i ^ OUTLINES OF GENERAL PHILOLOGY LONDON : PRINTED BY iPOTTISWOODE JlSD CO., NEW-STREBT SQUABE AKD PAELIAMEST STKBBT OUTLINES OF GENERAL OR DEVELOPMENTAL PHILOLOGY INFLECTION BT R. G. LATHAM, M.A., M.D. LATE FELLOW OP KING'S COLLEaE, CAiTBRIDGE ; AND LATE PROFESSOR OF EKGLKH, UNIVERSITY COLLEGE, LONDON LONDON LONGMANS, GEE EN, AND CO. 1878 All rights reserved PEEFACE. east The present work is the ' Outline ' rather than the full exposition of the subject upon which it treats. And it is an outline of the Method of investigation rather than a collection of any specific results. Moreover, it is a work on General, or Developmental, rather than upon Comparative, Philology. In the latter branch of study there is no want of valuable works, and a fair amount of well authenticated and generally recognised details. In respect to the growth, development, and a fortiori, the origin and general character of Lan- guage, there are fewer works, fewer recognised facts, fewer trustworthy principles of criticism, and, not un- naturally, much more speculation. The two principles which, in the present treatise, have carried the writer farthest are — (1.) That of beginning with the languages of the present time, and arguiug from them backwards, i.e., from the more familiar to the unfamiliar, from the more certain to the less uncertain. Theoretically, then, the best languages to besfin with are those in the most advanced state of development; inasmuch as they have the longest history, and, as a rule, the greatest amount of material 532 v^ vi PEEFACE. for its investigation. The language, however, which everyone best understands is his own, whatever it may be : and this is really the starting-point in every philological investigation, either actual or possible. To an Englishman, the difference fortunately is unim- portant, because his language is both the one which he knows best, and the one that belongs to the most ad- vanced stage of development. It is, always, it may be said safer to argue in this way, i.e., from the known to the unknown rather than vice versa : and there is no doubt but that such is the case. There are degrees, however, in the danger or difficulty of reversing the process : and if there is one subject of human knowledge which is more dangerous and difficult to investigate a priori than another, that subject is the growth and origin of language. (2.) The second point is the necessity of looking more closely to the idea than to the expression of it. The former is the thought itself, the latter the sounds by which it is communicated to the person spoken to. The thought is the sime throughout all languages and in every stage of each. The manner, however, in which it clothes itself in words or syllables differs with the conditions of time and place. In a word like ' jsypacjja ' and a combination like ' / have written,'' the expres- sion is different both in the form and the principle of its formation. But the idea is the same, viz., that of a certain action done in time past, but continued in its actual or possible operation up to the time of speaking, or time present. In like manner the differences be- PEEFACE. Vll tween one and more than one ; between male and female ; between time past, time present, and time future ; between certainty and contingency ; between being an agent and being an object in an action, are universal. Of the combinations of sounds that express them the name is Legion. With the idea, then, so far as we have a clear conception of its nature, we have a certainty and a unity ; with the guise it may take in language we have any amount of variety. In the familiar terms Person, Voice, Number, Case, Tense, Gender and Mood, we have these mental concep- tions not only recognised as such, but classified, named, and defined, converted, so to say, into categories. The signs of them are their Inflections. What those are taken by themselves, is taught as Etymology ; how they stand to one another, we learn from Syntax. Of these two divisions it is only the former that deals in single words. In Etymology, moreover, it is the chief questions which are connected with the Inflections. There are other details in this division besides these ; but Inflection gives us the most important of them : and to see oiu: way to the origin and structure of this is to get an adequate conception of the most difficult problem in Language, save and except the mysterious one of its earliest infancy and origin — and even this, when the general character of its later history is known, is no illegitimate subject of speculation. That the vast majority of Inflections originated as separate and independent words, first combined as com- pounds, and subsequently modified in form, in import. VIU PKEFACE. or in both, may safely be assumed as the dominant opinion ; so that the extent to which they are reducible to elements of this kind is the leading question in the present investigation. These elements we are bound to seek, though we may fail to find them. Nor are our data wholly inadequate. There are three stages in the development of Language which are generally recog- nised, and for which we have three (perhaps four) cur- rent names — the Analytic and Synthetic represented, in different degrees of progress, by the languages of the Indo-European class, the Agglutinate, and the Mono- syllabic. Any pretence to exhaust the data thus supplied is out of the question. A short general view of the characters of the three stages, with an exposition of the chief materials that serve for the illustration of each of them, and an occasional instance of an Inflection reduced, or shown to be reducible, to its elements, is as much as the writer, in a short work like the present, can attempt, and he doubts whether he could do much more in a larger one. It is with these limitations that he wishes his book, though it bears a somewhat ambitious title, to be read. It is to the principles of the processes by which languages are changed, rather than to the changes themselves, that he chiefly refers. In the way of detail he limits himself to those that may serve as examples or illustrations. Tliese he selects as he best can ; and when they either illustrate a rule, or explain a condition of thought, he makes much of them — some- times, perhaps, too much. But this is about all that can be done safely. CONTENTS. PART I. SrXTMESIS AXD AyALYSlS— MATERIALS FOR THE STUDY OF THE SYNTHETIC AND ANALYTIC STAGES OF LANGUAGE-AGGLUTI- NATION—SELECTED MATERIALS FOR THE STUDY OF THIS STAGE —MONOSYLLABICISU— MATERIALS. SECTION I. SYNTHESIS — ANALYSIS. SECTION' 1 — 4. Inflections in Latin and in English . . . . 5 —7. ' & yeypacpa, yeypacpa' — ' have written ' as opposed to ' Avrote "... ... 8 — 12. Analysis of the expression ; Perfect redujjlicate 13 — 26. History of the Greek form 27 — .32. Generality of the change 33 — 35. Rate of change .... 36—38. Continuity .... 40. Dialect 41 — 48. Incidence and order of changes 49 — 50. ' Old ' an equivocal term . 51 — 52. ' Synthesis ' and ' Analysis ' correlative 53 — 54. ' I am singing ' . 55. ' I do sing ' . . . . 56. ^ Synthetico- Analytic^ and ' AnalyticO' 57 — 58. Analj'sis and Synthesis . 59. Analysis, Synthesis, and Inflection 60. ' Infiections ' and ' Furinativcs ' e terms Synthetic ' PAGE 1 2 6 6 10 12 13 15 16 18 19 19 20 20 20 21 21 CONTENTS. Sub-section. materials for the synthetic and analytic stage of language. FECTION PAGK 61 — 63. Data and Materials, their nature . . . .21 64. The Greek 22 65. The Latin 22 66—69. The Sanskrit and its congeners 22 70—76. The German 24 77. The Lithuanian . . 26 78—79. The Slavonic 26 SECTION IL AGGLUTINATION— FORMATIVES. 80 — 87. Formatires as opposed to Inflections 88. Formatives for Nouns . 89— 9L „ Verbs 92. Characters of the Agglutinate class Fin Declension 93. Laplandic Declension . , 94. „ Conjugation 95. „ Verbum Nominale 96 — 97. Agglutination and Synthesis 98. Turkish Declension 99. „ Conjugation . 100. Absence of Gender 101. Yeniseian Declension . . 28 .30 31 32 35 36 3G 36 38 Sub-section. materials for the agglutinate state of language. 101—103. The Fin or Ugrian Family 38 104. The Turk, Mongol, Mantshu, Yeniseian . . .39 105—110. The Dravidian Family 39 111—112. The Brahui . . ' 41 113. The Basque 41 114 — 115. Other Agglutinate languages ... .42 116 — 123. Literary records of the Agglutinate period . . . 42 CONTENTS. XI SECTION m. MATERIALS FOR THE MONOSYLLABIC STAGE. BHCTION PAGE 124—126. The Bhotiya or Tibetan 45—46 127. The Burmese 46 128. The Singpho 46 129 — 132. The Siamese, Mon, Cambojian, Anamitic . . .46 133—134. The Chinese, &c 46 135—138. The Siib-Himalayan and Naga groups . . .48 139. Monosyllabicism and Analysis — Chinese and English 50 140. Amount of Inflection in English . . . .51 141 — 143. Monosyllabicism, Agglutination, Synthesis . . . 52 144. Formatives 54 145. Composition — Importance of Accent . . . . 54 147. Transition to Part II 56 PART II. PARTS OF SPEECH—MIGRATION— THE INFIiXITIVE—VERB OR NOUN- PERSON, VOICE, NUMBER, CASE, AND TENSE— GENDER AND MOOD —ORIGIN OF, AND GROWTH OF INFLECTION. MIGRATION, ETC. 148—157. Migration 57 158 — 159. The Infinitive Verb as a Part of Speech ... 60 160. Verbs and Verbals 62 161. Hunting and hunter 62 162. To err and error, kc. . . . . . . . 62 163 — 165. Leap and leaping, &c. 62 166 — 168. Possible Declension of the Infinitive . . . . 64 169. No Finite Verb with a Person 65 170. Apparent exceptions 66 171 — 172. The Finite Verbs a Verb and something more . . 66 173—174. The forms of the Verb from Stockfleth ... 66 175. Inflection 68 Xll CONTENTS. SECTION' 176—179. 180—181. 182—183. Addition and DifFerentiation Declension and Conjugation No Tautology in Language PAGB . 68 , 69 . 70 PERSON. 184 — 185. Person — Noun and Verb . . . . 186-187. Persons of the Verb .... 188 — 189. In what Case are they ?— Garnett's doctrine 72 73 73 VOICE. Voice — Noun and Verb 76 The Active and Passive Verbals 77 The Passive Verbal and the Passive Participle . 79 The Latin Supines . . . . . , . 80 Words like trustee, mortgagee, &c. . . . .81 Voice of Verbs — the Norse Redective . . . . 82 The Passive Aorist, in Greek 83 . . 84 190. 191. 193. 194. 195—196. 197. 201. 203 — 206. The Deponent and Reciprocal NUMBEE. 206. 212. 213—217. 218—219. 220—221. 222 226. 227. 228. 229. 230. 232—237. 238. 239. Number — Noun and Verb, Reduplicate Plurals 86 Plural and Dual, &c. ...... 89 The Dual 89 Exclusive and Inclusive Plurals .... 90 Declension of them in the Btihing 92 Traces of them elsewhere ..... 94 The Gieekrh ifj.hv^iyiii ..... 97 So-called Trinal and Quaternal Plurals 97 Pluralizing Duals ...... 97 Plural of the Finite Verb . . 98 Plural of the Verb and Noim .... 98 Collectiveness and Plurality .... 101 Place of the Sign of Number . ... 102 CASE AND TENSE. 240. Case-endings and Tense-endings 241. So-called Signs of the Genitive Case 242. Forms like whilom, cullo, ncllo, kc. 243. Place of the Preposit ion 244. Relation between Case and Tense 103 104 105 106 106 COKTE^'TS. XUl GENDER. SECTION PAGB 245—246. Gender and Sex 107 247 — 248. Personification 107 249. Sex in Natvu-e 109 250—253. Gender and Concord 110 254. The Sign of Gender m.fo7-m Ill 255. Limits of the system ; Animate and Inanimate . Ill 256. The Genders in Danish and Swedish . . . . 112 257 — 259. national and Irrational 112 260. Gender — Adjectival and Pronominal rather than Sub- stantial ......... 113 261 — 268. Its origin in the Demonstrative Pronoun . . . 114 269. Traces of the division into Animate and Inaninrnte . 117 270 — 272. Eeal Sex — Grammatical Gender . . . .117 273 — 274. Prerogative of the Neuter 119 275 — 276. Faint character of the Signs of Gender . . . 121 MOOD. 277—280. Faint character of its Signs . 281 — 283. Moods — What is the Number of them 1 284. The Optative 285. The Imperative 289—290. The Conjunctive .... 291. The Conjunctions — Causal, Illative, &c 292. ' If ' and ' if not '— ' That ' and ' that not ' 293 — 295. Concord of Mood or Tense 296—297. Original of Modality . 298. ' If amd' since' .... 299—303. The Optative .... 304— 305. Triple division of Logic . 306, « res ' and ' iVo ' . 308. Collocation 309 — 311. The JNTj/i-Indicativ^ Moods . 812—313. The Optative Tenses in Latin .... 314 — 316. Gender and Mood — Addition and Differentiation 318. The Non-Finite Verb in Latin .... 319. Spectnm, spectatu — amare, amari . . . . 320. Usse and essere — Roumanian Infinitives 321. True Division of Mood- — Negative and AfBrmative 322 — 323. Approximate Negative Inflections 324. Questions and Answers . .... 121 123 124 124 126 128 130 131 132 1.33 134 135 138 139 140 143 144 146 147 147 . 148 149 151 xiv CONTENTS. BKCnON PAGB 325. Strengthened Negatives 152 326. Figures of Speech — Gender, &c 153 327. The Relative 153 RETROSPECT. 328—329. Person . , . . , 153 330—332. Voice 154: 333—334. Nvunber 154 335—337. Case and Tense, Gender and Mood . . . . 156 TENSE — THE GREEK AORIST. 338 — 342. Forms like iQ-t\Ka and Hwko. 158 THE POSTPOSITIVE ARTICLE. 34:3. The Postpositive Article ; Definite . . . . 161 344. In Eoumanian 162 345. In Lithuanic 163 346. In Slavonic 163 347—348. In Scandinavian 164 349. The Postpositive Article ; Indefinite .... 165 THE AUXILIARY VERBS — 'HAVE' AND 'AM.' 350 — 351. ' ITm^e' as 'habeo' ' possideo' 166 352. ' Have' as 'est miW 167 353. ' Be ' and ' Haiw ' in Roumanian 168 354. Omission of the Participle 169 355. ' Have ' — auxiliar and JVon-auxiliax . . ..169 356. Ordinary Conjugation of the Verb Substantive . 169 357 — 358. Infinitive in Gaelic Participial 170 RETROSPECT — PERSON— VOICE — NUMBER— CASE AND TENSE — GENDER AND MOOD. 359. Person 173 360. Voice 173 361 — 363. Number expressed by reduplication .... 174 364. CoUectiveness 175 365. Signs of Number— their character .... 176 366. Case and Tense 176 CONTENTS. XV SECTION PAGB 367. Double division of Cases 176 368. Number of Cases Indefinite 369. Case and Mood 370. Indefinite character of certain terms . 371. Voice 372 — 374. Difference between Cases 375 — 379. What Case is in its origin 380 — 381. Method of investigation 382—390. General View of the Origin of Infections and Forma- tives — Conclusion 185 176 177 177 178 178 179 184 ADDENDUM. In p. 34, § 94. The meaning of the Laplandic verb Idnam is not given. It is in English loose, or loosen, Danish lose, Greek \iot, Latin luo and solvo. ANALYSIS AND SYNTHESIS. PART I. SYNTHESIS AND ANALYSIS— MATERIALS FOR THE STUDY OF THE SYNTHETIC AND ANALYTIC STAGES OF LANGUAGE — AGGLUTINATION— SE- LECTED MATERIALS FOR THE STUDY OF THIS S TA GE—MONO S YLLABICISM—MA TERIAL S. SECTION I. ANALYSIS AND SYNTHESIS. INFLECTION. § 1. "VVe know what is meant by the Numbers, Cases, and Genders of the Latin Nouns ; we know what is meant by the Voices, Moods, Tenses, and Persons of their Verbs ; and, finally, we know that all our examples of these parts of speech, and their like, are single words as ' homini,^ ' homi- nibus,^ ' scripseram' and dozens of others. Yet every one of these single words, when we translate it into English, i-e- quii-es more words than one to render it intelligibly ; e.g. for ' homini ' we must say ' to (a) man ' ; for ' hominihus' ' to men ' ; for ' scripseram,' ' / had written ' ; and so on thi-ough a large portion of the grammar. Nevertheless, there is no part of the words ' homini ' and ' hominibus ' that exactly translates the preposition ' to 5' neither is there any syllable in words like ' scripseram ' which translates ' had.' S 2. Yet we know what we are doing ; and we under- stand the import of the difference. We know that, though B /? 2 ANALYSIS AND SYNTHESIS. in form the two moles of expression may be widely different from one another, they are absolutely identical in meaning, and that they both convey the same idea, though the lan- guage in which it is clothed may be different. In short, the Latin expresses by single words what the English expresses by the combination of more words than one. In Latin, the Nouns have more Cases than we have in English ; and, for those signs of Case which are wanting, the English substi- tutes a Preposition. In Latin, the Verbs, moi-eover, have more Tenses than the English ; and, for those signs of Tense that are wanting with us, we substitute certain words — ' be,' ' ((,711,' ' have ' — and these ai"e called Auxiliai-s, or Auxiliary Verbs. § 3. The system of the Numbers, Genders, and Cases of Nouns, is called Declension ; and Nouns ai-e said to be de- clitied. That of Voices, Moods, Tenses, Numbers, and Per- sons with Verbs, is called Conjiigation ; and Verbs are said to be conjttgated. The Conjugation of the Verb and the Declension of the Noun are called Inflections ; so that both Nouns and Vei'bs are inflected. § 4. There is a great deal more inflection in Latin than there is in English, and there are many more combinations like ' / have written ' in English than there are in Latin. In fact, the two forms, or the two modes of expression, are in the inverse ratio to one another. Yet one form, by no means, excludes the other. The Romans said ' scripturus sum ' = 'I am about to write': we say ' fat hej-'s,' just as the Ro- mans said ' patris.' We also say, concurrently, ' of {a) father,' and sometimes use one form, sometimes another. It is doubtful, however, whether the meaning is exactly the same in both cases, or rather it is cei-tain that it is not so. Still, roughly speaking, the forms may be said to con-espond with one another. But be this as it may, it is certain that single forms like ' patris ' and * scripseram ' are inflections, and that combinations like ' of a father ' and * I had icritten ' are not. § 5. Such is the general view. It is })robably a loose, INFLECTION. 3 probably, an incomplete one. I believe, however, that the best way of amending it is to give a minute analysis of a single sentence. When Pontius Pilate said o yeypa^a, yi- ypatpa, he meant something different from what would be conveyed by o typa\pa, eypaypa. This last combination would have implied simply ' / wrote.' The reduplicate form, which is Perfect and not Aorist, means a great deal more ; viz.. not onlv a recognition of the fact that he had written some- thing, but that what he wrote had a present application of some kind or other : possibly to the effect that be would not write again ; possibly to the effect that he meant to hold by what he had written. The act of writing was past ; the de- tennination was present • the result of it was futm-e. It was all this because, instead of using the Aorist, he used the Perfect (prcetei'itum perfectum). § 6. So much for the tense. But the sign of it — the reduplicate yt — must also be noticed. The farther we go in the present researches the more clearly we shall see that in the expression of ideas wherein the repetition of an act, the plm-ality of an object, or the continuance of a state is an element, the repetition of either the whole word or a pai-t of it is one of the first resoiu'ces to which, in the earlier stages of language, the speaker resorts ; indeed, it is so common that it seems as if it were the first contrivance that presented itself. No wonder, then, that in a tense like the Greek Perfect, where an action done in past time crops up (so to say) and reproduces itself in the present, the sign of it should be a reduplication. § 7. In Greek it stands in strong contrast to the Aorist, as we have seen. In Latin the line of demarcation becomes fainter ; and, in English, the expression of the ideas conveyed by single words in a definite tense, like yeypacpa and others, is simply impossible. There are no such tenses. § 8. But there is something equivalent to them ; indeed, there is an admirable substitute. But it is a cii-cumlocution. The translation of yiypafa is / have written — three words B 2 4 . ANALYSIS AND SYNTHESIS. for one; an Analytic sentence instead of a Synthetic tense. This is a gi-eat change ; and we cannot be too methodical or minute in our investigation of its nature. Let us, therefore, begin with ascertaining, in due order, the exact details of (1) the Meaning of the word have; (2) its Regimen, or Government; (3) the Concord; and (4) the Collocation of the whole sentence. ( 1 ) Have means hold, possess, own, and the like ; and where there are holders, possessors, or oivners, of anything, there must be something that is held, possessed, or oioned. (2) Such being the case, it is an Active or Transitive Verb ; so that the Noun which it governs must be in the Accusative Case. (3) With this Noun any word in apposition to it must agi'ee ; and, so doing, must also })e in the Accusative Case. (4) In respect to the collocation, the Adjective in English generally precedes its Substantive : as a good man. But by a very slight change in the structure and an extension of the clause this order may be reversed ; as a man good and loise. In a,ll these cases it must be remembered that the syntax of the adjective is that of the passive ])articiple as well. § 9. All this is, probably, not only intelligible, but self- evident. Nevertheless, in the combination under notice, ev,ery one of the statements must be taken with a qualifica- tion. 1. There ai'e certain nouns incapable of being governed; in other words, thei-e are certain nouns which only combine with an i>i-transitive verb. All words expressive of the manner in which a thing is done are of this kind, e.g. mile, ho%ir, and the like ; and, Avhat is more, such expressions as ' / have ridden a mile,' or ' I have talked an hour,' are legiti- mate. We cannot, however, be said to possess either miles or Jiours. If so, where is the active, transitive, or goverrdng power 1 2. Again, the concord between the participle and its noun in the way oi gender, is not quite so simple as it appeal's to be. INFLECTION. In English, indeed, its true nature is of no practical importance ; inasmuch as the participles, like the adjectives, have dropped theii- signs of gender altogether. But there was a time when they had them ; so that theoretically the question as to the Concord presents itself. What would it be in English, sup- posing that, in English, we had two, or even three, signs of sender for the Accusative Case? The answer that first suasests itself, of course, that in ' / have ridden a horse ' it is masculine, and in ' / have ridden a mare ' it is feminine. This mai/ be the case : but it is not so necessarily. The commoner concord is that of such sentences as ' Triste hipus stabulis '=the ' wolf is a had thing for the falls.' If so, 'habeo equam equitatt(.m=I have ridden a mare' is, as good as 'habeo equam equitatam;' and even in 'habeo equum equitatum,' the ' -um ' is not masculine, but neuter. .3. Thirdly ; the adjective, as a general rule, precedes the substantive it agrees with ; and the same rule applies to the participle, / have a black horse, I have a white mare. Yet this is not the collocation in / have ridden a horse. § 10. But the explanation of these anomalies is easy; especially that of the third and second, which are closely connected with one another. In habeo eqxiam equitatum the -um is not masculine, but neuter; and the noun it agrees with is not the word eqiumi, but some word meaning thing or object, which is understood. Hence, the translation is / have a mare {as) a ridden (thinrj) ; and, as this is the concord, the participle is in its proper place. § 11. The first anomaly is somewhat more perplexing ; and in some languages it is, to a cei-tain extent, avoided. In English we say / have been ; whilst J'ai ete is what the French say, and Jeg har varet what the Danes say. But the Italians and the Germans, though with other participles they agi-ee with the English, seem to shrink from using such an extreme form as lo ho stato, or Ich habe gewesen (/ have been), and say ' lo sono stato ' and ' Ich bin gewesen (/ am been). Theoi'ctically, the Germans and Italians are right ; 6 ANALYSIS AND SYNTHESIS. yet, except in the case of the paitiei})le of the Verb Substan- tive, they do just as we do ; and just like us follow a natural tendency : the tendency to stretch an analogy beyond its due bounds, and cai-ry the influence of a legitimate precedent too far. We may call this abusive, or we may call it cata- chrestic ; but, whatever be its name, the fact is both real and common. Out of natural combinations, like ' / have written a letter i^ wow-natui-al ones, like ' / have spoken an hour,' are constantly developed. In some languages the process goes farther than in othei-s. In the English, Danish, and Fi'ench, where it goes so far as the term for simple J)eing or existence, it reaches its utmost limit. § 12. That the verb have, however much its full import may be disgiused is, still the verb which ti'anslates the Latin haheo, teneo, or possideo, is manifest from the Spanish tengo, word for word, the Latin teneo, and from the Old German eigan, word for woi-d, the English ow7i, which are used con- currently with the ordinary ' have.' S 1 3. Such is the sense or meaning of the combination, and such the principle that makes the relation between the two intelligible. What follows belongs to the history of the expression. A change in its application takes place ; and it first shows itself in the Latin, in wliich language there is a confluence in meaning of the foi-ms like scripsi and the forms like raomonlL Scripsi, in Latin, means either ' / have written,' or ' / vrote ; ' and momordi either ' / have bitten,' or ' / bit ; ' so that the difference of import no longer coincides with a difference of form ; as it does in the Greek i.ypa\pa and yeyijaipci. § 14. When this happens one of the two forms becomes su- perfluous, and, then, sooner or later, drops out of the language. The one that thus dies out is, almost invariably, the one repre- sented by yeypafa and momordi, i.e. the reduplicate perfect. Meanwhile the distinction in import,whether adequately or in- adequately expressed, or whether not expressed at all, still exists. § 15. It exists in both Latin and English, but with » I^'FLECTION. 7 difference. In English the forms like swam, sjwke, and others, where the past tense is formed by a change of vowel, mean, in respect to time, exactly the same as those ending in (I or t, as moved, slept, &c. &c., and there is only one mean- ing for each or both of them. That is the one of the Greek Aorist, not that of the Greek Perfect. In Latin each of the two forms scripsi and momordi can be used with either meaning; scripsi as '/ have written' or '/ wrote,' and momordi, as ' / bit,' or ' / have bitten.' This being the case, the Analytic form in have has a different position in the two languages. In English it is pre-eminently common and con- s])icuous. 'In Latin it is exceptional, rare, perhaps rudi- mentary, ' Comperticm habeo, milites, verba viris virtutem non addere,' may be translated, ' / am in possession of a dis- covery' but it is, really, word for word, ' / have discovered.' Even in Greek combinations like ' e'^w 'ypaipaQ= having written {something) I have it (now),' has the same import, and the same elements. There is a participle in the past tense, and a verb in. the present. The former places the action and object with which it is connected in time gone by ; the latter connects it with the time of the speaker's notification of it. When the participle is passive the past action is connected with the object (^x*^ yeypap^ifoy), when active with the agent (fx*^ ypa-^ae), i.e. ' I have — as a written thing^=having written, I have.' § 16. Why is ' have ' the particular verb which takes the place of the reduplication ' ye- ' in ' yeypacpa,' and the like 1 We have ah'eady seen what it suggests, viz. continuation repre- sented by rejietition. In ' / have ridden a horse ' what does ' have ' govern 1 Undoubtedly * horse.' And what does 'have' mean? Undoubtedly it means 'possess,' or 'own.' But what if the horse is not mine, and I am not the possessor, or o"«Tier, of it 1 Again : what is the notion of possession in I have written a letter'] I may, perhaps, have it in my hand, or hold it. If I do I, pro tanto, possess or have it. But I may have posted it, and possibly be willing to give 8 ANALYSIS AND SYNTHESIS. much to get it back. All this seems to be a long way from anything like possession or holdership. But it is connected with me, and, to that extent at least, is mine ; and if at the time of speaking I connect myself with the object spoken about, I am, in some sense, an owner, or possessor, or holder. At any rate, I have an intei'est in it, and am connected with a part of the transactions appertaining to it. If it were not so, why should I speak about it 1 § 17. The horse was ridden, and the letter written at a time anterior to the time of speaking about it. Hence the time of the verb have is present, the time of participle which implies the action is past. § 18. Let us now look to the changes which the i-edupli- cate form, like ye-ypafa, has undergone, as a sign of the Tense. There are some who believe that originally the whole word was doubled, i.e. tliat the reduplication was really a repetition. Thei'e is certainly some ground for this view. But, at present, it is suificient to take the ordinary Greek Perfect as we find it. This repeats the first consonant of the root, whatever that may be. It expands it into a syllable by sufiixing e — no matter what may be the vowel of the root. If a word begin with tioo consonants it takes the fii-st. only. But this is not the case in Latin, where from spondeo they make spo-spondi. Nor is it so in Mcesogothic, where from hlaij]a they make hlai-hlaup. § 19. All three languages sometimes change the vowel of the root — cipicio, hedopKa — parco, pe2)erci, and Jdaijm, hlai- hlaup, as we have seen. But this change of vowel probably arises from secondary causes. § 20. Thu-dly, in Latin frequently, and still more fre- quently in Mcesogothic, the reduplication di-ops ofi"; in which case the sign of the Perfect tense consists only in the change of vowel. § 21. As the language becomes more and more Analytic, the abandonment of the reduplication increases; but not that of the changed vowels. There are no reduplications in the INFLECTION. 9 languages derived from the Latin : a mere vestige (the word did is the chief instance) remains in the languages of the German family. § 22. Concurrent ^vith tliis change of form is a change of meaning. In Latin, words like cu-curri mean either ' / raw,' or ' I have run '; and words like scripsi, either ' / wrote,' or ' / have written ' ; and it is only from the context that we can tell, in the way of import, which is which. § 23. When Cfesar writes ' veni, vidi, vici,' the reader who looks at the three propositions as they stand by them- selves has two meanings to choose between ; for so long as the text stands isolated, there is nothuig to help liim in his choice. Does Caesar mean '/ have come, I have seen, I have coTiqicered,' or does he mean 'I came, I saw, I conquered' 1 If the letter, or message, which contained these words had contained more ; if it had given in extension his intentions in coming, seeing, or conquering, we might decide. If he means ' / have come, and have no intention of going,' then veni has the sense of a Perfect ; but if he means ' / came, and then went farther,' the import is that of an Aorist. Moie than this — we have no warrant from the mere form that the three words are in the same predicament, inasmuch as one may be Perfect in sense, while the other two are Aorist, or vice versd. The chances are that what applies to one applies to all ; but it is by no means certain tliat such is the case. Compare, or rather contrast, this with the definitude and precision of the forms like yeypafa and (.yiHi^/ci. § 24. In English we go farther than they did in Latin. In Latin, each of the two forms had two meanings. In English, both words like swam, (originally a Perfect) and words like moved (an Aorist) have only one meaning between them ; and that is the meaning of the Aorist, or the 7ion- reduplicate form. § 2.5. The Latin, as we have seen, lost in the way of de- finitude and precision by the ambiguity of each of its two Tenses; for though such combinations as ' comjiertum haheo ' 10 ANALYSIS AND SYNTHESIS. and ' satis Jictiim haheo ' existed, tliey existed only as rudi- ments of little influence or elEcacy. In English, however, the Analytical combinations ' I have/oiind,' ' I liav said,' and the like, are among the commonest expressions of our lan- guage. Nevertheless, the English sins on the side of redundancy. If forms like ' / moved,' ' / called,' are sufficient, what is the fimction of forms like ' I torote,' ^ I swam "I There are, as aforesaid, two forms, originally and really ; two Tenses with only one meaning between them. § 26. But of this redundancy the language is relieving itself. (a) No new word, or word of foreign origin, when it finds a place in English, forms its past tense by a change of vowel. It is never conjugated like wrote, &c., but always like moved, ho,. (b) When one of the two forms gives way to the other, and becomes obsolete, it is, as a rule, the form constructed by the change of vowel, i.e. the old Perfect, shorn of its re- duplication. The changes in the other dii^ection ai-e rare ; perhaps non-existent. When the last of these has become obsolete, the long line of the reduplicate Perfects, from words like yeypatpa downwards, will have come to its end, and the typical representative of the Synthetic system will have become abolished l»y the typical representative of Analysis. § 27. Such is our exposition of a single example of the difference between Analysis and Synthesis, and it cannot be denied that it has been a long, we may say a lengthy, one. But ' the whole is easier than the jiart,' and I believe that the example selected is the best one that can be found. How much is there beyond it 1 Is it a mere fact in the history of three languages, or is it a pai't of a systematic whole 1 How far can we invest it with generality? § 28. The principle, in nearly all its details, holds good for all the languages of the German family. EATE OF CHANGE. 11 The prmcii)''e, with nearly all its details, holds good for all the languages of the Latin family. The principle, as far as the liistoiy of the language goes, holds good for both the ancient and modern Greek. But the history of the Greek is not so varied as that of the Latin class. There are no separate and independent lan- ffuases that have grown out of the Greek, as the French, Italian, itc, have gi-own out of the Latin; and no such lan- guages as those of England and Scandinavia, that are the near congeners of the German. In short, we get fewer data in Greek ; and besides this, the rate of change has been slower. § 29. Of the Sanskrit on one side, and of the Lithuania and Slavonic on the other, we may say that the tii-st ends too soon, and that the last begins too late, to give us more than a single stage — or, at any rate, any series of changes — com- mensui-ate with the difference between either the Latin and the French, or the Moesogothic and the English. There is no doubt upon this point in respect to the Lithuanic and the Slavonic. Respecting the Sanskrit, more will be said in the sequel. § 30. The rate at which languages change varies with the language ; and it is very rarely that it is the same with two together. Tlie languages in which the stability seems to be at its maximum are the Greek, the German, and the Icelandic. The literary Romaic seems to be much less altered from the classical Greek than it really is ; but this is because it is written as much as possible on the classical models ; and it is beyond all doubt that, by this artificial adaptation, the natm-al tendency to change, even in the spoken language, has been modified. Of the dia- lects our knowledge has, of late, been greatly increased ; and it is certain that there is much to be learnt from them, § 31. The German of Germany, and to some extent of the Netherlands, is much in the condition in which it was nine 12 ANALYSIS AND SYNTHESIS. hundred years ago ; i.e. its inflexion is almost identical with that of the Old Saxon of Westphalia, and the West-Saxon (Anglo-Saxon) of Wessex : and though less Synthetic than the Mcesogothic, which, itself, is far less so than the Latin, Greek, and Sanskrit, it appears more Synthetic than it is when compared with the present literary English. This is certainly not only the most Analytic language in the world, but the one in which the change has been the quickest. § 32. The same applies to the Icelandic as compared with its derivative, the (literary) Danish and Swedish. These, less Analytical than the English, are Analytical as compared with the Icelandic. This last, since the twelfth century, when its prose literature begins, has been but slightly changed ; and the Icelandic of the newspapers of to-day is, practically, the Icelandic of the old classics, Snori-o Sturleson, and the Saga wi-itei's. Between this and the somewhat earlier language of the Ed da there is a slight difference ; and, again, there are certain archaic forms pecu- liar to the Runic period ; i.e. the time of the introduction of Christianity. These have a special value, inasmuch as they help us to connect the Norse with the Mcesogothic — the oldest specimens of the Scandinavian with the oldest speci- mens of the German proper, or Teutonic, group. § 33. Upon the s})ecial question as to the Hate of Change, and the conditions which seem to regulate it, far too little has been written. But the best contiibution to it is Peter- sen's work upon the relations, as Analytic and Synthetic, between the old Noi-se language and the new ; or the Icelandic and the Scandina\'ian of the Continent, i.e. the Swedish, Nor- wegian, and Danish. § 34. In these three countries a form of speech, which in the tenth century was certainly intelligible over the whole of Scandinavia, and which was always treated as a single lan- guage, seems to have preserved its original character in Iceland, the colony, rather than in Denmark and Sweden, the mother countries. With Norwav the case was somewhat CONTINUITY. 1 3 different. However, in all three there was change ; though in each at a different rate. It went upon the same principle, and affected the same inflections in the sjime way thiough- out. But it went on in Denmark quicker than it did in Sweden. Denmark, Petersen observes, seems always to have been a ceutuiy in advance of (i.e. more Analytic) than Sweden; so that the condition of the Danish in (say) 1400 was, there or thereabouts, that of the Swedish in 1500, and so on thi-oughout. Both, however, began to change early, i.e. before the date of the first specimens of their respective literatures. § 35. With Norway it was different. In Norway the oldest literatiire is Icelandic ; though, in Norway, it is more usual to call it The Old ^^orse ; and such it really is. It is in one well-known instance more archaic than the Norse of Iceland. And, as it began, so it contiaued ; little altered until the end of the fifteenth century. Then, concurrently with the Reformation, it changes so rapidly that, in less than a century, it is as Analytic as its sisters of Denmark and Sweden. § 36. We now retiu-n. Along with the question of rate we must take that of Contimdty ; clearly understanding that this is not so much the actual continuity of the language itself (for this is essentially continuous), but the continuity of our knowledge of it ; the continuity of the Kterature which em- bodies and fixes it ; the continuity which tells us what it was up to such a time, and what it was after that time. If gaps in the evidence present themselves, though the language is as continuous as ever, our knowledge of it, except so far as it is inferential, is interrupted. § 37. Now actual continuity of the history of any language in detail from the date of its earliest record to the time of its investigator is rare. Where there is actual continuity the chances are that the history of the language is a short one : and in respect to the two great classes — the Greek and the Latin — it is easy to see that, during one period at least, a break of 14 ANALYSIS AND SYNTHESIS. some sort must present itself. With a language long recog- nised as the standard of speech, long known as the only- exponent of a literatui'e ; a language that has for centuiies been colloquial among the men of any intellectual cultiva- tion ; a language wliich from the very fact of it being written had, to a cei-tain extent, fixed and established itself as national — with a language invested with such a preroga- tive as the concurrence of elements of this kind cannot fail to establish, no subordinate form of speech can well become fit for writing ; or, if so fitted, be applied to the composition of any work of importance, any work likely to survive the generation, or even the writer, that composed it. It must itself be established before it is written ; and before the time can come when this can take place a long series of unrecorded changes must be in operation, and something like a sepai-ate language be acknowledged. We know not when these changes first made their influence definitely perceptible ; still less when the earliest of them began. The germs of some of them — as in compertum habeo, &c, belonged to the parent language ; possibly in its eai'lier stages. But for their full development time is needed ; and during this period the change goes on silently, and the details of it are unrecoi-ded. With the dei-ivatives from the Latin this is self-evident. § 38. With the Greek it is somewhat different. The dialects of the Greek never developed themselves into sepa- rate and independent languages like the French, Spanish, Roumanian, and others; and, besides this, they were rarely written without some classical, semi-classical, or would-be- classical model in the mind of the writer. The Byzantine literature was continuous, no doubt ; but its vernacular cha- i-acter is moi-e than doubtful ; and even to the standard Romaic, as it is written at present, the same objection ajjplies. It is only in the true provincial dialects of Greece, its islands, and parts of Asia Minor, that the real growth of the language can be studied, and that but imperfectly. CONTI>'UITY. 1 5 § 39. In the German family tliis want of continuity is not so great : nor is it so closely connected with the history of the language in general ; but, at the very beginning there is a deplorable hiatus in the record ; for the Moesogothic stands by itself. The language was reduced to writing long before any other ; and it was not upon German giound that the translation of the Gospels was produced. It is due to the influence of the Greek Church, and the alphabet was founded on the Greek. It was not till more than two cen- turies later that Christianity introduced the alphabet into either Germany or Britain : and the earliest records here are chiefly from Britain and Southern Germany. Neither of these is anything like identical with the other, and neither is what we look for as a lineal descendant of the INIoesogothic. The German resembles it the most in its structure, the British in its phonesis. There is a gap, then, in respect to both time and space — both date and locality. § 40. This leads us to another point of view. The history of a language may be continuous ; but histories of its dialects may be difierent. One dialect may represent the language at one time, another at another. In Great Britain, the present literary English is that of the Midland dialects ; but for the Midland dialects we have no record earlier than a.d. 1175, if so early. The literary language of King Alfred's time was the West-Saxon, or the English of Wessex. By a.d. 1400 it was all but superseded by the IVIidland forms of speech. Its history, since then, is that of tlie provincial dialects south of the Thames, especially those of Dorset, Somerset, Devon, and Wilts, where the old characteristics are best preserved, or rather the least lost. The Lowland Scotch, a cultivated language from the time of the first Stuarts, is a dialect of the great Northiimbrian division. But it is only represented by small and doubtful fragments until the 14th century. Hence we find, for the language of England at lai-ge, one history ; for its dialects (all of which, under difierent political and geographical conditions, might pass for different Ian- 16 ANALYSIS AND SYNTHESIS. guages), three. Thei-e is continuity here ; but it is continuity of an imperfect and (/iscontinuous kind. It may be continuous in respect to a language at large, or in a general way ; but not continuous in respect to its several dialects. Nevertheless, even in respect to these, it has a continuity, as we expect a priori. But the records of it may be discontinuous ; and this is the case when one dialect at one time is the litei-ary language, and another at another. There are no ilata for the subsequent enquirer as to the lan- guage that has lost its prerogative. This is the case with no language more decidedly than our own. We know the West- Saxon up and down to a certain time ; and that well, for it was the literary language of the times of Alfred and Edward the Confessor. Then a change comes over the whole lan- guage ; we might say an eclipse. About a.d. 1200 the lan- guage revives, and it is only up to that time that we can trace the present literary English. But the language of Alfred is, at present, a provincial dialect ; while the English of Macaulay is unknown in its exact foim beyond (about) 1200, and the Scotch of Burns is but obscurely known till a century later. Yet both existed when the West-Saxon was the sole (or main) representative of the German of Britain ; in other words, the English language. The record, then, of the language, is uniform ; of its representative dialects, frag- mentary. § 41. We now return to oiir old friend, the typically Syn- thetic form yf'ypaa.' This is because it does not wholly eliminate the single-worded form from the language, for we still say ^I simj.' The sense, how- ever, is slightly different. '/ sing' means that / sing habitually, rather than that ' / am now in the act of singing.' In Latin, words like ' canto ' had both meanings. By saying ' sum in cantatione,' or ' sum cantans,' they might indicate the same distinction that we do. Combinations, however, like the fii-st were rare, and those like the second rarer — if existent. Here, then, we have Synthesis and Analysis, but without the displacement, or supercession, of the one by the c 2 20 ANALYSIS AND SYNTHESIS. other. Still ' / am singing ' may pass as an instance of A nalysis. § 54. But though it be this, it is notorious that it is not veiy easily analysed. We know that it is made up out of two words ; hut whether these are to be translated ' simi cantans ' or ' sum in cantatione ' is a point upon which there is no unanimity of opmion. § 55. Again — '/ do sing' — this is a form that no one would translate by 'facio' (or 'ago') 'ca7itare,' but rather by ' canto.' Yet, even when it is so ti-anslated, the combina- tion is anything but analytic, and canto is manifestly synthetic. But ' / do sing ' and ' canto ' are not analytic and synthetic to one another, for they by no means coincide in imjDort. In other words, there is no con-elation between them. ' Canto ' tells us that the speaker is either singing when he tells you that he is doing so, or that he sings habi- tually. In / do sing the speaker merely states that he does-not not-sing. It implies a previous state of denial or doubt as to whether he sings or not, and as such is used emphatically, as if in reply and contradiction to some one who said he did not. § 56. The full name of the stage of language under notice is either Synthetico-Analytic or Anal gtico- Synthetic. How- ever, when spoken of as such, and especially Avhen compared with the terms '■Agglutinate' and 'Monosyllabic,' 'Agglu- tination' and ' Monosyllahicis7)i,' it will simply be called ' Syntltetic,' and ' tSyntheticisni' or 'Synthesis.' § 57. Analysis by no means exckides Synthesis. Many forms of speech which, from the general amount of change that they have undergone, and fi-om the undeniably analy- tical character of the majority of their phrases, may reason- ably be called Analytic, preserve remains of their old and more original synthesis. This, however, only means that the change from the predominance of one foi-m to that of another is not sudden, but gradual, a fact of which we only need to be reminded. INCIDENCE OF THE CHANGES. 21 § 58, More important than these ai-e the Postpositive Ai-ticle of the Xoi-se and Roumanian languages, the so-called Passive Yoice of the ISTorse, and the Futures {jyarlero =2)arabola7-i haheo-=^I have to S])eak=loqitar) of the lan- guages derived from the Latin. These have been actually- constructed either within the Analj^tic period or during the more equivocal period of the transition. On these, more will be said in the sequel. § 59. It has manifestly been the Inflectional part of language which the foregoing criticism has most especially illustrated ; for it is this which the system of Synthesis and Analysis best explains. It, of course, when explanation is needed, can be made to do much more than this ; but the lan- guages which have been referred to explain it better than all the other languages of the world put together. Moreover, it is only partially that even Inflection has been explained, inas- much as next to nothing has been said as to what Inflection really is when considered by itself. All that has been said has been in respect to its connection with Analysis. ^^Hiat will soon follow will be the relation of the Inflectional system to that of the Formatives : and just as the Synthetic stage has illustrated the fonner, the Agglutinative stage will illustrate the latter. § 60. More, then, will be said upon the structure of the Formatives in the sequel. At present the difference between them is only indicated by a single instance. The -mus in both voca-mus and vocit-amus is Inflectional : the -it- in voc- it-amus is Formative. Sub-Section II. MATERIALS FOR THE SYNTHETIC AND ANALYTIC STAGE OF LANGUAGE. § 61. The forms of speech that illustrate this period are few, but veiy important. One and all they represent lan- guage in its latest known stage. Hence, they are the mate- 22 MATERIALS FOR THE rials with wliich we begin our investigations ; for the right method is that of proceeding from the hxter to the earher, from the more certain data to. the less uncertain, from the known to the unknown. § 62. Besides all this, the languages themselves are, upon the whole, those that ai-e not only the best worth understanding, but the best understood, viz. the Greek, the Latin, the Sanskrit, and the German ; and, in a less degree, the Lithuanic and the Slavonic. § 63. Again, these are not only the languages that have undergone an important amount of change, but those which have an adequate literature wherein the several changes are recorded. § 64. Among these the first place must be assigned to the Greek. The Greek records begin earlier than the Latin ; and, like the Latin, the literature descends to the present time. But, on the other hand, and as a drawback to its otherwise high value, there are no languages that stand in relation to the Greek as the French and Spanish, &c. do to the Latin. As a set-off to this, the Greek begins with a greater variety of dialects ; some of which, as the ^olic, the Doric, the Ionic, and the Attic, may pass for different lan- guages. Moreover, at the time when the structure of the two languages fii'st becomes open to comi^arison, the Greek is the more synthetic of the two. § 65. The Latin begins later than the Greek, and with fewer dialects. On the other hand, it undergoes far more change than the Greek, and that in all the languages derived from it, which are more than one : viz. the Italian, the Spanish, the French, the Proven9al, the Romance, and the Roumanian, not to mention numerous dialects and sub- dialects. § Q'o. The Sanskrit, both in its antiquity and its synthetic character, stands on the high level of the Greek ; but it has not the same undoubtedly continuous history. § 67. Again, the Sanskrit, both in its antiquity and its A^^ILYTIC AND SYNTHETIC STAGE. 23 synthetic charactei", stands on a higher level than Latin ; but here a similar question suggests itself. Has it an equal amount of difference and vai-iety in the way of dialects and derived languages ? § 68. There is no answer to either of these qxiestions which commands universal assent. There are those who con- sider that the present languages of Northern Inilia, the Hindi, the Gtijerathi, the JIarathi, the Bengali, the Uriya (or Udiya), and others are not the descendants of the Sanskrit ; that they are not in their relation to that language what the English and German are to the Moesogothic, the Italian and French to the Latin, or the Romaic to the Greek. There are those who think that the Hindi, kc, are only Sanski-it in the way that English is Latin, i.e. through a strong inter- mixtme of Latin words, whether directly from the language of Rome or indirectly through the French. They deny the gram- matical or structural connection ; and this connection they take (as do their opjionents) as the evidence and test of descent. § 69. Those who hold the last view are innovators ; for the doctrine in favour of the Sanski'it origia of the languages of Northern India is the older one, and it has also any amoimt of authority to support it. They are in a minority ; and a fair display of argument, whether sufficient or insuffi- cient, has been brought against their views. Still the doctrine thus strongly supported is not (so to say) Catholic. The belief that the Romaic is Greek, the French Latin, and the English Gennan is held hie, ubique, et ah omnibus. The same belief is not held respecting the Sanski-it and the HindL Hence, though at present I pass no opinion on the question, I abstain from j^utting a disputed fact on the same level as an admitted one ; for if the present work have any value at all, it derives it from the admitted certainty of the facts upon which its I'esults are founded. Hence the most convenient way in which I find it necessary to treat the term 'Sanskrit,' as a class-name, is to speak of it as the ' Sanskrit and its 24 ANALYSIS AND SYNTHESIS. congeners icliatever they may he' That the Pcdi and Zend ai-e such congeners no one doubts. § 70. The German, though far inferior to the three above-named languages in both its synthetical structure and its date, gives us a more varied, as well as a more continuous, histoiy than any of them. Its high value in this respect is absolute and intrinsic ; but to those who speak it, whether Germans proper. English- men, or Scandinavians, it is the mother-tongue, and as such the form of speech of which they know more than they know of all the other languages of the world put together. It is the language with which they compare, and by which they measure, all others. It is not only the best language to begin with, but it is the one with v,^hich we, as German, must begin. § 71. Of the German family of languages the history, which is scarcely half so long as that of the Greek, and full five hundred years shorter than that of the Latin, begins with what is known as the German of Mcesia, or the M(£So- gothic of the Ulphiline Translation of the Gospels, of which the date is, within a few years, a.d. 375. As compared with the present English, the Moesogothic is undoubtedly syn- thetic. But it is not synthetic after the manner of the two great classical, and the one great Indian, languages, viz. the Greek, the Latin, and the Sanskrit. § 72. The history of the German family of languages is both varied and continuous. It cannot, however, be added that its continuity presents itself in the first instance, inas- much as the Moesogothic, for more than thi-ee hundred years, stands entii-ely alone. Neither was the Mcesogothic, as we know it, adequately known in its historical or its geographical relations to Germany. It was for German emigrants rather than for the Germans of the mother-country that the Ulphi- line Translation was made. Their settlement was in the Koman pro\'ince of Mcesia, not far from Plxilippopolis. But THE GERJLIN FAMILY. 25 before they settled here they had occupied a part of Dacia ; and, before they crossed the Danube, had probably been soldiers in the Roman service on some part of the Marco- manic March. § 73. As to the part of Germany in which they origi- nated there is no unanimity of opinion. Michaelis refen-ed the language to Thuringia ; and suggested that the present dialect of that rlistrict was the best existing representative of the Mcesogothic. I am not awai-e that anything has been adduced that invalidates this opinion; and I am satisfied that there is much which tends to confirm it. But, be this as it may, there is a gap of more than three hundi-ed years between the Mcesogothic and the next representative of the German family. § 74. It is the paits south of the Danube, and on ground which made no pai-t of the Germany of Tacitus, that we find the Old High German, which is mainly that of Suabia {D ecu- mates agri), Alsace, Bavaria, and Switzerland. The earliest specimens of this ai-e older than anything on the imdoubted soU of Germany. These date from about the eighth century. § 75. It is in Britain that the German, now known as the Ewjlish, fii-st presents itself : a few years earlier than the oldest High Gei-man. And this is, for the third time, German, on ground other than that of the Germany of Tacitus. The German of Britain is represented by three literatm-es : (1) the West-Saxon, for the part south of Thames, from about a.d. 750 to 1400; (2) the Midland, or 2Iercian, which is the pi-esent literary English; and (3) the N'orthumhrian, of the pai-ts between the Humber and the Forth. Of German, on the undoubted soil of German, the Old Saxon of Westphalia gives us the earliest sj^ecimens ; about A.D. 900. The Frisian is specially akin to this, as well as to the West-Saxon of Britain, and, moreover, is considered to be transitional between the Teutonic and the Norse, or 26 MATERIALS, ETC. Scandinavian groups. Of this last, which is most especially illustrative of the change from Synthesis to Analysis, enough has been said in §§ 32-35. § 76. It is manifest that in this family we have a most valuable store of materials : a fair amount of continuous his- tory ; a fail* amount of separate and independent languages ; a very considerable number of dialects and sub-dialects ; and other elements of value besides. But with all this, the main element of its importance is relative rather than positive. It is the family to which our own language belongs, and, consequently, the family which we best understand. § 77. The Lithuanic Famihj. — In this there is nothing earlier than the sixteenth century, and this is in the lan- guage of the western parts of East Prussia, in which nothing is now spoken but Grerman and Polish. This is called Prussian, as opposed to the LWmanian of Lithuania proper, and Old Prussian, wliich, so far as it means anything, means the old language of Prussia ; for of anything that can be called Middle, or New Prussia, we have no vestige. In this Prussian the demonsti-ative pronoun is used as the definite article, whereas the allied dialects are without one. On the east and north-east of Lithuania pi-oper lie the Lett districts of Courland, Livonia, Esthouia, Vitepsk, and Polotsk ; and a division of some kind between the Lett and the Lithuanian, as dialects or languages, is rightly recognised. But what it amounts to is another question. The statement that Lithuanian is notably more synthetic than the Lett, and the two tongues are as widely separated from one another as Latin and Italian, is a gross exaggeration. What light the Lithuanic family thi'ows upon the field of the philologue is practically that of a single language in a single stage. But, thus limited, the Lithuanic is a language of inordinate interest. In the declension of its noims it is more synthetic than the Latin. § 78. The Slavonic Family. — The first point to notice in this is the number of its divisions and sub- divisions. Its THE SLAVONIC FA3IILY. 27 exti'eme forms, or tlie languages most unlike one anotlier, are tlie Polish, or the Bohemian, on one side, and the Eus- sian, or Moscovite of Great Russia, on the other. 1. T\iQPolish,^iih.th.eKassuhic; a fragment still spoken in Pomerania ; and the Linonian, of which a fragment remained in Liinenburg in the latter half of the last cen- tuiy. 2. The Sorabiaii, still spoken in Lusatia and Branden- burg. Intermediate to the Polish and 3. The Bohemian, Czech, or Tshekh. This is the name that the Slavonians of Bohemia gi^e themselves. In Mo- ravia the language is called Moravian, and in Upper Hun- gary, Slovack. These are more closely allied to each other than to the languages of the second branch. 4. The Slovenian, spoken by the Slavonians of Austria, who called themselves Sloventzi. It is a political, rather than a philological, di\-ision, inasmuch as it graduates into the Servian. § 79. All these nations use an alphabet of Roman origin, and, except when Protestant, are Roman Catholics. The Slavonians, however, originally used an alphabet founded on the Greek, and called Gla(jolitic, or Glagolite. 5. The Servian, spoken ia Bosnia, Servia, and INIonte- negro. Alphabet of Greek origin ; and creed, the Christi- anity of the Eastern Church. 6. Russian. This falls into the Ruthenian, Rusniak, or Red Russian, of Eastern Gallicia (originally Lodomeria) ; the Little Russian {Malo-russ) of Kiev and the surrounding Governments ; the Black Russian of Grodno and Minsk ; the Wliite Russian of Smolensko ; and finally, the Moscovite, or Great Russian, of the rest of the Empii-e. 7. The Bulgarian ; which probably differs from all the others as much, or moie, than they differ fi-om one another. Alphabet and creed of both Russia and Bulgaria. The same in Servia. 28 ANALYSIS AND SYNTHESIS. For so large an area, and so many divisions, the general uniformity of the whole Slavonic class is remarkable. SECTION II. AGGLUTINATION, FORMATIVES. § 80. The -0 in voc-o is an Inflection ; so also are the -amus, -atis, -avi, -avissem, &c., in voc-anius, voc-atis, voe-avi, voc-avissem, and the like. The -it- in voc-it-o is a Formative ; and so are voc-it-amus, voc-it-atis, voc-it-avi, and voc-it-avissem. A Formative is different from, and opposed to, an Inflection. § 81. The outward and visible difference between the two is that the -it- keeps its form throiighout the whole conjuga- tion of the verb to which it belongs ; whereas the termination -0- changes with the Person, Number, Tense, Mood, and Voice, becoming -amus, -atis, and the like, according to cii'- cumstances. Changing the expression, the -it- is a part of the theme, and, as such, is pei-manent ; whereas the inflection is acljuiict to the theme, and mutable. § 82. The -it- in voc-it-o belongs to the theme, but not to the root, and it is to the root that it is attached. So attached, it precedes the inflection in the way of order. The uiflection attaches itself dii-ectly to the I'oot only when there is no such intervening syllable as the one under notice. This is the view of the two adjuncts simply and solely in respect to their place and permanence. § 83. The logical difference between them, or the natiu*e of their respective meanings and functions, is another ques- tion. The -it- in voc-it-o expresses a difference in the nature of the act (that of calling) itself The inflection denotes cer- tain circumstances that attend the act ; viz., the agency, the time, and the conditions under which it takes, or may take, place. In this way the first of them indicates that the speaker, the object spoken to, or some object spoken about, is the agent, or, if there be more than one, the agents. In AGGLUTINATION, FOKMATIVES, 29 the way of time it indicates whether it is Past, Present, or Futiu'e; in the way of conditions, whether it he Indicative, Imperative, or Conditional. In other words, they give signs of Person and Number, Tense and Mood. With the character of the act (or state) itself these have nothing to do. On the other hand, the diiference between vuc- (the root) and voc-it- (the theme) relates to the act itself, which is, of course, so far as the voc- goes, one of the same kind as that denoted by voc-it-. The -it-, however, indicates a difference. Voc-o is translated / call ; voc-it-o, I call often : and as there are many words thus differentiated, there is a class of forms which are called Freqioentative. What -it- is in Latia, - efl- is in Greek, as ^Xty-w, fXey-ed-w. In vire-sco=I begin to grow green, and yepo-o-kw=/ begin to grow old, the -ak oabak Genitive jSgai gieSai oabai Infinitive j%aid gietiaid oabaid Allative jagaidi gieftaidi oabaidi Factive jSkkan gie^an oabban Locative jSgain gie'Sain oabain Comitive jagai-guim gie^ai-giiim oabai-guim Caritive jSg-ai-taga gie6ai-taga oabai-taga Persons. Singular Number. First. Second. Third. Nominative mSn (/) dSn (thou) san (he, ^-c.) Genitive muo dii su Infinitive muo du su Allative munji diinji sunji Locative must dust sust Comitative muin duin suin Caritive muotaga dutaga Ihuil. sutaga First. Second. Third. Nominative mai dSi sai Genitive mSdno daduo sad'^o Infinitive mSdno dSduo sSduo Allative mSdnoidi dadnoidi sadnoidi Locative mSdnoin dadnoin sadnoin Cotnitative mSdnoguim dSdnoguim s§duoguim Caritive mSdnotaga daduotaga sadnotaga 34 AGGLUTINATION. Plural. First. Second. Third. Nominative mi si di Genitive mill sin din Infinitive mill sin din Allative migjidi digjidi sigjidi Locative mist dist sist Comitative miuofuim dinguim singuim Caritive miutag-a dintaga siutaga t lajioin § 04. Active. . To In^dicative. Present. ^COS& Singular iSuom lauok iSdno Dual lodnu lonol)£ette ladnoba Fiural laduop laduokettet Preterit. lodniik Sinrpdar lodmim lodnuli lanoi Dual lanoime ISuoide lanoiga Plwal lauoimek • lauoidek OoifJTJNCTIVE. Present. lodnu Singular lonucam lomicak lonuca Dual lonuciedne loiiuc;eppe louudteva Plural loiiuciep louuctbppet Pi-eterit. lonucek Singular lonusim lonusik lonusi Dual lonuseime louuseide lonuseiga Fiural louuseimek louuseidek Imperative lonuseigje Singular , , lano lodnus Dual lodnu ladno lodnusga Plural iSduop iSduot loduusek THE FIN y^m^.V^F^B 35 Singula)- Dual Plural Singular Dual Plural Singular Dual Plural Singular Dual Plural Singular Ihial Plural Passive, ixdicatite. lodnuju\aim lodnujuvve loduujuTvup lodnujuTvim lodnujuTuime loduujuTuimek Present. lodnajuvuk lodnujiivvu lodnuj uvTubffitte loduuj uvaba lodnujuvTu'bsettet lodnuj ujek Preterit. lodnuj uTvik lodnuj uvuide lodnuj uvuidek lodnuj uvni lodnuj uTuiga lodnujuvve Co^^JTINCTIVE. Present. loduuj uwucam lodnujuvvucak lodnuj a^'vu(5a lodnuj uvTucsedne lodnuj uvvucaejipe loduuj u^'vu(5Eeve lodnujuvvoc'eep lodnuj uvvuceeppet lodnujuvvucek Preterit. lodnujuwusim loduujuvvusik loduujuvvusi lodnuj uwuseime lodnuj uwuseide lodnuj uvvuseiga lodnuj uwuseiniek lodnuj uvvuseidek lodnuj uvvuseigj e Imperative. lodnujuvvu lodnujuvvus lodnuj uwujsedno lodnuj uvvujselike lodnuj uvvusga lodnuj uvvujsedno lodnujuvvujsekket lodnujuvvusek § 95. Verhum Kominale ACTIVE. Injinitive ISdnot Factive {sujnne) ISuodet Comitative (gerund) lanodediu PASsrvE. lodnut and loduuj uvvut lodnucet „ lodnuj uvvucet loduudedin „ lodnuj uvvudedin C'aritive ISnoksetta lodnukEetta „ lodnuj uvvukajtta Noinen Verbale. ACTIVE. PAssn'E. Kumen Subst. C Abstract iSdnom Verbale \ Concrete iSdno Komen Adject. ( Pres. Part, ladnome Verbale \ Pret. Part, laduom lodnuj ubme lodnujuvve lodmijuvvume lodjuvvum; loduum D Z 36 AGGLUTINATION. § 96. This apparent Synthesis is only superficial ; for the doctrine that underlies the whole system of the growth, development, or evolution of language, presupposes that all Inflections, if we can only reduce them, are reducible to words originally independent. The mere amount of these is a secondaiy matter ; the question is the evidence in favour of their reducibility, and it is certain that in languages like the Fin, the Tui-k, and their congeners, this is notably improved. § 97. Roughly, and for the sake of illustration, we may compare the union of two or more words in KSynthesis to a graft, in which the union is so close as to be indecomposable. The union in the way of Agglutination is a splice, whereof the separation is comparatively easy. We cannot always — perhaps we cannot often — say what the incorporated element was when it stood alone ; but we can generally convince our- selves that, of some sort or other, it is a superaddition. § 98. Again, when we have two inflectional elements in combination, one of which suggests the notion of (say) num- ber, and the other of (say) case, we can, in Agglutination, separate the two signs with far greater certainty than we can in Synthesis. The Turkish grammar thi-ows a strong light on these points. Adeni : = man. Nominative adem lidemler man, men Genitive ademiin ademler-un of „ Dative ademeh ademler-eh to „ Accusative aderai ademler-i man, inen Ablative ademden ademler-deu from „ Compare this with the declension of Iwmo, or any Syn- thetic substantive, and see whether the separation of the signs of number and case ai-e anything like so distinct. § 99. The Conjugation of the Verb is still more sugges- tive of differences. How easily arc the several significant affixes separated from each other. TUKKISH XOUN AND VEEB. 37 Negative Impossible Negative Impossihle Negative Impossible Passive Neg. Passive Sumek = Zoie {Pi-esent Imperative). Active. siimemek not to love suelunemek suilmek suilmemek suilehmemek not to he able to love Passive. to be loved not to be loved not to he able to be loved Causal Active. sudermek pudermemek suderelunemek suderilmek suderilmemek to cause to love not to cause to love not to be able to cause to love to be made to love not to be made to love Ifup. Passive suderilememek 7tot to be able to be 7nade to love Passive Causal. Euildermek to cause to be loved Negative suildermemek not to cause to be loved Impossible suilderehmemek not to be able to cause to be loved Recipeooal. suishmek to love one anothe?' mutually Negative suishinemek not to love ^'c. Imjjossible suisheliiuemek not to be able ^-c. Passive suishilmek to be loved mutually Neg. Passive suisliilmemek not to be loved mutually Imp. Passive suisliilehmemek not to be able to be loved mutually Causal suisliderinek not to be able to be caused to love mutually PEESOJfAL. suinmek to love himself Negative Eiunmemek not to love Sfc. Impossible suinelimemek not to be able ^-c. Passive suinilmemek to be loved Causal siiindermek to cause to love himself 38 AGGLUTINATION. § 100. And last, not least, as an element in the more manageable character of the Agglutinate languages, most of them are without the distinction of Gender. That this facili- tates the separation of case-endings into their elements is manifest. Three distinctions are involved in the Synthetic, two in the Agglutinate, inflections. By the 6nal syllables in such words as hon-um and hon-as, three differences are indi- cated, that of Case, of Gender, of Number ; and the work is done by a single combination of two letters. § 101. The following is an equally illustrative instance, from a language now nearly extinct, but probably inter- mediate to the Turk and the Mongol. Yenisei AN, Bieng = hand. Bieng-en = hands. Singular. Plural. Nominative bieng bieng-en Genitive bieiig-da bieng-en-da Dative bieng-deug bieug-eng-den Locative bieng-gei . bieng-eng-ei Ablative bieng-denger bieng-en-denger Instrumental bieng-sas bieng-eu-siis Frosecutive bieng-bes bienff-en-bes Comitive bieng-san , bieng-eu-sau Sub-section II. MATERIALS FOR THE AGGLUTINATE STAGE OF THE LANGUAGE. § 102. In the amount of the material, or data, which it supplies the Agglutinate class shows to disadvantage. The languages are numerous, no doubt ; but there is scai'cely one Agglutinate language of which we have enough. Some begin early, but then they become extinct prematurely. Some (indeed, the majority) are in full life at the present moment ; but, on the other hand, we know nothing of them in any earlier stage. We have, in the Agglutinate class, languages DATA OK MATEEIALS. ' 39 which are in many respects semi-Synthetic, and others which are well-nigh Monosyllabic ; but we have no single language which, in one of its stages, is Monosyllabic, and, in the other Agglutinate. In many cases there is a fair amount of both duration and continuity ; but, even when this is the case, the amount of change is inconsiderable. § 103. The Fin of Finland, the Finlandic, the Finlandish, or the Fin proper of the Duchy of Finland, is a single lanr guage of a class : a class sometimes called Fin, sometimes Ugrian ; and it is the number and character of the different divisions and sub-divisions of the classes which, invest it with its recognised importance. Tavastrian, Karelian, Vod, Tshud, Vesp (or Ludin), Krivonian, Lief, Esthonian, Per- mian, Zirianian, Votiak, Lap (or Laplandish) of (a) ^^or- way, (b) /Sweden, (c) Russia, Tsheremiss, 2Iorduin, Yogul, Ostiak, Samoyed, and (1) Tshuvash, are all denominations which, either as separate languages or dialects, supply illus- trations of some kind of the difference and likeness between the several members of this very important class. § 104. Next in geographical order come the Turk, Ye- niseian, Mongol, and Mantshu, ftimilies, the last two of which are in contact with the Monosyllabic area. The Yeniseian is still implaced ; but it is probable that it belongs to the same division as the languages with which, it is here associated. Of the remainiag languages of Northern Asia, of which the affinities run either in the direction of Japan and the Kiu^ile Islands, or that of Kamtskatka, the Aleutian Chain, and America, it is not necessary to take cognisance. § 105. The Dravirian or Dravidian Family. — Spoken in Southern India, and opposed to the languages represented by the Hindi. This means that, ever siuce attention has been paid to it, it has been recognised as other than Indian of the Hiadi, Gujerathi, and Bengali type; i.e. has never, either rightly or wrongly, been considered a derivative of or from the Sanskrit. The foiu' best known members of this group, each with a literature, are— 40 AGGLUTINATION. 1. The Telinga, or Telugu, spoken by about 14,000,000 Dccupants of the Northern Cii'cars, and paits of Hyderabad, Nagpur, and Gondwana ; bounded on the North by the Oriya of Orissa, and on the side of the Pacific by Chicacole, North, and Pulicat, South. 2. The Tamul, spoken by about 10,000,000 natives from the parts about Pulicat to Cape Comorin, and on the West as far as Trevandrum. In the interior it extends to the Ghauts and Nilgherris. 3. The Canarese, spoken by about 5,000,000. Mysore is its centre; bounded by the Telinga N.E., and the Tamul S.E. ; on the North by the Marathi, other than Dra- virian. 4. The Malayalim, spoken by about 2,000,000 along the Malabar coast. § 106. These are the four old, or long recognised, mem- bers of the Dravii'ian ftimily ; but a fifth has recently been added, and that as an independent language. This is — 5. The Tulu, spoken by about 150,000 about N.L. 13° 30' along the coast, and, like the Canarese, bounded by the Marathi. The C'urgi, of Citrg, has scarcely been raised to the level of a language. As dialect it has been claimed for the Tulu, the Malayalim, and the Tamul. To these add not fewer than five barbarous dialects, or sub-dialects, of the Canarese, and spoken in the Nilgherris by the Tudas or Todavas, the Budugars, the Kohatars, the Irular, and the Kurumhas. § 107. Later researches have brought no less than four more groups within the range of the Dravirian family, though the position of one of them is equivocal. § 108. On the western frontier of the Uriya (Hindi) and the northern frontier of the Telinga (Dravirian), in the Circar of Gumstir, the Khond in several dialects is spoken : Gadaha, Yerukale, Savara ; and in Gondwana the Ghond, or Gtmdi. This is not so manifestly Dravuian as the Khond ; but still MATEEIALS. THE DEAVIEIAN FAMILY. 41 it is more Dravirian than auglit else. Little is said about it in Caldwell's standard Draviiian Grammar ; but he does not exclude it. § 109. But with the next group he hesitates. It is the most Northern (North-Eastern) of any ; and is Exirly repre- sented by vocabularies — Sontal, Slnghhum, Mundala, and Bhumij. Ho, meaning ')nan, is the name which the A'olehan apply to themselves. Ramgidir, Mongliir, Chuta ISTagpur, Gangpur, SLrgujah, Sumbulphur, Palamaw, and the Rajma- hal Hills, where the Sontals are stated to be intrusive, are, in the ordinary maps, the most conspicuous names for the Kol districts, and Kol is the usual name for the group. § 110. The Kol has recognised affinities with the Mono- syllabic family, and how closely the two groups approach one another geogi-apically may be seen in § 135. § 111. The Brahui. — This is the last language that has been connected with the Dravirian ; and, in respect to its geogi-aphy, it is other than Indian. It is Persian ; and, what is more, is separated from the most northern frontier of the Dravirian proper by the i\^o?i-Dravirian Gujerathi and Ma- rathi, not to mention the Persian and Indian dialects of the Lower Indus. It is in the mountain districts of Beluchistan that the Brahui, Brahooe, or Brahuiki is spoken, unwritten, and its Dravirian structure disguised by the inordinate amount of Persian with which it is overlaid. Those, how- ever, who have paid most attention to the subject, have pro- nounced in favoiu' of the Dravirian affinity. § 1 1 2. With this ends the list of those Agglutinate lan- guages of Eastern Europe and Asia which come imder the conditions and limitations of the present treatise; viz., (1) that they should be spoken in something like geographical contact with the languages of the Synthetic and Monosyllabic division, and (2) that the recognition of them as Agglutinate should be universal. § 1 1 3. Far apart from those, however, and on the farther side of the Synthetic area, on the extreme south-west of 42 AGGLUTINATION. EuroiDe, still sta.nds over the Iberian, Uuskarian, or Basque, whicli may reasonably be considered to represent, at least, the original language of the whole of the Spanish Peninsula. It is now limited to the provinces of Biscaya, Guipuscoa, Alava, and Navarre, in Spain, and parts about St. Jean do Luz and Labourdin, in France. Bounded on three sides by the sea, the Iberian Peninsula is isolated, whilst of the Pe- ninsula itself the Basque area is but a fraction ; and, fraction as it is, it is bounded on both sides bv intrusive lanofuases — the French and Spanish — both of Latin origin. After these, there is nothing nearer than the Celtic of Brittany, and the German of Switzerland, and the Slavonic of Bohemia. No wonder that its affinities are a mystery. § 114. Such^ are the chief Agglutinate languages of Eui'ope and Asia in respect to their geographical areas. We see, at once, that they are in contact with one another, and that some of them, the Ugrian and Turk most especially, are in contact with the North and North-Eastern members of the Synthetic class ; and we shall see, in the sequel, that as we move southward, others will touch the Monosyllabic frontiers of China, Tibet, and Northern India. This invests them with a special value, inasmuch as, over and above the inter- mediate or transitional character of the Agglutinate stage of gi-owth or development, theii- geographical position has a like intermediate chai-acter, suggestive of elements of affinity of a more specific character. § 115. And such is the case. There are many more Agglutinate languages than these — languages that ai-e to be counted by the score and hundred. Indeed, we may antici- pate and add that, with a few possible exceptions, all the world over, whether in Africa, America, Austi'alia, or Poly- nesia, whatever in the way of speech is neither SjTithetic, nor Monosyllabic, is Agglutinate, Nevertheless, it is only the few under notice that will be specially dwelt upon, § 116. So far, then, as mere material goes, the Aggluti- nate class is abundant to excess. But the material is nearly . ]\IATEKI.\LS. THE HEBREW. THE IRISH. 43 all of one kind; viz., a multitude of different languages, and an average amount of dialects and sub-dialects for each. To set against this, we must state that, as a rule, the material is of the later, or latest, date. § 117. The great exception to this statement, the oldest language of the class, and, perhaps, of the world at larg^, is the Hebrew, the only language in the Agglutinate division that can be compared, in respect of its antiquity, to either the Greek and Latin of Em-ope, or the Sanskrit and its congeners of Asia. This, or the Jews' language, is that of the Old Testament. But the language of the New Testament is Greek ; and it is in the earliest translations of the New Testament that we get the earliest monument of the languages that are now about to present themselves. § 118. To the time between the Bii'th of our Saviour and the Flight of Mahomet (i.e. the Christian and the Mahometan epochs), the New Testament (wholly or partly) is translated into Syriac, Cojytic, Uthiojyic, and Armenian. Then follows the Amharic and Georgian versions ; the former suggested by the Ethiopic, or Gheez, the latter by the Armenian. The earliest Irish belongs to the same class. It is safe to say that before the Flight of Mahomet it was a written language, and it is possible that some of the existing specimens of it may be of that age. This is the period represented in the Synthetic stage by the IMoesogothic translation of Ulphilas. Out of the Sp'iac alphabet grew the Ai-abic ; and with this, the Ai-abic itself, before the time of Mahomet, became a written language. § 119. The date of the oldest compositions in the Geor- gian is uncertain. Probably, they date from the introduction of Oiristianity ; but are not so old as the oldest Armenian. § 120. Lastly comes the Tamul, the representative lan- guage of Southern India. It is not pretended that this, in either its cultivation or antiquity, rivals the Sanskrit. Nor has it been so carefully studied. The old, or classical Tamul 44 AGGLUTINATION. — tlie Shen Tamxil as it is called— is considered to have flourished as the language of a rich and vaiied literatiu'e as early as the ninth century, and to present itself in inscrip- tions earlier, Caldwell specially states that it was com- paratively free from the influence of the Sanskrit, and that it was richer in forms than the present language. I have not, however, met with any definite exjDOsition of these diSerences ; nor has any considerable proportion of the literature itself been printed. § 121. With the discovery of America came a great flood of light upon what we may call the unwritten lan- guages, or the languages which having no literature of their own had to be made known to us through the investigations of missionaries, travellers, traders, slavers, and, finally, professed philologists. This means the languages of the New World, as a matter of course ; all Africa, with the exception of Egy]Dt and part of Abyssinia ; Siberia, Caiicasus, North-Eastern Asia, and all the islands, whether of the Atlantic or Pacific, with the exception of Sumatra, Java, Bali, Celebes, and the Philippine Islands. Some of the gi'ammars of this period were composed before a.d. 1600, for in Mexico, Guatemala, and Yucatan the zeal of the mis- sionaiy had already found its vocation. Elsewhere, the accumulation of material was somewhat later. Enough, however, has been said to show that the languages of the Agglutinate division of which we have any precise knowledge are nearly all, in respect to their data, less than four hundred years old ; and all, with the gi-eat exception of the Hebrew, later than the Christian era, or, compared with the classical languages and the Sanskrit, later than the Augustan age. § 122. Later than these, in respect to their oldest records, come the Turkish, the Mongol, and the Mantshu. Of these the alphabets are of Syrian origin, and, in all probability, some centuries older than the earliest specimens of the litera- tui'e to which they gave origin. The Turkish is the oldest, the IMantshu the latest ; and it was by the Christian mis- MATERIALS. THE TIBETAN. 45 sionarles of the iSTestorian heresy tliat tlie Tiirldsh, the earliest of them, was introduced. It was to the Uiyhur dialect that it was ih'st applied; and, though manuscripts in it still exist, and represent the oldest form of the Turkish language, it has long been supei-seded by the Arabic, the alphabet of Maho- metanism. § 12.3. It is in Eastern Europe and Western Asia that the languages which most especially lie betioeen the Synthetic and the Monosyllablic areas present themselves. These, when we take the whole of them, are the languages of Caucasus and Ti'anscaucasia, all of them falling into dialects, and many of them forming small groups of distinct languages : the Cir- cassian (and Abkazian) ; the Midzhdzhedzhi or Tshentsh ; the Lesgian (of many denominations) ; the Georgian (with well-marked dialects for Guiiel, Mingrelia, Swaneti, and Lazestan) ; the Iron and Armenian (of which the ethnolo- gical affinities are not imiversally admitted), and an outlying obscure form of speech, called the Ude. SECTION III. MONOSYLLABICISM. ACCEXT. MATERIALS. § 124. The Bhot, Bhotiya, or Tibetan. — Beginning in the West, and on the confines of the Persian of the North- East, and the Tiu-kish of Chinese, and Russian Turkestan, we find the fii-st Tibetan in Bultistan, or Little Tibet ; then in Ladakh, Tiljet Proper, the Sub-Himalyan ranges of India between Cashmir and Kumaon, Butan, the obscure tract of country between the Eastern frontier of Tibet (or Butan) and China, and, more or less, along the whole Himalayan range, (ISTorth) fi-om Cashmere to Assam. § 125. The age of the oldest litei-ary records of the Tibetan is uncertain ; biit we kxiow that the spoken Tibetan of the present time is notably more modem than the written lan- guage. This is because the older orthography is still kept 46 THE MOXOSYLLABIC CLASS. up, and in (I believe) the great majority of woi'ds there is a considerable difference between the two; indeed, it is a common practice of writers on the language to draw special attention to the difference between the ' spohen ' and the ' written ' Tibetan, and to give each word in both forms. I cannot state the precise natvire of the difference between the two extreme forms. It is pi'obably as great as any that is to be found in any of the iV'o?i.-Semitic languages of the Agglutinate class; but, nevertheless, wholly incommensur- able with diffei-ences that present themselves in the Synthetic and Analytic forms of speech, where the rate of change has been quicker, and the material to act upon greater. § 126. 1. In respect to its dialects little is known ; the central parts being unexplored by Europeans. On the East and South it is difficult to say what is a dialect of the Tibetan and what a separate language ; especially along the Southern slope of the Himalayas. § 127. 2. The £ur7uese of Burma Proper, Arakan, and Manipur ; falling into the Ivhen, Ivai-yen, and other un- doubted dialects. On the frontier of Assam and India it is difficult to say when the forms of speech cease to be dialects of some supei'ior language, or when they constitute separate and independent languages. The denomination Naga, applied to a long list of tongnies spoken along the mountain-range on the south side of the Assam valley is a notable example of this. Some of the forms of speech have one, some another, set of affinities. § 128. 3. The Singplio. Spoken to the East of the Burmese, and to the North and East of the Siamese. That the Singpho, Bui-mese, and Tibetan are members of one and the same group, family, or class, is, pretty generally, admitted. "VVTiat the value of the group may be is another question, § 129. 4. The Thay or Siamese, with the Aliom, Khamti, Laos or Lau, and the Shan dialects of the Burmese Empii-e. §130. The J/c)», or language of Pegu. CHINESE. 47 § 131. The Kha, or Ka, of Cambojio.. § 132. The Anamese, or Anamitic, of Cochin China, § 133. The Chinese. — This has long been considered to be the tyiiical language of the Monosyllabic class ; and that, probably, on good grounds. Of its paramount importance in both politics and literature there is no doubt. The Mandarin, or classical langviage, is uniform throughout the empii-e ; though its history shows that it has luidei'gone several changes — artificial rather than natm-al. The opposite to this, as a form that has been modified by the influence of commerce, and contact with other languages, is the Chinese of Canton. Of independent dialects, the best known is that of the Province of Fokien. § 134. On the antiquity of the Chinese literature little need be said ; and as little about the method of writing it. It is not one which favours the exact representation of old and obsolete forms. Nor is there any warrant for any com- position of any very great antiquity being exactly in the form it took under the hands of the original author. The matter may be as old as the time of Confucius or older ; but the antiquity of the form — ipsissima verba ips>\simis Uteris — is another question. We are sure of nothing in this way older than the oldest manusci-ipt, and who knows the date of this 1 Of the changes in the Mandarin dialect we know something ; but this refers more to the tones than to tlie literal and syllabic striicture of words and their combinations. In the opinion of the present writer — for which he claims very little — it is pi'obable that more knowledge of the changes that any monosyllabic language has undergone is to be got from a comparison of written Tibetan in its older forms than from anything attainable fi-om the older com- positions in Chinese. It is not likely, however, that the comparison will be made for some time to come. Of independent languages spoken within the area of China Proper the Miaoutse of the Province of Yunnan has the best claim to be considered other than Chinese. 48 THE MONOSYLLABIC CLASS. From a single short vocabulary which the present writer has seen in manuscript it seems to be equally unlike the Chinese, and the languages of its western frontier. Of the languages of the Andaman and Nicobar Islands it is only necessary to state that they are not Malay. In the Malayan Peninsula, the Siamese, which runs farther south than any member of the Monosyllabic family, but which appears to be of more northern origin, is succeeded by the Malay — other than Monosyllabic, § 135. 9. 10. The St(.h- Himalayan and JSfaga groups. — Kunaweri, where the Sutlej, the most eastern of the Five Rivers of the Pimjab, flows from the higher to the lower levels of the Sub-Himalayan slope, and first becomes an Indian river, is the point where, in the details of this large and mixed division, it is the most convenient to begin. Here it is where monosyllabic forms of speech, and those of Northern India which are other than monosyllabic, come in contact. 1.2. For the Kanet of Kunawur we have, at least, two varieties, the Sumchu and the Thehurskud, with sub-dialects : unwritten. Then Eastward, and in Nepaul rather than in Hindostan, — 3. 4. The Magar and Bramhu. 5. 6. The Gurung and Murnii. 7. 8. 9. The Vayu (or Hayu), Kusunda, and Che- pang. 10. 11. 12. 13. 14. The Denwar, Ktiswar, PaJcsya,Thak- sya, and Tharu. 15. The Xepalese (Pahari). 16. 17. The Kiranti, or Klrata, with which we m;xy associate the Limhu, which, until lately, was treated as a separate language. Of the Kiranti dialects Mr. Hodgson gives us vocabulai-ies for no less than nineteen, of which the Limbre is the only one that has an alphabet. 18. The Lcpcha — spoken in Sikkim, and the most eastern member of this range; inasmuch as, cast of Sikkim, we find, SUB-HIiLlLAYAX GROUP. 49 in Butan, tlie Butani, or Lhopa, form of speech, which is Tibetan. The Lepcha, also, has an alphabet. 19. 20. 21. Bodo (Borro) and DhimaL In Hindostan ; spoken in Kooch Bahar, and Mymansing. 22. 23. 24. 25. Garo,Mikir, Ktiki (or Lu7icta),a,ndCasia. 26. 27. 28. 29. 30. 31. Abor (Aka), Dofla, Miri, Mishimi {Tablu'iig), Mithan, Deoria Chutia, and Jili(1). It is in the southern members of this last division that the chief matei-ials for the transition from Monosyllabicism to Agglutination are the best studied ; and of the raw material no one has given us so much as Mr. Hodgson, and that both in the shape of vocabularies, and of grammar, or, at least, of valuable grammatical outlines. The number of dialects and sub-dialects is one element of theii- value. Another is the fact that lying, as many of them do, well within the frontier of India, they come into actual contact with languages other than Monosyllabic ; the Hindi and Bengali on the West, and, on the East, in approximation to the Dravirian forms of speech of the Khond class. In the parts about the Eajmahal Hills membei-s of this last class are spoken up to the banks of the Ganges, which the most southern congeners of the Dhimal, Bodo, and Garo approach ; though it is doubtful whether there is ever any absolute contact. § 136. 10. The ]\^aga group, to which allusion has been made, under the head of ' Burmese,' must be taken along with another ; and both connected with the Valley of Assam. In this the Bengali is rapidly encroaching; but the original speech of the whole district was Monosyllabic. § 137. Of these languages and dialects less is known in the central valley than in the mountain-ranges on each side of it. On the ISTorth they are the Abor (and Aka), the Bojla, and the Mishimi (Tablung), and Mithan. These take us up, or near, to the Eastern extremity of the valley. For the Hne on the South we find, first, a series of dialects akin to the Burmese and Singpho, and also ceitain ISTorthern ofisets of the Siamese ; and, running from East to \Yest, a list E 50 THE MONOSYLLABIC CLASS. of not less than twenty languages, dialects, or sub-dialects, known luider the general name of ' Naga! These, of course, lie nearly parallel to the Abor and Dofla line from Tibet ; tlie affinities being with the Northei'n and Southern Tibetan and Burmese respectively. Dr. Brown, however, to whom we owe the valuable series of illustrative vocabularies, has pronounced that, between the two we ai-e justified in throw- ing the Burmese and Tibetan into a single class. Of the other gi'oups, the Siamese is the most isolated. § 1 38. In respect to Accent and Tone more will be said in the sequel ; for it is with the Monosyllabic languages that we may hegin, and argue upwards, i.e. from the earlier to the later stage. We may now ask not only whether we can go farther back (which will be seen in the sequel), but whether we can go forthei- forward ? We can certainly speculate on the tendencies of the moi-e advanced languages, and point to the portions of them in which change is going on at present, and also seems likely to proceed. But there are points which are of a more practical character ; points which are less connected with either of the two extremities of our chain than with the i-elation of these extremities to one another. Let us remind ourselves that, in way of gi'owth, 01- development, the two most opposite languages are the Endish and the Chinese. The English is the most Analvtic language in the world ; the Chinese the most typically Monosyllaljic. If we compare the two, we shall find some- thing like the old adage — ^Extremes meet.'' IS^or is it difficult to see how this may be the case ; indeed, how, to some extent, it must 1)6 so. § 1.39. The chief and most conspicuous process in the Analytic stages of language is to divest the words it affects of all such additions, affixes, or appendages as constitute Inflec- tion : indeed, if we conijiare the present English with the Greek, Latin, or Sanskrit, we may say that the function of Analysis in Britain has been to denude all English words of their inflections ; and, if they leave any standing, to make SUB-HIMALAYAN GROUP. 51 them as little distinctive as possible. We have, notoriously, veiy few real inflections of any kind, and it will be shown that of the majority the action, or influence, is incomplete. At present, it is enough to suggest the similar vininflectional cha- racter of the Chinese, and, at the same time, the A^ery opposite reason for it. In Chinese the inflections have not been de- veloped. In English they have been developed and dropped. It is, probably, more important to understand this than to trouble om-selves about giving it a name. It is more like a Circle than an}i;hing else ; and we may, if we choose, call it a Cycle — oh differentiam. This is what we do with ' Xoun ' and ' Name,' and, to some extent, with ' Mode ' and ' 3Iood ' ; though the only diflerence is, that the words (identical in meaning) ai-e taken from diSerent languages. It would be somewhat technical to call it a Cycloid. It is, probably, more of a Sjnral than aught else. But, be this as it may, there is certainly a tendency in the more advanced stages of Analysis for the non-inflectional character to retm-n upon itself. § 14-0. It may now be sho-rni that the English system of inflection, limited as it is at its best, is even more limited than at fii'st it appears to be. This is because it is less dis- tinctive ; and ' distinctive ' means as follows : — If two or more paits of speech have inflections of the same form, and, toti- dem Uteris, identical, however they may differ in import, their inflection is only distinctive in reference to words with different endings, and not distinctive in respect to one another. Thus the -s \n fathers is distinctive as opposed to father ; but whether it represent the Latin patris or patres is doubtful. In -wTiting we can indicate the difference by the apostrophe : fatlier" s=2)ntris, fathers=patres. But writing is Language, and something more. In Speech the -s, as between the Geni- tive Case and the Plui-al Number, is not distinctive, and, so far, is deficient in the function of an inflection. Again, from love, which is a verb as weU as a noun, we have love's = amoris, loves=iamores, and loves^^amat ; words which, in E 2 52 THE THREE STAGES. meaning, are three, but in form, one. The context, no doubt, tells us what is meant. But, if so, what is the function of the iullection t Again, we know that, originally, such in- fiections were dilFerent ; i.e. that the Genitive Singular ended in -es, the Plural in -as, and the Verb in -eth. But, now, we must take them as they stand ; so that out of the few inflections which we have in English, the three in -s are impeifect. So are the two in -ed (-d or -t), as loved =■ am avity and loved =amattts. So is the termination -hig in words like sjKaking, moving, and the like. Do they translate woi'ds like tiiotio, or words like movens 1 Originally, the foi'ms were ditferent ; for the Participle ended in -nd, the Verbal in -ung. So were, oi-iginally, the two forms in -d. Later on, they be- came confluent : and of the Analytic stage the tendency to confluence is a character ; and, in proportion as inflections become confluent, their operation as distinctive signs is im- paired. When these imperfect inflections ai-e duly considered, what, in the way of distinctive inflection, remains in our language? The -t in the words tcha-t, tha-t, and i-t (hi-t {vom. he). The form in -en for the Participle Passive is no inflec- tion, but a formative. Hence, it is, manifestly, of primary importance to know whether a language is on its way from Agglutination, or some lowei- stage, to Synthesis, or from Synthesis to Analysis : and between extreme forms, like the English and Chinese, it is not diflicult to do so ; with the intermediate ones it is otherwise. There are some combinations, however, such as ' / have been,' that scarcely can be other than Analytic. § 141. Such are the stages and such the materials. The stages are recognised, and the names have an adequate amount of currency. They ai'e not unexceptionable ; but it is better to take them as they are, than to try to improve upon them. Nevertheless, both the names and their distri- bution must be taken with a certain amount of reservation. It is probable that, in their two extreme forms the EETROSPECT. 53 Analytic and Synthetic constitute two distinct stages. But they are languages for which we have a long series of records ; languages which belong to the same great Indo-European class ; and, above all, languages which, to a degree not found elsewhere, can be studied continuously. Hence, independent of other reasons, the two classes may be taken as one ; and the stage which they constitute is here treated as such, not so much because it is a stage, but because it is a stage specially and pre-eminently susceptible of illustration. § 142. The Synthetic, or Analytico-SjTithetic, stage stands fii-st in order, because it is the one which, in Great ' Britain at least, is the best imderstood. It is the last of the series ; and it is from the newer to the older, from the more developed to the less developed, from the more certain to the less uncertain, that we must reason. The special elements of language that the above-named languages illustrate ai'e the Inflections. They are faii-ly illustrated ; but when we, also, see how they transmute themselves into Analytical equi- valents, the combined illustration of the two classes is voT-y valuable ; and it is only from the languages of this period that we can aiTive at it. But the difference between Syn- theticism and Agglutination is not absolute ; nor are all the Agglutinate languages equally in contrast to the Synthetic. In the mere number of cases the Fin surpasses the Latin and Sanskrit. Neither is there much difference in the reducibility of the signs. In the Turkish, however, which is probably the nearest congener of the Fin, the signs both of Case and Number (Gender there is none) are easily divisible into separate syllables, independent of one another, and, in short, more like reducible compound words than irreducible signs of Inflection. On the frontier of the Monosyllabic area, i.e. on the parts between the Himalayas and the Ganges, the Agglutinate class is nearer in structure to the Monosyllabic than it is to the Synthetic. Nor should this surprise us. Between the spread of the Russian eastward and the Turkish westward, the displacement of (probably) intermediate forms 54 TEE, THREE STAGES. is inordinate. It is probable that nine-tentlis of tlie area which is now Turk, IMongol, or Russ was originally other than Turkish, Mongolian, or Russian. Of one of these hypothetical lanrcuasjes — the Yeniseian, between the Tui-kish and Mon- golian — a fragment still remains. § 143. In the Agglutinate stage thei'e is, undoubtedly, Inflection, and that to a notal^le extent; just as in the Synthetic stage there was Agglutination. Neither of the two forms excludes the other. On the other hand, however, it is certain that the pi'oportion which the two processes bear towards one another varies. By this, (like the other two stages) i.e. Ijy its general character it must be judged, and by this general character each of the remaining stages will be tested. Agglutination will mainly coincide with the pre- dominance of Formatives ; Monosyllahlcism by the import- ance which it gives to Accent. § 144. The characteristic elements which the Agglutinate stage most specially illusti'ates are the Formatives ; and, even if it were not so, the system of Inflections is more adequately illustrated in the section on Synthesis. That they should characterize the Monosyllabic series, is a contradiction in terms. But it is not merely because the Formatives, if not assigned as charactei-istics to the Agglutuiate class, have no other place of allotment, that they are thus associated. Fait* reasons will be given in the sequel for enabling us to refer the origin of some Inflections, at least, to a corresponding Formative. § 145. The Inflection, the Theme, and the Root, are, until we come to Phonesis, the three parts of a word in its latest stage of development; or, if we begin from the be^'inninar, the Root, the Themative Afiix, and the Inflection. Thus voc- is the root of both vac a and voc-it-o ; it is in the latter the thematic element ; and -amus in both voc-amus and voc-it-amus the inflection. The -amus may become -atis, -ant, -avi, and the like. The -it-, and voc- remain unchanged. Words without any superadded elements, and RETROSPECT. 55 that in their shortest form, are the elements of our ideal of a Monosyllabic form of speech ; which we safely believe to be older in its structure than any wherein we find additions, sounds, or syllables. But words are not used singly. It must be an exceptionally short sentence in which there are not two or more than two in contact ; and certain words may not only come in contact with others, but coalesce with them. This gives us Composition ; which brings us to ask where and when Composition begins, i.e. when out of two separate words we get a compoiuid. § 146. The Monosyllabic stage precedes the Agglutinate, and, a fortiori, it precedes the Synthetic. In its extreme — perhaps we may say ideal — form it gives neither inflectional nor thematic elements ; but simply and solely the root. That this root should be monosyllabic is what we believe a priori, and, on the whole, rightly. But it does exclude dissyllables. A third syllable, however, would, in general, imply an addition of some kind. § 147. The paramount question in the languages of the Monosyllabic type is, the point at which two separate and independent words so coalesce as to become one. This means, when does Comjwsition begin ] There is no better example of which the difference actually is than the familiar instance of ' All black birds are not blackbirds.' It is the difierence of the Accent that indicates the change. But Accent underlies the Monosyllabic system, and belongs to Phonesis ; under which head it will be considei'ed. § 148. And now the treatment of the subject is changed ; and, from the data supplied by the materials which we have indicated and classified, an attempt wiQ be made to explain some of the chief details in the way of Inflection and its antecedents. It must be premised, however, that the whole mass of available material has not been treated exhaustively. For the Synthetic and Monosyllabic classes all the materials have been noticed. In the Agglutinate class there has been a selection. This is because, though there is no language, 56 THE THKEE STAGES. dialect, or sub-dialect on the face of the earth which may not illustrate some other, there is a gi'eat difference in the extent to which they may illustrate the questions under considei-ation. The languages and classes which have been enumerated are (so to say) on the line. Europe, except so far as the Sanskiit is other than European, is exclusively the domain of the Synth etico- Analytical class. In like manner the South-Eastern jjarts of Asia are Monosyllablic ; and that, not here and there, with a mixture of monosyllabicism and agglutination, but in contact with one another, en bloc, and to the exclusion of anything polysyllabic. But the Agglutinate languages are ubiquitous ; indeed, we may say, that whatever is neither Monosyllabic nor Synthetic is Agglu- tinate. It is manifest, then, that there must be a wide difference between some of these languages and others in respect to their value as data : and that those languages which, like the Mantshu, the Turk, the Yeniseian, the Fin, and the Dravirian, lie between the Monosyllabic and Syn- thetic areas, must have a special value of their own. 57 PART II. PARTS OF SPEECH— 3IIGRATI0X— THE INFINITIVE- VERB OR NO CN— PERSON, VOICE, NUMBER, CASE, AND TENSE— GENDER AND MOOD— ORIGIN OF, AND GROWTH OF INFLECTION. (1.) MiGEATION. § 149. * Omne verhum, quuni desinit esse quod est, viigrat in Adverbium.' This is from Serrius ; and Home Tooke remarks upon it that, when a man has a word of which he can make nothing, he makes an Adverb of it. It is doubtful whether he meant this to be compHmentary to the gram- marians. On the contrary, it looks like a grim sneer on either the word ' Migration,' or the goal towards which the gram- marians directed it. Be this as it may ; if one fact in lan- guage is more real than another, it is that of Migration. And equally real is the fact, that a large proportion of the fnigrants become Adverbs. Nor do they always stop at the Adverbial stage; but, on the contraiy, go onwards and become Prepositions, and subsequently Conjunctions : indeed, we may fairly say, that until we have got a word as a Con- junction we have not got to the end of it. § 150. But migrations are of many kinds. Some carry the emigrant to a distance; some over a mere strip of country. And so it is with the Parts of Speech. Cases and Tenses may migi-ate. The Second Personal Pronoun Singular has long been migrating ; so long that, in some cases, it has not only abandoned the domicil which it had in such a phrase as ' thou speakest,' to the Plural form ' you ' but has, itself, gone no one knows where. But ' you,' itself an Accusative form, has replaced ye, which was the Nominative 58 MIGRATION. so that in the sequence thou speakesf, ye speak, you speak, we have a double change. But, along with itself the Pronoun takes its Verb ; for speuk, like 'ye' or ' you,' is a plural form. The result of all this is, that the Second Personal Pronoun Singular and the Second Person Singular of the Verb are simply eliminated from the current language. § 151. In German, they say Sie — they; so that, in German, the migi-ation is that of the Third Personal Pro- noun. But it is possible that this is not a true migration ; or rather that it is not the migration of a word. It may be that it is a change of the subject-matter, and that, as a point of courtesy, a simple person spoken to may, by a fiction, be considered as more than one. Of fictions of this kind the Oriental languages supply numerous instances ; instances of ceremonious, reverential, or, at any rate, of an artificial, language. § 152. There is another kind of approximate migi-ation. Two words may be used to express what is sometimes expressed by one. We get an instance of this in the com- binations ' I have spoken', and the like. But let 'have' be omitted. In that case, the Participle alone expresses the same idea. If so, it then becomes a Tense ; in which the difierence, as in Latin, is either ' / Jtave spoken,' or ' / spoke.' Now this formation of a Tense out of a Participle is neither a hypothesis nor a fiction. It exists in the Slavonic lan- guages ; and that throughout each of them, and through all the divisions and sub-divLsions of the Family. And, as if to show that the so-called Tense is merely a Participle with an omission of its Verli, the Genders of the Participial form are i-etained in the Tense. § 153. The same is the case with the Second Person Plural in the Latin Passive. Words like ' aviamini,' ' regimini,' and the like are merely participles ; as truly participial as such forms as ' rervpiJErui e'irc~iv=invidia ; rov (j>Ooi'E~ii'=invidice ; rw (fnin- v(~u' — inmdid. In Latin, this indii-ect inflection is wanted. But in Latin there is no Article. § 167. I do not know that this particular consti'uction is found in the Plural number, or that we can readily produce such words as rh or rwr TVTTTtir. But there is no logical reason why we sliould not ; and, if we could, there would be no difticulty in translating them. They would simply give us equivalents to such words as heatings, risings, findings, and the like. There would be a shade of difference in their VERB OK NOUN. '65 import, just as there is one between to err and error — but that would be all. § 168. Can Infinitives have Gender 1 Are such com- binations as (I (pdoveiv, and // (piXoacxpelr possible 1 All that need be said on this poiat is, that words like 6 (})d6}nQ and ?; (piXo(TO(j)ia are common, and that, as Substantives, they cor- respond with the Infinitives in sense. The accidents of the Verbal Substantives are, also, the accidents of the Infinitive Mood in import. But there is one exception. However, in this there is nothing, either in form or import, which brings the Infinitive and the ISToun within the same class. This one exception is that of Person. The Infinitives have no Persons. If they had, they would thereby cease to be Infinitives, and become Finite Verbs. This, at first, seems but a trifling distinction. In reality it is a very wide one. It is just the distinction between two words and one ; just the difierence between a single word and a sentence ; just the difierence between (in Logic) a ProjMsition and a Singh Term. It is a distinction of no small import ; indeed, it is the difierence which justifies us in sapng that the relations between the Infinitive and the Finite Verb are incommen- surable. § 169. The Person of a Finite Verb is either a separate word, as ' /' in '/ stoim,' or its equivalent the -o in '7io :' both of which are, in Grammar, Sentences, in Logic Proposi- tions ; short ones, no doubt, but still propositions and sen- tences. Ko Infinitive has a Person. No Finite Verb without one. § 170. The exceptions to this last statement are only formal ; inasmuch as the Personal element may be implicit in the sense, though not explicit in the expression. This hap- pens when the Subject is either so manifest that it needs no sign, or too indefinite to be represented by any particular name. Instances of the first alternative are words like die and F 66 INFLECTION. fac, Imperative Moods, where the personality of the iutli- vidual addressed is self-evident. Of the second we find examples in the so-called Imper- sonal Verbs, tcsdet, pcenitet, and the like. In some cases the context sii|)plies what is missing. But in short sentences the case is different. In combmations like ' it rains,' it is hard to say who, or what, is the agent ; yet this is not because there is no agency, but because its nature is inde- finite. In the Infinitive, however, there is no name for an agent any^vhere, or of any kind, no name whether explicit or implicit, no name eithei- ciira dictionem or extra dictionem. § 171. It is only then in the Imperative Mood, where there is no possibility of doubt as to the agent, and in some few Impersonals, where the Nominative Case is too indefinite to be expressed, that the Finite Verb thus loses its characteristic. Whether it take this name in the shape of a separate Noun, as ' / think' ' John spealcs,' or by the way of inflection, as '■ puto' or by both processes, '■ E(jo puio,' and '/ thinh,' is immaterial. § 172. The Finite Verb then is a Verb, and something moi'e. It is a Verb ^'^'MS a Pronoun, or the equivalent to one. It is the name of an action plus the name of an agent. It is not one word, but two. It is not a single Part of Speech, but two Parts of Speech united as one. § 173. I know no better exposition of these details than the following, from Stockfleth's Laplandic Grammar, §§ 231- 233. A. There are two capital forms of the Verb : 1. the Sub- iective form, with the conception of Personality and Time; 2. the Objective form, without any other conception, especially of Perfcouality. B. The Subjective form has three Moods (modi) : 1. The Indicative. 2. The Conjunctive. VEKB OE NOUX. 67 This last has two divisions : a. Subjunctive (jtrceBens subjunctiuus) ; b. Optative {prcsteritum conjunctivuni). 3. The Imperative. C. The Objective form, independent of Person, and to some extent of Time, has two divisions : 1. Verbum Nomhiale ; 2. Nomen Verbale. To the Verbum Nominale belong — 1. Infinitivus, which answers to the ordinary Infinitive. 2. Factivus, which answers to the Latin Supine. 3. Comitattvus, which answers to the Gerund. To the Nomen Verbale belongs — a. Nomen Substantivum Verbale. b. Nomen Adjectimim Verbale. Each falls into two subdivisions : 1. Nomen Substantivum Verbale Abstracium, which corresponds with the two [Danish] forms in inff and -n, Lasning (i.e. lectio) and LcBsen (i.e. legere). It is inflected in both numbers. b. Nomen Adjectivum Verbale Concretum, which answers to the Latin Participle. 1. Frcesens Participium, without the conception of any definite time, and therefore, in conjunction with the auxiliary verb, it ap- plies to any time. 2. PrcBteritum Fai-ticipium, with the conception of past time ; iininflected. § 174. This last word, ^ univjlected,^ appKes to the La}> landic language, not to the Past Participle of the Indo- European class. The name Prceteritum Conjunctivum assigned to the Op- tative indicates its mixed character, and its relations to the Tense on one side, the Mood on the other. Of a Negative Gerund, which is called ' Caritivus," no notice has here been taken. It is necessary, too, to draw attention to Stockfleth's notice of the negative form of the English Infinitive. In the present language there is no sign for the Infinitive Mood, so that in English the Infinitive is equivocal, i.e. a form like rids and hunt, in which the Substantives a ride and a hunt are abso- lutely identical in form with the verba to ride and to hunt. F 2 68 INFLECTION. (3.) Inflection. § 175. Tlie Inflections are a natural class; and their import is tbe expression of what, in Logic, is called the Accidents of a Substance, Action, or State. As such, they are signs of Person, Voice, Number, Case, Tense, Gender, and ood. Of these. Person, Voice, and Tense are the most reaucible, i.e. the most readily analysed, and the most mani- festly derived from separate and independent words. Those of Gender and Mood are, perhaps, the last to develope them- selves and the first to become obsolete. Number and Tense seem to be both the earliest in origin and the most permanent in duration. § 176. The two principal methods of forming these Inflections are Addition and D Iffei'cntiation. By Addition we get a wholly new sound, vowel or consonant as the case may be. By Differentiation we get either the substitution of one sound (vowel or consonant) for another, or the same vowel in a modified form. Sometimes, indeed, we have nothing more than a change of either tone or of accent ; as, / heat you if you do tliat-=.I shall heat &c. § 177. Upon the difierence between the results of Addi- tion and Differentiation some important inferences depend. The termination -as in words like hon-as differentiates the Femiidne form from the Masculine honos ; but between the two we get three distinctions — those of Gender, Number, ■ and Case. It is not necessary to go into the explanation of the mechanism of this, and to show how it can be ; but it is clear that when such forms present themselves, there must be one of the three inflections for which we have no extrinsic, audible, or definite sign. It is convenient to call these inflections either ' modified,' or Ulifferentiated ;' and it is evident that such abbreviated or modified forms are much more difficult to reduce than inflections in exte)iso. Now, the Inflection of Gender is generally expressed by cither cUf- ferentiatioa of this kind, or by the omission or non-existence DECLENSION AND CONJUGATION. 69 of some increment which is present in one gender, hut absent in one of the others. In the Synthetic class of lan- guage, where Gender prevails, reduction is drfficult. In the Agglutinate languages, on the other hand, where there is no Gender, it is comparatively easy ; and we can, in many cases, lay our finger on the respective signs of Number and Case, and determine their sequence. Moreover, they often amount to syllables of notable length. Nevertheless, even here it is not easy to ascertain what they were as separate and independent words. It is enough, however, if they improve the presumption that such is what they were originally. § 178. The signs of the Conjunctive Mood are also, as a rule, differentiations or modifications ; and the Conjunctive Mood, like Gender, is late in. its development and earlv in its disappearance. This means that both look very like secondary formations, or formations from some earlier in- flection — with a difference. § 179. It is submitted, then, that the distinction is natural ; and that inflections formed by change rather than by the addition of some special separate element wiU be found to correspond with other characters, and with certain classes that such characters determine. When this is the case, we may presume that the arrangement is natural ; though, of coiu'se, its value must be measui-ed by the extent of the correspondence. How far this is the case will be seen in the sequel, especially under ' Gender' and ' Mood.' (4.) Declexsiox and Conjugation. § 180. Each of these terms is used in two senses. When we say that the Genitive Case is a part of the Declension of a Noun, and that the First Person Singular is a part of the Conjugation of a Verb, we mean that each belongs to the class for which laflpction is the generic name. This is a mere matter of classification, and requires no further notice. 70 DECLENSION AND CONJUGATION. But when we say that nuisce, clomini, lapidis, gradHs, and fnciei, all of which are Genitive Cases, belong to five difFei'ent Declensions; and that ayyio, 7noneo,re(jo, and audio belong to four different Conjugations, we mean something which, when we come to understand what is involved in the difference, is perplexing. § 181. Declension and Conjugation, in this sense, mean classes under which, when we find more forms than one for the savie Case or Tense, the different paradig-ms of the inflections are arranged. Sometimes words are placed in different declensions on the strength of very sKght differences ; sometimes a considerable amount of them is required before a second class is recognLsed. But this vaiies with the gi-ammar and the language ; and, in many cases, there may be different systems within the same language. This depends on the grammarian. The old number of the Greek Declensions was Five. They were subsequently reduced to Three. On the other hand, by giving the Simple and the Contracted Nouns a Declension each, they are raised to Ten. Again — the Conjugations of the Verbs in the Eton Grammar- are (or were) thirteen — ' sex Barytonor^mi, tres Gontracto- rum, et quatuor Verborum {71 -^a.' But these are manifestly reducible to three higher classes. Are we then to apply the same term to the Genus and Species 1 Or are we to have names for each class 1 These are questions which are not asked for the purpose of being answered. They are rather asked for the pui'pose of showing that the number of Declensions and Conjugations is, to some extent, a question of names rather than one of realities. § 182. The real question is — VHij should there be any difference of Declension and Conjugation at all 1 If a given meaning is adequately represented by one form, what is the use of another 1 Either rightly or wrongly, we believe that Natui'e is not tolerant of superfluities. Still, if there are not a great many ideas which have more forms of expression than one, there must be sometliitig exceptionable in many of VERB OR NOUN. 71 oiu' common modes of speecli. Two sucli are the ■words before us ; and ' synonym,^ or ' synonymous loord,' is a term of the same kind. Are there such things in Rerum Natura as two words with exactly the same meaning? § 183. There are not. Neither is there any method of forming a second inflection absolutely identical with one already existing, and along with which it can be used con- currently. There is, doubtless,, something Kke it. There are diflferences of structure in the fundamental parts of the word, which may produce small changes ; there are differences in the effect which two, or more, inflectional elements in combination may have upon each other ; there are differences in the permanence of certain elements, (t.e. an archaic form may become obsolete in one class of words earlier than in another) ; and there may may be confluence in meaning between two forms originally distinct. All this we may expect, and all this we actually find ; but they fail to give us a single instance of Language being, in the strictest sense of the word, tautological. Wherever it appears to be so, we must suspect the rule, or formula, which embodies it. We need not condemn it. Grammars are written less for comparative philologists than for learners of foreign languages. We therefore must frame the rules in the way that is most convenient. We cannot make them so stringent as to be fol- lowed by a long list of exceptions ; nor yet so guarded by qualifications and conditions as to include all possible varia- tions. We must do the best we can for the purpose in hand. But the philologue must not mistake this for anything more than a special exposition of a special form of sj^eech. From the Grammar of Language as it exists in natm-e he must exclude tautology. There are no such things as absolute synonyms ; no such things as declensions and conjugations that give us double forms for exactly the same cases and tenses ; and no such things as exceptions to a rule, provided the rule be expressed in terms sufficiently general. It is not, however, meant by this that, for the ordi- 2 PERSON. nary purposes of teaclimg, rules of less general character are inadmissible. Neither is a certain amount of exceptions. But we must be very cautious in admitting tautology as a real element in Speech. (5.) Person. Noun and Verb. § 184. The Persons are the Inflections with wliich it is the most convenient to begin. No other inflection plays so impoi'tant a part in classification ; and in no other inflection is. the structure so readily analysed. It is the Pronoun, moreover, which constitutes the broad and primary distinction between the Noun and the Finite Verb, In the Infinitive, whether Verb or Noun, we deal with a single word. In the Pinite Verb we deal with two words; one a Verbal name or the Abstract name of an action with- out the name of the Agent, the other a Pronoun, Pei'sonal or Demonstrative as the case may be. The Finite Verb, then, consists of two Nouns, and constitutes a sentence, or (in Logic) a Proposition. Hence it is a Term, and something more. § 185. The Noun as well as the Verb has Persons, though not in English and its Synthetic congeners. The Magyar, which is Fin and Agglutinate, is only one of the numerous languages which sujiply us with examples of this. Verb. 1 . 01 vas-owi = I read = reading-my. 2. 0\\&i-od = th(m rendest = reading-thy. 3. 01va.s-i(^ = we read = readiny-our. 4. Olvas-w/Jo/c = ye read = rcading-your. Substantive. 1. Alma-m = rrt?/ apple = aiyple-my. 2. Alm(a)-f^ = thy apple = apple-thy. 3. Alm(a)-n7i; = our apple = apple-our. 4. \lm{ii)-to7c = your apple = apple-yuur. ' NOUN AND VEEB. 73 The Celtic does the same; and, in both the construc- tion or parsing, is the same. It is not that of the Government of the Yevh by a Nominative Case, but that of the Concord of a ISTotm with either a Possessive Case or a Pronominal Adjective. It is upon this difference that the little which need here be said about the Person of the Yerb depends. § 186. The basis of everjiihing connected with the Pro- noun is the all-comprehensive dualism of the logicians — Bgo, with its counterpart JS^on-ego. Everj-thing in the Universe is either Ego or Kon-ego. Everjiihing, too, is either A or Not- A, and whatever is not Xot-A must be ^. In Philo- logy this dual becomes a triple partition. Every Name in Language is that of (1) the Speaker ; (2) the Object spoken to ; and (3) the Object spoken about. Out of this Triad we get the Three Persons of the Verb. § 187. The very term ' Person ' suggests such an origin ; and with such terminations as the -fui of ridrifxi, and the o- and T of the Second and Thii-d Persons, nothing suggests itself so readily as a connection between them and the Pronoun ; due. allowance being made for wear and tear. There are diffi- culties in detail, but not more than we expect. The Persons are old inflections, and have undergone very many changes. Transitional forms, however, that we miss in one language often present themselves in another ; so that the doctrine that the Persons of the Yerb are Pronouns in disguise has grown up spontaneously, and in the main is correct. § 188. In what Case are there incorporated Pronouns? This is a question which vmtil lately has scarcely been thousrht worth the asking. Does not ' / love ' translate ' amo,' and is not / the Nominative Case to its Yerb — its own special verb which it governs] It is all this, no doubt. On the other hand, however, the -/i- in the Yerbs ia -/ui manifestly points to the m- in me rather than to the -g- ia ego. It was the famous Welsh lexicographer, Lhuyd, who 74 PERSON. indicated the m?:eliliood of tlie signs of Person having been ohlique cases; and it is due to Garnett that its full importance has been recognised, and the details of the change been investigated. Garnett, who, with his usual justice to his predecessors, makes special reference to Lhuyd, main- tained that the Verb to which the signs of Person were attached was not so much a Finite Yerb as a Verbal, that the Pronoun which invests it with personality was not a Nominative in concord with that Verb, but one in the Possessive Case, the Verb being in a state of regimen or government ; in other words, we held that scriho was not write-I, but writing -my. He proceeds : The second person plural does not end with the nominative cluci, but with ech, wch, och, ych, which last three forms are also found coalescing with various prepositions — iwch, ' to you ; ' ynoch, 'in you;' torthych, 'through you.' Now the roots of Welsh words are confessedly nouns with an abstract signification ; as, for instance, dysg is both ' doctrina ' and the second person imperative ' doce.' Di/sy-och, or wch, is not, therefore, docetis or ducibiiis vos ; but doctrina xcstrum, ' teaching of' or '■by you.' . . . Doctrina ego is a logical absurdity; but doctrina met, ^ teaching of me,' necessarily includes in it the pro- position ego doceo in a strictly logical and miequivocal form. § 189. For some years no one found this1^doctrine harder to believe than the present writer. There was no denying the oblique character of 7n-, and -och, or -ych. (1) But the identity of form between the verbal dysg, and dysg as the second person imperative, and (as such) an undoubted verb, is not the same m English as in Welsh. In English teach, which translates dysg, only translates it as the equivalent to doce. The English equivalent to doctrina is not teach, but teach-ing. Now this so manifestly invests the Verb with a radical character that, to an Englishman, the abstract noun is reduced to a derivation or secondary idea. There are many words, no doubt, in English where the two foi-ms have the same meaning. Thus we say, I took NOUN AND VERB. 75 a swim, a walk, a dip, or a rest, &c. &c. ; but we do not say, * / got a teach.' Per contra, however, there was a time when there was no such sign as -ing for the verbal abstract, and when the aflixes my, thy, &c. might be affixed to every verb or verbal in the language. The more we think upon this, the more we feel that this must have been the case. As the exposition of the doctrine is not my own, I may say that it is thoroughly con\'incing. There must have been a time when words like swim and move existed, and when words like swimming and moving did not. Perhaps the difierence between terms like nare and natatio was not recognised ; but even in that case it must have been implicit in the root, and destined in due time to take a special and distinctive form. It may do this in some languages, in others not. When it fails to be developed, there is not much tendency to demur to such a doctrine as Garnett's. When it presents itself, it is a stumbling-block. There is visible evidence of the for- mative being of secondary origin. ; and we in England have to divest it of this, and to treat it as equally primitive with the simpler one. In the Welsh this is unnecessary. Still the doctrine is pre-eminently intelligible, and we, with no great difficulty, understand it. But when the existing Abstract forms, with these unmistakeable mai-ks of a secondary origin and derivation, are presenting themselves in almost every sentence, we are slow to become familiar mth it. We are like beginners in arithmetic. We can understand the principles of computation, but are slow and dubious when we do a sum. So it is with those whose languac^es abound with the sisns under notice. It takes a long time to get rid of their influence, and a long time to shake off its effects. It may be different with others, but this is what it has been with the present writer. Again, Most of us are in the habit of thinking that nothing but a Nominative Case can be the Subject to a Verb. Upon this more will be said in the notice of the Nominative Case. Hence, then, the real difficulty is the confii'uied habit of thinking 76 VOICE. otherwise. Some minor points lie in the details of the con- struction. Is the affix a Personal Pronoun in its racUcal form (me, te, &c.), or is it the mei, or the tuus, which is Possessive] Something depends on the difference. The forms like met and tuus suggest a later, those like me and te an earlier, period for the language in which the change took place, and when the Infinitive Verb first became Finite. The Personal element in any form is less evident, and the elements of the inflection are less easily distinguishable in the Dual than in the Plural, and in the Plural than the Singular. But this lies more in the complex character of those two Numbers than in the principle itself. If it only held good for the Singular Number, it would be suflicient to separate the Finite Verb from the Infinitive, and make them incommensurable. (6.) Voice. Koun and Vekb. § 190. Nouns as well as Verbs have Voice, though it is not recognised in the ordinary Grammars. Nor, for the purposes of special teaching, need it be. Indeed, in many languages it belongs wholly to Sjmtax. In some we have different sorts of expression in different stages. In some cases there is no sign at all, no formative, no inflection, notliing whatever in esse. Nevertheless, whether adequately or inadequately represented by signs, tliere is in the Noun a far more elaborate system of conceptions in the way of Voice than there is in the Verb. These, from the very nature of the language, may, with a certain amount of data, be determined and arranged a priori. This is because in the notion of Voice there are always an antithesis and a correlation. Hence, if we have only one form in esse, we are sure of another in posse, and, without saying exactly what it will be in language, can construct it for ourselves in thought. Thus, given such a word as hunter, i.e. the Verbal Agent from a transitive Verb, we know that there must be something that he is supposed to hunt ; in other words, an object. There may or may not be a name NOUN AND VERB. tl for this ; and in intransitive verbals, such as sleep-er, it is not wanted, though it is sometimes used, almost tautologi- caUy, to strengthen or specify the character of the state, e.g. he sleejjs the sleep of the righteous. If the verbal element, however, of the name of the Agent be transitive, there must be an object, and the name of this object is the correlative (signified or not) to the Verbal of the Active Voice. Thus far there is absolute certainty. On the other hand, however, it is well known that in EngKsh, at least, no such correlative exists. In a foxhunt both the horsemen and horses are hunters, though with a difierence. The huntsman, too, is a hunter ; but, as opposed to the whipper-in, he is a hunter of a special kind. For these then we have, either by com- position or derivation, two forms deduced from the verb hunt, with certain distinctions between them. The horses, then, and the men are htonters, and take theu' class-name from the verb involved in then- agency. But the fox has surely something like a similar connection. He is the object, or the thiTig hunted. Yet he has no single name as such. The thing hunted, the hunted animal, or the object of the hunter has to be expressed by a combination of different words. There is no necessity for this. One syllable is as easily attached to words like hunt as another. Nevertheless, we have no such word as the one suggested in English. § 191. Yet there is no scruple as to other modifications of the -er, (the -er agentis) imder notice. One female takes to the hunting-field and becomes a hunt-r-ess, another to the hotel and becomes a waitress. We have got, then, one form ; but the one that we most especially expect we have not got. We have not got a correlative to the Verbal Agent. Such is the case ; but it is necessary to add, that the affix -ess is other than English in origin. And this implies more than it states. We did not in English often apply the affix till it was specially wanted for the expression of a practical distinction ; and when we had to apply it, we took it from the French. 78 VOICE. Again, notwithstanding what has just been said, the formative that gives us a correlative to the -er in hunt-er is not wholly and absolutely un-English. Mortgag-or and niortgag-ee, gra7it-or and grant-ee, trust-or and trtist-ee, ai-e correlative to one another. But, again, they are words of a particular class; and again, they are words of un-Enghsh origin. They are French, as we all know. But even in French they are not exactly what we want, not exactly the counteii^arts to the verbals like hunter. They are rather Passive Participles with a special application ; and, as such, of secondary origin. But in a context they correspond exactly. They cnn stand as Nominative Cases to the Verb without either an expressed or an understood Substantive with which they have to agree ; and they (in English) can take the sign of the Plui-al Number, which the Passive Participle can not. § 192. Two points are now transparently clear : (1) that the ordinary Passive Paiticiple is not exactly the correlative of the Verbal in -er ; and (2) that, at the same time, it is something very like it. And with this likeness the history of the Inllection begins. The words of which hunter is the type have always had their opposites or counterparts, but not their exact ones. They never seem to have stood quite alone in impoi-t ; this meaning that they seem always to have had ceitain approxi- mate parallels, i.e. words of different form, but of like mean- ing. It was a diiference, however, which did not amount to much. It might requii-e a special and separate sign to ex])ress it, or it might not. But it was by no means indis- l>ensable. One language might adopt it, another not. One language might liave it in one stage, but not in another. At any rate there has generally been for the two ^c^ive Verbals the Participle proper, like amans, -antis [=qii{ amat), and the Verbal proper, like amat or (!ilso=gwi amat) ; but only one form in Passive sense. It was more Participial than Sub- stantival, and its Greek sign is the -d- in rvf-tl-ur. Whether NOUN AND A'ERB. 79 it was always thus participial, and never substantival, in other words, whether it prevailed to the exclusion of the more substantival counterpart, will soon be seen. § 193. But what if both amans and dmator can be trans- lated by qui amat, and are yet expressed by words as distinct as amaifis and amator % what is the difference between this pair of words that in the Active Voice is recognised and in the Passive is ignored 1 There are three ways of discover- ing it. a. In the first we must think it out. We must clearly see what we mean by coiTelation, and also that each of the forms under notice, no matter whether it exist or not, is correlative. They may exist in the shape of formatives, or they may not. This is the d, priori, or theoretical method ; and by it we get what we may call the ideal con-elative to such words, as hunter and its congeners, i.e. the class of Active Verbal Abstracts. h. We may work it out. We may write a sufficient amount of Latin prose, and see when and what distinctions we make ourselves ; or we may read so much classical Latin, and see what distinction the writers make. c. We may get what we can from such gi-ammars as touch upon the subject. From which we learn that words like amator are concrete, substantial, and substantival rather than adjectival, and that they imply the actor rather than the act ; whereas words like amans (amantis) apply to the act rather than to the agent, and are more general. By substituting for the single term a combination of the relative pronoun with the verb, we get farther into the nature of the difference. Amator and amans are equally qui amat. But what doas qui mean 1 Does it equal ille, or iste, qui, denoting some particular individual whose name is implied by the context, or does it equal quisquis, or any qui whatever ? If it mean the latter, it makes something like a class of lovers. A hunter is a certain one of a class which htmts ; one of more than one, perhaps one of many. Now tlus brings out the 80 VOICE. element of Plm-ality ; and we shall see, as we proceed, that Plurality combined with the Suhstantival nature of the Verbal under notice implies a great deal. But why not have a pair of Verbals on both sides, or why should there be any difference at all 1 The explana- tion of this lies in the history of the foi-m more especially under notice. At present it is enough to say that it exists. There is no doubt about it. It exists ; and the only question that stands over for considei-ation is that of its extent and in- fluence. Does it absolutely and entii-ely exclude any second form 'I Does it make any second form impossible ? It certainly does not do this. In the Lap language we have a recognised Nomen Szcb- stantivum Verhale : Active. Abstract iSdnom = Ai'etv. Concrete \i\.^-participial portion of the dominant inflection in Greek — i.e. that of the forms like tviztohui, TETv^fxai, kc. — may be is no easy question. They are certainly not Passives, like eru^y/jv. § 203. The Deponent. — Of Deponents there are two kinds : (1) those like vapido, and others where there is a Passive, import but an Active form ; (2) those where the form is Passive, but the sense is Active. The expLr VERB. 85 of the first of tliese generally lies on the surface, and is involved in the meaning of the Yerb itself. Being this, it requires no sign, whether inflectional or formative. Vapulo may as easily be interpi-eted ' / take,' or ' stiver a heating,' as * / am beaten.' Still better are the instances of audio, and cluo=' Jupiter audit '=' answers to the name of Jupiter ; ' ' KUKuig K\vei'=' bears a bad name.' Here the sense is reflective, and often means hears {him)self called. A name which a person gives himself, or even takes to himself as permanent, is of this kind ; e.g. KK7fOi}lus the name for 2. It is this sometimes, but only sometimes. It is this in Lithuanic ; but then it is only a gi-ammarian's inflection, i.e. no inflection at all, but simply two words — just as hi duo, lice, duo, horum duoncm, are two. This is the chlemma which is always presenting itself. When •we can reduce any inflection to its elements it ceases to be an inflection. But this is really no discouragement. On the contrary, it shows the reducibility of certain inflections, and this is all we need attempt. § 214. In the German languages we get such forms as wi-t and gi-t-=-we two, ye two. This is somewhat more to om- piu'pose. Bo-th — through its derivation from bd-twd= Danish begge to, German beide, has been objected to — is probably the same combination. Our-two, or o/tis-two, and your-two, or of you-two, are also respectively incer and ungker ; the first of which, and perhaps the second, are in use in provincial Westphalian at the present time. Incit and ungkit are the Accusatives. This, however, goes for very little. They all seem second- 90 EXCLUSIVE AND INCLUSIVE ary formations ; though it is only the first wMch is, like the Lithuanic, a simple compound. § 215. It is highly probable that there was in some guage or Numeral 2. language or other a Dual Number before there was the § 216. But tliis is not conclusive. There may jDOssibly have been some natural dyad ; which suggested not only the inflections subservient to the expression of Duality, but pos- sibly the Numeral =2 itself; some word like '^9«ir,' ' hrace^ ' couple^ and the like. The words quoted are of foreign origin, and probably artificial. Hence they throw no light on the special details of our enquiry. But they illustrate the character of it. It is suggested, then, that in natural dyads of this kind the elements of the Dual Number are the likeliest to be found. But these are numerous ; and who can say, d, priori, which was the one originally pitched upon for the purpose of being made up into a combination express- ing ino7'e than one and. less than three, for this is what the Dual representation comes to 1 This is likely. But it does not carry us far. § 217. The Greek and Latin o/i^w and ambo look as if they meant one on each side, and so applied to a pai't|of a triad. Again, there is in the West-Saxon a strange form, ivit Skillinr/= toe-two Skillin[/=I and Skillimj or SkilUng and I. This looks like something that carries us farther, and leads to the consideration of — The Exclusive and Inclusive Pronouns. § 218. It is patent from mere inspection that in our own language, as well as in its congeners of the Synthetic Stage, there is a difference in the Pronouns of the First and Second Persons, between the Singular and the Plural Numbers, wholly opposite to that which presents itself in the ordinary Nouns. The Plural of ' /' is ' we,' and the Plural of ' thou ' PRONOUNS. 91 is ' ye' Now there is a process in the less advanced lan- guages of the Agghitinate and Monosyllabic stages which is beginning to be duly recognised as an important element in the explanation of this anomaly, viz. the system of the Inclusive and Excliisive Pronouns. Thus — FkOM the AuSTEALIAlf, Ngaia = l\ nffuUe = wQ two = thou and I; ngullina = we two = he and I. Fkom the Melakesian. Ainyah = I ; akaijan = you two + I ; ajumrai = you two — I ; akataji = you three + I ; ajeimtai = you three — I ; akaija = you + 1 ; aijama = you — I. § 219. These two examples may suffice to give a general idea as to the import of the two terms. But much mofe than this must be said about them. The Double Plural, though not equally prominent in all languages, is present in most of them ; sometimes in its rudiments, sometimes in its frag- ments. Even in its fullest foi-m the system is widely spread. We have it conspicuously in the languages of the Sub- Himalayan range ; and this, as we have seen, is the district in which the Monosyllabic class comes in contact with the Agglutinate. Within the Agglutinate area Melanesia, Poly- nesia, and America abound in Persons of this kind ; so that these give us the chief mateiials for oiu- criticism. But there is both Exclusiveness and Inclu&iveness elsewhere, as will soon be seen. § 220. In some languages they play a very notable part, in others, they are almost latent ; and between the two extremes there are differences and degrees of all sorts. When at its maximum the proportion of these two Pliu-als to the Accidents of the other Parts of Speech is inordinate. Of its development in a single dialect, the following paradigm speaks for itself : 92 bXhing pronouns, DECLENSION OF BAIIING PRONOUNS AND OF NOUNS. I. — Of PRONOirjfs. 1st Personal Pronoun. 1. No7n. I, Go. 2. Gen. Of me, | ^o^y rmc^- Wa = my. I Disjunct. \\akQ = mine. ^ fDat. Tome, j (-,^_ ^^ ^.^^^^ lAc. Me, J 4. Loc. I Jf, ™®' 1 Wake gware (interior) . L U itlim me, J 5. „ i -^^^^^ ™®' 1 Wake di (entering, resting in). L In me, J 6. Abl. From me, Wake ding (removal). 7. ^//. Towards me, AVake la (uearing). 8. „ From towards me, Wake lang (departing). • 9. „ Towards me, Wake taure (behaving). 10. Soc. With me, P^^'-^^^^^^^^^ I (society). L Oonuug J 11. Priv. Without me, {^^^^^^^^^^^*^^} (privation). 12. Inst. By me, Go mi. 13. Loc. M, by me, Wa pumdi (proximity. Hind. pas). Dual. 1. Gosi, inclusive. Gosiikii, exclusive. ^ f Conjunct. Tsi, incl. AVasi, excl. "^'X Disjunct. I'sike, incl. Wasike, excl. 3. Gosi, incl. Gosiikii, excl. 4. I'sikegware, incl. AVasikegware, excl. -5. I'sike di, incl. AVasike di, excl. 6. I'sike ding, incl. AVasike ding, excl. 7. I'sike la, incl. AVasike la, excl. 8. I'sike lang, incl. AN'asike lang, excl. 9. Gosi taure, incl, Gosidiu taure, excl. 10. Gosi nung, incl. Gosiiku nimg, excl. 11. Gosi manthi, incl. • Gosuku mantbi, excl. 12. Gosi mi, incl. Gosuku mi, excl. lo fisi- 1 T fiucl. ^^•{wasi-}l"^"'nexcl. NOUNS, AXB VERBS. 93 Plural. 1. G6-i, incl. Gokii, excl. o f Conjunct. Ike, incl. "Wake, excl. ' I Disjunct. Tkke, incl. Wakke, excl. 3. G6-i, incl. Goku, excl. 4. ITiegTvare, incl. Wakegware, excl. 5. I'ke di, incl. Wake di, excl. 6. I'ke ding, incl. Wake ding, excl. 7. I'ke la, incl. Wake la, excl. 8. I'ke lang, incl. Wake lang, excl. 9. I'ke taure, incl. Wake taure, excl. 10. Goi nung, incl. Goku niing, excl. 11. Goi manthi, incl. Goku mantW, excl. 12. Goi mi, Lncl. Goku mi, excl. ^'•{wrke-}P^^Mexd. § 221. How these Plurals combine witli the Verb we may see in the following : — 1. SrNGTJLAE OF AGENT. Ja-wo, eat it. CONJUGATION OF BAHING VERBS. I. — Paradigm of Verbs Transitite m 'wo.' Root, Ja, to eat. Imperative, ja-wo, eat. AcTiYE Voice. Imjjerative Mood. 1. DUAL OF AGENT. Ja-se, ye two eat it. 1. PLTTRAL OP AGENT. Jd-ne, l/e all eat it. Ja-wosi, eat them two. 2. DTTAL OF OBJECT. Ja-sesi, ye two eat them tioo. Ja-nesi, ye all eat them two. J a-womi, eat them all. 3. PLtTEAL OF OBJECT. Ja-semi, ye two eat them all. Ja-n(?mi, ye all eat them all. 94, EXCLUSIVE AND INCLUSIVE Negative Form. By ma prefixed Ma ja wo, &c., and so in all the subsequent Moods. Indicative Mood. Present and Future Tenses. SINGULAR OP AGENT. |: DUAL OF AGENT, First Person. .Ja-sa, iucl. ■ Ja-gna, 1. -{ J eat or will 1. < Ja-suku, excl L eat it. L ice two eat it. PLTJKAL OF AGENT. " Ja-ya, incl. Ja-ka, excl. . ive all eat it. '{ DUAL OF OBJECT, Ja-gna-si, r Ja-sa-si, iucl. r Ja-ya-si, incl. I cat them 2. < Ja-suku-si, excl. 2. J Ja-ka-si, excl. tivo. I tve two eat them two. I we all eat them two. Ja-gna-mi, / eat thein all. PLURAL OF OBJECT. ' Ja-sa-mi, incl. Ja-suliu-mi, excl. •' . we two eat them all. ' Ja-yami, incl, Ja-ka-mi, excl, .we all eat them all. Preterite Tense. 1. Ja-tong 2. Ja-t-6ng-si 3. Ja-t-6ug-mi I f Ja-ta-sa, incl. I Ja-ta-siiku, excl. ) f .Ta-tii-sa-si, iucl. I Jcl-ta-siiliii-si, excl. , f Ja-ta-sa-mi, incl. I Ja-ta-siiku-mi, excl. 1 r Jan-ta-yo, incl. L Jak-ta-ko, excl. c) f Jau-ta-yo-si, iucl. I Jak-ta-ko-si, excl. Jau-ta-yd-mi, incl. excl. o J Jau-ta-yd-mi, I Jak-ta-kd-mi, § 222. This forms but a small part of the space devoted to these remarkable Plmals in Mr. B. Hodgson's Grammar of the Bdhing Language ; itself one out of the nineteen dialects of the Kiranti, one of the ten, or more, languages of Nepaul. It belongs to the Monosyllabic class, and tliis is indicated by- he hyphen between the several syllables. But this must not load us to suppose that each syllable is a separate word, PRONOUNS. 95 inasmuch as at tlie very point where our extracts end the author writes, that he has thus far separated the syllables ' merely to facilitate the student's comprehension, and that I shall do so no fui-ther, for the genius of the language is averse to any such treatment of its finely blended elements.' It is certain then that we are on or within the boundaries of the Agglutinate class of languages. § 223, Beyond all doubt there is a very gi-eat deal of this inclusion and exclusion expressed in the later and more advanced languages, and that in more ways than one. Wit Skilling=we-tioo Skilliwj=Skilling and I, has been already noticed. Let us now take a text from a different quarter : Tecum Philippos et celerem fugam Sensi ; relicta iion bene parmula, Cum fracta virtus et minaces Turps solum tetigere mento. Here the Inclusive Plural is as clearly expressed as is possible, the person addressed being in the ablative case, governed by a preposition denoting association, and the verb being in the singular number. It might be Plural, as might also the Pronoun ; and this would give sensimus and vobiscum. We know, here, that this is the parsing of the expression ; but in an uninflected language, or one of which the inflections were known only by bits and fragments, tecum might pass for a nominative plural. As for such combinations as tu- cum, it is probable that they are common ; the -cum being translated adverbially, or as a/ua ; and so giving those simul- taneously, at the same time, or accompanying. § 224. For tecum sensi write mecum sensisti. The result is the same. Two persons are indicated, and to some extent two persons are named ; i.e. one by the pronoun as a separate word, and the other by the person of the verb. Which of these two is the Included, which the Including, term? Perhaps there is no actual inclusion at all, but only some- thing akin to it. Which is the original, primary pronoun ? 96 EXCLUSIVE AND INCLUSIVE Is it / and you, or Yoii, and II It is hard to say ; and, at present, there is no need to enquire. But in the languages where the system is at its full development no distinction is more diificult. Perhaps it is one of the same kind as that which perplexes us in the case of shall and will ; but tliis amounts to an explanation. In the extract just given the change of Person in tetigere suggests the excluded agents. § 225- This indefinite number of individuals involved in the affair or action spoken of, and therefore in some degree connected with the speakers, is what we have in the Spanish nos otros, the Italian noi alti'ui, and the French nous autres. Let A. ask B. what C. thinks of D., and let the answers be, ' Comme nous autres,' or, conversely, ' Comme votos autres.' The fii'st means 'As I and yo%i and others like us;' the second, ' As you and others like you, hut not as /' (or * like of me '). § 226. It is not easy to get an insight into the details of this system ; and less easy for an Englishman than for a native of many other countries. Beyond those who, as Frenchmen, Italians, and Spaniards, are familiar with the European use of otros, alfrui, and aiitres, the only Europeans and Americans who know much aboi;t the system of nider languages are the missionaries ; and these, as a rule, admit that the right use of the form is difficult. So good a general philologist as Norris observed, that in all the explanations he had met with the so-called /jiclusive was really ^tcclusive, and the so-called .^'ccclusive /wclusive ; and that most men had confounded the two, and that he himself was only in the process of learning which was which. The point from which the line is drawn has, in my mind, something to do with it. Is it / with or without you, or yott, with or without me 1 Is it I + or — you, or yoti + or— 11 For the exclusion of a tliii'd object, the methods are numerous; and some of them, such as the Fi'cnch &c., require an additional word — autres &,c. PRONOUNS. 97 § 227. Then comes the question of Case. "We have seen that what is generally called the government of the Verb by its Nominative Case was, in the earlier stages of language, either a combination like speaking -ray, or speaking- from-me. If so, these Pei'sons, notwithstanding their cor- respondence with the Casus Rectus in import, may really be Oblique cases — perhaps Possessives. At any rate, there is no great difference between the Possessive and the Personal expression of the Pi-onoun, especially in Greek, where to ilioy=iyij, to g()v=(7\i. Let then (to) !}fj,iTfpoy = riijie~iQ ; and then compare form for form, and meaning for meaning, and place for place, the Greek hep- with the French autre. Mark, too, that the spiritus of {ifiirepoQ and iifielg is not the spiritus lenis of ii-ioc, but the aspirate of e-spor. § 228. Again : from the Melanesian group of the New Hebrides we get such sequences as the following : Inau-=.l ; kha-im — jon\ nau=:^h.e; namuhl-=^yi^Q two exclusive, or I 4- some one not the person spoken ; arivan-=-we two inclusive, or I + the person spoken to ; kha-mulh=^je two; wa^arw =you three ; dra-ti7i=we three ; dra-tavats= we four ; 7ia-iavatz=you four. These look like trinal and quaternal Numbers, and it has been suggested that such they are in reality. They are rather so many Numerals, the numei-ical value of which varies. Of the speaker or the person spoken to, one or both may be included. § 229. It is little, then, that the Dual Number does in the way of explaining the Plural. It rather complicates the question. Even when we have got an intelligible Dual, we are by no means sure of keeping it. What if it encroach upon the higher denomination, and so doing plivralise % What, if having done this, it eliminate the oiiginai plural % Now if word for woi'd the Greek ricL and o-^wt are the Slavonic nos andj'ws, and the Slavonic nos and jus be the Latin nos and vos % It has done this. If so, we are tausht that definite and precise as the sti-ucture and the formation u 98 PLURAL. of the Dual seem to 1)6, they helj) us but little in our enquiries. The Duals here adduced are those of the Pro- nouns. But the whole subject of the Pronominal Numbers is peculiar. § 230. Plural of the Finite Verb.- — The Verb here means the Finite Verb ; for the Infinitive and, a fortiori, the theme and root are Nouns. The ordinary Grammars, how- ever, recognise the Nvimber of the Verb, though they ignore it in the Infinitive. They can scarcely do otherwise. The most that a grammar for teaching pmposes could tell us would be that in the Finite Verb the verbal portion of it has no sign of Number of its own, and that what it has it takes from the inflection of its pronouns. Practically this would be enough. But in general gi-ammars it is necessary to add, that the Finite Verb, when divested of these adjuncts, is a Noun ; and that though the Infinitives have no outward sign of Number, the Verbal Abstracts, which are almost identical with them in meaning, and sometimes identical in form, have one : liunting, huntings ; fast, fastings, and the like. § 231. We can speak of twenty acts of dining, i.e. of twenty (linings, as readily as we can of twenty dinners ; tliose who partake of them being the ditiers. But twenty (linings may be made by twenty difierent partakers of a single dinner; and twenty different diners (agents) may dine off a single di7iner. Of the Seven Captains against Thebes every one may have fought in seven, or seventy-times seven, fghtioigs ov fights. Biit it was no part of any inflection to express any number beyond that of the agents. Plural of the Verb and Noun. § 232. Is there then no such thing in verborum naturd as a Plural Finite Verb ? As a definite sign of Plurality there is nothing. Is thei'e anything like it 1 There is some- thing not unlike it. There is the -d)- in (jAiy-id-M, or the -It- in voc-it-o. There is the Frequentative Formative. It NOUN AND VERB. 99 is not exactly a Plural, but it is sufficieutly akin to one to enable us to perceive that when we say tu voc-it-as there is more than one act of calling ; and that, even though the caller may be a single indiAT.dual. In vos voc-it-amns we get a plural import for both numbers ; the -it- in the Verbal part of the combination, or the theme, and the -amus in the Pronominal part, or the Inflection. But it is only the -amus that is an Inflection. The -it- is Formative. All this is written partly to illustrate the statement that the vei-bal part of the Finite Verb has no sign of Number ; but partly with a far more important object. It is submitted that these Frequentative formatives are vii'- tually the Verbal Pliirals, i.e. the signs which denote tlie plurity of the calli-ngs, the dinings, the Jightings, and the * omne quod exit in -ing.' § 233. It is now submitted that, mutatis miitandis, what applies to the Frequentative Verb aj^plies also to the Collective of the Noun. There is no question of any double Plui'ity in the Nouns, so that the Collective Formative has nothing to prevent its becoming an actual sign of Plurality ; and that, as an Inflection, i.e. a changeable affix, rather than a permanent part of the theme. The plurals, too, that ai-e required for such words as eye, ear, nose, tongue, hand, and the multitudinous concrete, individual, separable, and nu- merable substantives of the same kind, ai-e far more numerous than those required for such abstractions as seeing, hearing, smelling, touching, and the like. Nouns, then, have more Plurals than Verbs, and keep them longer. Probably, too, they took signs for them earlier. But for all this, the Noun must take its jilace in its typical form, i.e. the Concrete, and the visible, audible, or palpable Substantive. With Abstract Nouns and Adjectives the case is somewhat different ; and the Personal Pronouns have, as we have seen, a system of theii' own, § 234. If thLs be the origin of the Plural, the Collective formatives of the Noun, now existing in language, should be H 2 100 PLUR.iL. less numerous and less deianite than tlie Frequentative forma- tives of the "Verbs — (pXiy-ed-u), voc-it-o. This is because they are all by hypothesis converted into plurals. And such is the case. There are but a few collectives, just as there are but few successful traitors. When the usurper becomes a king he is no longer traitorous, and when collectives become plurals, they are no longer formative. § 235. But this does not carry us far. Nine-tenths of the details concerning formatives and inflections in the present work have been applied to the illustration of the difference between them, and not to the likeness. Again and again the -it- in voc-it-amus has been opposed to -amus, as a part of the theme, and not as a pai-t of the inflection ; as a permanent part of the word, and not as a changeable or moveable one. There is no doubt upon this j^oint. But behind this lies the leading doctrine of the work at large, viz. the hypothesis that the inflections are, as a i-ule, dedu- cible from, and reducible to, formatives ; which themselves were originally separate and independent words. If so, it is only in these extreme or typical forms that they can be definitely separated. "When we come to the boimdaries, it must be by Type rather than by Definition that we classify. § 23G. But we need not at present do this. Definition by type, though valid, necessary, and in some cases indis- pensable, is deficient in validity and precision. The following is from the Hindostani, in which ma7Y/=mau; as it also does in the Persian. 1. Hindostani. Sinr/ular. Plural. Noininative mard mard Genitive mard-ka mard-on-ka J? mard-ke mard-on-ke !> mard-ki mard-on-ki Dative mard-ko mard-on-ko Ahlafiva luard-se mard-ou-se Ini'ssive iiiard-nien mard-on-men Adessive mard-ne inard-on-ue NOUN AND VERB. 101 2. Persia:?. Singular. Plural. Koniinative mard raard-an Genitive i-mard-i i-mard-an Dative mard-va mard-an-ra Accusative mard mard-an Ablative az mard az mard-an Tlie -on- here is uuiversally treated as a sign of plurality. Yet in place and stability it is more of a formative than a plural. Are we then to deny that it is a pliii*al of any kind 1 Are we to say that it is a plural, but not an inflection 1 Are we to say that the distinction between the permanent and the changeable parts of a word is unimportant, or that Plurality is only a modified form, or Variant, of CoUective- ness 1 We may certainly say all this, and more, and not be very far from the truth. § 237. I submit, then, that Pliuals are not naturally changeable or movable, and that naturally they are parts of the theme — just like the -it- in voc-it-o, &c. ; and that it is not till we get to a certain stage of development that they are other than formative, or thematic. It is only in the Synthetic languages that they become generally and mani- festly movable, and not only this, but indistinct — indeed, it is because they are indistinct that they are movable. They are this because, in the Synthetic languages, there is a great development of a new element— Gender ; and with 'Gender Number becomes fused, combined, or amalgamated. Hence, in the main, the Formative Inflections, or the Col- lective Plurals, are mova^ble, non-radical, or non-thematic only when these signs are complicated by those of Gender. I do not then say forms like the -on- in 'mard-on-kd and the like, are not Plural, for we must take the term Plural as we find it. I only say that under certain circumstances the word Plural may become ambiguous. § 238. Collectiveness, however, seems to be an older idea, 102 CASE AKD TENSE. and as such it prevails as language advances. There are, and must be, in the domain of Thought more plurals than col- lectives : just as there are, and must be, more units than pairs, more twos than threes, more parts than wholes. Every collection is what it is because, as a unity, it contains more than one individual. In the earlier starves of language it may not be very necessary to separate such unities into their elements ; and, moreover, it may not be very easy to do so, for it must be remembered that the art of countins is of comparatively recent origin. It seems, then, that befoi-e the series of Numerals became adequately developed (and how low it is in some countries even now will be seen in the sequel), the notion of Collectiveness prevailed over that of Plurality. As definite Numeration became easier. Plurality would prevail over Collectiveness. How closely they are allied, and how naturally we mix them together in thought, and how difficult it is to say which, in certain cases, should determine our choice of an expression we well know ; for we know how naturally with a Collective Substantive we connect a Plural Yerb ; and how the majority of grammarians warn us not to do it, and how, nevertheless, we continue to do it. § 239. The sign of Number follows that of Person, and precedes that of Case ; and from this it is readily separated. With Gender (or Conventional Sex) it takes the same place ; but, whether it be that one of the signs is omitted or that the two coalesce, there is often but a single syllable ; sometimes only a single sound or letter serves for the expression of both — e.g. hon-us, bon-ce, hon-os, hon-is. (8.) Case ( t Nouns). — Tense (of Verbs). § 240. Signs of Case, or Case-endings, fall into two classes : (1) those of the Greek, Latin, and Sanskrit type ; (2) those of the Fin and Turk type. In the first the signs of Case, Gender, and Number NOUN AND VERB. 103 become confluent ; or, perhaps, two cbop out. There is nothing to be done with these without a great deal of preliminary analysis, which is by no means easy. In the second we get an independent and separable syllable. But this goes for little until it can be shown what it is as an independent word. We cannot often do this ; and when we succeed in doing it, the result may be that the combination is one of simple composition rather than a true inflection. Short of this, however, we get a fair amount of in- formation so far as it goes ; for it is something to know in a word like the Turkish adem-ler-un-=ih.e Latin hom-in-um, that the -ler- denotes IsTiimber only. But this depends upon a fortunate coincidence; viz. the fact that where we get what we may call syllabic signs of Kumber, we also fail to find the signs of Gender wanting. It is certain that in the Synthetic languages it is a very diflicult matter in many cases to say what represents Grender, what Number, and what Case. It is not so drfficult to do this in the Agglutinate forms of speech, and it is often very easy. The inflections are fewer, and the signs of them are more distinct. We get thus much, no doubt. But, iintil Ave get more, Case-endings look Like mere arbitrary signs ; being what they are solely through theu' i-elation to the main words. What they are of themselves we have to discover, if we can. The earliest doctrine was (and is to a gi-eat extent) that they wei'e ^prepositional in origin and in character ; modified in form ; and rt/fixes rather than prefixes to the words with which they were connected. It is not worth our while to incriminate the compound Postpositive Pre-position as a contradiction in terms. We must take it as we find it ; j ust as we must recognise Dis-junctive Cwi-junctions and the like. They are this, doubtless, to some extent, and in a certain way ; but it is probable that the extent to which the doctrine presented itself to the philological world (we might almost say obtruded itself) has depended on an accident. Nine philologists out of ten are familiar with some synthetic tl 104 CASE AND TENSE. and with some analytic forms of speech ; and as words like lapid-is, lapid-i, lapid-e, &c. are translated of a stone, to a stone, from a stone, the notion that words like of to, from, kc. correspond to the terminations like -is, -i, -e, &c. forces itself upon them ; or, at least, presents itself spontaneoixsly. ~B\\t this is little more than a i-eproduction of the translation of yi-ypafci, or r(Tv(i by I have written, or / have beaten; wherein no one thinks that, word for word, or rather syllable for syllable, the foi'mal elements of the two combinations are identical in origin. The words vohis-cum, nohis-cum give us true postpositive prepositions ; but in these it is the syllable -bis-, and not -cum, which is the sign of Case. The real fact is that, as a rule, the actual Prepositions of a lan- euase do not coincide with the Case-endings in form, what- ever they may do in import. § 241. There is no better instance of this than the history of the so-called Genitive Case in English and French. What the Case which, both in French and English, comes next to the Nominative may be in import, is equivocal. The English has a sign for it-s. But the French has nothing of the kind. Taken by itself, the French Genitive (tete=head) is no more Genitive (or Possessive) than it is Nominative, Dative, Voca- tive, Accusative, or Ablative. But it has its sjjecial pro- noun, c?^= of; at least, so far as ' de )!ea ;' only, however, so far as the import goes. Word for word, de and of no more stand for the -is of capitis than the have in I have written stands for the ye- in ye'ypo^a. But, at any rate, we may say, the import is the same. Not so in all cases. In English a picture of John is a very different thing fi-om John's picture. At least, however, de and of govern a Genitive Case. This is just what tliey do not. The Case that both de and oj NOUN AND VERB. 105 govern is the Ablative. To come to the worst, the Preposi- tions are Genitival in sense. This depends on circumstances ; and it does this occasionally because what we now call Genitive is partly Genitive and partly Ablative ; Genitive when it is possessive in import, Ablative when it is partitive. Now it is with the notion of partitivity only that both of and de are naturally connected. The real history of the two pai'ticles and the inflections with which they become con- nected, is that the meanings of the two Cases are sufficiently akin to become confluent. In like manner, the Greek Ablative is the Latin Genitive. But this is a question connected with the Affinity of Cases, which is here only alluded to. The reason, however, for alluding to it is much to the present point — -viz. the maccuracy of identifying the inflections of the Synthetic parts of the Noun, with the Prepositions which represent it (more or less approximately) in sense. There is confusion when we call the French de a sign of the Genitive Case; and when we say that the Genitive is the Case that it governs, there is ' confusion worse confounded.' A combination of separate words with the same meaning as a single word with which it corresponds, and has superseded, is one thing ; the identity of the respective elements of the two combinations, is- another. Between ' / have written ' and ' yiypafa ' we see it clearly ; between capitis and de tete we see it indistinctly. § 242. Notwithstanding all this, it cannot be denied that, in some instances, we have what to all intents and purposes are Cases formed fi'om actual Prepositions. Such a word is whilom=at or about some time (while) ; some time ago. The Preposition is obsolete ; the Substantive obsolescent. Such, however, is the structure. Here the Preposition is a post- positive. On the other hand, in Italian, such combinations aa pello=per ilhcm ; collo = cum illo ; nello = in illo and the like ; where it is truly /?re-positional. These are manifestly secondary fonnations. The Gaelic, which gives us the lan- guage of an earlier stage, has also words like it, which are, 106 CASE AND TENSE. in like manner, ;:>r(?-positional. But in Gaelic and Italian we have the several Prepositions as separate words. These are, also, secondary. Are they inflections, or are they mere com- pounds 1 They are certainly the latter ; perhaps the former. But in no case are they the inflections of which we have to analyse the structure. We see this at once. The inflections that we have here to examine are those of which the origin is problematic. And here comes the diflficulty which always besets us. When the case is plain, there is no problem at all. § 243. Between the combinations like whilo7n and those like collo we get both />re-positions and ^osi!-positions. The real Preposition is neither one nor the other. Wherever there is a Preposition there are tvjo ISTouns, and its place is generally between them ; thus being inter-positional, i.e. pre- positional to one, jJos^positional to the other. § 244. Such is the criticism that applies to the doctrine of the connection between the Prepositions and the Cases of Nouns. There is a connection of some kind, but the one that has just been noticed is too much of a sectional, or partial, character. Reasons will be given in the sequel for believing that the signs of Case are of earlier origin than the Preposi- tions ; and that, where case-endings can now be replaced by Prepositions, they were not Prepositional in origin, though akin to the Prepositions in import. And this is as much as need be said at present. And even less than this will be said about Tense. All that is, at present, said about the tw^o is, that Case and Tense belong to the same system of Inflection, and are analogous to one another; i.e. that, mutatis mutandis, Case is Tense in respect to Time, and Tense is Case in respect to Space ; and that words like ' hehind,^ ' here,' and ' hefore ' correspond with words like 'jmst,' 'jyresent,' and 'future.' It is, however, from Space that the principle of direction is deduced. This is be- cause the relations in Space are the most definite, i.e. those that most specially present themselves to senses. Such, at least, is the presumption. GEKDER AND MOOD.— XOLTN AND VEEB. 107 (9.) Gender (of Nouns). — Mood (of Yeebs). § 245. Gender is one thing ; Sex is another. So is the import of the word Masculine and Feminine as opposed to Male and Female. Sex with its Males and Females is a natural division of objects : Gender, with its Masculines and Feminines, is an ai-tificial, a conventional, or a grammatical one. Artificial, however, as it is in the nature of external objects, it is real in Language. This reality we must recognise and explain as best we can. Of forms like Qeoc, Xeawa ; genitor, genetrix; duke, duchess ; freund, freundin ; fox, vixen, and others, where there is no change in the inflection, and where the gender coiTesponds with the sex, nothing will be said. The forms and combinations that specially illusti'ate gender are those like bonus clypeus, bona ijanna, and bonum scutum, where we get well exemplified the two chief difficulties connected with this part of grammar. These are : (1) the little agreement there is between sex as we find it in nature and gender as we find it in language ; and (2) the incomplete and indirect way in which the gender, thus characterised, is exjjressed, Clypeus, parma, and scutum have very nearly the same meaning, viz. that of a shield of some kind : a shield which is not only invested with a sex, but which, as clypeus, is Masculine, as parma. Feminine ; and, as scutum. Neuter, i.e. neither one nor the othei-, or else both. § 246. There, certainly, is something in this which is widely difierent from what is meant by Sex, Male, and Female ; or, in other words, thei-e is a variance between the Ganders of Thought and Language and the Sexes of external Nature. And it is this that perplexes us. What is this Gender which has no counterpart in Sex ] What is it in Thought 1 What is it in Form % How far does it extend % What is the history of its development 1 § 247. ^Y^lat is it in thought 1 The principle that has commanded the most attention is that of Personification. 108 GENDER AND MOOD. In the earliest stages of language and in tlie newest we can always find some inanimate or some immatei-ial object which is invested with Sex of some kind. It may be as material and common-place as the scythe or spade of a day-labourer, or as high and abstract as the Seven Cardinal Yii-tues of the philosopher. It need not be indicated by an inflection, nor by any such formative as the -ess in duchess, &c. &c. It may show no sign of its existence in the substantive to which it i-efers, any moi'e than it does in such words as sldp and scythe. Nevertheless, if, when these are represented by their Pronovms, the sailor calls his ship * she ' or ' her,' while the mower calls his scythe ' he ' or * him,' the principle of personification asserts itself; and with it comes the notion of Gender, or conventional Sex. There are other subordinate and partial explanations of this contrast ; and these, so far as they go, may be valid. But Personification carries us the fai-thest. § 248. Yet it does not carry us far. It is the most con- spicuous in the domain of Mytholog)^ with its Gods and Goddesses, Here the Personification does not so much personify individual objects already existing, as it creates, or makes for itself, imaginary persons. For the more inde- finite objects of thought it makes Abstractions ; which, in general, are Feminine. Sometimes there is a physical element manifestly suggested ; as when in certain American languages the Sun is the Man in the Sky, the ]\Ioon the Woman in the Sky. Here the difierence of bulk and power originates the image. But in the German languages there is a Ifan in the Moon, and the Moon is Masculine whereas the Sun is Feminine. Here the Sex is determined on different principles ; but, in both there is purely physical element. In respect to the habit of unlettered Englishmen calling one implement a ' he ' and the other a ' slie,' it may bo merely a remnant of the older system of Genders preserved; though of Geiuler in the Noun itself all conspicuous traces ax-e lost. It may in some cases be what NOUN AND VERB. 109 is suggested by Cobbett, viz. becavise cei-tain objects are more specially connected with himself — lite a wife. It may be this. It may be that in some cases a number of individual objects may have only the value of one of a higher denomina- tion ; so that if many females are reqiiired to make up the value of a single male, the question is as much one of Number as of Sex. There are some forms in Tutshek's Grammar of the Galla Language that favour this view, and in a very low state of language something of the kind is not unlikely. Then there is the whole of the Vegetable Kingdom, which so far as it is productive of seed and fruit is Feminine rather than Masculine. Nevertheless, we may take all that has been thus suggested, and making the most of it, ask how far it carries us. The answer is, ' Not very far ; ' farther in some languages than in others, especially in those where it connects itself with M}i;hology. But it nowhere gives us a general principle. We have only one foundation for our argument, and this we must get from the actual nature of the objects themselves with the nature of the words, or formatives which symbolise them in Language. With these we can consider the two essential questions. How did the idea of Sex fb'st present itself 1 and how fai' did it extend itself beyond the domain of Natural Sex 1 § 249. The fundamental fact in this is the undeniable existence of such a difference as that of Sex, not only in external Nature, but in Thought, and in the Language that expresses it. All that we have to do with this is to i-ecognise it. In words like boy and girl its expression is clear enough. In words like man-servant and maid-servant it is equally so. In words like genit-or and genet-r-ix we can, also, understand it, though not so readily. But all this is intel- ligible. We may wish to know more ; to know, for instance, the exact origin of the formative syllable like -ix ; but why they are used, and what they imply, we know. There is a difference in Nature with which, as forms, they correspond. It is in the Substantive that we comprehend it best. 110 GENDER AND MOOD. § 250. In the Adjectives and Pronouns, especially the Demonstrative and Relative, we do not see our way quite so clearly. A male or a female ' good,' a male or a female ' tills' a male or female ' which,' we can only understand when we know what they mean as Substantives. But this we can discover fi'om the context. In a language where there are neither formatives nor inflections all this is mysterious. But when we get these, and, along with them, Concords, we can realise the character of the change. They adapt themselves in form to the Substantives with which they agree. This agreement is of two kinds. The Adjective gives us a Quality, i.e. when attached to its Substantive, tells us that a good man is a man and something else. The Pronoun gives us the actual Substantive with which they correspond under another name. In fact, it gives us a Spionym. There is the element of a Concord in all this. § 251. If neither the Adjective nor the Pronoun have any sign of Gender whatever, there is no positive Concord ; i.e. thei'e is no sign whatever by which we get either agree- ment in form or disagreement. Bvit there is no Discord ; nothing by which the notion of Concord can be violated. There is nothing of this kind in the present English. We cannot, in English, violate the rule of Concord (which is a natui-al one), because the Adjective has no sign of Gender. We could do so when it had one. We can do so, even now, with the Pronouns of the Third Person and the Eelative, because they have one. If speaking of a female we cannot say ' he did this.' The Concord, then, when there is a sign to expi-ess it, is a matter of common sense rather than aught else. We must avoid a DiscovA. § 252. Given, then, a language which has arrived at a state in which we get Parts of Speech, Substantives, Ad- jectives, Participles, Demonstrative, Relative, and Personal Pronouns with signs denoting difference of Sex, there must be a Concerd in form when any two words applying to the same object come together. Wc may dispense with sucli NOUN AND VERB. Ill Concords altogetlier ; but, if we have tliem, we must use them in their proper places ; and just as we have Concords of Person, Number, and Case, we may have one of Sex. In short, we have it ; and we * hiow the reason why! If we have words like genitor and genetrix, and bonus and bona, we must not say bona genitor, or bonus genetrix. There is no reason why we should have such forms ; but if we have, we must not combine them cZiVcordantly. § 253. But Gender is not exactly co-extensive with Sex ; and the pei-plexing fact now confronts us, that we have signs of a difference of Sex extended to Substantives where there is no Sex at all. In other words, we have Masculines and Feminities in Gender and Language where we have no corresponding Males and Females. This is oui- fundamental difiicultj^ § 254. Wliat is the sign of Gender in Form ? An In- flection rather than a Fo7-mative. The -ix in genet-r-ix is a formative; for its inflections are exactly those of genitor. The Plural of both is -es ; their Cases -is, -i, -em, -e, and so on, the Plural. The -or and -ix are parts of the theme — unchanged throu"-hout the Declension. In words like bonam, bonas, bonarum the Vowel changes with the Case and Number. In combining these we must avoid a Discord; and the sign which makes this Discord possible is an Inflection rather than a part of the theme. § 255. How far does this system extend ? It does Qiof extend to the Neuter. We may call Nominatives like regnum or bonum Neutei'S. But they are not Neuter as opposed to Masculine and Feminine, or to 2Iale and Female. They are not Neuter as to anything in the way of either Grammatical Gender or of Natural Sex. They are the Negatives or Opposites to something of a very different kind ; and they represent a contrast which oi-iginated in an earlier stage of language. At present, the Neuter may, conveniently, be treated as a third Gender ; but, in origin, so far as Gender is equivalent to Sex, it is no Gender at all. 112 GENDER AND MOOD. It was not in the difference between Male and Female that the Neuter oi-iginated, but in the difference between Animate and Inanimate., for that is the difference which was the ffrst to be i-ecognised. Thei-e is no doubt about this. The Dual division of Animate and Inanimate preceded the Triple dis- tinction of Masculine, Feminine, and Neuter, § 256. More than this. As this Dual division was the fii'st to be recognised, it will probably be the last to be abandoned. In the Old Norse (or Icelandic), the Genders are the tliree ordinary ones — Masculine, Feminine, and Neuter. But in the Danish and Swedish they are reduced to two ; the one that is abandoned being the Feminine, the one that is retained the Neuter, Of something akin to this we have an example nearer home. In the inflection of an Interrogative and Relative Pronoun the division is, also, Dual : (i) Mascxdine or Feminine, i.e. Common, and (2) Neuter — (1) who, (2) ivha-t ; (1) who, (2) whi-ch ; though the -ch in the last example is not exactly an inflection, but the -Ik of the Scottish whi-lk, which is the Adjective like. § 257, In some grammars, instead of Ariimate and In- animate, we find such words as Rational and Irrational. And we see our way to this when we consider how readily the lower animals may be looked on as tilings. But this is not all. There is a great deal which makes the division a natural one. Whatever may be the exact difference between States and Actions, and, however clearly we may make out a case for every state being, to some extent, an action, there is a great practical difference between the agent or subject on one side and the patient or object on the other, especially when we consider the proportion in which names are likely to be allotted to them. The more conspicuous the agent the greater will be the likelihood that he will take a name ; especially when he belongs to a class that gives names ; and this is the class Avhich the Animate beings of the world best represent ; and of the Animate beings the Rational most exclusively. IS^OUN AND VERB. 113 § 258. The relegation of the lowez- animals to the class of Things rather than Persons, Ls diametrically opposite to the later system of CTender. In Gender we invest objects which have no natiu-al with a non-natural Sex. The older process, which Kmited the Animate world to the Rational, divested of Sex such animals as had it. § 259. This leads us farther. It is probable it was neither Sex, nor Life, that determined the original division. It is more probable that it was the difference between the class of Agents and the class of Objects. § 260. So much for a limitation of one kind ; a limita- tion which reduces the number of Genders. But there is another, and one of another kind, that we must, also, take under consideration. To what Parts of Speech does the system, as we have it, extend 1 Does it extend to the Sicb- stantives % If it do so, does it manifest itself by the same clear, and definite signs that present themselves elsewhere ? Supposing it to do so, does it do so to the same degree throughout the class ] To all this we must answer, that the Substantive is not the Part of Speech wherein the system of Gender is most typically represented. In the typical Sub- stantives there is no inflection to express it; no sign or mark whatever. The words, nevertheless, are Masculine, Feminine, or Neuter, as the case may be. They are sup- posed to be this always ; and, when the difference of Sex is real, they are so ; but Sex is not Gender, neither is Gender Sex. When the Gender, then, neither coincides with Sex, nor has any distinctive sign of its own, how far is it Gender at all ] How far are words like lajns INIasculine, and words like nubes Feminine, when they stand alone 1 The sense tells us that, in the way of Sex, they are neither the one nor the other ; while in their declension they are identical. In fact, the terms in question have neither real Sex nor any sign of grammatical Sex or Gender when we take them as single words ; and, when we have to do this, we find them I 114 GENDER AND MOOD. somewliat tronblesome. In onr lanqaiage they are not so, because in English there are pi-actically no signs of Gender ; but in French and German there are scores and hundreds of words which when found for the first time have to be looked out in a dictionary before we know their Gender. In short, except foi' the fact that an Adjective or Pronoun, which has a sign of Gender, agrees with them, and they with it, they have none. § 261. The SubstantiA'es to which this most specially applies are those of the Third Declension in Greek and Latin ; and these it is which are most typically Substantival. The inflection of words like dominus and domina is that of the Adjective. § 202. Gender, then, is the result of a Concord, rather than of aught else ; and it is in the Adjective rather than in the Substantive that we find it. § 263. It is in the Adjective rather than in the Sub- stantive that we find the Gender as opposed to Sex. It is in the Substantive, however, that we find Sex as opposed to Gender {(jenitor, genetrix). The Concord, too, is a strange one ; with Nouns of the Thiixl Declension it is the Substan- tive tliat agrees with its Adjective or Pronoun rather than. the Adjective or Pronoun that agrees with its Substantive. Are we, then, free to say that the signs like the -or and -ix in geiiitor, &c., the formatives expressive of Natural Sex, originated in the Substantive 1 and that those like -us, -a, -tmi, the signs of Gender or conventional Sex, are adjectival in origin 1 The first statement we may safely make ; the second is not so clear. AVe can only say something like it. It by no means follows that because the Substantive may have taken its Gender from the Adjective, it is the Adjective in which we are to seek its origin ; for this is by no means the case. I submit that the Adjective took it from the Pronoun. If so, it was in the class of Demonstratives and their nearest congeners that Gender took birth. By ' nearest congeners ' I mean the Relatives, the Interrogatives, the PROXOTJN. 1 1 5 Definite Aiiicles, and, above all, the Prononns of the Third Person. § 264. It is in the Demonstrative Pronoun that Gender seems to have originated. Demonstratives are Pronouns, and Sex is a Qualitij. Pronouns, however, have no Qualities. They have only Relations. But every Pronoun has its duplicate; i.e. a Noun which corresponds with it. And this Noun ha^ Qualities ; and when the Pronoun is, so to say, translated, interjyreted , or specified, by its correspondent Noun, it, in its adopted form, takes to itself Qualities ; and amongst them that of Sex. It has not these in itself; inas- much as words Kke ' this ' and ' that ' imply nothing but a greater or less distance between the objects they apply to and the speaker. If a man, a woman, and a t7'ee present themselves at different distances, the man being the nearest and the tree the most distant, nothing but the difference of the intervening space is suggested by such terms as ' this,' 'that,' and '■you' Nevertheless, '■tins' is the name of the Masculine man ; ' that ' the name of the Feminine woman ; and • you ' the name of the Neuter tree. But change their pla<«s, and the Sex of the object changes also ; and each name may have any one of the three imports. Relations are changeable ; and so is the power of their signs in Speech ; but so long as words like ' tlds^ ' that^ or ' you ' mean man, woman, or tree respectively, so long are they Male or JMas- culine. Female or Feminine ; or neither one nor the other, which is Neuter. It is only, then, for the time being that these Pronouns have Sex. But this time is always changing, § 265. The Demonstrative Pronouns, then, have Sex just as much as the Nouns. But only /or the time being. They have the sex (for the time being) of the Noun with which they correspond. This depends upon the Attribute of Rela- tions ; but Relations are always changing. It is impossible under these conditions that such Pronouns can have any permanent sign of them ; for any integral part of their natural I 2 116 GENDER AND MOOD. structure ; anything that would not be every instant liable to be converted into something else ; anything essential to, and inherent in, their sti-ucture ; anything like an Attiibute that might be attached to them as Pronouns. Such an Attribute must be a Quality — an inherent Quality, if we like ; but it is just because a Pronoun has the Attribute of Relation, to the exclusion of that of Quality, that it is what it is. On the other hand, the counterpart Nouns had the Attribute of Quality, to the exclusion of that Relation ; and, because they had it, they were Nouns as opposed to Pro- nouns ; and, because the Attribute of Quality is permanent, they were permanent names. Some change must now take place in respect to the Pronouji. There must either be more of them, i.e. new words, like he, site, or the, or else existing wox'ds modified, as heo, heo, hit, is, ea, id, and the like. Or both processes may be adopted. In the Synthetic language we have both forms, but with differences. In our own language we get the Neuter of both he and the ; or it (hit), and that. Still, the fii-st is used in a Personal, the latter in a Demonstrative sense. But the West-Saxon Personal form is se, and its Feminine seo ; of this the only Gender or Number is, extant in the present English, she. In the Lithuanian, and I believe in that language only, it is declined throughout — in six Cases and three Numbers ; for it has its Dual forms, though they are only compounds formed by the addition of the term for Quantity, or rather, Quotity. § 266. But it is not this that we need enlarge on. The point to which we must most specially attend is the fact, that in the 2)assage from the Demonstrative to the Personal Pro- noun, we get not only Gender in a Pronoun, to which, as an Attribute of Quality rather than Relation, it is naturally and essentially foi'eign, but a reason for its being adopted, and an argument in favour of its being the Pi'onoun in which it originated ; or rather the Pronoun in its transi- tion fiom Demonstrative to Personal. PEONOUN. 117 § 267. It is in this transition that we find the germ of. the triple division into Masculine, Feminine, and Neuter. The dichotomy of everything in the Universe into, ' Every- tliing is either I (Ego) or Not-I {Fon-ego), is that of the logicians ; and it is exhaustive. In Language it becomes, for the Pronouns of the Third Person, a trichotomy ; viz. / {the Speaker) ; Thou {the object spoken to — generally Rational, and always Animate) ; and ' This,' ' That; or * The other: For the first two no sign of either Animation or Sex is needed. For the third, something of the sort is I'eqiiired. The special name of each of the three objects in a narration cannot be continually repeated. Neither can we ring an indefinite number of changes on ' This; ' That; and ' The other.' A change of form takes place. § 268. When this is completed, the Inanimate, or Iri-a- tional division of the older stage of language, and the Neuter of the newer, become, in a way, confluent. But only par- tially ; and so as to meet the classification of the grammarians of the Synthetic languages. The older division includes the Neuter, a7id something more. Hence it is only to a certa.in extent that they coincide. There is a paitial confluence, but there is no co-extension. § 269. Traces of the older division still survive. The Danish and Swedish Genders have already been noticed. So have the Inten-ogatives and Eelatives of our own lan- .guage. In the Slavonic, the Masculine Gender falls into two divisions : the Masculine Animate, and Masculine /^animate. With the Masculine /?ianimates it is the Nominative (as else- where) with which the Accusative agrees in form. With the Animate, it is the Genitive. § 270. We now see something to which such an apparent anomaly as Gender, or wow-natural Sex, can be traced. {a) There is real Sex in Nature, and in Language there is a Formative to sign-ify it. It is in the Substantive that this originates. 118 GENDEK AND MOOD. (ft) In another direction there is, also, a conciiiTent pi-ocess. The Demonstrative and (oi-) Personal Pi'onouns take to themselves a sign of Gender, and, this being done, agree — or stand in Concord — with the Noun to which, as names, they correspond. (c) This tendency to sign-i£j the Concord extends itself. When a Substantive has an Adjective in combination with it, there is, whether sign-i&ed or not, a Concord — at least, so far as a Discord would involve a contradiction. There need not be any formal, external, or positive sign of it; in other words, it is not necessary, nor is it always the case, that there is any sign of Gender for the Adjective. If, how- ever, there be signs, they must not be discordant. In some cases, the Substantive being understood, a single word such as bonus=good man, bo7ia=good woman, stands by itself. When this is the case, the Adjective is, pro tanto, Substan- tival. § 271. Then comes the extension of such signs of Con- cordance beyond the bounds of Natural Sex. ' Gender ' now becomes a convenient term ; and, when once the change has begun, thei-e is much to promote it. The Demonstrative Pronoun itself is something more than a Demonstrative. It is also Interrogative and Relative ; also the Definite Article. This being the foundation of the Third Person, it presents itself in the Verbs more often than the other two Persons put tosether. Then there is older division of Animate and In- animate ; and, still more important, that of Rational and Irrational. This would make the whole of the lower animals Neuter ; and, so doing, contribute greatly to the confusion of Sex and Gender. Then there are the minor elements which, in detail, would promote the same intermixture — Personifi- cation, especially when connected with Mythology, being the chief of them. But, with all this, it must be remembered that tlie extension is by no means very wide ; or, at least, that, except with the Pronoun and the Adjective, it is of an equivocal and imperfect character. In words like dominus PEONOUN. 119 and domina, the distinction is real. In -words like mensa and sylva, it is conventional ; but in both cases the declension is Adjectival. With the t}^3ical Substantives of the Thii-d Declension there is no sign of Gender at all. There are words like genitor and genetrix ; but, in these the difterence is real, and the formative signs are in theii- proper place. In Neuters like opus the Nominatives and the Accusatives are alike ; and, in the Plural number, they end in -a {ojyer-a), rather than in -es {lajnd-es). But this is not a difference of Masculine and Feminine, it is the older difference of Animate and Inanimate. § 272. Of the extent to which the present class of Neuters is something more than a mere collection of the names which are neither Masculine nor Feminine, we have already had evidence. We have seen that the present triple division of Gender is of secondary origin. The lower languages tell us that the older division was a Dyad, or Ducdism ; and that this was the first in origia. The later languages — the Danish and the Swedish — tell us that the double division survives the triple, and that throughout the whole domain of Gender. The present English tells us that ia the Interro- gative and Relative Pronouns there is no Feminine ; whereas, there is a Masculine and a Neuter. The Slavonic tells us that, concurrently with the ordinary triple division into three Genders, the Masculine draws a distinction between Animate and /^animate. § 273. From another point of view, however, the whole gist of the present observations on Gender have been to the effect that, over and above its tendency to present itself late and to disappear soon, its signs are of an indefinite character. To the Neuter, however, neither of these criteria appertain. This is because the Neuter is not an instance of Gender. The Neuter shows itself early ; and has yet to become either obsolete or obsolescent. And it is just the same with its Inflection. Its signs are definite ; no mere differentiations of some vowel or diphthong, but con- 120 GENDEE AND MOOD. sonantal in form and permanent in duration. The dif- ference between !> and >/, e%wi and earn, the, and theo, se and SCO, is indicated by an evanescent vowel. The signs of the Neuter are Consonantal ; and, if not exactly ever- lasting, permanent. The Neviter sign, in short, is always the most positive of the three. It is, certainly, in the Pronouns that this compai-atively definite, permanent, ard undoubtedly consonantal sign presents itself. I do not use this as a reason in favour of its havinsf originated in the Pronoun. I merely indicate it as a pre- sumption in favovir of its having done so. Against this view I know nothing. In the Gei-man Family of Languages, however, the Adjectives follow, in its signs, the Gender of the Pronoun. Fro tanto, this is in favour of the view. Of course, practically, the Neuter is a Gender ; but, historically, it is connected with a system in which Sex was only a secondary and accidental element. § 274. Of the extent to which the earliest Inflections are the most persistent thei-e is no better evidence than that of the Interrogative Pronoun. The Greeks said oc, h, f>, ', the Latins qui, qu(B, quod ; the English who and what, having nothing that corresponds to >/ and quce. We may, or may not, call this Gender. It is cei'tiiin that it is not Gender founded on Sex. We may, or mny not, call this Gender Neuter ; but it is certainly not the Neuter of a triple division, or /j-ichotomy. It represents one of tioo classes, not one of three ; and it is not expressive of a mere residue, consisting of something that is neither ' tlns^ nor ' that.'' It is a term like ' Tlie other ' when it applies to one object out of two ; but it is not ' The other' in such a phrase as ' This, That, and Tother.' It is ' The other ' as opposed to either ' This' or ' That ' singly ; but not ' The other ' as opposed to ' This ' and ' That ' collectively. The difference, for the purposes of practical grammar, is of no great importance ; but the fact of being retained, to the exclusion of the Feminine, in the Interrogatives, is not without its import. PRONOUN. 121 § 275. We may now add, that the sign of Gender is one of a peculiarly faint and indistinct chai-acter ; being formed by modifying some sign previously existing, rather than hy the introduction of any new element. Few, if any, of these signs contain a consonant, neither can they always be separated from the sign of the Number and Case ; indeed, as a rule, the three become confluent, and one syllable may stand for all. When inflections are of this indefinite character, they are rarely either old or permanent. They come into language late, and they fall out of it early. This is mentioned — not for the first time — to show that the two characters may be taken together ; and that so they give a natural character to any class founded upon them. § 276. It was not, then, for nothing that the §§ 172, 195 were written upon the difierence between Inflections formed by the separate and definite addition of what, by hypothesis, was oi-iginally a distinct and independent word, and Inflections formed by Differentiation ; and that, at the first opportunity, it was laid before the reader ; nor is it without a purpose that the term Variant has elsewhere been suggested. More in the special notice of the Mood will be said about this. At present it is sufiicient to state that in Inflection the differentiated form Ls common to both Mood and Gender. Mood. § 277. The remarks with which the notice of ' Gender ' concluded are those with which it is convenient to begin the notice of ' Mood.' This is because, in the present treatise, some pains are taken to establish a comparison between them. In both divisions thei-e is a ininimum of Inflection in the way of addition ; the chief changes being fi-om one vowel or diphthong to another, i.e. by differentiation. In both divi- sions the result is short-lived. The distinction of both Gender and Mood come into language late and drop out early. 122 MOOD. § 278. Of the way in whicli a Mood like the Conjunctive dies off, we have a good instance in our own language. In very con-ect English we are supposed to say ' If this he the case, you are right,' rather than ' If this is the case,' &c. But ' be ' and ' is,' it may be objected, are two difierent verbs, and not the same verb in two different Moods. Hence, ' if I were you ' as opposed to ' if I was you ' is nearer to the point. Now this is the only instance in English of a positive distinction in the way of Mood. In sentences like ' if he choose, he will comply with your request,' the form ' choose ' is preferred to ' chooses ' ; but we know that the distinction is often neglected. What, however, when we get it, is the sign of the Conjunctive 1 Simply the omission of that of the Indicative. With the exception of the word were as opposed to was, all the English Conjunctives are of this negative character. The signs, then, of Mood, in English, are obsolescent, those of Gender obsolete. § 279. It is easier to say what Mood is not, than what it is. Etymologically, it is a Mode of some kind or other ; but it is certain that, in the divisions and subdivisions of the class to which the term applies, there is very little regularity of relation. In words like Vocito^I call often, ■yr/paffK:w= / begin to grow old, we express the kind of action, but none of the circumstances, conditions, or accidents under which it presents itself to the person spoken to. In words like rvirrtii-, rinr-oj, tv-te, or rvirroifiL there is no change whatever in the character of the action, which is neither frequentative, nor inchoative, nor anything else end- ing in -ive, but purely and simply heat. In words like vocamus and vocitamus we have two different verbs; in words like vocamus and vocemus, vocitamus and vocitemus, we have identical verbs in different moods. § 280. In Greek and Latin, Erequentatives like vocito and fXeyiOio, and Inchoatives like viresco and yripacricio — in conti'ast to sucli forms are amarem or TinTT(>ij.n — ai'e not likely to be mistaken for anytliing but what they are. But VEEB. 123 in many languages all these formations are considered to constitute Mood ; so that the Moods of this may amount to more than a dozen. These, however, we may wholly exclude ; and the Infinitive Yerb we have excluded already. This last may, of course, be called a Mood, if a definition of that term be framed so as to allow the Mood to Nouns. But this is another question. The Moods now under notice are those of the Finite Yerb exclusively. § 281. How many of them are they 1 This is a question which partly belongs to Logic, and paitly to Grammar ; and when this is the case, each illustrates the other. It is the Finite structure of the Yerb that makes that Part of Speech what it is in the eyes of the logician. The Infinitive portion of the compound supplies the Predicate, the Personal part the Subject. Between the two we get a Proposition ; some- times a very short and compact one— as eo = I go, no=:^I sv;im, and others. But it is not with all these forms that the logician has to deal. It is only when the verb delivers a statement, an assertion, or a predication to the effect that some one does, or does not do, somethiiig, or that some one is or is' not, something, that the logician recognises the Yerb, The simple irnion of the name of an agent, or being, with that of an action or state, is not enough for him. He may or may not allow that such combinations constitute sentences ■ — perhaps even propositions ; but such propositions and such sentences are excluded from the province of Logic. One class of this kind is that of Comviands ; another, that of Questions. We can found no argument on either. The Command must be either obeyed or disobeyed, and the Question must be answered in some way before the first datum in any inference can be recognised. § 282. The logicians, then, ignore Commands, and that rightly. They also ignore Questions, but not so rightly. This is because in such sentences as ' What is this ? ' there are two terms ; viz., ' what ' and ' this ; ' and between the two the result is, ' this is something concerning which in/or- 124 MOOD. mation is required.'' But this is incommensurably between two propositions, and one is what we have ah-eady seen in the case of the Finite Verb and the Infinitive ; nor need it detaia us now. What we more especially need notice is the extent to which the element of uncertainty and conditionality is common to the Verbs of the Conjunctive Mood and the class of Questions. § 283. Hence, between what the logician admits, and what he excludes, we get something like a rule, by which we may ascertain the number and nature of the Moods ; for the triple division is, undoubtedly, meant to be exhaustive of the different forms under which a Subject and a Predicate in combination can present themselves. Hence, also, the more we bring these divisions of the logician into harmony with those of the grammarian, the more do we give precision and definitude to our conception of the function and number of the Moods. In respect to the last point, we may safely say that, for the Finite Verb, they are not more than four — not more ; possibly not so many. § 284. Foiir in Greek, and Three in Latin, are the num- bers with which we are the most familiar. The names are ( 1 ), Indicative ; (2), Imperative ; (3), Conjunctive, for both languages; (4), Optative, peculiar to the Greek. The In- dicative is the one which the logician adopts. The Impera- tive he has recognised ; but only to exclude it. With the Conjunctive and Optative the case is different. These are, by no means, so evidently the counterparts to the class of Questions as the Imperative is the counterpart to the class of Commands. § 285. As a Mood (for nothing need be said about the In- dicative), the Imperative is, on the first view, a very simple one ; and if its character depended on the purely imperative part of it, there would be little to be said about it. The ob- ject addressed must, of necessity, be in the Second Person, and his name may be dispensed with ; for its presence and its relation to the speaker are self-evident. Hence, we may VEUB. 125 have such words as die, and fac, and others, where the verb presents itself in its simplest form. As more individuals than one may receive a Command, the Imperative may have more Xuvihers than one; but as a Tense, it is essentially Future. The time when a command is delivei-ed is Present ; that of its contemplated fulfilment, Future. An Imperative Mood in a Past Tense is a contradiction in terms. § 286. These are the characters of the Infinitive which give us as formulas. But if a formula is one tiling, the practice of language is another. Eamus — let us go, and ea.nt = let them go, are in form Conjunctive. But it is in sense wholly unconditional ; so that by the double test of form and import it is neither Imperative nor Conjunctive. But its sense is Precative, and as such, implies permission, which is akin to an order. Again, in Moriamur at in media arma ruamus, the sense is Hortative. "With a secondary verb, like let or may, we may find these differences expressed. But then the combinations are analytic. § 287. Wishes are of the same class with orders, suppli- cations, and exhortations ; but in wishing, a man need not go beyond himself. When he says, ' / wish you well,^ he does not so much give vent to the wish, as he tells us that he wishes, or that he is a wisher. The pui-ely Optative ex- pression partakes of the nature of an Inteijection. Between the Interjection and the Verb, as Parts of Speech, there is a wide interval ; but in combinations like utinam, elO' tjipeXe, ' would I could,' we have, between the elliptical forms of the expression and the suggestive, rather than indicative, cha- racter of the sentence, something that is truly Interjectional. In English, when we use ' may,' we get the difference be- tween the Conjunctive and Optative sense by a transposition — ' you may be hajypy ' — ' 7aay you be happy.' And so it is in the Latin sis — Sis licet felix, uhicuRque mavis, Et memor nostri, Galatea, vivas. 126 MOOD. In the Greek — 'i2 Trai, yivoio jrarphs evTvxea-repns, T«S dXA' ofJioios Kai yiVoC av nv kukos, the same woi-d is both Optative and Future ; in the latter case with av superadded. § 288. Then, m the way of Tense, we have such abnormal forms as the Imperatives of the Greek, and extravagant forms as the Imperatives of the Greek Aorist. In the Active Voice, the form may as easily be considered Future as Aorist : Tv\poy, ^paffnv, ■n-air^unv, &c. ; but in the Passive there is no doubt as to the tense of rvijiOijTi, -vfOiimr, TvtpOj]- Toj', TV(pd}iTU)i', Tv(j)di]rE, TV(i>6tiTojff(n'. This is telling a man on Saturday to take a beating on the 2}i'eceding Friday. We cannot do this ; though, in saying, ' consider yourself' beaten,' we can get at something like it ; and that loose likeness of the kind may i)ass for real ones, is a fact that we must recognise in language, and take it as it presents itself. There is a great deal of it, and that in several languages. Thus, in the Finpai-adigm of § 92, we may notice the entry of a ' Factive ' case ; the ordinaiy explanation of which term is, not so much that something is actually made, converted, or trans- formed into something else, as that it is made, or becoming like, or as something else. In languages where there either is or has been the recognition of the system of ^aiclusive and /^elusive Plurals, the paradoxes connected with the Person of the Imperative can scarcely be said to exist; for it is often hard to say how far the designation of the object addressed is that of any particular person at all. It certainly applies, in the first instance, to a second person ; but if, when either ex- cludes or includes the speaker and the objects spoken about, its import becomes equivocal. That the ideas, notions, or conceptions, connected with that of Command are numerous, we have already seen. They have, also, a tendency to gi-adu- ate and run into one another. § 289. The Indicative and Impei\itive Moods ai'e commen- surable. In each there is but one term, and only one proposi- VERB. 127 tion. Neither of them, however, is commensurable with the Conjunctive; because in the Conjunctive there are tico propositions. The cUtFerence here is the difference between the Finite Verb and the Infinitive repeated ; except that, in the question between the Finite and Infinitive it was one of terms, whereas, here it is one oi propositions. § 290. There are no two propositions which may not stand in juxtaposition to one another. They need not be connected ; inasmuch as, to use a familiar simile, they may touch one another like mai'bles in a bag ; nor even, when connected, need they have any sign of the connection. But, given some such sign, the possible series of them must be co- extensive with the possible ways in which connection of any kind can be expressed. It is not impossible that the number and natui'e of these may be calculated a priori ; and that each of them may have a class-name ; even though the class be limited to a single conjunction. This is what is possible. Actually, however, we have a limited series of classes, each with its distinctive term, which is amply adequate to illustrate the meaning of the term Conjunctive. They fall into two classes : those which exercise no influence upon the Verb in the secondai-y pi-oposi- tion ; and those which influence, or modify, it. All of these, however, except the first two, have some effect or other upon it ; in respect to either its form, or the interpretation of the connecting link. It is this influence which makes the dif- ference between two propositions in mere juxtaposition (the marliles in the bair) and propositions with a sign of con- nection, or Conjunctives. There is, however, only one form of the conjunction that affects the Mood, exclusively. There is another that affects either the Mood or the Tense ; but which of the two be the one that is so affected depends upon the language luther than the power of the conjunction taken by itself. This, of course, anticipates the contrast between the Greek and the Latin ; in which, between the two, we find a Tense in one language where a Mood presents itself in 128 COXJUXCTIVE. the other ; a point of no small importance in showing how unstable is the character of the Mood, and how readily it is converted into something else. § 291. The principal conjunctions, each with its own recognised class-name, are as follows :— («) Causal. The day is hue ; because The sun is bright. ih) iLLATrVE. , The sun is bright ; therefore The day is fine. (^) OOPULATIVE, The sun rises in the east ; and The suu sets in the west ; or The sun rises in the east and sets in the west. Here the single word sun serves as the Subject iov two Propositions : The sun shines ; and The moou shines ; or The suu and moon shine. Here the vfov^L shine serves as the Predicate for two Proposi- tions. VEEB. 129 In like manner : (d) Disjin\CTrvE. John will do it ; or Thomas -will do it. This may take the more compendious form : John, or Thomas, will do it. (e) Adtersatite. All fled but Hector. Here the import of but is equivocal. This is because we can imderstand the Proposition either as — All fled except Hector ; In which case but is a Preposition ; Or, All fled but Hector (did not fly) ; where but is a Conjunction. (/) COMPABATIVE. You are fairer than she. This we must parse as if it were written : You are fairer than she (is fair) ; or You are fairer than her. Here than is a Preposition, and governs her in the Ob- jective Case. Thou art a girl as much fairer than her As he is a poet suhlimer than 7ne. K 130 MOOD. Here we may write she instead of her, and / instead of me ; but the lines ai-e by Prior, and this is how he wrote them. Wliich gives the better OTammar of the two is another question. Both are current in the average English. It is clear that, with the exception of the first two, all these conjunctions affect either the form or the parsing of the second proposition, supposing that they are two, or, as in the case of the Adversatives and Comparatives, raise the question as to whether there are two propositions, or one. On the other hand, it is equally clear that none of them affect either the Moods or the Tense of the Verb. § 292. With two other conjunctions, however, the case is different — these being ' if^ and ' that.^ The first of these affects the Mood, the second the Tense, in English. In Greek and Latin it is different ; as will be seen in the sequel. At present, it is enough to say that the difference is an important one. For this a single illustration of each is sufficient. (a) If, and If — not. Here it is considered better to say: If he do this he will repent ; than If he does this he will repent. "Whether every one who speaks English does, in ordinary conversation, so speak, is another question. Such, however, is the rule ; and the older the stage of the language the more valid it is ; but, as language advances, it becomes obsolescent. Still the difference is, fundamentally, a real one. We may call the import of this ' if Conditional. [b) That and That — not, or Lest. Here we say : I do this that I mni/ succeed. rather than also rather than I do this that I iui(/ht succeed ; I did this that I might succeed, I did this that I mat/ succeed. THE CONJUNCTIVE. 131 We may call tliis ' that ' (or lest) Intentional, or, perhaps, more conveniently, Potential. § 293. In each of these instances we get a new structure. In the fii-st we get what we may fairly term a Mood with its special sign. It is a negative one in the present EngHsh ; nevertheless, between the two there is a dilferentiation. So there is in the second ; but it is a differentiation of Tense rather than of Mood; and its element is a Concord— a Concord of Tense. But it is one of Jlood as well. If the two propositions did not depend upon one another, and were merely unconnected sentences, there would be no call for a concord. Both, however, are Conjunctive ; inasmuch as they involve not only propositions, but imply a link between them. In this way, however, all the other conjunctive pairs of propositions were in the same predicament. With ' if ' and ' that ' we have something more ; viz. a special structure for the subordinate one. § 294. ' Conjunctive,' then, and ' Mood,'' are passable names ; but neither is imexceptionable. There are Conjunc- tive sentences in which the verb undergoes no change ; and in the case of ' that ' the Concord was one of Tense. Still, the terms may pass ; and now we may take the two foregoing cases sepai-ately. § 295. Tlie construction that depends upon ' if,' and words of a similar import, is the one that we must first in- vestigate. This is because it is pre-eminently a construction in the way of Mood ; for the other, or the one indicated by ' that,' (fee, is partly one of Mood, partly one of Tense. Words like '■if- are conditional. The propositions which they give us in Logic are called hypothetical ; but both terms point to the same relation between the two sentences or clauses. Something will he, take place, or he done, pro\aded that something be done, or take place in reference to, and iu connection with, it ; in other words, one of the two proposi- tions depends upon the other ; and the conditional Conjuno- K 2 132 MOOD. tion connects them. In the mind of the Speaker there is a manifest diffei'ence between the two propositions in respect to the certainty or reality of the action oi- state indicated by them. It need not be absolutely certain in both, inasmuch as we can say ^ If I do this I may succeed,' as readily as we can say, ' If I do this I shall succeed.' But one of them must be more or less wncertaiu. This difference is a difference per se ; or, at any i^ate, it is not one of Person, Voice, Num- ber, or Case. Nor is it one of Frequentativeness, Inchoa- tiveness, or anything indicated by the Formatives. There is a Negative suggested by it ; but to express this in extenso would be to do away with the element of uncertainty upon which the conditional character of the clause or proposition depends. It does not, however, so much matter what it is, as how it suggests itself to the speaker. Nor is the exact form of the suggestion important. The real question is whether the difference between them be sufficiently striking to suggest a different form for the less certain term as opposed to that of the certain one. § 296. It is submitted, then, that it is the perception of this difference that suggests the process of differentiation — when it takes place, which it may or may not do. The con- trast, however, in the speaker's mind between the less certain and the more certain is the primum mobile in the change. It is by no means necessaiy that there should be one at all ; but if there be, this is its origin. It is rarely, if ever, of a very decided or conspicuous character, and it rarely, if ever, suggests the notion that it was once an independent word, or even a biliteral syllable. It is something more than an ordinary accent, but to the class of Accents, Tones, and Modified Vowels, it belongs, rather than to that of substan- tial and definite increments. It effects such changes as those from Tvirru) to TviTTt], from rirvfa to rirvcpuj, or from sum to sim, and not much more. There is a little more than this, in the difference in Greek between the Conjunctive and Opta- tive, and this will be noticed in the sequel. CONJUNCTIVE. 1 33 § 297. In the Agglutinate languages there is also some- tliing more, as may be seen in the Lap, § 94. But, so far as the present writer can form an opinion, these are not In- flectional. The sign, which in the Lap is consonantal, and, as such, formed by Addition rather than by Differentiation, probably belongs, like the -it in voc-it-o, to the original verb. This we must suppose to be defective in the other moods, and, in the Conjunctive, suggestive of some kind of action or state which is less positive or definite than that suggested by the Indicative. A Consuetudinal import would meet this condition ; for in sentences where the Verb implies merely that such or such an act or state is habitual, would make its connection vdth the associated proposition more or less doubtful. In other words, it would give us the elements of a cowlition. There might be a habit, but there might also be exceptions to it. § 298, This is how we get the Conjunctive Mood as governed by the Conditional Pronoun; i.e. by some word like ' if.' But ' if,' and its equivalents, are words of a very wide import, and they are this because, in all kinds of doubt or uncertainty, there are degrees. In many cases we can measure these, and hence calculate the probabilities. Some- times an ' if' gives us Httle more than a mild form of denial, and is all but negative. Sometimes it is the ' if duhitayitis ' =' as may or may not he.' Sometimes it is the ' ifconcedentis,' or the 'if'of one convinced (perhaps against his will), by what he has learned, in favour of a more affirmative view. In this case we often say ' since,' a word referring to past time, and indicating a change of opinion. When we do this, the Verb that follows is properly Indicative; i.e. the word ' if means ' since,' and the doubt is dispelled. But on the doubtful ground of admission or denial, ' if itself may be used as since. If so, it may be followed by a verb in the Indicative. 134 MOOD. The Optati\t:. § 299. In strong contrast to wGrcls Kke ' i/"/ stand words like ' tliat.'' In sentences like — I do this that I may succeed, and I did this that I might succeed, we have Tense, where before we had Mood. This leads us to the consideration of the 02)tative forms in Mood or Tense, as the case may be. There are two sorts of Optativeness ; and, between tliem, the Greek makes a special Mood of many- Tenses, whereas the Latin gives many Tenses to a single Mood. 1. The Optative element of the first is simply that of a vague wish, vaguely expressed. The wish is one in which the element of intention, or the will to do anything to forward it, is at its minimum. The accomplishment depends upon un- certain circumstances, rather than on any action of the wisher. It may, perhaps, partake of the natvire of a prayer, or, when mixed up with past time, of a regret. May something occur favourahly ! Would that something had happened, or (what is commoner) that sometldng had, not happened ! If we trans- late ' opto ' by ' -wish,' rather than by ' choose,^ the first gives a purely Optative expression. The second may be called De- siderative. In tbe first the time involved is Future ; in the second, Past — past and irremediable. In Thought this is, perhaps, as much akin to the Interjection as it is to the Verb; indeed, expressions like ' icould I could' are verbal only in form, especially when compared with such interjec- tions as lo, ecce, proh (vocative or deprecative), and the like. When phrases like ' loould, I were ' are changed for phi-ases like ' 0, that I were' the mixture of the two elements, as con- nected in thought, is conspicuously clear. § 300. In this no second proposition is involved, except so for as the word expressive of a wish may be verbal. In ' would I xoere^ the woi-d * loould ' is, doubtless, a verb ; but THE OPTATIVE. 135 it can, witli a very slight difference of meaning, be replaced by tbe Interjection. We may not, however, say that this kind of meaning excludes a second preposition. "VVe may only say that it can dispense with one. § 301. The more we look into the import of the Optative, the more clearly do we see that it is not to the Conjunctive Mood that it is more especially akin. It is far more so to the Precative variety of the Imperative ; and the more it re- sembles this the more it partakes of the natiu^e of an Interjec- tion. We, doubtless, have it in Conjunctive form, as in — Sis licet felix, ubicunque mavis, Et memor nostri, Galatea, vivas, »&c. ; and Serus in ccelum redeas, diuque Laitus intersis popido Qnu'ini ; and in innumerable other cases. But the Imperative has a similar connection. On the other hand, both in Greek and Latin, we have particles of an Inteijectional chai-acter in such elliptical expressions as — ' Ei^' a>(f)eX' 'Apyoiis fif] bioKTuaQai (rKd(f)os KoXi^wi' e'f aiav Kvaveas ^V[i7r\r]ya8as, &C., when the Optative sign is et, whereas uxpeXe represents some- thing that the ship Argo ought not to have done. In— O utinam subitse rapei-ent mea poma procellse, Vel possem fructus excutere ipsa meos, the force of utinam is Present, while that of raperent is Past ; a construction impossible if there were any element of inten- tion in the exclamation. The past import, then, lies less in the wish that the winds would do something that they will not do on the sti-ength of the prayer, than in the feeling of regret that both, for the time subsequent to it, and the time anterior to it, the winds did not do such things at all. The wish, by implication, had been made before, and that in vain. 136 MOOD. § 302. We may now look to certain Latin forms. The forms like amarem, amavissem, amaverim, are Tenses of the Conjunctive Mood. In Greek, forms like rvwroifii, rerixpoiiiij Tv\poi^i, &c., are Tenses, and something more ; i.e. they are Tenses of a Mood which, in Latin, is wanting, viz. the Op- tative. Hence, to a cei-tain extent, what is Tense in Latin is Mood in Greek, and the Latin Conjunctive Tenses represent a great deal of what, in Greek, is represented by the Optative as a Mood. To some extent, too, these Tenses and Moods are complementary to one another. § 303. It matters little whether we take the Greek Ojitative, as a Mood, or the Latin combinations as Tenses, sejiarately ; or whether we take the two systems as a whole. Little, too, does it matter what we call them. In any case, the purely optative element is at a minimum. In such exclamations as — 'Q jTcu, yevoto narpos evrvxearepos, Ta S' aXK' onoios, koX ye vol' av ov kokos, it is only one of the forms that expresses a wish. The remaining forms are for the greater part other than optative — i.e. they are (with the aflSx ai') Futui-e ; with 'iva and the like intentional ; and, with the Latin avi, if not actually Interro- gative, a near appi'oach to an interrogation. § 304. This brings us back to the triple division of the Logicians ; which, by what it iwcludes and by what it ex- cludes, is held to be exhaustive of the three primary forms in which the combination of a Subject and a Predicate may be conveyed. We may call this a Proposition in reasoning, which, consideritig that Logic is merely a part of Language, specially applied, we are free to do. Or we may call a Command, or a Question, by some other name ; though I submit that this limitation of the term Pro2)osition as deter- mined by the logicians is in no way valid in Language. In either case, however, the division of the Logicians into State- ments, Commands, and Questions is sufficiently near to that THE OPTATIVE. 137 of the gi-ammarian, to limit the numbei' of his Moods ; and it must be admitted that, as a Mood, the Optative is one of a very mixed chai-acter. § 305. The bearings of this approximate relation, and partial correspondence, between the division of the Logicians and that of the Philologists now explains itself. Between what the logician admits, and what he excludes, we get three different ways in which a Subject and a Predicate in com- bination may present themselves to the speaker. Definite statements correspond with the Indicative Mood ; Manda- tive or Imperative with Commands ; questionable cases, or cases which involve a question, with Questions. What this way of presentation may be, is not so easy to say. But it is easy to say what it is not. The relation of voca or vocato is not the relation of voc-o, and voc-it-o, inasmuch as vocito forms as good a Proposition as voco ; whereas neither voca {vocato) or vocita (vocitato) forms any proposition at all. And the same is the case with the other Formatives. Viresco and yrfpacTKb) comport themselves, in a logical proposition, exactly as vireo and yepaw do. The difterence, then, between the Imperative, or Conjunctive, on the one hand, and the Indi- cative on the other, is not one connected with any difference in the character of the verb itself We may call differences of the last kind Modes. We may even call them Moods ; but they are not the Moods represented by the Indicative, Imperative, and Conjunctive denominations. As little are the differences of Voice, Tense, or Number ; because all these are present in the Moods which correspond with Commands and Questions (Questionables). With Person there is an apparent difference ; but it is only partial. The Imperative is, formally, limited to the Second Person, But the Second Person is common to it and to the Indicative. It is no dif- ference, then, of this kind that constitutes a difference of Mood. Besides this, we have seen that the Imperative Mood, in Thought, has more Persons than one. Eamus, in Latia=^e< us go, which is Imperative. This is, in Thought, 138 MOOD. ' may we go,' a Precative Imperative, if we choose to call it so. In like manner, ' morarmir et in media arma ruamus ' is what we may call JIortative:=' do as I do.' We may call these Mandatives; but, whatever we call them, they are Variants of the Imperative. § 306. What, then, is the difference which the N'on- Indicate moods signify 1 It is certainly a peculiar one ; one per se. It is the difference between ' Yes ' and ' No.' But it is not the exact and absolute difference. Combinations like ' This is not good' are as truly Indicative as * This is good.' But the Negative when\lefinitely expressed is as positive as the Affirmative ; i.e. as free from any uncertainty. It is not, then, to this extent, a difference of ' Yes ' or ' Ho.' An absolute denial is only a contradiction ; and as such simply the affirmative of an opposite. In the Command, however, of the logician, and the Imperative Mood of the grammarian, we get something that is neither one nor the other, but a mixtm-e of the two ; viz. an explicit affirmative with the suggestion, or recognition (either consciously or unconsciously), of a possible, or probable, denial. And this is something wholly different from anything like any other Formative or Inflection. It is, also, one of a more svxbtle and less definite character than any of them. It is also one which is not very necessary to language. It is long before it is recognised by any sign ; so that language does, for a long time, without it. It is also, early in its disappearance ; so that language, hereafter, is likely to dispense with it altogether. Its sign, like that of Gender, with which it agrees in the last respect, is either differential or negative — at any rate, it is indefinite and indistinct. It coincides but loosely and iaidifferently with its distribution ; for we have ' seen Imperatives replaced by Conjunctives, and Optative Moods by Conjunctive Tenses, in the Synthetic languages; while, in the Analytic we have such purely Imperative combinations as ' let us go ' superseding forms like the Latin 'eamus.' Finally, we have no separate syllable, no con- OPTATIVE. 139 sonantal soiuid, nothing in short beyond the modification of a vowel to which we can assign its iiaflection. The -m in ame-m as appended to avi-o is no instance to the contrary ; inasmuch as the most that we can get from it is the fact, that the original sign of the Person was retained ia one Mood after it had become obsolete in anotlier. § 307. We can go farther in this dii-ection. We not only get no definite sign, or additament aliunde, as we do in the signs of Person, Yoice, Number, Case, and Tense, but we cannot very readily find a probable one ; in other words, we not only fail to discover an additamental sign, but fail to see our way to the making of one. The two quarters in which we are most iaclined to look for it, are the Particles that indicate either a denial or an interrogation — the Nega- tives and the Interrogatives. But this is just where we fail to find them. Of the Particles themselves there are plenty ; but even though they may partially coalesce with their verbs, they never help us to the origin of the difierences that separate forms like tvt^-t-f, tv--tw, and -vTrroifxi from tvttt-w. The likeliest word to enter into an inflection is the Greek el ; but we have seen how it comports itself. In 'EXQ^ &ei'addecl to tlie theme or root in a definite and sepai'able form. It is not suggested that this rule is ia- variable ; it is only stated that such is usually the case. Except in the instances of Gender and INIood, the general sign is made by an addition of something ah extra. If, in Gender and Mood, there is no such material to apply, it is useless to try to reduce the signs of them to any originally independent and separate word. But their forms can scarcely be said to stand alone ; even though they be ever so alien to the class of Persons, Number, Cases, and Tenses. There is a system towards which it (so to say) the formation of Mood sravitates. The difference between an Indicative and a Conjunctive is not exactly one of Accent, or Tone, or Emphasis. But it belongs to the same order of mutations, neither is it, in Thought, independent of the rules of Colloca- tion. In the present English, we may get three different im]3orts from the same combination, word for word : (1) ' You may he hap2^y ' = You may, {or not) be, hajypy — may, possibly or prohahly, he happy. (2) ' You may ' kc.-=You are free to he happy, or ' / let you he happy ^ (3) ^ May you he happy V — This conveys a question — Are you free, or ahle, to he happy 1 (4) ' May you he happy ! This expresses a wish — / wish you may &c. Here we have four Moods ; in the 5i(/?i-ification of which the only two elements are Accent and Collocation ; and in Collocation we get a new element, and one of a different kind. § 309. Of the two, or three, iVoji-indicative Moods (for this is the safest general term for them) four characteristics have been notified : (1) The differential and (as such) indistinct nature of their signs. (2) The lateness of their appearance, and the eai-liness of their disappearance, in language. (3) The difficulty of seeing our way to any elements out OPTATI\^. 141 of which any sign in the way of an additament can be made. (4) The comparative closeness of their afiinity with the system of Tones and Accents (though they are not exactly either one or the other), as opposed to the moie distant con- nection with the system that gives ns the signs of Person, Voice, Number, Case, and Tense, as to hifiections, or that of the Formatives of any kind. The first two of these characteristics are, probably, re- lated to one another as cause and effect ; because what is faintly sign-i&ed may easily be neglected. Of the third it may be said, that we have, at the present time, a conspicuous series of equivalents in such familiar verbs as may, can, might, would, should, ought, and (to some extent) will and shall. It is by these that, in the Analytic stage, they are superseded. That such is the case is a very good reason for considering that, in Thought, the two systems have the same import. On the other hand, the very fact of the system of Analytic combinations of Auxiliaries being later than that of synthetic Inflections (and originating in its decay) is conclusive against its being the origin of it. § 310. The system that gives us the three iVow-indicative Moods is a system per se. And it is not this by accident. Its typical illustration presents itself in the Conjimctive as governed by ' if.' Its basis, in Thought (we may almost say Feeling) is the contrast between a statement conveying a certainty (or what passes for one) and a statement conveying an ^tncertainty ; the two being connected with one another — if not exactly as Cause and Effect — as a result and a condi- tion. We know how important little like ' if are in matters of this kind. We know that when one word assumes a ' Yes,' the other suggests a ' iV-o.' We know that, though in some cases, the contrast may fail to impress itself on the mind of a speaker, there are others wherein it makes itself painfully apparent. We know this from our own experience ; and we know, from the history of language, that there is a special 142 MOOD. way of indicating it. And now we may remind ourselves that it is only the Moods under notice, only in those equi- vocal sentences which involve an uncertainty, and recognise in its expression, that we have anything of the same kind — anything that has its foundation in the intention, hope, fear, wish, or curiosity of the speaker who utters them. The nearest approach to it is in the Future ; and it is with the Future and the XJonsuetudinal Present that the Con- junctive is most closely allied. In the Conditional Conjunctive we have this influence of imcei-tainty in its typical form, and the conjunction governs the Verb ; i.e. determines a change in its form. In this change it has the difference of Mood. In the use of words hke ' that,' words implying an inten- tion, it is not so much a case of Government as one of Con- cord. In the Greek it is one of Tense and Mood ; ia the Latin it is one of Tense and Tense. In Stockfleth's Lap Grammar it seems to he the same. At least, he writes — 1. Modus Indicntivus (experience). 2. Modus Conjunctivus (feeluig). This falls into two divisions : (a) Modus Subjunctiims (prasem conjunctivus). (h) Modus Optativus i^prceteritwn conjunctivum). 3. Modus Imperativus (will). Here, as in Latin, the Optative is the Past Tense of the Con- junctive. §311. In the Imperative no second clause is needed. Yet it is implied. The command may be obeyed or disobeyed. The same secondary sentence, in the form of an answer, is implicit in the delivery of a Question. In the simple Optative there may or may not be a second verb, the difierence between a simple wish and an intention being the fact that, in the latter, something is done with a view to a result. But in all, the secondary element is a question of ' Yes ' or ' ]\^o,' or the OPTATIVE. 143 equivocal * Perhaps,' in which the original uncertainty pre- sents itself. § 312. The faint and negative nature of the inflection of the Optative 2Iood, as we find it in Greek, is by no means characteristic of the Teiises by which it is replaced in Latia. In— am-arem mn-ares am-ai'et am-aremus arn-aretis am-arent am-avissem am-avisses am-avisset am-avissemus am-avissetis am-avissent am-avero am-averis a7n-averit am-averimus am-aceritis am-averint we have an inflection of a more than ordinary character ; i.e. one wherein the additaments are of more than the average length. § 313. The structure of these forms is doubtful, inasmuch as it may be got by two processes. a. One of these gives us the doctrinal fact that the will appended simply the same as those of vell-em and noll-em, &c., volu-isse and nolu-isse, &c., and that these are the same as those of the Yerb Svilastantive, si-m and ess-em, the ques- tion as to which is oldest being beyond the range of our criticism. h. The other would make them the tenses of the Yerb Substantive, bodily attached to the tense-signs of the ordinary Yerb, in which case the system of expression must be held to have originated in the Yerb Substantive. Against this lies the difficulty of seeing how such a combination of tense-signs as voc-av- or scrips-, and -essem or ero (erim), can be made to give any sense ; in other words, how they can be parsed. It cannot be done in any instance but one. But with one it can be done, viz. with possum. If the pos- stands for potus or pate sum, we can parse the compound throughout, allowing only for the diflerence between an Ad- verb and Adjective. By identifying the pas- with the -pos in com-pos we can parse it without any abatement whatever. 144 MOOD. If so, the process of formation whicli is here intelligible may have become extended to combinations which are not so ; and how far this extension of a false analogy {catachrestic) may affect a language we know from the existence of such combi- nations as ' / have been,' and other less extreme instances of the same kind. But possum is only one word, and a single word is a very insignificant precedent for a system of cata- chresis to be established on. This is true. But, when we consider that the ' pos-' represents the auxiliar ' can,' with its special Potential import, and that the ' -sum ' is the Verb Substantive — a verb moi-e in use than any hundred words put together — this latter view becomes less improbable than it is at first sight. It is not one, however, that is pressed upon the reader very decidedly. In any case, however, the Latin equivalents to the Greek Optative Mood are Terises ; and that, not only because they are classed as such in the Latm Grammar, but on the internal evidence of their structure. §314. In §§ 176-179, attention was drawn to the difference between Inflection by Addition and Inflection by Differentiation. It is in ' Gender ' and '■Mood ' that the latter principle most especially displays itself. Both Gender and Mood have an inflection of a less defi- nite, decided, and reducible character than that of Person, Voice, Number, Case, and Tense. § 315. Both are move or less exceptional in language, and disappear early. It is now added, that in the case of Mood the signs seem to be more closely connected with the system of Tone, Accent, and similar forms of emphasization, in which there is a change of existing elements, i-ather than the introduction of new ones, than with the system of the other recognised Inflections, which, by hypothesis, are either actually or possibly reducible to words originally separate and inde- pendent. § 31G. ^071-Indicative was the general but negative term THE NON-FI>'ITE VERB. 145 which best seemed applicable to the Imperative, Conjunctive, and Optative denomiuations ; and, upon the same principle, it is convenient to call the Infinitives ^Vo?i-fimte. As such, they must be considered as Verbs, inasmuch as the Finite forms are treated as Yerbs, so that the term JVon-Gnite invests the words to which it applies with a Yerbal element. § 317. In import, however, the Infinitives are Nowns; nouns .giving the name of either a state or an action. They are nouns, then, but not substantives. But neither are the Ab- stract Nouns substantival in sense, whatever they may be in their declension. At any rate, they are not substantial. The Abstracts, however, are substantival in their declensions, which the Infinitives are not. But neither are they Verbal in this respect. Between the two, the Non- Finite Verbs are a class per se ; but a class on the borderland between Noun and Verb. That, so far as Voice and Tense, they have an undoubted and actual inflection, which is best seen in the Greek forms like -Erv(pivai, rvxperrdni, and the like, is evident. In Greek, too, they can be invested with an approximation to Case ; i.e., by considering that in combinations like to ' ' Whom say they I am ? ' These are familiar examples, the result of well-known processes, or, at least, of processes that are recognised under the well-known names — Ellipsis, Pleonr- asm, Personijication, Attraction, irpog to (7T]p.aiy'6jj.evov, and others of less importance. All these express conception, and indicate differences widely foreign to the character of the differences expressed by Person, Voice, Case, and Tense ; for in these last the accidents of the conceptions belong to the external world, and not to the world of Mind. In this respect even Gender — i.e. conventional as opposed to natiu'al Sex — is in the same gi-eat dichotomous class of Mood and its numerous congeners. Finally, we may note the connection between the Conjunctive Mood and the Relative; the consi- deration of the place that Mood takes in Language being a more complex question, and one involving a general classifi- cation of the other denominations — \\z. Person, Voice, Num- ber, Case, and Tense. (10.) Retrospect. § 328. The order in which the different denominations of Person, Voice, Number, Case, and Tense, along with Gender and Mood have been examiued, will now be repeated in a general and retrospective view of these relations in the way of classification. Person. § 329. In every Finite Verb there are two names ; either separate, or combined into a single word ; as, / love, am-o. 154 EETEOSPECT. The result is a statement, Negative or AfBrmative, as the case may be. This means that in the Finite Yerb we have a Verb and something more. It is a Verb plus its Noim ; and when we have this we have not so much a single word as a sentence. Voice. § 330. In Voice, as in Person, we have more words than one. But though there are two names, there is but one individual to which they apply ; inasmuch as the Agent and the Object are identical — / strike me [myself). Here the second pronoun by no means constitutes the whole Predicate. It is only a part of it. This is what we find in the Middle Voice ; as considered only in respect to its structure. In import it, having been purely Reflective, becomes Passive, Reciprocal, Deponent, and, occasionally. Active. Still the structural characteristic is the incorporation of the Personal Pronoun as the name of the Subject, and the Reflective Pronoun as that of the Object, with a Transitive Verb. The elements then of the Middle Voice are tliree in number ; so that combinations like amatur and rinrrETai are even more compound, or composite, than words like amat and rinrrei. § 331. The structure of the true Passive is, also, one that gives lis two words rather than one, the second element being the Verb Substantive — as e-Tvfd-rjv. The -Tjr here is Finite ; and it is from this that the sign of Person is derived, Number. § 332. Whatever may be the minor differences between Person and Voice, as denominations, they belong to the same Class, Order, Section, or whatever we choose to call it. § 333. So far as Vei-bs are Personal, or Finite, they are, in respect to Number, in the same class as their Pronoun; so that words like amamus, etc., give us ttvo words. The JVow-finite portion of it, or the Infinitive element — except in u few possible exceptions — has no PEESOX AND VOICE IX XUMBER. 155 signs of Number. Yet in every Finite Terb there are two elements : (1) tlie Number of Agents, or Subjects; and (2) the Number of the Actions, or States. Of these, as a rule, the Finite Yerb gives us the first only. It tells us that more captains than one fought against Thebes, but it does not tell us how often each captain fought against either Thebes or any other place or person. Logic, in questions of Language, is a good servant but a bad master ; and some of its obsolete terms are useful. Of the Quotity, then, of the agents, it tells us in the Plural Number that they were more than one. Of the Quotity of then- fightings it tells us nothing. It could not well do so ; though in forms like voc-it-amus, ^Xiyojjiev, and (possibly yeypacpafxev), it gives us something like it. Agaiu ; the Infinite, or Non-Finite, Yerb has no signs of Plurality. But it might have them, as has been indicated elsewhere. It is not difiicult to see our way to this. It is not in the nature of language to express too much in a single word and at once. The recognised Plural, then, of the Yerb is the Plural of the Person ; that is, the Plural of a term which is other than Yerbal. The Plm-al of Infinitives exists in reality, and if need be could be expressed by a special sign. But the Yerbal Abstract, in such words as ' erring J or ' error ^ is sufficiently akin to the Infinitive in Thought to make a Plural for the Infinitive superfiuous. § 334. Having thus put the Plural of the Finite Yerb in its proper place, and having accounted for its absence in the Infinitive, we get the question of Number in its proper form. This means the Number of Agents, ii-respective of their States or Actions ; but not the number of States or Actions, irrespective of the Agent, or the object in such or such a state. In this we get but one name, or term ; i.e. a single word rather than either a sentence or a proposition. It is this exclusion of any second noun which separates Number, as a denomination, from Person and Yoice. We need not be very careful to measure the exact difference between the 156 EETROSPECT. groups, or to ask whether it be one of Class or Order. It is certainly not one of Section or Subsection. It is one of a broader — a much broader — description. But just as it is removed from Person and Voice, it approaches Case and Tense. We have now done with the Inflections that convey, in a single word, both a Subject and a Predicate, or those which give us sentences rather than single words. Between the ideas of Rejyetition, expressed by Eeduplica- tion ; of Collectiveness, or Number with Totality ; and Flurity, or the concej^tion of more than one irrespective of the extent to which they constitute a whole, we see our way clearly to the development of the signs of Number. Case and Tense. § 335. Case and Tense are more closely connected with one another than either of them is with Number. Never- theless, as opposed to Person and Voice, they are allied. Person and Voice belong to the division in which there is either a Subject or a Predicate, or both. The undoubted Noun has Case, but not Tense. The Finite Verb has Tense, but not Case. The Infinitive Verb has both Tense and Voice. But its Cases, as has been shown in § 1G6, are of an equi- vocal, indirect, or constructive, character. In respect to their connection and relation with one another, Case and Tense are analogous. What Case refers to in Place, Tense refers to in Time ; though, with Time, Tense is more exclusively connected than Case is with Place. As inflections. Case and Tense are the most important and typical of the class ; and in the way of criticism there is more to be said about them than all the others put together. § 336. Number, though diflerent in many respects from Case and Tense, has been shown to be more akm to them than it is to Person and Voice. We cannot exactly measure the value of the classificational divisions thus suggested. We need only insist on the wideuess of the diflcrence involved in CASE, TENSE. GENDEE, MOOD. 157 the primary Classes, Orders, or whatever we choose to call them. Between combinations which, in a single word, give us loth a Subject and a Predicate, and those that, in a single word, give us but one of the two, the distinction is a very broad one. It is now submitted that just as these two higher classes stand apart fi-om one another, so do Gender and Mood, subordinate members of a third Class or Order, stand apart from both. § 337. It is, also, submitted that, so fai- as the changes of form in individual words are concerned, these three classes exhaust the different ways in which Inflection presents itself ; whether its character be determined by the structure of a word or by its import ; out of the elements by which it is made up, or through the idea upon which it is founded. If so, all the forms that belong to neither of the preceding classes belong to this. And such seems to be the case ; the links by which they are connected being the internal, mental, or subjective character of the conception that determines the expression. In the other division, i.e. in those where it is founded upon the name of the Agent or the Object in a state or action, this is often an individual of the most material sort, while, even when it it abstract, it is founded on something pre-eminently cog- nisable by the senses. In Number, names of this kind are implied, and the only question is as to the nature of them. In Case and Tense we have, in each, the notion of direction, and that in its three most general forms : to, from, at ; hither, hence, here ; thither, thence, there ; this, that, yon, and the like \ and in time, noxo, then, hereafter ; at the present time, in past tim.e, in time to come. These, especially the latter, which deal with the States or Actions and (as such) Abstractions, though not material, or substantial, belong to, and originate in, the external world ; and they are what they are because they do so. In Mood the question is not so much what the object of any conception actually is, but whether it is or is not ; whether it exist at all, and, if it do, under what conditions. 158 . TENSE. (11.) Tense. The Greek Aorist. § 338. We may now notice one or two instances of a more special kind than those which have already presented hemselves. These, as a rule, have been introduced for the sake of illustrating some general principle. The present criticism ls submitted to the reader for a different reason. It applies to a particular Tense, and to a very important one ; one, too, in a very important language. In the mind of the present writer, the current doctrine concerning it is founded on a misconception. Up to about 1830 this doctrine was genei-ally admitted ; mainly, or perhaps exclusively, because the reasons against it were unknown. Since that time they have been recognised ; and, by no lower an authority than Bopp, have been authoritatively rejected. That the question to which the criticism applies is a wide one may be seen on the surface. The special Tense in question ■■■ is the Greek Aorist (Aoristus primus). What it is generally held to be is weU known. What the present writer believes it to be is as follows. The doctrine he defends is laid down somewhat dogmatically ; but this is because the question is easily presented imder its different divisions, and because it is thus that the issue is most clearly stated. Hence, the true sign of the Perfect is the reduplication ; with or without a change of vowel. The true signs of the Aorist are the augment and the affix -/v«. The true sign of the Future is -a- ; and between the Future and the Aorist there is no connection ; though 0- is common to the two. The -a- of the Aorist is the ~K- of words like £6r]Ka and fCioKu assih\lated ; in the first in- stance through the influence of the small vowel (eypaxbe) ; from wliich it affects the other persons by extension. That WrjKa and icwKu ai'e Aorists which do not assume the characteilstic of the Future is true ; but it is by no means true that the -a-- in the other and ordinary Aorists is the -a- of Tv^u) and yijuilw. On the contrary, it is the -a- of the THE GREEK AOEIST. 159 Aorist itself, developed out of the -ic- by Assibilation. The -o-- of the Future is, probably, the -n- of the Aoristus ^olicus, Tv^iEui ; but this is no true Aorist ; the -(T- being the -rr- of the inchoative, or desiderative Verbs in -iteiw ; in which, signifying something desired, or something about to be begun, it involves a Future element. § 339. The -v-, in forms Like cf/v-Xjjca, Ulopica and others is not the true, original, and normal sign of the Perfect ; but the -K- of the Aorist with the reduplication of the Perfect, and generally (but by no means always) its import. It has extended itself from the Aorist to the Perfect ; and there is reason for its having done so. § 340. This must, in the first instance, be investigated ■with a view to the Verbs of which the fundamental form ends in a vowel — ( or v (not contracted), and a, e, and o (contracted). All these are in the same predicament. These have no consonant wherewith to close the syllable to which the sign of Person is affixed ; and, as this sign is vocalic, two vowels would be broiTght into contact with one another. It is suggested that, in this, there is an element of instability, and that words Kke ri-Ti-a from tiu), would be likely to change their own termination for a stronger one. To a change of this kind the -k- of the Aorist woidd readily lend itself. § 341. With the Contract verbs the same condition presents itself ; with the exception, that the vowels a, e, and o have less of a radical character than the t in r/w. They come, however, at the end of the word to wliich the personal affix is to be attached. § 342. With the Barytone Verbs ending in X, p, r, or p (a liquid) the contact is, at the fii'st view, identical with that of the -a in. ri-TVcp-a, yi-ypcKp-n, 7re-7rrjy-a, and iri-Troid-a. But this is not the case. Liquids, more especially X and y, have in most languages a tendency to take after them a semivowel ; in which case they stand much in the same position as the true vowels. Except in these two divisions the rule that the sicTi of the Perfect (with or without a change of vowel) is 160 TENSE. nearly absolute ; the chief exceptions being the preterites of the comparatively few words which end in r, o, or 6, i.e. the dentals. The primum mobile for this hypothesis lies in the Slavonic ; which gives us two points of support. 1. In the Slavonic the sign of the Past Tense is h (k), just as it is in £di]Ka, and Uwku. And this Bopp allows to be the a in eypa\pa. Omit the Augment and the sign of the First Person, and the two forms are identical ; Greek, -dj]K- and -hwK-; Slavonic, hih=fui, tshitah=I read. The Persons run thus : {a) I icas, i^c. 1. bih bismo 2. bi biste 3. be bisheh (i3) I teas, Si'C. 1. biyali biyasmo 2. biyashe biyaste 3. biyashe biyas/m (a) I read, i^c. 1. tsbitah . tsbitasmo 2. tshita t-^hitaste 3. tshita tshitashe (/3) I read, *e. 1. tphitab tsbitasmo 2. tsbitasbe tshitaste 8. tsbitasbe tshita^u It is submitted then that it is from the -//-{/.•) that the -s- has been deduced ; and that, in accordance with one of the most general rules of Philology ; viz. that, when the palatals g, k, change their sound, their tendency is to become sibilants. The change is one of the commonest in language, and it almost invariably takes the same direction. The Pala.tal becomes Sibilant; the Sibilant rarely, if ever, becomes Palatal. Indeed, the process is known as that of the Assibilation of the Palatals. THE GREEK AORIST. 161 In this Assibilation, then, lies the h^'pothesis as to the origin of the -a- oi'iypa^{Tra)a and the Aorists in general ; and thus far it looks as if the authoi'itative sanction of Bopp recom- mends it. But such is not the case. He stops at the identifica- tion of the Greek -0-- with the Slavonic -Z;; but with the belief that, of the two, -n- is the older form. Yet no one knew better than Bopp how decidedly tb e change between the Guttu rals and the Palatals ran exclusively in one way ; and how thoroughly his explanation was exceptional. He seems, however, to have ignored the objection. Few, too, know better than the same illustrious scholar the difference between the antiquity of the structure of language and the antiquity of its early records. But this he seems also to have ignored, or at least to have made the Sanskrit an exception. Be this as it may, be certainly takes no notice of the change from ^ to s iinder the influence of the small vowel that succeeded. If this view be considered correct, it explains something more than the mere structure of the Aorist ; for it enables us to account for something that is, otherwise, almost imac- countable, viz. the suggested affinities between the Aorist and the Future. The -n- is the characteristic of each ; and, the artificial rule which deduces the Aorist from the Future * mutando w in a et prseponendum augmentum,' is so con- venient that it almost looks natural. But Tense for Tense, the Aorist and the Future, except on the assumption that ' Extremes raeet^ are antijiodes to one another ; one a Tense of Past, the other of Future time. The true history of the Aorist gets rid of this anomaly. Whether the -cr- of the Future be the -tr- of the Desiderative Verbs and the Aoristus -i^olicus, is a question that must be investigated on its own merits. At any rate, however, the investigation is simplified by the riddance of the Aorist connection, (12.) The Postpositive Aeticle. § 343. The Postpositive Article means the Article that not only follows its noun, but is absolutely incorporated with M 1G2 THE POSTPOSITIVE AETICLE. it ; so that the two words become one, and the Article takes the form of an inflection. In some cases this second element in the declension of nouns (for it is by no means uncommon) may be overlooked ; but in three languages, at least, of Europe they are conspicuous, § 344. Roumanian — Latin Family. — The Roumanian of Wallachia and Moldavia, or the ancient Dacia, is a derivative from the Latin. In the Latin there was no Article. In its later stages the Pronoun of the Third Person became its substitute ; i.e.; il in Italian, el in Spanish, le in French. In Roumanian it is as follows : — Latin, homo ille ; Roumanian, dm-ul. Nominative Genitive Dative Accusative Vocative Ablative Latin, par Nominative Genitive Dative Accusative Vocative Ablative 6m-ul 6m-ului oma-uhii 6m-ul oin-ul 6m-ul oameni-i oameni-lor oameni-lor oameni-i oameni-lor oameni-i •ens ille ; Roumanian, pei-int-ele. perinte-le perinte-lui periute-lui periute-le perinte ' perinte-le perinzi-i periuzi-lor perinzi-lor perinzi-i perinzi-lor periuzi-i Latin, casa ilia ; Roumanian, kas-a. Nominative Genitive Dative Accusative Vocative Ablative kas-a kas-ei ka,s-ei kas-a kas-e kas-a kase-le kase-lor kase-lor kase-le kase-lor kase-le INDEFINITE. 163 Latut, mvlier ilia ; RouMAifiAlir, muer~ea. muer-ile muer-ilor muer-ilor muer-ile muer-ilor (?) muer-ile (?) Latix, sacerdos (jpapa) ille ; RouMANlAK, pop-a. Nominative nnn-a Nominative muere-a Genitive muere-i Dative muere-i Accusative muer-a Vocative muer-a Ablative muere-a Genitive Dative Accusative Vocative Ablative pop-a pop-i pop-i pop-a poi>a pop-a popi-i popi-lor popi-lor popi-i popi-lor popi § 345, Lithuanic Family. — Here the Demonstrative Pronoun is JiS:= Latin is {ea, id). 3. lasculine. Feminine. Nominative jis ji Accusative jj J? Locative jame jei Dative jam joje Instrmnental jflme J9 Genitive jo jd3 In combination with the Adjective it runs thus- Masculine. Feminine. Nominative gerasis geroji Accusative gerg-ji^ ger^je_ Locative geramjame geroroje Dative geramjam geraijee Insti-umentd geraju geraje Genitive gerojo geroses And so on through the Dual and Plural. 11 2 164 THE POSTPOSITIVE AETICLE. § 346, In the late Slavonic there is no Article eo nomine. But there is a Demonstrative Pronoun — Vruc=ivarm. Indefinite 3fasculine. Feminine. Neuter. Nominative vruc vruca vruce Genitive vruca vruce vruca Dative vniku (-(5em, -mu) vrucoi vrucu Accusative vruca (vruc) vrucu vruce Locative vrucu (-cem, -mu) vrucoi vrucu Sociative vrucini (iyem) Definite. vruconi vrucim (iyem) Masculine. Feminine. Neuter. Nominative vrudi vruca vru(5e Genitive vrucega Vi'uce vru(5ege Dative vrucemu vrudoi vrucemu Accusative vrudega vrucu vrude Locative vrucemu vruL^oy vrucemu Sociative vrudim vrucom ■vrucfm § 347, The Scandinavian, or Norse, Family. — In the Danish sol-en=the sun, bord-et=the table; the first form being Masculine (or Feminine), the second Neuter; the Genitive Case is sol-en-s, bord-et-s, the Plural sol-ene, bord- ene. With a slight change of the vowel, the same is the case in the Swedish, In the Icelandic, however, we get with an older form a fuller inflection ; and en becomes hin ; et, hitt. In Danish there are only two Genders. But the older Icelandic had the ordinary three — Masculine, Feminine, and Neuter. All the signs of the Definitude, Number, Case, and Gender were, totidem Uteris, the Numbers, Cases, and Genders of the Demonstrative Pronoun he, heo, hit ; save and except the omission of the Aspirate in the compound Article, along with a few minor points of euphonic accommodation. DEFINITE. NjaUTEE. Nominative -it Accusative -it Dative -nu Ablative MASCtTLESE. -ina Nominative -inn -ina Accusative -inn -ina (-na) Dative -nmn -inui (nni) Ablative -in -iunar(-unar) 165 And so on in the Plural. Tlie order of the Cases and Genders here given is Rask's ; for Rask was the philologue who of all others strove to combine, so far as was practicable, the natural order, in the way of affinity, of the Cases and Genders, with the ordinary paradigms of the Grammars written for the purposes of teaching. § 348. To show how the extreme forms of Declension may meet, the newer Norse, as compared, or contrasted, with the old, supplies a curious example. For the purpose of practical teacliing, a grammar- wiiter for the Danish might write thus : — ' The Definite Article is the same as the In- definite, except that its place is transposed. The Indefinite Article precedes, the Definite follows, its Noun,' For teach- ing pui-poses this rule is both short and true, and, as such, safe as a recipe for the formation of the two, one being known. But the coiTCspondence is a pure accident, and, in the way of representing either the agreement or the diflerence between them, as wi-ong as wrong can be. It is good as a formida. The Masculine of the Demonstrative Pronoun is han, the Neuter hit = he and it [h-it). The Masculine of the Numeral s is -en, the Neuter -ett. Between the very definite he, the, and the very indefinite one=:some one, rig, the difference, in thought, is a wide one. However, between the two forms the original distinction, small at first, becomes as language advances obliterated. 166 THE AUXILIAEY VERBS. § 349. So much for the incorporation of the Definite Article, and the extent to which it takes the guise of an Inflection. It is the languages of the Synthetico- Analytical stage which best exhibit it. The incorporation of the Inr definite Ai-ticle is commoner in those of the Assrlutuiate class. OllVi GALESE. Balla = -dog. Nominative ball a ballek Genitive ballage balleliiige Dative ballata ballekuta Accusative balla balleku Ablative ballagen ballekugen Here the ' ek ' is the numeral one just as ' un ' is in French ; and, as we expect a p7'io7-i, it is not extended to the Plural number. This, however, is only because the sign of it is a Numeral. In Thought, there is no such limitation. A collection of objects can be just as indefinite as a single one, and, analytically, it has the word ' so7ne ' to express it. In Speech, however, it is chiefly in the Singidar that its inflection occurs ; though in Coptic and some other languages we have such combinations as ou-moui=lion, Jia-moui= lions ; where, by the way, the Article, or Numeral, is pre- positive. In French this would be des lions, i.e. a combina- tion denoting indefinite plurality ; although, as an abbreviated form of dc les, an Oblique Case, it takes, in Syntax, the place and import of a Nominative. This is because, * as ' some '= ' m.ore than one,' it is plural, and because it also=' less than all,' it is partitive. (13.) The Auxiliary Yeebs. Have and Am. § 350. The remarks upon what we may call the wow- natural sense of the word have in such expressions as * I have written a letter ' (which has gone out of my possession), * / have ridden a horse ' (which is not mine), and the extreme HAVE AND AM. 167 anomaly involved in such a phrase as ' / have, been ' (when the notion oi possession is wholly out of the question), which were made in the analysis of the parsing, and general cha- racter of such and suchlike substitutes for the Greek Perfect, were only made so far as they illustrated the identity of the word ' have ' in these cases with the ordinary verb denoting possession. They showed that the word eigan (=own) and tengo (teneo) were used in the same sense. But they connected the use of the word ' have ' itself somewhat too closely, or rather, too exclusively, with the special notion of possession. A more comprehensive term, or rather series of terms, might have been used. But they would have been unfamiliar ones. The better words would have been such coinages SiS my-ness, thy-ness, his-ness, her-ness, or their-7iess ; inasmuch as these would have given a wider and more genei-al connection between the object on one side and the speaker, the person spoken to, and the object, or objects, spoken about on the other. Such coinages, however, are useful at times ; for there are instances in which they actually suggest themselves in the expression. The single phrase, in Latin, est mihi pater=haheo pair em, tells us this. And what we find in Latin we find, also, in the Gaelic : ta caraid agani, ta sgain agam. This is, syllable for syllable, ' est amicus me 4- ad,' ' est culter me + ad' ; and so on throughout the Pos- sessive Pronouns. This use of the Substantive Verb with the Personal Pronoun is carried to a great length in Gaelic ; at least, as compared with the extent to which it is carried in Latin ; indeed, iu Gaelic it seems to exclude the word ' have ' or its equivalent. § 351. In words like 'own' and 'possess,' then, it is possible that the term for ' my-ness ' and the like occiU'S only in its more special and definite sense, so that the more indefinite form represented by ' have ' may have been the older one, and an outgi-owth from the circumlocution * there is to me,' or something like it. This would be a somewhat unnecessary refinement, if it 168 THE AUXILIAEY VERBS. were not for the fact of ' have ' being so closely allied to ' a7n, ' in its history ; and that, in respect to its prevalence as an Auxiliary Verb, in connection with, or rather in contrast to, the Verb tSubstantive, This latter is essentially Intransitive. The former is essentially Ti-ansitive; and this is the dis- tinction involved in the difference between the Active and the Passive Voices ; and it is also probable, considering the functions of the two verbs in the Analytic stage, that, as Auxiliaries, they nearly represent what is represented by the two Voices. § 352. At any rate, the Irish use of the Substantive Verb with the Possessive Pronoun is eminently conspicuous. IMhaitli liorn go raibh me = J wish I were &c. Bfear Horn go raib me =1 tvish I wei-e &c. Is eigin daimh a blieit = I must be Is feidir liom a beit = I may he Ba choir damli a bheit = / should he Ni tig liom a blieit = / cannot he Caitfld me a bheit = I vmst he Is tniach liom nhach raibh me = / am sorry I am not Here nihait=diQSvce ; &/ear= advantage ; ei^i'M:=necessity ; y*eit< M'= possible ; coir=right; tigim=l come (tig=:imj)er- sonal); cai(;?(;Z:= constraint, obligation; ^?-wac/i = sorrow ; le =with ; da=to ; m and mp=m.Q and I ; &e^■^!=be ; 5ro= that; and raih, a compound of ro (=very or verily) and the Verb Substantive. Finally, nhach is a Negative. § 353. This is how the two ideas (' est mihi ' and ' habeo') are connected in Thought. How the two are intermixed in their history may be seen from the Eoumanian. Fi = &e {esse). Yo^T = heen. 1. Sint = sum sintem =sumus 2 Eshti = es sintetsi = eis Z. Er = est siut = sunt HAVE AND AM. 169 Ate = {habere) . Avtrio = habit us. 1. Am =habeo arem =habemus 2. Ai =habes avets =habesis 3. Are =habet an =habent I have been Sj-c. 1. Eu am fost noi am fost (also r/s/i) 2. Tu ai fost voi atsi fost (also au and a) 3. El au fost ei au fost Als (eale). 1. Am fi fost laudat am fi fost laudat 2. Ai fi fost laudat atsi ti fost laudat 3. At fi fost laudat ay fi fost laudat The Infinitive here is that of 'fuo ' ; of which the Passive Pai'ticiple would be 'futus,' but is, here, 'fost.' ' Slnf,' the Latin ' sicnt '=' sum.' ' Ushte ' is a variant of the Plural * estis ' ; ' sint etsi,' is ' sunt + estis.' ' Am '=' habeo,' is, as a word, as like to ' sum' or elid-l as it is to 'habeo'; whilst ' are ' is much more like the Substantive ' er ' than any pro- bable form of * habet.' In ' Eu am fost ' we have ' / have been ' ; but ' voi atsi fast' is 'vos estis fut-i '^=-' ye are been.' 'Amfifost laudat ' =' habeo [possihlj sum) fore futus laudatus' ; while 'atsi fifost laudat '=' estis fore futi laudati.' This certainly suggests that as auxiliars the two verbs are mixed up into one another. But, aiixiliar for auxiliar, this is a mixture of the signs of two Voices ; and the corollary from this is that, in Analytic combinations, the Verb Sub- stantive (am &c.) may combine with an Active Participle, and the Verb Possessive (for such habeo, or its equivalents in other languages, may conveniently be called) with a Passive one. § 354. But what if the Auxiliai-y Verb thus combined be omitted, and the Participle alone stand for a Tense (in Past Time) ? In such a case a combination like ' {el/xi) TtTvtpioQ =' I am one who has beaten,' may bear the import of {tifi^) Tf-vfi^ivoQ, and vice versd. Then let this Participle 170 THE AUXILIAEY VEEBS. have signs of Gender, and retain them after it has been used as a Finite Verb. The result is, a Finite Te^ise with Gender ; and in the Slavonic languages, and in more than one of the Agglutinate stage, this is what we actually find, and recognise ; i.e. a Verb with Gender. Such, at least, is what we find it in the Slavonic paradigms. But the interchange of the two Auxiliaries is all that we are now considei-ing. § 355. The more general way of stating this is to say that when the Vei'lj Substantive {am &c.) and the Verb Possessive (liaheo &c.) become auxiliars, they, to a very con- siderable extent, lose the power they have as ^^ow-auxiliars, or mere ordinary verbs, and become to some extent fused into a single or equivocal auxiliar with a variable import. § 356. We have seen, then, that in one language, at least, the Auxiliary Verb answering to ' have,^ as it presents itself in the ordinary grammars, is made up out of tivo verbs. Is there an}i:.hing unusual in this? Nothing. In the German Family it is made up out of three — is, he, was ; the last, I believe, being peculiarly German. In Latin it is made up out of hvo^sum and fiit. In Greek, each of these verbs has something like a complete Conjugation. But elsewhere it is defective in respect to one of them. What, however, is wanting in the one is made good by what presents itself in the other. This gives us a whole, on the principle of Defect and Comj)lement. In the Greek we get an inkling, and something more, of the diiference, in sense, too, between f(/xt and (twu) (v^w/jt) ; inasmuch as we translate the former by he, the latter by become. That the latter involves a Future sense we infer a jwiori, and we also know that in the Gei-man Family he had the sense of a Future. But in Futurity there is Contingency ; and he, rather than is, in strict grammatical English, follows, or is supposed to follow, the Conditional Conjunction ' ?/.' What the import of was, as an independent Verb, may be is not so evident. It is only certain that three Verbs may pass for one in the Conjugation of the Auxiliary Verb Substantive. We know, HAVE AND AM. 171 too, that ' is' is the one which is the most purely and espe- cially aivsiliar, because when we use it as an ordinary Verb, we identify it with ' exist.' This gives us Existe^ice as a state and Entities as beings. Of these we think simply and solely as to whether they are or are-not. What they are, or may be, we do not consider at all. In this we have the Summum Genus of things thinkable. In ' Deus est ' the Verb has a double import. As an ordinary one, it impKes the existence of Deus. As an Auxiliary it denies its non- existence. § 357, We now revert to the difference between the pure and proper Infinitive Mood and the Verbal Abstract, which is illustrated in § 320 from the double forms ' laude ' and ' laudare ' of the Roumanian. Of these, that in -re is not so much the Infinitive Verb as the Abstract Noun. If this be true, and there be no second form, the Latin must be considered to have no true Infinitive -at all ; but, instead of one, the closely alb'ed Verbal Abstract. There is nothing improbable in this ; notwithstanding the universal and reasonable use of the term ' Infinitive Mood ' in all our Latin Grammai-s. A difference of the same kind is not only con- spicuous in the Gaelic, but is recognised by the grammarians, who have long considered that, in origin, it is a Participle, though, for the purposes of ordinary gi-ammar, an Infinitive. Here, then, the substitute for the Infinitive is not the Verbal Abstract, but the Partici^ile. Active. Imperative 1)1 = 66 {thou) Irifiyiitive do {or a) blieith = tohe = at, or to being Participle ag {or a) bheith = being iarm beith = having been ar ti bheith = about to be Imperative buail = strike thou Infinitive do {or a) buladdh = to strike Pa7-tici2)le a buladh = striking iar mbualadh = having struck. le {or ar) ti bualadh = about to strike 172 EETROSPECT. Passive. Infinitive a lolieit buailte = to he struck Tartieiple buailte = struck iar mbeith buailte = having been struck le ua bualadh, or ar ti bualadh = about to be struck Upon this series of examples Nielsen writes thus : Some writers on Irish Grammar deny the existence of an in- finitive, and say that the place of it is supplied by a verbal noun ; but this is onlj' ciuibbliug about names ; the infinitive and parti- ciple imply the force of nouns in Irish, as in aU other languages. Again, in giving the paradigm of the recognised Parti- ciple, he wi'ites : These, and the like, may more properly be considered as parti- cipial phi-ases, composed of the infinitive and a preposition, than as simple participles. — Irish Grammar, Notes 28, 29, p. 148, pub- lished 1808. § 3.58, Between the Participle and the Verbal Noun, the least we get from the two is that the recognised I n fin i tive is more akin to either one or the other than it is to the normal Infinitives of the grammars. This, perhaps, may be condemned as a refinement. But it is a necessary one. We must think twice before we assume, for exactly the same conception, two different signs. If we fail to do this, as is often the case, we are generally misled by the ambiguous use of the terms Declensioii and Co7ijugation. Twice, too, must we think before we assign to a single form two meanings. When we fail to do this, we are misled by such statements as ' The Present and Imperfect Middle are the same as the Present and Imperfect Passive.' What is really the fact in the first case is that two closely allied forms may, in a certain stage of the language, come to have but one sensible import ; and, in the second, that one form under the same conditions may come to have two allied senses. They may, and do, at some time or other, come to this. But PERSON. VOICE. 173 they do not originate in it; and, in the philology which treats of growth and development, the difference or identity of origin is of more importance than the difference or identity of either forpa or meaning at any subsequent period. (14.) Retrospect. Person — Yoice — Number — Case and Tense — Gender and Mood. § 359. Person. — This, in the main, means the Person of the Finite Verb ; though it has been shown that the Noun as well as the Verb may have the sign of Person. In the first Person, especially in the Verbs in -fii, the -m of the Personal Pronoun is easily and has long been recognised. In the later forms, such as am-o-:=I love &c. its structure is, to say the least, obsciu-e. But, to a certain extent, we see that the second element is a Personal Pronoun. How either this or the Verb with which it is connected came to be what they are, is another question. What we can see clearly is the structure of the combination. We may or not see this d, priori ; but between what we get from the presumptions and from what we get from the history of the compound we see our way to the nature of its elements. All this, however, is the analysis of a sentence, and not that of a single word. § 360. Voice.— The Reflective {Middle) Voice is the combination of the name of the object with that of the Verb, sometimes with that of the Agent as a part of the compoimd, sometimes as a separate word — am-o, I love. We see our way also to this. In the Passive we may do the same ; the elements here being the Verb in its Passive form, either as a Participle or as the Passive Abstract, and the Verb Substantive. In forms like amare and aviari we get the same difference of sense by means of a difference of Case ; but, as has Ijeen suggested, these formations may belong to the declension of 174 HETROSPECT. the Passive Verbal Abstract rather than to that pu^ Infinitive Verb. § 361. Number.— (1) Of the Fer?).— This is that of the Person, or Pronoun ; not that of the Verb itself. (2) Of the Pronoun. — This is pecuKar ; viz. that of the Exclusive and Inclusive Personals ; and, until we know more about these, our speculations are unprofitable. (3) Of the Noun. — Of the nature of this, in Thought, we have an inkling <^ j}''"^'^'""^ > ^^ least, we can, d, priori, get a presumption in favour of what it is likely to be ; in other words, we know where to look for it. We may, however, fail to do this ; for it does not suggest itself quite so readily as the signs of Person. The least, however, that can be said about it is that, when we get an example of what it actually is, we can understand the process that gives it to us. When we learn that, in the eai-lier stages of Language, we find it to consist in the reduplication of either the whole or a part of the word to which it belongs, we see that this is a natural — perhaps the natural — sign of Plurity. It is this, as an element in the sign of Plurality for the Noun, as truly as the Personal Pronoun is of Personality in the Finite Verb. Still, the evidence of its being this is less patent. In the Finite Verb we have the actual Pronoun attached to tho main word ; so that there is no mistake as to what it means. It means the Speaker, the Person spoken to, or the Object spoken aboiit. The Reduplication, however, even though it means this, does not express its meaning so definitely. It means what it does only so far as Plurity is a form of Repe- tition and Repetition a sign of Plurity, On the other hand, howevsr, it has a I'emarkable prerogative. It is not an addition, or increment ah extra. It is a part of the word to which it is prefixed ; and, being this, it is permanent so long as it is recognised as a sign ; and is wholly independent of any difierence of language. It is just the same whatever the form of speech may be ; and in being this, is either wholly or almost wholly alone. NUMBER. 175 § 362. For all this, however, it is only in the earlier stages of Language that it occurs. But this is not so much because the Reduplication was a sign that could be improved, but because there were other forms of Thought to which it ■was equally applicable. "We get it in the less developed languages (not to mention other more indirect associations), as a sign of the Superlative Degree ; in the Greek as the sign of an action that is Past continued, by a (^wa^i-repetition, up to the Present, viz. in the Greek Perfect as opposed to the Aorist. Again : we get it, in the Turkish, as a sign of what passes for Mood. The simple negative gives us a Negative form. Double it, and it represents an Ivipossible — some- thing which has no very definite denomination. § 363. This is not exactly Plurity, nor is it exactly Plurality. Nevertheless, in all the cases, it is something that implies repetition. Flurity, then, in Nature is not necessainly Plurality in Grammar ; for of Plurity, as of Repe- tition, there are many forms, and a sign that suits one, and which, so doing, is an adequate one when it stands alone, is something very diiferent when it applies to more conceptions than one. And when it does this, it is likely to become obsolete. It sir/w-ifies too many ideas to be distinctive. § 364. By adding one, or more than one, to a unit we get Plurity ; and the sum of the additaments is an Aggi-e- gate, or Collection. But this is, itself, a Unity. By resolving it, however, into its pai-ts we get, by a reversal of the process of Thought, a Plurity. In the Welsh, not to mention other languages, we get certain Singular Numbers derived from the Plural ; as if the Plural (Collective) was the fundamental or radical form. We can understand this when we look either at the stars or the inside of a peasbell. We take cognisance of the Aggregate before we distinguish the Indi- vidual. It is not, then, too much to say that, for certain objects, the Plural form may not be older than, or as old as the Singular. Here, then, there is a conflict ; not, how- ever, between the actual idea of ' one or more than one,^ but in regard to the aspect in which we see it. 176 RETHOSPECT. § 365. That out of the signs of Collectiveness, the later signs of Plurity have been developed, has already been suggested ; but between the extent to which we see our way to the nature of the two inflections, there is a notable dif- ference. The reduplicate form explains itself; and, also, contains, within itself, the elements of its structure. The sign of Collectiveness, like the signs of Case and Tense, which are now coming under notice, belongs to a different class, § 366. We can see, then, our way, partially, to the origin of the signs of Number ; though not so clearly as we see our way to the structure of Person and Voice. But both Person and Voice are scarcely Inflections. They are cer- tainly not the inflections of single words. They give us not so much a single word with an additament as a sentence with two terms. In Gender and Mood we may also neglect the inflective element ; at least in the words where the diffei-ences are signified by Differentiation rather than by Addition. This is because they are noc made out of any combination of separate words ; so that thei-e is nothing to which we can reduce them ; and this reduction is one of the main objects of the present treatise. Such being the case, there is now no other outstanding division of the system of Declension and Conjugation but that of Case and Tense. § 367. Case and Tensc. — These we may take together, since they are analogous ; Case being to Place what Tense is to Time. Tense, however, is more exclusively connected with Time than Case is with Place ; a fact of no small importance, because it tells us that the analogy is only partial, and, what is more, it suggests the division of the class into, at least, a dichotomy ; viz., (rr) Cases based upon the conceptions of Space, and (5) Cases not so based. § 368. It is certain that several inflections are, rea- sonably, called Cases, in which the notion of either fixation or direction in Space is, to say the least, obscure and indirect. The Nominativ", Vocative, and Accusative in the Synthetic, the Factive, Comitative, Instrumental, and Caritive in the CASE. 177 Afjfliitinate languages, are of this kixid. Between sucli a Case as the Caritiv? and the Dative or Ahlafive we have, probably, the maximum amount of difference in the impoi't of any two inflections which come under the same Glass ; and it is not too much to say that in neither Person nor Voice, in Kumber nor Tense, any difference of the same amount can be found. In other words. Inflections bearing the name of Cases may be found which are more widely separated fi'om one another in impoi't than they are fiom Inflections belong- inc to another class, denomiuation, category, or whatever we like to call the heads under which the Pai-ts of Speech, and the elements of the system of Declension and Conjugation, ai-e arranged. Nor should this surprise us. In Person, Voice, Number, and Tense, the amount of possible inflections is limited by the nature of the idea. In Gender there is a like limitation — Male, Female, and Neuter. In the older divisions of Aniaiate and Inanimate, the range of difference was even less. Gender, no doubt, is a more difficult class to explain than Case ; but this is not because it embraces more numerous and heteregeneous conceptions, but because it deals with and creates a fictitious, conventional, and non- natural distinction. § 369. Mood, on the other hand, agi-ees with Case. It mainly deals with the diffei-ence between the Positive and the Conditional ; but when we find that this inA'olves all the possible questions of Afiirmation, Negation, and Doubt, its range of application becomes indefinite. § 370. It is not difficult to see what this leads to. Case and Mood, as categories, or something akin to them, aie, as much as anything like a category can be said to be so. Negative, and, as such, indefinite in their extent. They take — in Case more especially — besides the expression of certain conceptions more characteristically their own, such others as, either expressed oi- capable of being expressed by inflection, are not included under the heads of Peison, Voice, Number, N 178 KETROSPECT. Tense, or Gender ; and this is why they are only partially and imperfectly commensurate with them. Take away their accessory and extraneous details, and Case and Mood liave their own pi-oper sei-ies of conceptions which it is the function of the truly Casual and Modal signs to indicate. Limited to these they are like the Classes with which, both in the ordinary gi-ammars and in the present work, they are asso- ciated ; and, accordingly, Case, with its primary equivalents, ' tli. — Lectures on Light. Crown 8vo. 1.?. sewed, li. 6rf. cloth. — Lectm-es on Light delivered in America. Crown 8vo. 7s. Gd. — Lessons in Electricity. Crown 8vo. 2.?. Gd. Woodward's Geology of England and Wales. Ci-owu 8vo. 14s. Wood's Bible Animals. With 112 Vignettes. 8vo. 14i. — Homes Without Hands. 8vo. 14s. — Insects Abroad. Svo. 14s. — Insects at Home. With 700 Illustrations. Svo. 14s. — Out of Doors, or Articles on Natural History. Crown Svo. 7.v. Gd. — Sti-ange Dwellings. With 60 Woodcuts. Crown Svo. 7s. Gd. CHEMISTRY &. PHYSIOLOGY. Auei-bach's Anthracon, translated by W. Crookes, F.R.S. Svo. 12.v. Buckton's Health in the House ; Lectm-es on Elementary Physiology. Fcp. Svu. 2*. Crookes's Handbook of Dyeing and Calico Printing. Svo. 42s. Select Methods in Chemical Analysis. Crown Svo. 12s. Gd. Kingzett's Animal Chemistry. Svo. [//( t!i<' pivss. History. Products and Processes of the Alkali Trade. Svo. 12s. Miller's Elements of Chemistrj'. Theoretical and Practical. 3 vols. Svo. Part I. Chemical Physics. 1 6s. Part II. Inorganic Chemistry, 21s. Part III. Organic Chemistry, Kew Edition in the press. Watts's Dictionary of Chemistry. 7 vols, medium Svo. £10. 16s. Gd. THE FINE ARTS & ILLUSTRATED EDITIONS. Doyle's Fairyland ; Pictures fi-om the Elf-World. Folio, 15s. Jameson's_ Sacred and Legendary Art. 6 vols, square crown Svo. Legends of the Madonna. 1 vol. 21s. — — — Monastic Orders. 1 vol. 21s. — — — Saints and Mart-yi-s. 2 vols. 31s. Gd. — — — Saviour. Completed by Lady Eastlake. 2 vols. 42s. London, LONGMANS & CO. 8 General Lists of New Works. Lougman's Tliree Cathedrals Dedicated to St. Paul. Square crown 8vo. 21*. MacaiUay's Lays of Ancient Rome. With 90 Illustrations. Fop. 4to. 21.s. Macfan-en's Lectures on Hamiony. 8vo. 126'. Miuiatm-e Edition of Macaulay's Lays of Ancient Rome. Imp. 16 mo. 10s. Gd. Jloore's Irish Melodies. "With 161 Plates by D. MaclLse, R.A. Super-royal 8vo. 21*. — Lalla Rookh. Tenniel's Edition. Witli 68 Illustrations. Fcp. 4to. 21*-. Redgrave's Dictionary of Artists of the English School. 8vo. 16«. THE USEFUL ARTS, MANUFACTURES &.C. Bourne's Catechism of the Steam Engine. Fcp. 8vo. Gs. — Handbook of the Steam Engine. Pep. 8vo. 9«. — Recent Improvements in the Steam Engine. FciJ. 8vo. 6*. — Treatise on the Steam Engine. 4to. 42*. Cresy's Encyclopaedia of Civil Engineering. 8vo. 42*. Cnlley's Handbook of Practical Telegraphy. 8vo. 16*. Eastlake's Household Taste in Furniture, &c. Square crown 8vo. 14.<. Fau'ban-n's Useful Infonnation for Engineers. 3 vols, crown 8vo. 31*. 6w,or '^mediate recall. ■^^^^S^/pi ^/^2lA-50w.ii -62