
 
 universitat 
 
 142 
 
 uns 
 
 ... 20,164 
 
 unser . . . 
 
 168 
 
 unter ... 
 
 ... 40,49 
 
 ur- 
 
 96 
 
 urtheil... 
 
 96 
 
 Vater ...25,48,49,51, 
 
 53, 54, 146 
 
 ver- ... 19,65,69,99 
 
 verderben . . . 174, 181, 
 
 206, 224 
 
 ver/aulen 107 
 
 vergebens 99 
 
 vergehen ... 10,99 
 
 vergeigen 99 
 
 vergessen 182,183,206 
 verhehlen ... 83,182 
 
 verkau/en 99 
 
 verlieren 19, 45, 61, 
 99,180 
 
 verlust 45,61 
 
 vermdgen 226 
 
 verrecken 99 
 
 versehren 31 
 
 verwirren 181 
 
 verzehren 41, 82, 182 
 
 vieh 12,34,54 
 
 viel 12,34 
 
 vier 121 
 
 vogel ...41,48,50,146 
 
 volk 139 
 
 voll 17,23,42,77,157 
 
 vor 23,99 
 
 vorder 79,97 
 
 vorhanden 146 
 
 Wachen 
 wachs ... 
 
 36, 50, 184 
 50 
 
 wachsen 184 
 
 wagen 206 
 
 wahren 106 
 
 wahrsagen 118 
 
 wald 41,147 
 
 Walsch 23 
 
 wamme 48 
 
 wange 142 
 
 wann 66 
 
 war (pf.) 17, 20, 61, 
 183, 217 
 
 wdre 194,217 
 
 warm 21, 50, 56, 75, 
 
 106 
 warmen . . . 106, 107 
 
 warnen 106 
 
 was ... 50, 54, 161 
 waschen 17, 84, 184 
 
 was /ur 161 
 
 wasser 5, 17, 37, 49 
 
 wasserstoff 6 
 
 weben 182 
 
 wechsel 104 
 
 weaken 30 
 
 weder 79 
 
 weg ... 50, 96, 146 
 wegen (vb.) ... 206 
 
 weib 48 
 
 weichen 179 
 
 weigand 54 
 
 weigern 54 
 
 weile 17 
 
 wein 17,37 
 
 weise (adj.) ... 223 
 weise (manner) 111 
 weisz (adj.) 37, 50, 54 
 weiszUh.) 31,37,53, 
 
 223 
 
 weizen 50 
 
 welch 161 
 
 welt 21,27,65 
 
 wenden ... 187,219 
 
 wenn 66 
 
 wer 161 
 
 werben 181 
 
 werden 45, 77, 181, 
 
 229,232,239 
 
 wer/en 48, 57, 181 
 
 wer geld 21 
 
 werk 50 
 
 werth 21 
 
 werwolf 21, 27, 115, 
 
 117 
 wesen ... 61,183,217 
 
INDEX OF GERMAN WORDS. 
 
 391 
 
 wetter 17,49 
 
 wetzen 50,54 
 
 wider 79 
 
 wiedehopf 41 
 
 wieder 79 
 
 wi'.gen... 14,180,206 
 
 wiese 88 
 
 wiesel 88 
 
 wild 17,49 
 
 will(vh.) 17,221,228 
 
 wilh 36,141 
 
 ■willen 141 
 
 willkilrlich 180 
 
 wind 17,26,37 
 
 winden 17, 181, 219 
 
 winter 146 
 
 wir 63,164 
 
 wirken 83 
 
 wirr 181 
 
 wissen... 45,82,223 
 
 wittib 37 
 
 wittwe... 27,37,142 
 
 witz 223 
 
 wo 168 
 
 wolf ...84,37,42,48, 
 54,146 
 
 wblfin 54 
 
 wollen ...221,228,229, 
 232 
 
 warden 229 
 
 wort 20,137,143,147 
 
 wund 78 
 
 wunde 78 
 
 wilnschen 84 
 
 wurde 181 
 
 wilrde (subj.) 181, 194, 
 235 
 
 wurf 95,146 
 
 wurm 147 
 
 ZaJil . 
 z&hlen . 
 zahm . 
 zdhmen 
 zahn . 
 zahre . 
 zange . 
 zart 
 zaun . 
 zehe 
 zehn 40 
 zehren . 
 zeichen. 
 zeigen . 
 zeit 
 zer- 
 zerren . 
 zeugen . 
 
 47 
 
 ... 23,49 
 ... 23,49 
 ... 12,49 
 
 49 
 
 79, 144, 146 
 30, 76, 146 
 39 
 20 
 18 
 50 
 54, 57, 121 
 41 
 77 
 77 
 49 
 100 
 82 
 29 
 
 ziegel 50 
 
 Ziehen ...29, 53, 55,82, 
 
 180,206 
 
 zimmer . . .47, 57, 76 
 
 zimmermann ... 47 
 
 zmn 
 
 zistag ... 
 zog (pf.) 
 zbgern . . . 
 zorn 
 
 zu (to)... 
 zu (too) 
 zucht ... 
 zug 
 
 49 
 
 17 
 
 32, 55, 180 
 
 106 
 
 182 
 
 66, 100, 229 
 ... 66 
 ... 55 
 ... 55 
 
 zunge ...13,39,47,50, 
 57, 74, 137, 140, 142 
 
 zur 66 
 
 zwanzig . . . 121, 122 
 
 zwar 65 
 
 zwei ... 37, 49, 121 
 
 zweifel 121 
 
 zweig 50 
 
 zweite 124 
 
 zwiefach 121 
 
 zwirn 124 
 
 zwirnen . . . 121, 124 
 
 zwint 124 
 
 zwo 121 
 
 zwblf 121 
 
INDEX OF TERMINATIONS. 
 
 N.B. — This index does not include the declensional endings 
 •of the demonstratives (articles, adjectives, pronouns and pos- 
 sessives), which will be found in their proper place, namely, 
 128-133, 155-157, 160-162 and 164-168. 
 
 The references are to the sections (1-240). 
 
 -ability 
 
 -able 
 
 -age 
 
 -an 
 
 -ance 
 
 -ar 
 
 -ation 
 
 69 
 69 
 50 
 
 105 
 69 
 
 102 
 69 
 
 -bury 65 
 
 -ce (adverbs) 111, 124 
 -ce (plural) ... 145 
 
 -d (participles) 19, 54, 
 90, 186-189, 229 
 -d (adjectives) ... 90 
 -d (weak pf.) 186-189 
 -d (abstract nouns) 78 
 -dom ... 67, 75, 109 
 
 -e(or no ending at all 
 
 = infinitive) 19, 77, 
 
 89, 107, 229 
 
 -e (or no ending at 
 
 all = past part.) 19, 
 
 179-185, 187 
 
 -ed (past part.) 19, 54, 
 
 90, 186-189, 229 
 
 -ed (adjectives) ... 90 
 
 -ed (weak pf.) 186-189 
 
 I. — English. 
 
 -en (past part.) 19, 77, 
 89, 179-185, 229 
 -en (adjectives de- 
 noting material) 89, 
 105 
 -en (verbs in) 93, 106 
 -en (plural) 19. 140 
 -er (nouns denot- 
 ing agent) ... 102 
 -er (nouns denot- 
 ing instrument) 90 
 -er (nouns denot- 
 ing origin) ... 102 
 -er (nouns denot- 
 ing males) ... 102 
 -er (comparatives) 
 
 61, 80, 90 
 -er (verbs in) ...106 
 -es (plural) 19, 139 
 
 -es (sg. 3) 202 
 
 -ese 69 
 
 -ess 5,69,87 
 
 -est (superlatives) 80. 
 90 
 -est (sg. 2) 202, 209, 
 210 
 -eth (ordinals) ... 124 
 -eth(sg. 3) ...202 
 
 -fast 
 
 110 
 
 392 
 
 -fold . 
 -ful .. 
 
 -ham .. 
 -head .. 
 -hood .. 
 
 ... 110, 124 
 110 
 
 65 
 
 109 
 
 67, 108, 109 
 
 -ian 105 
 
 -ing (nouns de- 
 noting action) 91, 
 103 
 -ing (diminutives) 91, 
 103 
 -ing (in the first 
 term of a com- 
 pound) 117 
 
 -ing (present part.) 
 
 103, 229 
 -ing (gerund) 103, 229 
 
 -ior 102 
 
 -ish ... 67, 91, 105 
 
 -ism 67 
 
 -ist 69 
 
 -kin 
 
 103 
 
 -1 (verbs in) ... 106 
 -le (nouns denot- 
 ing instrum.) 88 
 •le (verbs in) ... 106 
 
INDEX OF TERMINATIONS. 
 
 393 
 
 -ledge 50, 109 
 
 -less 110 
 
 -ling (derivatives 
 
 in) 103 
 
 -ling (adverbs)... Ill 
 
 -long Ill 
 
 -ly (adjectives) 17, 50, 
 
 67, 110 
 
 -ly (adverbs) 110, 111, 
 
 124 
 
 -ment ... 
 
 69 
 
 -n (past part.) 77, 
 
 89, 179-185, 229 
 -n (adjectives de- 
 noting material) 
 
 89, 105 
 -n (verbs in) ... 106 
 -ness 68/104 
 
 -or 
 -ous 
 
 102 
 69 
 
 -r (nouns denot- 
 ing agent) ... 102 
 
 -r (nouns denot- 
 ing instrum.) 90 
 
 -r (comparatives 80, 
 90 
 
 -ric 109 
 
 -ry 69 
 
 -s (adverbs) ... Ill 
 
 -s (plural) 19, 63, 139 
 
 -s (genitive) 19, 20, 
 
 63, 111, 150, 151 
 
 -s (sg. 3) 202 
 
 -se (verbs in) 93, 106 
 -sh (adjectives) 67, 91 
 
 -ship 67, 109 
 
 -som 110 
 
 -some 110 
 
 -st (superlatives) 80, 
 
 90 
 
 -st (sg. .2) 202, 209, 
 
 210 
 
 -t (past part.) 19, 78, 
 90, 186-189 
 -t (weak pf.) 186-189 
 -t (abstract nouns) 
 
 78,90 
 
 -t(sg. 2)202,217,222- 
 
 227 
 
 -teen 121 
 
 -ter (nouns in) ... 79 
 -th (abstract nouns) 
 
 78,90 
 -th (ordinals) ... 124 
 
 -th(sg. 3) 202 
 
 -ther 79 
 
 -ty ... 50, 54, 122 
 
 -ward Ill 
 
 -wards Ill 
 
 -wise Ill 
 
 -y (adjectives) 17, 50, 
 91, 105 
 
 -y (nouns; 69 
 
 -yer 102 
 
 -at 90 
 
 -bar ... 104, 110, 117 
 
 -chen 6, 23, 103, 146 
 
 -der 79 
 
 -e (abstract nouns) 23, 
 74, 142 
 -e (plural) 23, 143, 
 144, 146 
 -e(dat. sg.) ... 152 
 -e (pres. subj.) ... 192 
 -e (pf . subj.) 23, 193 
 
 -et 69 
 
 -el (nouns denot- 
 ing instrum.) 88 
 -el (diminutives) 23, 
 88,103 
 -eln (verbs in) ... 106 
 -en (nouns in) 39, 141, 
 146 
 -en (infinitive) 19, 23, 
 77, 89, 107, 229 
 -en (past part.) 19, 77, 
 89, 179-185, 229 
 -en (adjectives de- 
 
 ll. — German. 
 
 noting materi- 
 al) ... 23, 89, 105 
 
 -en (plural) 19, 87, 
 140-142 
 
 -en (ace. sg.) 87, 149 
 
 -en (gen. sg.) 87,150 
 
 -en (dat. sg.) 87 152 
 
 -end (pres. part.) 90, 
 229 
 
 -end (abstract 
 nouns) 90 
 
 -ens 124 
 
 -er (nouns denot- 
 ing agent) 102, 146 
 
 -er (nouns denot- 
 ing instrum.) 90 
 
 -er (nouns denot- 
 ing origin) . . . 102 
 
 -er (nouns denot- 
 ing males) . . . 102 
 
 -er (comparatives) 23, 
 61, 80, 90 
 
 -er (plural) 23, 80, 
 138, 147 
 
 -era (adjectives 
 denoting mate- 
 rial) ... 89,105 
 
 -era (verbs in) ... 106 
 
 -es (gen. sg.) 19, 20, 
 
 63, 150 
 
 -est (superlatives) 23, 
 
 80,90 
 
 -est (sg. 2) 202, 209, 
 
 210 
 
 -et (past part.) 19, 90, 
 
 186-189, 229 
 
 -et (sg. 3) 202 
 
 -et (pi. 2) 203 
 
 -ete (weak pf.) 186-189 
 
 -fach 124 
 
 -fait 110, 124 
 
 -faltig 124 
 
 -haft 110 
 
 >hdlb 124 
 
 •halben 124 
 
 -heit 67, 68, 108, 109, 
 136, 142 
 
 -icht 23, 105 
 
 -ieren ... 6, 107, 229 
 
 -ig 17, 22, 23, 50, 91, 
 
 105, 168 
 
394 
 
 ENGLISH AND GERMAN GRAMMAR. 
 
 -igkeit . 
 
 ... 107, 109 
 
 109 
 
 23, 87, 142 
 
 mg 91, 103 
 
 -inn 87 
 
 -iren ... 6,107,229 
 -isch22, 23, 67, 91, 105 
 
 -ismus 67 
 
 -ist 69 
 
 -keit 68, 109, 136, 142 
 
 -lei 124 
 
 -lein 6, 23, 103, 146 
 
 -lich 17, 22, 23: 50, 67, 
 
 110, 111 
 
 -ling 23, 103 
 
 -lings Ill 
 
 -los 110 
 
 -mal 
 
 124 
 
 -n (infinitive) 77, 89, 
 229 
 
 -n (past part.) 77, 89, 
 179 -185, 229 
 
 -n (adjectives de- 
 noting material) 89, 
 105 
 
 -n (plural) 19, 87, 140- 
 142 
 
 -n (ace. sg.) 87, 149 
 
 -n (gen. sg.) 87, 150 
 -n (dat. sg.) 87, 152 
 -n (dat. pi.) 39, 152 
 
 -n (sg. 1) 217 
 
 -n (pi. 1) 203, 209, 210, 
 
 211 
 -n (pi. 3)203,209,210. 
 
 211 
 
 -nd 90,229 
 
 -nen (verbs in) ... 106 
 -ner (nouns de- 
 noting agent) 102 
 -nisz 68, 104, 136, 143, 
 
 146 
 
 -rei 69 
 
 -rich 102 
 
 -m (adjectives de- 
 noting material) 89, 
 105 
 
 -s (gen. sg.) 19, 63, 
 
 150 
 
 -s (adverbs) 111, 124 
 
 -sal 104, 136, 143, 146 
 
 -sam 104,110 
 
 -sch 67, 91 
 
 -schaft ... 109,142 
 
 -sel 104 
 
 -st (abstract nouns) 70 
 
 -st (superlatives) 23, 
 
 80,90 
 
 -st (ordinals) ... 124 
 
 -st (sg. 2) 23, 202, 209, 
 
 210, 211 
 
 -szig 122 
 
 -t (past part.) 19, 54, 
 
 78, 90, 186-189, 229 
 
 -t (adjectives) ... 90 
 
 -t (abstract nouns) 78, 
 
 136 
 
 -t (ordinals) ... 124 
 
 -t (sg. 2) 209, 222-227 
 
 -<(sg. 3) ... 23,202 
 
 -t (pi. 2) 203, 209, 210, 
 
 211 213 
 
 -te (weak pf.) 186-189 
 
 -te (ordinals) ... 124 
 
 -tel (partitives) . . . 124 
 
 -ter (nouns in) ... 79 
 
 -thum, -turn 67, 75, 
 
 109, 147 
 
 -ung ... 
 -ut 
 
 -wcirls . . . 
 -weise ... 
 
 -zehn . . . 
 -zeln . . . 
 -zen 
 -zig 
 
 91, 103, 142 
 90 
 
 Ill 
 
 Ill 
 
 121 
 
 106 
 
 ... 93, 106 
 50, 54, 122 
 
 ERRATUM. 
 P. 236, 1. 16, read Saxones. 
 
 OF THE * X. 
 
 VERSITT) 
 
 Butler & Tanner, The Sel wood Printing Works, Frome, and London. 
 
3~rf?t 
 
/ 
 
 £ 
 
 X 
 
 t\ 
 
 
./ 
 
 (?• *«, 
 
 .# 
 
 
 & 
 
 
 X s 
 
 <%k 
 
 
 .& 
 
 % 
 
 <£ 
 
 *) x @ 
 
 *& 
 
 
 ^ 
 
 '-& 
 
 
 /• %& 
 
 & 
 
 \, 
 
 *o. 
 
 rh i x - y 
 
 X 
 
 <bptete& \ y M2te£& \ 
 
 rl 
 
.<? 
 
 
 f 
 
 
 ^ 
 
 ^ 
 
 **. 
 
 <*>. 
 
 
 £ 
 
 
 ,<r 
 
 "*>■ 
 
 <*> 
 
 # m$m£^ 
 
 • \ 
 
 V 
 
 **>> 
 
 X / 
 
 s \ 
 
 v 
 
 • 
 
 \ 
 
 %> 
 
 °* 
 
 % 
 
 
 
 .*> 
 
 .# 
 
 v 
 
 •<& 
 
 * (Mli 
 
 \ 
 
 ./ 
 
 .^ 
 
 ^ *'