IMAGE EVALUATION TEST TARGET (MT-3) i 1.0 I.I 1.25 140 .1^ <^ 73 ^J. ^^♦- >^ ^^ V <^ y HiotograiAfic Sciences Corporation 23 WKT MAIN STRICT WfBSTER,N.Y. 14SS0 (716)872-4503 # \ 5V <^ < -s^V v\ Si. CIHM/ICMH Microfiche Series. CIHIVl/ICMH Collection de microfiches. Canadian Institute for hiistorical Microreproductions / institut Canadian de microreproductions liistoriques ^ O^ * ^ Technical and Bibliographic Notas/Notas tachniquaa at bibliographiquaa T ti Tha Instituta has attamptad to obtain tha bast original copy available for filming. Features of this copy which may be bibiiographlcally unique, which may alter any of the images in the reproduction, or which may significantly change the usual method of filming, are checked below. D □ D D D Coloured covers/ Couverture de couleur I I Covers damaged/ Couverture endommag^e Covers restored and/or laminated/ Couverture reataurte et/ou peillcuMe r~n Cover title missing/ Le titre de couverture manque Coloured maps/ Cartes gtographiques en couleur Coloured ink (i.e. other than blue or black)/ Encre de cou'eur (i.e. autre que bleue ou noire) Coloured plates and/or illustrations/ Planches et/ou illustrations en couleur D Bound with other material/ RaliA avec d'autras documents Tight binding may causa shadows or distortion along interior margin/ La re llure serrde peut causer de I'ombre ou de la distortion la long de la marge int^rieure Blank leaves added during restoration may appear within the text. Whenever possible, theee have been omitted from filming/ II se peut que certainaa pagea blanchaa ajouttea lore d'una restauration apparaissant dans la texte, mais, lorsque cela itait possible, ces pages n'ont pas M f ilm^aa. Additional commenta:/ Commantairea supplAmentairas: L'Instltut a microfiimA la mailleur exemplaire qu'il iui a AtA possible de se procurer. Les details da cet exemplaire qui sont peut-Atre uniques du point de vue bibiiographlqua, qui peuvent modifier une image reprodulte, ou qui peuvent exiger une modification dans la mAthoda normala de filmaga sont indlquAs ci-dessous. D D D Q D D D n Coloured pages/ Pages de couleur Pages damaged/ Pages endommagtes Pages reatorad and/or laminated/ Pagea restauriaa et/ou pellicultea Pagee discoloured, stained or foxed/ Pagea dteolortes, tachatAes ou piqudes Pages detached/ Pagea d^tachtes Showthrough/ Transparence Quality of print varies/ Quality in^gala de I'lmpreesion Includes supplementary material/ Comprend du materiel aupplAmantaira Only edition available/ Seule Edition diaponlble Pagee wholly or partially obacured by errata slips, tissuaa, etc., have been ref limed to ensure the best possible image/ Les pagea totalement ou partiallement obscurcies par un fauillet d'errata, una pelure, etc., ont At A filniAaa it nouveau da fapon A obtanir la meiileure image poasibla. T P o fl b tl Si fl Si o This item is filmed at tha reduction ratio checked below/ Ca document est film* au taux da rMuction Indiqu^ ci-dassoua. 10X 14X 18X 22X 26X 30X y i2X 16X 20X 24X 28X 32X Th« copy fllmad h«r« has b««n r«produc«d thanks to tha ganarosity of: Douglas Library Quaan's Univarsity L'axamplaira filmi f ut raprodult grAca A la gAnirosit* da: Douglas Library Quaan's Univarsity Tha imagas appaaring hara ar^ tha baat quality posslbia considaring tha condition and lagibility of tha original copy and In kaaping with tha filming contract spaclflcationa. Laa Imagas sulvantas ont At* raproduitas avac la plus grand soin, compta tanu da la condition at da la nattatA da I'axampialra filmA, at Bn conformitA avac las conditions du contrat da fllmaga. Original copias In printad papar covara ara filmad baginning with tha front covar and anding on tha laat paga with a printad or illustratad impras- slon, or tha back covar whan approprlata. All othar original copias ara filmad baginning on tha first paga wKh a printad or Illustratad Impras- sion, and anding on tha laat paga with a printad or illuatratad Imprassion. Las axamplalras originaux dont la couvarture an papiar ast imprimAa aont filmAa an comman9ant par la pramiar plat at en tarminant soit par la darnlAra paga qu! comporta una ampreinta d'imprasslon ou d'iliustratlon, soit oar l« sacond plat, salon la cas. Tous las autras axamplalras originaux sont filmAs an commandant par la oramlAra paga qui comporta una amprainte d'imprasslon ou d'illuatration at an tarminant par la darnlAra paga qui comporta una tails amprfflnta. Tha last racordad frama on aach microfiche shall contain tha symbol ^^ (moaning "CON- TINUED"), or tha symbol V (moaning "END"), whichavar applias. Maps, platas, charts, ate, may ba filmad at different reduction ratloa. Those too large to be entirely included in one exposure are filmed beginning in the upper left hand corner, left to right and top to bottom, aa many framea as required. The following diagrama illustrate the method: Un des symboles suivants apparaltra sur la darnlAre image de chaqua microfiche, selon le ces: le symbols — ► signifie "A SUIVRE", ie symbols V signifie "FIN". Les cartes, planches, tableaux, etc., peuvent Atre filmAs A des taux de rAduction diff Arents. Lorsque le document est trop grand pour Atre reproduit en un seul clichA, 11 est filmA A partir da Tangle supArieur gauche, de gauche A droite. et de haut an bas, en prenant le nombre d'images nAcessaire. Les diagrammas suivants illustrant la mAthoda. J 1 2 3 1 2 3 4 5 6 l«wppi"P ARYO-SEMITIO SPEECH: A STUDY IN LINGUISTIC AECHAEOLOGY. BY JAMES FREDERICK M^CURDY. WARREN P. DRAPER. LONDON : TRUBNER AND CO. 1881. ^^■r^ip.R)iKv, iiinpii iHtp 1/ i^r-5oiA.'^^.^"i Entered uccording to Act of Congress, In the year 1881, by WARREN F. DRAPER, In the office of the Librarian of Congress, at Washington. 1/ "i TO MY REVERED IXSTRUCTOR AXD FRIEND, PROFESSOR WILLIAM HENRY GREEN. 11943 Wi PREFACE. The following work is substantially a reprint of articles contributed to the Bibliothcca Sacra, the last of them having npiiearecl in January of the present year. Inasmuch as the ground taken up l)y it has been regarded by many influential scientists as ju'operly closed to the practical worker, and an attempt to compare the Aryan and Semitic systems of lan- guage is often ^poken of as mere dilettanteism, it seems necessary to say a few words by way of apology. The investigotion whose results are here presented has been carried on under the conviction that the field should not be abandoned until inquiry should be proved to be a search for the undiscoverable, or, in other words, until true scientific methods should be proved to be unavailing. If it appears that hitherto the full resources of science have not been called out, investigation would seem to be not only legitimate but necessary. The following considerations may be adduced as having controlled the purpose of the work : (1) It is possible to compare the forms of the two systems better than has been done hitherto. Proto-Aryan forms, which are used on the one side of the equations, have been brought out of late yearfe with ever-increasing definiteness and accuracy. A dictionary of Proto-Aryan roots, generally reliable for comparative purposes, not only may be, but has been, constructed, and the processes which are involved in its grand results are to be commended to the study of every trained and cautious etymologist. What the principles are which should lead to equally valid results in the search for Proto-Semitic roots cannot long be a matter of doubt ; and a sound and sure Semitic morphology is certainly within reach. In Chapter IV. I have presented an outline of the (T) —w- ^l» VI rKKI'ACE. inorpliolofry of l)()lli Aryan and Soinilic roots, drawn up with- out reference to any liarnjonizinu: of tlie (wo syslenis, eillior in llieir principles of structure or in individual forms. Criti- cism from competent jud^,'es upon these attempts is earnestly desired, especially upon tlie .Semitic invest i erbs ; three of these classes are apparently secondary, 94-98. — Development of secondary Semitic roots in general ; general observation as to the use of inflective ele- ments in the formation of secondary roots, 98, 99. — Prcdetermlnative letters discussed in the order of the Hebrew alphabet, 99-102. — Indeter- minatives, 102-106. — Fostdeterminatives, 106-111. — General results of the inquiry : some of the predeterminatives originally vowels, others mere breathings, and the rest inflective formatives, 111-113. — Indeterminatives partly breathings and partly the result of inner vowel expansion, 113.— Fostdeterminatives most frequent ; any consonant might be so used ; some of them were once vowels, 113. — Many roots show no determinative letter, having three consonants from the beginning; the composite origin of multi- literals Is apparent, 113. — Forir.s with first radical the same as the last simply repeated the first radical, 114. — Roots consisting of but one con- sonant and a vowel, 114. — Semitic roots before the consonantal stage shoved as great variety in formation as the Aryan, 114. — Scheme of possible and actual root-forms in Froto-Aryan and in Froto-Semitic, 114-1 li. V. — Comparison of Rootb, 117-171 Difficulty at the outset in reconciling the different functions of the Towels in the roots of the two systems, 117. — Considerations leading towards a reconciliation, 117-120. — Roots adduced for comparison must be ideally represented by their consonants alone, 120. — Kinds of roots to be used ; exclude alleged or suspected onomatopoetic or interjectional roots ; take only those which express primitive notions, 120-121. — Value of the evidence according to these conditions, 122. — Use of etymological analogy, 122. Comparisons made of words relating to fire, 122-125 ; words for shining, 125-128 ; words for cutting or separating, 129-136 ; words for rubbing and bruising, 136-140; words for uniting, 140, 141 ; stretching or extending, 141- 147; bending or curving, 147, 148; various kinds of movement, 149-152; position, 152-154; shutting or enclosing, 154, 155 •.ff'xcrding against or fearing, 155, 156; binding together, 156, 157; pressing or crushing, 157, 158; carving or graving, 158, 159; piercing or infixing, 159, 160; wetting or pouring out, 160; being cdd, 161 ; thinking, 161, 162 ; knounng, 162-164 ; being or exist- ing, 164. — Noun-fimna of less evidential value, 164, 165. — Words for horn, 165 ; tor field, 166 ; for wine, 167. Frocominal or demonstrative roots not to be profitably treated, 167, 168. Tabular view of the comparable forms, 168, 169. — Closing remarks, 169-171. 1-- I if :H !!« RELATIOI^Q OF THB ARYAN AND SEMITIC LANGUAGES. CHAPTER I. PAST AND PRESENT TREATMENT OF THE QUESTION. The subject-matter of the Science of Language has been rescued from the confusion and uncertainty which marked its superstitious and mythical treatment in pre-scientific times. The general methods and principles of its right comparative study are well ascertained and universally acknowledged. In accordance with these principles and methods certain families or classes of speech have been clearly established ; and the work of classifying the various dialects of the world is steadily ad- vancing with the progress of exact knowledge and critical investigation. There are two main tests whereby the rela* tionship of languages, or families of languages, may be dis- covered or confirmed : the comparison of structural features, and the comparison of roots. The former criterion finds its application in the attempt to show that the languages in question have in common their leading types or modes of expression as these are revealed in their flectional and syn- tactical characteristics ; its principles are those of Comparativt Grammar, in the strict sense of the term. The latter cri- terion is employed in the endeavor to prove that the idioms compared possessed in their primitive state the same working T I !i J . i H 2 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. vocabulary, by reducing their current vocables to their rad- ical forms and primary meanings ; its principles are therefore those of Comparative Etymology. Both of these methods are legitimate in their respective spheres, for they aim, with equal deference to established laws, to reach fundamental forms of expression as a basis of comparison. The grammatical test is naturally surer than the etymological ; since forms of thought as expressed in the categories of grammar are more directly and palpably indicative of a common mental history among the speakers of language. This is so mainly for two reasons: First, grammatical features are found by experience to be more permanent and less easily transferred than verbal expressions ; and this distinction we are bound to regard as valid for the pre-historic as well as the accessible forms of any groups of languages which may come up for comparison, so that it must hold equally good for the hypothetical proto- grammatical and proto-radical periods of them all. Second, the conditions of the rise and vicissitudes of grammatical features are better understood than the conditions of the pro- duction and early fortunes of roots. Tiiere is, to be sure, a great deal that is obscure in the former sphere ; but in the latter nearly everything, as we shall see later, is a matter of dispute. For these reasons, and because such rapid and tri- umphant progress has been made in the province of compara- tive grammar as a test of linguistic relationship, it has lately become widely the fashion to uphold the exclusive validity of this criterion, and to declare that the resemblance, or even the identity (it it could be proved), of the stock of roots in different families of speech is of itself no proof of real affinity. Of course, it is admitted that where grammatical analogies prevail, etymological coincidences furnish valuable confirm- atory evidence of ultimate identity, and so far they may be regarded as accrediting relationship. But investigators are seriously warned against regarding such evidence alone as being of any value whatever in this department of the science of language. The present essay is an attempt to remove some of the odium which attache* to the theory thus impugned. ;; RELATIONS OF THE AYRAN AND SEMITIC LANGUAGES. 8 '■\ Before going further, however, It will be necessary to lay down two principles upon which tlie validity of all the sub- sequent reasoning will largely depend. First it must be understood that all comparative linguistic reasoning furnishes only probable evidence, not demonstration of the kind that m said to be mathematically certain. Tlie conviction of the <;arlier identity of forms compared may rise to the height of moral certainty, but this can only hapi)en through tiie accumu- lation of probabilities. Even in the strongest kind of proof, namely that afforded by the analogies of grammatical forms, there is a " metaphysical possibility" that accepted conclusions may be erroneous : and the invincibility of the arguments in their favor is only due to the extreme unlikelihood that early speakers, from any chance or combination of chances, or through any occult operation of consentaneous intellectual causes, should have been led to employ similar types of ex- pression for the same forms of thought, without any co-opera- tion in the production of such linguistic phenomena. The facts to be considered in making up the case are, (1) the ultimate phonetic identity of the forms compared, (2) together with the degree of resemblance in the ideas expressed by these forms, and (3) the number of cases in which such re- semblances are traceable in essential forms as compared with the extent of the whole field of investigation. Precisely the same classes of facts are to be adduced when the roots of two or more families of language come up for comparison. In both kinds of investigation we have to do with the weighing of probabilities. The evidence may differ in degree in favor of the former sphere of comparison, but it does not differ in kind. The methods of science are equally applicable to both departments, in the processes of selecting, sifting, analyzing, restoring, and re-adjusting. In both provinces the final comparison must be made only with the residuum of the last analysis, and then the decision rests upon the inherent prol}a- bilities in each case. The second position to be upheld as a necessary preliminary is that in all the processes of the investigation we must hajre RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. ^1 |i If t 'I i' regard only to the well-established facts and conclusions of science, and not to any theories and hasty assumptions that proceed from the philosophizing that is rife upon such subjects. The success of the laborer will here depend mainly upon the caution and discrimination whicli he exercises in settling the limits and the conditions of comparison, and the patience and judgment which he employs in tracing each current form and idea to be compared to their fundamental expression. His business is simply to ascertain facts ; if those facts are established, the conclusions to be drawn from them will meet with acceptance or rejection according to what may seem to each critic to be the antecedent probabilities of the case. Then only can current theories as to the necessary conditions of primitive speech be admitted into court, and the testimony thus received may pass for what it is worth. It might seem to be unnecessary to state so formally what ought to be ac- cepted as one of the common-places of all science. But the statement comes to be a necessity, when it is found that some of the most influential writers on the science of language maintain that the field of comparison is absolutely limited to those families of speech in which grammatical affinity can be shown to exist. They assert that the inflectional and syntacti- cal features of any system of languages necessarily prevailed from the very beginning, and that idioms outside of the limits just designated must have been separate from the very first, from the very peculiarities of their structural type. They maintain that all language starts with the sentence and not with the word, and that single terms are therefore not eligible for comparison. They say, moreover, that as single sounds are liable to constant change, phonetic agreement among cur- rent roots would be a sign rather of a primary difference than of identity. These and other objections to the admissibility of the comparison of roots alone as a test of relationship will be considered in the next chapter, and will be shown to be either half-truths of science or mere hasty assumptions of a pre- mature linguistic philosophy. What it concerns us now to maintain is that the field is the whole world of speech, and RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. that judgment is to be passed not upon attempts to go beyond restrictions arbitrarily laid down, but upon reaults arrived at after a strict application of the methods of science to the materials chosen for comparison. The statement of the true method of procedure in this sphere is very simple. What the investigator has to do is to make the comparison of Aryan and Semitic roots after the forms chosen for the purpose have been reduced to their sim- plest expression. That is, they must be proved to be actual roots in their respective idioms, and they must be treated as expressing the root-idea. This, however, involves a careful study of the principles of root formation and development in the two systems in their primitive individual history. That is to say, we must deal not with current roots found in the Aryan and Semitic families of speech, but with Proto- Aryan and Proto-Semitic roots ; and these must be elim- inated according to the laws which are found to prevail in their respective spheres. In the following brief review of the efforts heretofore made to harmonize the Aryan and Semi- tic languages, the theories will be judged according to the canons just laid down. That most of the theorists have failed to secure even a patient hearing from many lead- ing linguistic scientists is due in great part to the fact that they have almost vvhoUy disregarded these axiomatic principles. The whole period covered by attempts to settle the general problem before us might be properly divided at the point of time when comparative philology was established as a science. Previously to that epoch the question cannot be said in strictness to have had a history; for there is no history where there is no law of progress. But even in the later era we shall have to distinguish between those theories which have been advanced without regard to the just demands of sci- ence, and to those which show more or less deference to its methods as well as its spirit. Before the science of lan- guage was founded, even in its broadest outlines, it was im- possible that any intelligent view of the subject could •i t; 6 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. be readied. Even the very conditioiw of the investigation could not be ai)i)rehondc(i. Theories the most vaguo and unsujiportod were lield as to the relations of tho various dialeets of human speech. Previously to the close of the last century, (he comparative treatment of languages was usually only a sort of philological alchemy, in which Hebrew roots playetl the part of the philosopher's stone. Instead of regarding the several idioms of the world as developed from decayed and germinal forms, one language, accessible only in the literary and cultivated [periods of its history, was venerated as tho common source of all tho rest, and lan- guages the most diverse in structure and in typical character were believed to have been developed naturally and gradu- ally from one of the least flexible and versatile of all forma of speech. This notion was based upon tho persuasion that the oldest records of the race must have been composed in the earliest language, and that the most sacred of all tongues in its history and varied associations must have been the form of speech bestowed upon man at his creation by the gift of his Creator. Originating among the teachers of the synagogue, we know not how early, it was embraced by the Fathers of the Christian church,^ and held almost undisputed sway until the comparison of languages became a subject of sober inquiry .2 During the Middle Ages, when the rabbins engrossed the study of the sacred languages, and continued to illustrate the congenial theory of the antiquity and origi- nality of the Hebrew tongue, there was not the interest or the knowledge in the Christian church that would hav^e been necessary for its intelligent criticism. In the period between the revival of learning and the development of the science of comparative philology, there was, indeed, occasional ob- jection to this venerable doctrine ; but it was based rather 1 Gregory of Nyssa, however, surmised that the Hebrew was one of the lan- guages that arose out of the confusion at BabcL Orat. contra Eunom., xii. Quoted by Franz Deiitzsch, Jesurun, p. 48. 2 Theodoret, Philo Judaeus, and some of the rabbins regarded the Aramaic as the more ancient idiom. This, however, is only a sort of collateral theory. Theodoret supposed that the Hebrew waa a special divine revelation to Moses. RELATIONS 0^ THE ARYAN AND SEMITIC LANGUAGES. upon its general improbability, than upon definito scientific evidence. During this period, also, a modification of the old opinion grew into some favor ; according to which the He- brew was held to be, if not the source of all other languages, at least the most ancient, and the one which preserved with the least degree of change the original stock of roots, and therefore the standard with which the verbal forms of all other tongues should bo directly compared. The doctrine, in the one or Mie other of its general forms, was held very tenaciously ; and, etymology being rather an art than a science, or rather an art founded upon no science, the task of comparison and assimilation was a very simple affair. For, as the expounders of the theory could not bo refuted by an appeal to established laws of relationship between the various forms of speech, they were free to cite at pleasure mere coincidences and fanciful analogies as proofs of true affinity, and thus to vindicate the suppoeed sacred prerogatives of the Hebrew tongue ; l)eing opposed only by the smiles of an incredulous few, which they could afford to ignore, as having the support of nearly all who were interested in the subject. This dogma, so long and widely and firmly held, has now no more than a historical significance, and needs no labored or formal disproof. It is sufficient to remark that the Hebrew has no claim to consideration, in this connection, above its Semitic sisters or reputeil Indo-European cousins, and that its long ascendency has been due, under the conditions of erroneous linguistic principles, simply to its high antiquity and the circumstance that it is the best known and the most highly venerated of its ancient family, by reason of its sacred associations. The Highlander and the Welshman, who affirm that their respective dialects have also a claim to be con- sidered the primitive languages, have much of the same kind of evidence to adduce as that which has always been advanced in behalf of the Hebrew ; and they, in their turn, might be met by a strong array of striking analogies, presented with equal confidence, as proof that the idiom of the Sandwich Islanders should not be left out of sight in any candid examination of the question. 'H 8 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. It will perhaps be proi)er to illustrate the methods of this system of comparison by a few instances selected from the works of writers in recent times, and even in the j)re8ent century. They will forcibly suggest the great advance made in linguistic science within the last sixty or seventy years, and may also serve as a warning to any who may still insist on a radical afTinity between verbal forms on the evidence of mere external resemblance. Wo find the acute and learned Moses Mendelssohn * among the later serious advocates of the doctrine that the Hebrew is the parent of all other idioms. Matthias Norberg,'' a re- spected scholar of the early part of this century, after close scrutiny, detected in the Greek language the inherited linea- ments of the same venerable and prolific parent. According to him, idvo^ arose from oy , a people, by the inseilion of ; Xo7(o9) was transposed from iip , a voice ; fivdeta, was changed from bttJia , to liken. But the most frank and hearty exposition of the theory that we have seen is a little book by the Rev. Alexander Piric,^ a man of considerable linguistic attain- ments, but of still greater ingenuity. We cite some of his numerous derivations. Ho supposes that our word bog" comes from nsa , to weep or run with water ; that boggle (bogle) is connected with bna , as inspiring terror ; and that tar is derived from "ikh , to mark, as being much used for marking sheep, sacks, etc. From Da"i he would deduce the Latin rego^ because stoning was an exercise of the supreme authority as a judicial punishment. In his opinion, boa, " to retribute," gave rise among the Hebrews to the word camel^ on account of the revengeful disposition of that animal.^ 1 In prolegomena to his edition of the Pentateuch, cited by Delitzsch, Jesurun, p. 46. * See Friedrich Delitzsch, Indogormaniscb-Semitischo Wurzelvurwandtscbaft, p. 3. * A Dissertation on the Hebrew Roots, intended to point out their extensive influence on all known languages (Edinburgh, 1807). The introduction, written by another hand, says of the author, " had he never lifted his pen on any other subject, the following pages would establish his character as a scholar and a Christian." * This derivation, however, it should be remarked, was once quite common. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 9 asj, to Htcal, givofl the origin of our word knave, wluch " at picscnt i8 uHcd in a bad sciiso, the same in which tho Ilcbrcws used it." Comparing Solomon's dcHcription of hlH hj)ou80 oa '' a garden incloHcd," ho imagines that 133 includcH llie idea of beauty, which is guarded with peculiar caro, and tlm*^ hcnco arises the Greek 'yvvrj ; while '• the cognate Latin g-enita, a doughter, is plainly the source of our Janet.'' t'I or in, to judge, gave birth to a numerous progeny. Beov, what is just, and 8«i/ov, skilled (in judging), do not surprise us very much ; ])ut wo are further asked to accept Blvrf, whirlpool, or whirlwind, " from the idea of vehemence in pleading." And, as tho judgment-seats of antiquity were often groves, BevBpov is added to tho family, which is next increased by the accession of our English t/67t, because oracular judgments were frc(iuently delivered from caverns. For a similar reason any hollow vessel came to bo called a tun, " the d being changed into t, as usual." As a judge held a dis- tinguished station, tho Spanish Don is next admitted to the domestic circle ; and since i^i also means to dispute, and " as people in angry dispute are still said to be teethy, or 1.0 show their teeth," it was thought inhospitable to leave tho Latin dens chattering outside in the cold, nba , with other meanings, has the sense of carrying away captive. "Now the m prefixed forms a noun ; before a it sounds ang-, hence the Teutonic anffel, with its cognates." The confusion of tongues at Babel arose, ho says, from a defect of labial utterance. When one would have said Bel (bya), he said Babel. Hence also our word babble. Ho is very sparing of onomatopoetio affinities ; but he would probably concede to that class of analogies the relation he holds to exist between the Hebrew ^39 , sorrow, and och hone ! These instances, though perhaps more whimsical, are not more unreasonable, than many of the combinations that have \png been held, and are still to be met with in current litera- ture. We find a writer so recent and influential as the late Albert Barnes stating, in his popular commentary on Job, that our word evil comes from the Hebrew V^i» It II 10 RKLATION8 OF THE ARYAN AND SEMITIC LANOUAOEa. I U surely nc(!(!Hrtnry, in view of such facts, that tlio gnneral |)riuci|.i(\s t)f Iho Hcionco of langui'i,'o shoulti bo mmlo an CHscntiul |)urt of u lil)erul education, ut UniHt to such au extent that one will not need to l)0 a 8i)eciali»t to bo able to detect and disprove such inaccuracies as theso. But we must n(»w consider the more safe and sober at- tcmi»ts lliat have been made to compare tho two great famili(!s Imfore us. The study of the .Sanskrit, which afforded a clew to the mazes of tho varied forms of Indo-European H|»cech, was also the occasion of a more just appreciation of the condilious of the problem we arc considering. In that ancient language, so {Hjrfcct and intclligiblo in structure, largo numlters of Aryan words wore detected in their most clenHMilary accessible form, revealing to tho acuto and delicate l»crception of such men as W. von Humboldt, Grimm, and IJopp the laws which determined their modification into other varieties of expression. Science having thus vindicated her claim to this vast province of sjMiech, it was felt that other districts — nay, tho whole realm of human language — must also bo subject to her of right. Henceforth tho reign of fancy and caprice in these all'airs was at an end ; and their intrusions would always bo unwelcome to the new r/^ij^ime^ though they could not always bo re|X)lled. In tho treatment of tho relations between tho two great families of s|)eech, now clearly established and defined, as well as between the several languages in each, it was felt that laws regulating all changes of form must be sought and assumed to exist, and hence also that the utmost caution must bo used in the comparison. This, we mean to say, was the tendency of tho method of inquiry, and tho professed aim of tho several in- vestigators. Some, however, while recognizing the necessity of this principle, have failed, unconsciously, to act upon it, being frequently led to violent and capricious assumptions through their eagerness to attain the final theory of solution. Others, again, influenced either by dogmatic prejudices or by a conservative temper, have refused to indulge in any speculations upon the subject, or go so far as to assert RELATIONS OP THE ARYAN AND SEMITIC I.ANOUAOF.S. 11 that the lanp^mgos thcmsclvoH, aH well an the rmoH thoy typically roproncnt, can iiovor ho proved to luivo \>vvn origi- nally identical. Willi rci^ard to tho earliest portions of tlio present period, wo have chiefly to remark v. tendency to hrinj^ Semitic; words into cloHO connection with tho widely-related ami liospitahlo Sanskrit. Adolung's Mithridatcs, tho n»onnmontal l>v)undary- mark between tho old and tho new regions of philological rosoarch, holds alRO a certain dividing-placo in the history of tho j)rosent qnestion. Its learned anthor was tho (Irsfc to compare, to any extent, tho Sanskrit with the Semitic; vocal)- ulary. As to his method, however, ho is to ho [daced wholly within the old unscientific period. Not being himself a Sanskrit scholar, ho was tho more inclined to the prevalent error of comparing full-grown words, and not roots, or even stems, in tho languages discussed. lie connects, for example, tho Sanskrit ddima, first, with the Hebrew ens , Adam. Some of tho greatest pioneers of philological science, also, with all their sagacity and penetration, wore carried Ijoyonci tho limits of probability in their theories, or rather conjec- tures, upon this subject. Being not, in general, Semitic scholars, and their survey being necessarily rapid and super- ficial, their analysis was not sufficiently profound to deter them from assuming close relations to exist between forms which had only a casual and external resemblance. Tho tendency to assimilate tho two idioms, excited by the mag- nificent results of tho comparison of tho several Aryan languages, may bo inferred from tho fact that even W. von Humboldt accepted a multitude of tho most superficial com- binations as proving an essential affmity between the forms compared. Bopp, also, attempted to establish a number of analogies which must be called forced and arbitrary ; though that great pliilologist was unwilling to guarantee the absolute correctness of all his conclusions on this subject. As we are now approaching tho latest period of the invest- tigation, and shall have to speak of the comparative value of theories largely influential at the present time, we may 'I I » 12 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. refer in passing to an opinion advocated at one tine by Lepsius and Benfey and more positively asserted by Bunsen (in his Outlines of the Philosophy of Universal History, and elsewhere). The view held by them was, in general terms, that the Semitic and Aryan families are related to one an- other, and have as intermediary the Coptic, or rather the an- cient Egyptian, as representing the North African group of lan- guages. The leading arguments were, that the striking resem- blances and analogies between the grammatical forms of the Coptic and the Semitic pointed clearly to a connection be- tween those languages, while the fact that many coincidences were found between Indo-European and North African vocables created a presumption in favor of an early relationship be- tween these also. In the defence of these positions Lepsius* and Benfey ^ wrote special treatises, and the same theory was maintained by the Egyptologist Schwartze in his work on Ancient Egypt (Vol. 1. 1843). It is doubtful if the survivors of this group of theorists would now maintain this doctrine, at least as far as the Indo-European family is concerned.^ And it must be allowed that the verbal resemblances be- tween the Indo-European and North African families of speech are too sporadic, and apparently too superficial, to war- rant any serious attempt to compare them in the present state of science. The question of affinity between tae Semitic and North African families is still undecided.* Gesenius, the great lexicographer, and inaugurator of sci- entifi^^ Semitic studies in Germany, maintained, in general, a neutral attitude towards the problem before us. True to the empirical principles of his philosophy of language he re- frained from dogmatic generalizing while he could not make * Zwei sprachvergleicnende Abhandlungeu (1836.) ^ Das verhaltniss d. agypt. Sprache zum semitischen Sprachstamm (1844). 8 With regard +0 Lepsius it may be inferred from his last work, Nubische Grammatik (Berlin, 1880), p. iii ff., that his present views on the question .f the classification of languages exclude the above theory. * The theory is discussed unfavorably, from his philosophical point of view, by Renan, Histoire gene'rjile des langues sdmitiques (4th ed., Paris, 1863), p. 80 fF., 456 f. Cf. the moi 3 intelligent and liberal remarks of Sayce, Introduction to the Science of Language (London^ 1880), p. 178 ff. m RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 13 certain progress towards fir.ed underlying principles of unity. It is tru3 that both in his Manual-Lexicon and in his Thesau- rus he has instituted a vast number of verbal comparisons with Indo-European forms, which have helped more than all else written upon the subject to bring the question before the minds of ordinary students, and to affect their opinions regarding it. But he refrained from presenting dogmatically a theory of these analogies, being inclined to believe, until further light should be thrown upon the problem, that they were the result either of an early contact of the races leading to an exchange of vocables, or cf onomatopoeia, or of mere accident. It should be remembered, however, that his senti ments on this subject were formed before modern science had reached those of its grandest conclusions which might well justify still broader assumptions. Yet he adopted and amply illustrated a theory whose establishment would tend towards the solution of the problem — the doctrine, namely, that the triliteral Semitic stems were reducible to significant and fundamental biliteral roots contained in the first two consonants; the last letter exerting the special modifying influence that determines the meaning of the word. In large numbers of these ultimate roots he discovered close correspondences vdth Indo-European forms, which, however, he declined to accept as conclusive proof of internal rela- tionship. We come now to consider the opinions of two authors whose opinions have been so fully elaborated as to entitle them to be considered the founders of a special school * of Semitic philology. We mean Julius Fuerst^ and Franz 1 The "Analytico-historical," so-called, because, on the one hand, according to its principles, the various elements of language &nd of individual words are held to be endowed with inherent significance which is to be determined by a profound analysis, and because, on the other hand, they call to the aid of their investigations a body of Jewish tradition, such as the Targums, the Talmud, the Masora, and the later Rabbinical writings. The name serves to distinguish their system from the so-called " empirical " school of Gresenius, and the " critical " or philosophical school of Ewald. These terms have now little significance, as they •erve to designate tendencies or principles rather than well-defined sects or parties. ' Lehrgebftude der anun&isohen Idiome mlt Besng auf die indo-germanischen lu 14 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. Delitzsch,^ theorists whose vast learning and patient industry it is impossible not to admire, but whose philological system it is equally impossible to accept. In it the process of verbal analysis for the purposes of comparison with analogous forms is carried to its greatest extreme. The chief monuments of this system are the Jesurun of Delitzsch and the Woerter- buch oi Fuerst ; the former an exposition and defence of its principles ; the latter, the repository of its practical results. Their leading positions may be summarized as follows : (1) That all languages have been developed from one common stock of elements, all of which, in every part of speech and in every word, have a significance, definite and divinely imparted. (2) That this innate idea is to be educed through a minute analysis of each form, and the widest comparison with the forms of other dialects of the language of mankind. (3) That the Sanskrit is the master-key to unlock the secrets of all Aryo-Semitic speech, there having been originally one " Sanskrito-Semitic " idiom, from which proceeded six families of speech — the Sanskrit, the Medo-Persian, Semitic, Graeco- Latin, Germanic, and Slavonic. They thus annul the ordinary classification, and make all the Semitic dialects together a sister idiom to each member of the great Aryan division. (4) That, accordingly, the chief resort for purposes of com- parison is the Sanskrit, while the other related languages should also be consulted as supplementary and illustrative. (5) That all Semitic triliteral forms can be traced to original biliterals, parallel to the most numerous class of Sanskrit roots, and being the significant element in each form, as containing the original and typical idea. (6) That the remaining portion, the determinative modifying element, con- sists of a suffix, or, far more frequently, a prefix, corre- sponding in meaning. fl.nd as nearly as possible in form, to the Sanskrit prepositions. In the elucidation of this system they have subjected a vast number of forms to examination Sprachcn. Leipzig, 1835. Librorum Sacroram Concordantiae. Leipzig, 1640 HebiHisches und chaldaischcs Woerterbuch, Leipzig, 1857-61. 1 Jesurun ; sivc Isagoge in grammaticam et lexicographiam linguae He- braicae, contra O. Qesenium et H. Ewaldum. Orimmae, 1838. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 15 and have besides illustrated their conclusions by citations chiefly from authors of the rabbinical school, the products of whose fancy they have elevated to the dig-nity of scientific demonstration. The objections to the whole theory are ob- vious: (1) The reduction of the triliteral Semitic roots to biliterals is too through-going and mechanical. Analysis does not always yield biliteral roots ; nor is it to be expected that it should. Triliterals, as well as biliterals, have existed from the beginning. (2) The combinations attempted with the so-called sister tongues are not made upon any sound etymological prin- ciple, nor ?''e the forms reducible, in many instances, to any- thing like even external resemblance. The following com- parisons may be cited as evidence (Jesurun. p. 175), -in-o , to be pure, with Sanskrit qrd and Lat. cremare, to burn ; '•{C'Xi , to conceal, with Gr. (liveiv, and Lat. manere, to remain ; tia-s , to subdue, with Skr. pad, to go, and Gr. waTeiv. (3) The pre- positional additions which are supposed to have been prefixed to the biliteral roots do not preserve any fixed and certain meaning in the various instances cited as illustrations. The views of Ewald, the greatest grammatical and his- torical genius among the Semitists of the last generation, are deserving of consideration. As might be expected, they are original and unique. Employing his special faculty of inves- tigating the nature and relations of grammatical forms, he endeavored to prove by researches in the Indo-European, Semitic, North African, and (oo-called) Turanian families of speech, that these are outgrowths of a common stock, which is most nearly represented now by the Indo-European. A discussion of this view will have to be made in the second chapter, when we come to consider the question of criteria of affinity particularly. It is sufficient to say in the meantime that ^;he evidence adduced in its favor is precarious in its very nature, and therefore inconclusive. He made subordinate the question of the relations of special words or predicative roots, though he maintained the possibility of such combinations, and a considerable number of comparisons may be gathered from his various linguistic writings, all of them ingenious, H 16 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. but none of them convincing, because not based upon a sys- tematic tlieory. Singularly enough, for a man of his insight, he failed to trace such words to their primary expression. An ambitious and laborious effort was made by Ernst Meier, in his Hebraisches Wurzelworterbuch (1845), to con- struct a dictionary of Hebrew etymology, upon a theory which must be pronounced extravagant and on all grounds untenable. His main position was, that the stock of roots in the two families might be reduced by analysis to a mere handful ; that the Semitic forms, which are currently larger than those of the Aryan division, might be brought to a pri- mary conformity with the latter, by throwing qff from each of the triliterals a letter which was regarded as secondary. Such letters were supposed to havo been developed in accordance with an assumed principle of reduplication in the formation of verb-stems, analogous to that which pre- vails in Aryan perfects. That is to say, one of the primary letters might be repeated in the formation of the stems of the Semitic perfect tense, and this was followed by the adoption of the developed forms as current roots. Since, however, the repetition of the same sound was felt to be dis- agreeable, the secondary letter was dissimilated from its pri- mary in most cases, though the limit of choice was confined, in each instance, to its own class of sounds,^ The most laborious and persevering investigator of the subject in recent times is Rudolf von Raumer, who is also well known through his Indo-European researches. The reader will find his theories succinctly stated in his latest contribution.^ He has considered it a necessity to establish laws of phonetic representation regulating the changes u»der- gone by roots that appear in both families. These are as follows : (1) The hard Semitic explosives or mutes are rep- resented etymologically by the corresponding Aryan sounds ; (2) the soft Semitic explosives are mostly represented by the 1 Cf. Friedrich Delitzsch, op. cit. p. 8. " Zeitschrift fiir sprachvergleichende Sprachforschung, xxii. p. 235-249. Compare also D. Pezzi : Glottologia aria recentissima (Turin, 1877), p. 37-41. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 17 hard Aryan sounu^ of the same organs. On these assump- tions it is obvious to remark that there is no regularity in the alleged correspondences. It follows from his principles that the Aryan ^, for example, may be represented either by Semitic t or d; khy k or g ; p hy p or b. We do not main- tain, in the meantime, that this cannc*^^ be a fact ; but it is evidently not in harmony with the observed facts of other languages that are mutually related, in which the mutes as well as other sounds are either equivalents, or are differ- entiated according to laws normally invariable. As Von Raumer's scheme is without observed analogy, very strong evidence should be adduced in its support before it is entitled to acceptance. But in the combination which he makes for the purpose of proving his assumed laws he does not advance much beyond his predecessors. He seems not to have kept in mind the consideration that, if the two families were ever one, they must have separated before the full-grown noun and verb stems in each system were developed ; for he com- mits the error of failing to search for Proto-Aryan and Proto-Semitic roots, as furnishing the only basis on which lawful comparisons can be made. His combinations are in general only a little less improbable than those of Fiirst and Delitzsch, referred to above. His assumed phonetic laws are, therefore, still unproved. The Italian scholar, G. I. Ascoli, has given the weight of his great name to the general theory of an ultimate relation- ship of the two families. He has, in letters addressed to Bopp and A. Kuhn and in contributions to scientific jour- nals in Italy, also attempted to bring forward special evi- dence for this doctrine based upon the resemblance between certain formative elements (case-endings, etc.) in the re- spective systems. In this he follows close upon the track of Ewald, though in a narrower field, and the nature of the proof is equally uncertain with that adduced by the latter. He also deals with the well-known similarities between some of the numerals and most of the pronominal stems, a subject to which Lepsius had before him given special attention. 18 RELATIONS OF THE ARYAN AND SEMITIC LANOUAGEg. il The inherent difficulties of this branch of the investigation are that we do not know the roots of the numerals, and that the further back we go to their primary forms the less resem- blance they seem to show ; while as to the pronouns, as we shall see later, the phonological investigation is somewhat uncertain. The testimony from this source is, moreover, too general to be universally satisfactory, since several pronouns are alike in a great many other families of speech. Ascoli has also formulated laws of phonetic change. To Von Raumer's rules he adds a third, to the effect that an Aryan g- is represented in Semitic by p. The evidence given for this is scanty and precarious. The most scientific and also the most satisfactory attempt to prove an Aryo-Semitic relationship is undoubtedly that of Friedrich Delitzsch, in his Indogermanisch-Semitische Wurzelverwandtschaft (1873). As to his general attitude towards the question, he is fully convinced of the hopeless- ness of attempting to reconcile the divergent grammatical systems ; but holds it to be a possibility, that at some remote period, before any flectional tendency was exhibited in either, they possessed a common stock of roots. In seeking to ascertain the roots which may be shown to have once been the same, he recognizes the principle that we must aim to draw them only from the original languages from which the two families arose respectively. In making up the list of dialects from which the original Semitic language must be constructed, as far as its roots are concerned, he rejects the Old Egyptian rightly and the Assyrian wrongly. His view as to the latter appears (p. 29) to have been, that for lexical purposes Assyrian roots could not afford any essential help in the solution of the problem. But his own valuable labors sir ce then in the interpretation of the cuneiform inscriptions have only confirmed the justice of the claim, long since put forth, for the independent character and essential importance of the Assyrio-Babylonian branch of the Semitic family. When Dr. Delitzsch comes to the treatment of the roots that are eligible for comparison, he shows an advance upon RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 19 his predecessors in tho endeavor to employ a systematic theory as to the constitution of those roots. Taking advan- tage of the labors of Indo-European investigators, such as Gurtius and Fick, he assumes as valid the distinction made by them between primary and secondary roots, according to which the latter differ from the former through the posses- sion of one or more determinative letters, which represent, according to a sort of phonological symbolism, modifications of the radical notion (p. 33 ff.). His views on this branch of the subject seem to be philosophical and sound. In taking up the Semitic roots, lie shows evidence of not having made a careful and thorough analysis. He proposes to throw off the old limitations occasioned by the theory of biliteral as distinguished from triliteral roots, according to which the former are eligible for comparison with outside languages, while the latter are not. But that he is really controlled by that theory is plain from his classification of Semitic roots (p. 43 ff.). He draws the line broadly between roots with " weak " and those with '' strong" letters. In the former class the weak letters are claimed to have little or no essential significance, while in the latter each letter is primary and autonomous and the forms containing it may be put directly on a level with the Aryan roots. Now what are these insig- nificant weak letters ? We find that along with k , ^ and \ the same unimportant part is assigned to n and ». Why n should be excluded does not appear. It is no more and no less an original Independent sound than fy or 9. But the radical error here is the assumption that because roots containing these letters are " weak " in tho inflections, the sounds themselves must be adventitious and unmeaning. The fact is, however, as we shall show in our fourth chapter, that the weak letters are as independent and significant in their orig- inal forms as the strong, and that the determinative letters are no more taken from the latter than from the former. Again, the letter s alone of the strong consonants is put in the same class with the weak letters when it appears as the first sound in roots. But the true view of the matter is that 20 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. the weak letters may be used as true predeterininatives along with the strong letters 5, la, and p, since each of these is found to occur at the beginning of secondary roots as the modifying element. The views of Dr. Delitzsch as to phonological representation should also be subjected to some criticism. In seeking to prepare a scheme of correspondences in sounds (pp. 82, 83), he commits the error of neglecting to reduce the phonetic stock of both systems to the limits that obtained in the original languages. With regard to Aryan sounds, indeed, he confines himself to those wliich have been accepted by phonologists without dispute, as belonging to the primitive idiom ; but in the Semitic family he takes the sounds just as they stand, only grouping together, for the purpose of bring- ing out a set of equivalents to the Aryan sounds, those which are organically the most closely allied, without investigating the question of their true historical relations. For example, he assumes that the Hebrew x, where it answers to the Arabic ^j6, and the Ethiopic ^, is to be classed with n, all of them representing the Aryan d and dh; while the Hebrew x, an- swering to the Arabic ^Je, and the Ethiopic /\, is to be grouped with 6,\i},andto, as representing in common the Aryan s. The fact is, however, that the Arabic ^j6, md its Ethiopic analogue appear to have been developed far more frequently from a radical :e than from a radical x The question of the production of these letters is surrounded with great obscurity, but this much is plain.^ Further, his system divides sharply between the different kinds of Semitic gut- turals. Thus, the Hebrew n, with its Semitic representatives, historically corresponding, as he claims, with the Aryan g-h^ leads one division of sounds : while N and » with their rep- resentatives form another, a'nswering to the Aryan spiritus lenis. Here ti is thrown out altogether, although it is a 1 Thus, 9 as the first letter of Ethiopic roots (see Dillmann, Lexicon Aethiop. col. 1322 ff ), appears to come in only one case certainly from an original 1, while it is sometimes actually developed from a primary VS. RELATIONS OF THE ARYAN AND SEMITIC LANOUAOES. 21 sound at least as fundamental as n. To be consistent he would have had to represent it also hy the Aryan fffh which would have been self-evident ly erroneous. The true view is that all the Semitic gutturals (except m , which is common to all languages), are of pure native origin, and are capable of an organic classification which precludes the possibility of the theory we have criticised. That these errors should detract in many cases from the value of the comparisons of roots made by Dr. Delitzsch was inevitable. Moreover, there is a want of consistency observ- able in the application of these laws. Thus, in comparing the Heb. ci"»n^, a thin board, and its Arabic hometyma, with Gr. o-zeoTT-Tw, to shave off, he remarks (p. 76), that "there is nothing surprising in the agreement of the aspirated h with the Indogermanic A:," though, as we have seen, he sets forth the same n sound as being the representative of the Aryan g-h. In the choice of roots for comparison it is unfor- tunate that so many of them, perhaps the majority, arc liable to be objected to on the ground that they may be of onoma- topoetic origin, and therefore more likely to have arisen in- dependently in the separate history of each family. In spite of these and minor defects, the work is of much value from its stimulating and suggestive character, as well as from the actual contributions it makes to linguistic learning. Another German scholar, J. Grill,^ has taken up, with much acuteness and ingenuity, the question of the relations of the two families from the stand-point of the constitution of their roots. Recognizing the divergence ujt merely of flectional characteristics, but also of root-structure in the two systems, and emphasizing the fact that in the Aryan root the vowel is coordinate with the consonant, and that in the Semitic it is subordinate, he seeks to harmonize the two by carrying the view back to a hypothetical period when a so-called " Alpha-sprache " prevailed, whose peculiarity was 1 Zeitschrift der dcutschen morgendlandischen Gcsellschaft, Vol. xxvii. pp. 425-460. Ueber das Verhaltniss der indofjermanischen und der semitischen Sprachwurzeln ; ein Beitrag zur Fbj'siologie der Sprache. m S9 RELATIONS OF THE AKYAN AND SEMITIC LANQUA0E8. that a was the only vowol sound employed in either. This view will l)C taken up, and shown to be improl)al)lo, when the same proWleui i» to be dealt with in the course of our own investigation. In addition to the names of sj)ecial investigators already cited, general mention should be made of some of the greatest liglits of linguistic science, who with more or less confidence favor the doctrine of the possibility of a real re- lationship between .he two families, though they have not attempted to formulaic any special scheme for harmonizing their divergences. On this side may be put the names of Eugene Burnouf, Max Miiller, Pictet, and Steinthal. The opinion of the last-named is specially valuable, because he has discussed ' the question on general linguistic principles more thorouglily than any other of those who have not entered into an analysis of vocabularies. On the other hand, the probability, and even the possibility, of such affinity is rejected upon general principles by an influential, and per- haps at present the dominant, school of linguistic philoso- phers, who eitlier hold to the theory that languages of differ- ent inflectional types are necessarily of diverse origin, or on general anthropological evidence favor the doctrine of the diversity of human species. Among the most pronounced of the opponents of any scheme of reconciliation are Pott, Schleicher, Renan, Friedrich Miiller, and Sayce. Their views will necessarily be considered in the next chapter, when we come to talce up more particularly the question of the criteria of relationship. 1 Zeitschrift d. deutscben morgenlind. Qesellschaft, xi. 396 ff. CHAPTEll II. CRITERIA OF RELATIONSHIP. In passing now from the more critical to the more con- structive portion of our Essay, it will be well to throw some light on the nature of the task before us, by exhibiting the more obvious points of contrast between the two families of speech.* Bringing thus into view the distinguishing features of each idiom, we shall be the more able to propound the conditions of a just investigation, and to establish the true criteria of evidence as to their relations. In every language, or group of languages, there are three elements, whoso peculiarities determine its special character, and help in different degrees towards its classification. These are, its sounds, its structural principles, and the con- ^ ' V""J >1 tents of its vocabulary. In the case before us the numerous points of dissimilarity seem at first sight radical and indica- tive of a diverse oriccin, while the points of agreement appear accidental and superficial. As regards the first element, the sounds of the respective languages, great divergence is apparent among the dentals, in which the Semitic family has developed a strong tendency to multiply sibilant and lisping sounds, and a wider differ- * Comp. Ewald, Ausfiihrliches Lohrbuch der hcbraischen Sprachc (8th ed.), 1870, p. 26 fF. ; Eenan, Histoire gen^ralo des langnes S^mitiques (4th ed.), 1863, p. 18 ff., 454 ff. ; Whitney, Language and the Study of Language, p. 300 ff. 23 >-:::-:J^ A' 84 RELATIONS OF THE AIIYAN AND SEMITIC LAN0UAOE8. Ill I enco »till among tlio gutturuls, in which the Bamo foraily ox- hiI)ifH an UHtoiiishing vnricty of |)h<)iiel'iO cxproHHion. Oil cxiiuuiiiiiK tiio roots iinil the gciicrul ntriicturo of tho wonlH, wo an; at onco Htrucli hy tho Htrajjgo and uniiiuo prin- ciplcrt that i-ontrul the iSemitic dialects. While in tho Aryan family, rootH may consist of a consonant and a vowel, or of two or more cunsoimnts accompanying or groujied ahout a vowel, it is an almost invariable Hemitio law, that tho roots of nouns anil verba, so far as tho analysis of living forms can testify, are based npon three consonantal sounds. As to Semitic words in actual (Speech, we see exemplified as univer- sully the [leculiar jjrinciple that the vowels are used to express subordinate, modified, or accessory notions, while tho con- sonants, which form the framework of tho word, embody its fundamental idea. Again, this family has only to a small extent the habit or capacity of compounding words, a circum- stanco which tended to multiply tho number of its roots, while tho Aryan languages, having developed that principle largely, were enabled to economize their original stock. Further, the moro strictly grammatical features of tho two idioms appear to bo no less radically divergent. Beuan characterizes the Semitic grammar as a sort of architectural and geometrical structure, as contrasted with tho latitude and flexibility that mark the inflections and syntax of Aryan speech. In tho Semitic verb there is a great variety of forma (" species," quasi conjugations) to express modifications of its genera^ motion, which represent chiefly simple subjective conditioii;;, .j.g. causative, declarative, desiderativo forms ; while in is tenses, which arc few, tho moro mctaphvsical idea of time is vague and indeterminate, and in those dialects which in a more reflective stage in the history of the race, attained to greater precision in expression, could only be definitely indicated by the help of limiting words. In the same way its moods are also few and entirely foreign in typi- cal structure to those of tho Aryan languages. With regard to its noun, the prevailing absence of case-inflections, and the formal modification before a limiting noun, called the con- ■'«i.s. '■■'»• RELATIONS OF THE ARYAN AND SEMITIC LANOUAOES. 25 struct sta. j, are Diuoiig tho more obvious {)cculiai'itics. Tho objective perHoiml 8uffixo8 of verbs, and tbo poHrtCHHive per- sonal suffixes of nouns are further important charuuteristics of the Semitic family. Within the sphere of the lexicon, also, we are not led, im- mediately at least, to unmistakable marks of real affinity. If the stock of roots in the respective vocabularies was origi- nally tho same, the evidence of this does not ai)pear on tho surface. The leading differences between the two families being thus iiiiicated, the character of the problem to bo solved becomes more intelligible. Tho following mode of procedure will perhaps be the most natural and serviceuble. After a glance at the question of phonetic phenomena, the grammat- ical features of the respective systems will be taken up and it will be considered particularly whether there is a possi- bility of reconciling the divergences outlined above. After estimating the results of this inquiry it will be necessary to decide whether any other criteria have a right to be admitted, and an attempt will be made to show that the comparison of roots alone is not opposed to the true methods and {principles of linguistic science. These discussions will form the sub- ject of the present chapter. It will then be proper to take up the two systems separately, without reference to the question of harmonizing individual words, the object in view being the obtaining of primary forms that may be legitimately compared. This will involve, first, a reduction of the sounds of each family to their original limits and expression, and, second, the presentation of a scheme of phonetic representa- tion. The treatment of this subject will comprise another chapter. Then it will be necessary to treat of the constitu- tion of the roots of the respective systems according to the laws that prevail in each. The concluding portion of the work will then be taken up with the comparison of roots, chosen and dealt with according to the principles that are found to underlie their production and development. Taking up now the subject of the criteria of relationship, 6 26 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. a few words will have to be said on the subject of the sounds of the two families. As they stand, they do not accord, in so far as sounds are found in either system which do not appear in the other. Tlife question arises : Are we to regard these differences as precluding any attempt to compare the stock of roots in the two idioms ? Certainly not. The vari- ants may tiot be original. Sounds are often found in lan- guages in their modern or literary form which did not exist in their early condition; and sounds frequently appear in one or more of the branches of a linguistic family which the parent tongue did not possess. In the Aryan family, for ex- ample, there is not one of its branches which does not con- tain sounds foreign to the primitive speech, from which all in common sprang. It follows, therefore, that phonology is not a primary criterion of linguistic relationship at all. If, after reducing tho phonetic stock of each system separately to its primary range of sounds, there are found in one sys- tem some which do not appear in the other, this fact is still not decisive of original diversity of idiom. An examination of the structure or of the verbal forms of each language may prove beyond a doubt a primitive unity, in spite of the pho- nological differences. Thus if we take the sounds of the Keltic group, as they are found to have existed in the origi- nal Keltic language, the gutturals which belong there are not represented in the Indo-Eranic idiom ; nevertheless these two branclifcs of the Aryan family undoubtedly came from one common stock. So the cerebrals in Sanskrit have not pre- vented the harmonizing of that language with the dialects of Greece. It appears, then, to be a false and arbitrary restric- tion which those scholars make who would prohibit any attempts to harmonize the Aryan and Semitic idioms, on the ground that the phonology of the two shows such distinct features.^ It is not well to lay much stress on such differ- ences ; for that would be to appeal to an unsound source of ^ E.g. Prof. A. H. Sayce, who tells us in his Principles of Comparative Phil- ology (1874), p. 101 f. ; Introduction to the Science of Language (1880), ii. p. 176, that the phonology of the two systems opposes the idci of their relationship. RELATIONS OF THE ARYAN AND SEBUTIC LANGUAGES. 27 comparison. Nor is it hard to account for the notorious fact of important changes in the phonology of any people. ^ I The influence of climate, food, habits of life, and external < ^^ conditions in general upon the or-^ans of speech, is both < ( extensive and familiar; and it is easy to perceive why, through the course of ages, and long separation under different skies, each of the branches of one original language has often developed sounds quite unknown to the phonology of the other. The comparer may reduce the stock of sounds in each system to its limits as they appear to have been fixed in the original languages. If the sounds are then found to have been at one time the same in each, this settles nothing decisively as to the original relations of the systems compared. If, on the other hand, sounds are found to have existed from the earliest accessible period in eitlier idiom which are not found in the other, this also proves nothing as to primitive relationship. It is the business of the comparer, in either case, to seek for laws of phonetic representation by the comparison of roots, not directly of sounds, according to which certain sounds in the one system may eventually be found to correspond with certain sounds in the other system. These sounds thus harmonized may be either approximate equivalents, or they may be such as analogy shows to be capable of representing one another through permutation in human speech. The main point to be insisted on here is, that sounds are not at all a primary criterion of linguistic relationship. It is sometimes for- gotten or unperceived by glottologists that sounds are com- pared with one another only as they become the outward form in which ideas are clothed. Significant terms are the proper material of comparison, and the sounds are traced out, classified, and compared secondarily according to the history of the embodied thoughts. The direct and indopendent comparison of sounds is, properly speaking, a department of physiology. Those who put forward the theory just criti- cised might not maintain in general that a striking diver- gence in the phonology of any two systems necessarily 1 1 I ,; 28 RELATIONS OF THE ARYAN AND 8EMITIC LANGUAGES. precludes their original identity. This would be to contradict history. But when the two idioms in question are thought, on other grounds, to be radically separate, the phonological objection is us natural as it is fallacious. We must now turn to the structural peculiarities of the two systems of speech. Here we shall have to regard the lan- guages just as they appear in actual use, and inquire whether anything can be inferred as to their early condition. In other words, we must, by analyzing and comparing the verbal and syntactical forms, endeavor to reduce them to common primordial principles. In our previous Article we had hinted at the general value of grammatical comparison in this field of inquiry ; but here it will be necessary to con- sider the question more at large. The conditions for this investigation are hot! Tivorable and unfavorable. On the one hand we find the two groups based upon fully-developed inflectional systems. There is also abundant material, in the form of a large literature in both idioms, bequeathed to us by a long line of intellectual ancestors. Moreover, the internal laws of each of these types of human expression are sufficiently intelligible ; foi' the principles of Aryan speech have furnished the more familiar elements of Comparative Philology, and the Semitic dialects, in their simple and regular structure, reveal easily the process through which their vocables are built up. But, on the other hand, we have this disadvantage, that we do not possess in either idiom literary remains that throw any direct light upon its primitive form. Go back as far 418 «ve may, we meet with only full-grown words, in whose complex sounds we seem to hear no more than a faint echo of the simple language of the world's childhood. Taking up now the word and the sentence as the two maiii elements of human speech, and regarding the structure of both as tlie surest distinguishing features of a language or linguistic group, the inquiry naturally divides itself into two branches. First, as to the word, we may assuir\c its special character to be exhibited in its typical form, as this is associated with the process of its development from the RELATIONS OF THE ARVAN AND SEMITIC LANGUAGES. 29 root. In this way, e.g. we may contrast the structure of dictum from die with that of boj?? from bcp , or dicens, dicentes with boip, D^be'p; noting such matters as the part played by the vowels in each set of words, as related to the function of the consonants, and the significance of the prefix or affix as entering into the inflectional system of each type of language. Secondly, w? have to compare features of syntax; the Semitic sentence is placed side by side with the Aryan, and the endeavor should be to determine whether the existing forms can be reduced to a common system of expression. Now, it must be acknowledged that hitherto such inquiries as these, conducted, as they have been in some cases, most acutely and profoundly, have had but ill success so far as their main object is concerned. The result, at best, has merely added to other presumptions in favor of an organic relationship, through the exhibition of a few analogies in the more fundamental structural principles of the word and sentence, which have, however, arrayed against them numer- ous divergences, apparently no less radical and essential. Our more definite conclusions, however, must be reserved until we have analyzed the evidence. If we consider the structure of Semitic and Aryan voca- bles, we find the following to be, perhaps, the most striking difference : in the latter class the radical portion of the word is almost always modified by additions at the end, whether in the base forms of nouns and verbs, or in the various inflections to which these are subject ; while in the former the principle of augmentation at the beginning is also followed, as, for example, in the formation of the species (conjugations) of verbs, of the future (imperfect or aorist) tense, and of a large portion of the derivative nouns. This fact is seized upon by Ewald,^ who compares it with the 1 Abhandlung liber den Zusammenhang des Nordischen (Tiirkischcn), Mittel- landischen, und Koptischen Sprachstammes (aus dem Zchnten Bande dcr Abhandl. der konigl. Gesellschaft der Wisscnschaftcn zu Gottingen). Gottin- gen, 1862. The full title of Professor Pott's treatise, in which it was severely criticized, is as follows : Anti-Kaolen ; oder mjrthische Vorstellungen Tom „ ■ • I HI 1 1 , \ 30 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. predominance which the Coptic gives to prefixes in the for- mation of words, and infers from this, among other eviden- ces, that the Semitic holds an intermediate position between that language and the Indo-European. He ascribes to this strong inclination for prefixes in the Semitic dialects the absence of terminal inflections in the nouns, or of cases, properly so called.* Yet from the circumstance that such elementary inflections as those that express person, gender, and number are formed through affixes, he assumes this to have been the original principle of formation. On this ho rests one of his pleas for the acknowledgment of an original affinity with the Indo-European stock.'' Not a very strong case, surely. Yet when we consider the intermediary rela- tions which the Semitic seems to bear to the Aryan and the Coptic, the presumption upon this ground does not seem worthy of being slighted altogether. "We need, however, to look a little more closely into the structure of such forms in the respective types of language. When we examine an Aryan word, and arrive at what is considered the root, we find that the latter is transferred to a a derivative or to an inflected form without internal modificar tion. In all cases, certainly, the principle is clear that the parts of the root are inseparable, and that its vowel as well as consonantal elements must enter into the combination. But the Semitic principle is totally different. The consonants XJrsprunge der Volker u. Sprachen. Nebst Benrtheilung der zwei sprachwissen- schaftlichen Abhandlungen Heinrich von Ewald's, Lemgo u. Detmold, 1863. Although Professor Pott made an efibctire presentation of the mor<) obvious difficulties of Ewald's system of comparison, neither his arguments nor ours have any tendency to lesson the merit of the permanently valuable portion of the treatise, in which, starting from fundamental principles common to both families (which appear to us probable, though to him as scientifically estab- lished), he has traced with unsurpassed penetration and ingenuity the structural development of the two idioms. i The accusative and genitive in Arabic, and the accnsative in Ethiopic bear no true analogy to the cases of like appellation in the Aryan tongues. On the indefiniteness of the like endings in Assyrian, see Schrader, Assyriach-Babylon- ische Eellinschriftcn, p. 230 ff. 3 Comp. \ 107 c. in his Ausf. hebr. Sprachlehre. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 31 wliicli form the root v.x stem, while remaining themselves unchanged in their new relation, arc separable, and may admit between them any of the whole stock of vowel sounds. Each of them, in fact, seems to be the centre of functional activity for itself within a certain range. Now, this diver- gence from the Aryan system seems to be even more radical than would be the assumed primitive correspondence in formative methods which we have just considered. It seems to be nearer the sources of the individual life in each system of speech, and therefore to be a more important element in determining their early relations. Thus we find that while from one plausible analogy we would be led to hope that a bond of union had been discovered, we are warned by a more searching analysis that the breach is wider than we had thought.^ From this one point of view, therefore, we seem compelled to abandon the expectation of proving a structural relation- ship, and unless stronger evidence is forthcoming from other 1 Ewald does not seem to have recognized this necessary priority of more essential to more formal characteristics in these languages. He thinks that the formative elements in the Semitic family, where prefix and affix were both em- ployed, largely determined the principles of " inner mutation in the roots " (Zweite sprachw. Abhandlnng, p. 64). He says that these appendages, press- ing equally before and behind, tended at last to force their way into the body of the root, thus favoring the internal play of the vowels as modifying elements. To this, he adds, the original divisibility of the root lent its influence. We would suggest that the relations between the formal appendages and the inter- nal structare of the word are as follows : — The greater freedom in the location of these appendages in the Semitic words is n secondary influence, due to the independent existence assigned to each radical of the triliteral root, so that not the whole body, but the individual members decide the place of the external additions. Hence, while in the Aryan languages the influence of analogy would of itself be sufficient to cause these appendages to appear uniformly at the place first chosen, namely at the end, the same tendency could not be equally felt in the Semitic vocables ; for each letter would assert its autonomy, and claim its rightful share of the tributary elements. Naturally the force of the middle radical was kept in abeyance by the two others, one on each border. But that this was due merely to the exigencies of its position, and not to its own quies- cence, may be inferred from the fact that in the most highly developed of the Semitic tongues — Arabic and Ethiopic — this letter assumed a pow. ?>2 IJRLATIONS OF THE ARYAN AND SF3IITIC LANGUAGES. sources, we must only fall back upon the hope of establishing an ante inflectional affinity. Wo have now to inquire whether there is anything in the syntactical features of the two forms of speech to justify us in holding to a radical affinity between them. This task seems even less promising than the one just attemiitcd. The general aspect of the Semitic mode of expression seems to have nothing whatever in common with the typical character of an Aryan sentence. They are as divergent as the mental characteristics of the two families of which they are the expression. The thought in any given case seems to be cast in entirely different moulds.^ In the Semitic period we are struck with the absence of qualifying and sub- ordinate clauses ; its parts are simply co-ordinated. There is nothing complex in its structure ; all is simple and direct, both in the construction of the members of the sentence and in the arrangement of its words. The specific distinctions of importance are, the relative positions assigned in each to the subject and the predicate, the modes in which the sentences are united, and the ways in which they express the relation of dependent words. Now, the same difficulty meets us in this comparison as that which we encountered in considering the structure of verbal forms : as far back as we are able to trace the two idioms we find that they have preserved essen- tially the same modes of expression. Thus it is character- istic of the Semitic syntax, throughout its history, that in the ordinary, direct, simple sentence the verb precedes and the subject follows ; while in the Aryan languages the re- verse order is as prevailingly the rule. It may be surmised that the actual order in the Semitic idiom was not the original one, and that there, as in the Aryan sentence, the subject, as being the leading word, was in earliest times placed first. But this is incapable of proof. Ewald insti- tutes a subtle parallel ^ between supposed changes in the 1 The cardinal distinctions are delicately discriminated by Renan, Histoire g^n^rale, etc., p. 19 ff. * Zweite sprachw. Abhandlnng, p. 57 ; comp. p. 28f. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 83 verbal and in the syntactical structure of the Semitic language. Ho believes, as we have seen, that the formative elements in Semitic words were originally placed at the '^nd, and that the principle of prefixing them was of later origin. He then affirms that in conformity with this process there was an early but gradual change in the order of the parts in the sentence, so that what seems to us to be the natural arrange- ment was inverted.* The same ill-success seems inevitable in examining another leading distinction. The mode in which a dependent is joined to a governing noun in the Semitic, and which is found in all its dialects, bears no analogy to anything known in pure Aryan grammar.^ That the first of the nouns should be modified, instead of the limiting one, ip a principle essen- tially Semitic. Whatever may have been the origin of this construction ; whether or not the vowel termination of the construct state, which is universal in Ethiopio, and has sur vived besides in archaic forms in Hebrew,^ was the original bond of union between the words so related, the impossibility still remains of bridging over f i linguistic interval between this and the Aryan usage, according to which, the second or limiting noun must undergo inflection, or be governed by a preposition. With regard to the third leading distinction in the sphere of the syntax, we think that the simple co-ordinated structure of the Semitic sentence with the prevailing use of merely copulative particles, is not so radical or so inherent in the system as to furnish even the external conditions of linguis- tic comparison. It is due, as it appears to us, almost entirely 1 That the Indo-European order is the most natural may be inferred fh>m such primitive types of language as the Chinese. See Max Miiller, Science of Language, i. p. 118. 3 The employment of a similar construction in modem Persian, and in Arme- nian, being a usage borrowed from the Semitic, is no exception to this rule, any more than is the tendency to separate the letters of a word by the insertion of a Towel, which is shown sometimes in ihe first-named language, and has the same source. * For opinions as to the origin of this termination, see Green, Heb. Gram 4 198 a ; Ewald, Ausf. hebr. Spl. {211 a. 84 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. to the intellectual character of the people at the formative perioclH of their language. The Semites, as a race, have not been given to haliits of reflection or to logical reasoning, deligiiting rather in the contemplation of the external fea- tures of the objects of sense and the more lively emotions of the soul. IJcncc the absence of inferences, of close defi- nitions, and of special qualifications. The discursive faculty was but little employed, and required no special instrument for its expression. 1 But the comparison of the two idioms in this spliere would soon lead us from the study of the lan- guage to the study of the races themselves, and take us beyond our province. Having thus attempted to outline a system of structural comparison between the two families of speech, it remains for us to sum up the meagre, yet instructive, results of our inquiry. 1. The two families are conspicuous among the languages of the world, through the possession of fully developed in- flectional systems, as distinguished from the idioms called agglutinative, isolating, polysynthetic, and partly inflectional. * The early inversion of the natural order of the elements of the simple sen- tence may have contributed its influence to the formation of Semitic style, as Ewald maintains (Zwcito sprachw. Abh., p. 59), but probably only to a slight degree. Pott seems to be in error when, in criticizing Ewald, he says (Anti- Eaulcn, p. 281), that the brevity and uniformity of the Semitic sentence are due to the paucity of adaptable conjunctions, and of moods and tenses, which would subserve a like end. For, if wc look merely at Ethiopic, a Semitic dia- lect which does possess a marvellous capacity for the expression of logical and connected thought, wc see that it possesses those grammatical elements to the requisite amount. The inference is then near at hand, that, at the time of its growth into a distinct language, these parts of speech were evolved from its quickened resources, in order to servo the purposes of an exceptionally active intellectual life among the people ; there being also no doubt that much mental activity did once exist. See Dillmann, Aethiop. Gramm., p. 6 f. ; Ewald, Ausf. hcbr. Sprachlchrc, p. 34 f. This conclusion, as confirmed to a certain extent by the history of the Arabic, would go to show that the Semitic type of expression was conditioned by the mental antecedents of the race, and not by an inherent inadequacy of the language. Of course, when the cruder dialects became old and fixed, they lost the capacity of development, and when employed for unac- customed purposes, had to borrow the necessary expressions from foreign idioms, as is proved from the history of Aramaic and Talmudic Hebrew. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 35 2. Without considering the question whether what aro ordinarily called roots in the Semitic dialects aro really ultimate significant elements, it is i)lain that the bases of verbal forms in the two families are essentially distinct in their structural principles. This dissimilarity is marked not simply in the phenomenon that in the Semitic idiom they aro generally composed of three consonants, but more fun- damentally, in the independent activity assigned to each of these letters. 8. With regard to the formative elements of living words, wo saw that there was some reason to believe that in the most essential, and presumably the most primitive, of inflected forms, they were attached at the end of the roots, as in the Aryan languages. This, however, does not furnish, by itself, a very strong argument in favor of a grammatical affinity. 4. The syntactical peculiarities of the two systems, as would naturally be expected, do not yield more favorable results, following, as they do, upon structural principles themselves divergent. We are thus left without any direct demonstration of re- lationship from this source of evidence. The question then recurs : What, if any, is the residuum of testimony, from a structural comparison, in favor of the theory of the original unity of the two systems ? It is to be feared that no answer, universally satisfactory, can be given. In some minds the common possession of an inflectional system would of itself create a strong presumption of an identity of origin. And when to this fact is added what has been alluded to with regard to the intermediate position of the North African family of languages, whose inflections hardly rise to the dignity of a system, but betray, when they do exist, a marked resemblance to the Semitic, the inference seems proper that the families last named went hand in hand in the earliest stages of their history, and after their separation followed in very different degrees the structural impulses which all three idioms had received in a common home. Bnt apart \n. ti i' 86 RELATIONS OF THE ARYAN AND SEMITIC LANOUAOES. i! from this, and on j^enoral lin^istlc considerations, it docs not Hccm likely that two such highly and fully developed systems of speech would have originated without a strong, even thoun^h very early, bond of relationship. They repre- sent a supremely great achievement of the human mind, something unique in the history of men ; and one is led to attribute a common impulse to the beginnings of each, as in the contemplation of the worship of the synagogue and of the cathedral we arc led back to the one supreme religious idea that the world has known. The theory of an original diversity in the two families appears, in fact, to raise a moro formidable difficulty than those which the doctrine of their unity occasions, because the psychological phenomenon which it would imply is less credible than the assumption of a divergence from a common idiom, which, before the separa- tion, contained the germs of a grammatical system. Yet this kind of evidence is both too general and too sub- jective to command universal assent. At best it affords a presumption, and not a demonstration. Although, therefore, we think that the two families of speech were still united when the first manifestations of the inflective impulse were felt, yet, as we have very little scientific proof to present, based upon grammatical comparison, it is only left to us to see whether there is not another kind of evidence available in the inquiry. We are thus led to compare the verbal forms possessed by the two families, and thence to determine whether analogies between separate words are obtainable in sufficient number to justify us in regarding them as something more than mere coincidences. But at the outset we are confronted by arguments urged against the admissibility of such evidence by those who hold that the two idioms are radically distinct. It will be necessary to test the validity of such objections before proceeding further. We are first met with the general plea that, as grammatical features are the proper marks of linguistic relationship, it is unscientific as well as futile to go behind them, and to com- RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 87 pare tho lexical contents of the two groups.^ This declara- tion is sweeping and iniix;rious. Against any plausiljlo coincidences already brought forward it is always \irged that they must bo tho result of chance or of onomatopoeia, or of some subtle intellectual analogy in the formative processes of early speech. Against those who make any systematic attempt to compare the two idioms on tho basis of their respective vocabularies it is maintained that they begin at the wrong end. The failure of Bopp in his attempt to compare the Indo-Euroix)an with tho Caucasian and Malayo- Polyneslan families of speech is paraded ^ as a proof of the exclusive sulTiciency of the method of grammatical compari- son, of which ho had been the originator and expounder. Now, before considering the special difficulties raised by these theorists in tho way of adventurous and irreverent investiga- tors, we should say that these vehement protests against an alleged unscientific method are themselves not at all in the spirit of true science, inasmuch as, if universally heeded, they would stand in the way of all progress in the further com- parison of languages. A stop would at once be put to all efforts to co-ordinate into special families those language? of the so-called Turanian group, which agree only in the agglu- tinative or combinatory character, just as the Aryan and Semitic families agree in being inflectional. And so for the the classification of other types of human speech. It may also be assumed that if the same spirit had been dominant at the beginning of the present century, those bold but happy * So Renan, Friedrich Miiller, Sayce, and other opponents of tho theory of an original affinity. 2 Sco Friedrich Miillor, Grundriss dor Sprachwisscnschaft, I. Band (Vienna, 1876), p. 58. Comp. Bcnfcy, Gcschichte dcr Sprachwisscnschaft u. der orlcn- talischcn Philologie in Dcutschland. Munchcn, 1809, p. 511 ff. It is very likely that Bopp was inaccurate in many of his combinations with the above-mentioned languages ; but on this general question of the admissibility of verbal corapari sons, wo cannot but respect very highly the judgment of tho immortal foundei of Comparative Philology. Here, as in his Glossarium Sanscritum (within the Aryan family), ho was too hasty and liberal in tho admission of analogies. But this was due to his method in practice, and not necessarily to the unsoundness of his theory, into whoso conditions he probably saw as clearly and deeply u any dogmatic obstructionist of the present hour. li ■ I : J I 88 RELATIONS OF THE ARYAN AND SEMITIC LAN0UA0E9. goiioralizations without wliich, porlmps, comparative grammar itsi'll' might not have been croutt'il, would have been do- iiounccMi m unrtcientilic. The great discoveries within tlie sphere uf the Indo-European family have made it faHhionahle to believe that glotto jgy liat* unfolded all its fundamental principles, while it is forgotten that only small districts ol hiunan speech have been explored and annexed to the domain of science. The reaction against the old lawless methods of comparison which now prevails is no doubt wholesome and just; but it is a «jucstion whether this ono of its present forms ought, or is likely, to be jKn-mancnt. JJnt, more particularly, it is alleged that wo arc bound to forego any attempt to assimilate the two groups, because (it is said) science has established the fact that the various types of speech now known rest upon a primitive diversity of origin — that language was develoi)cd at first from num- berless dialects, and not from a commo»i source. Now if this dictum were conceded to bo indisp ly true, it would not settle the question at issue ; for wc -Id next have to determine what constitutes the primitive type in any given case ; in other words, whether the two inflectional families of the world's sjieech may not have arisen from one original dialect. Such an issue is not necessarily excluded by the conditions of the supposed fact of linguistic history. For the limits of each early typo or dialect must be settled in ono or both of two ways : by appealing either to the evidence of the science of language, or to that of comparative ethnology. If we refer to the former, we find this at least, that these two families are the only ones that have a fully developed inflectional system ; a fact suggestive of a possible primitive bond between them. If we appeal to the latter, the evidence is decidedly unfavorable to those who maintain a diversity of origin. The Semite differs but little physically from the Aryan, and resembles the European more than the latter does a Hindoo. This is acknowledged by Renan, one of the most influential of the class of writers alluded to, who admits that the current distinction is based chiefly upon language, and RF.LATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 80 afTirniH tlmt, vicvvcil from tho phyHical Hide, tho Scmito and the Indo-EuroiKJun form but one raco.^ Tho ('onHulcrutioii that tho two HyMtemH of niKJceh toj^clher now occupy ho much of tho carth'rt Hurfucc does not eomo into conlUct wilh tho QHSumptiona of the theory wo arc conHiderinj? ; aH 1 hough tlio doctrine necessarily involved a certain ratio Ixilwccn tlio primitive extent of a language and tho numluM' of its present sfieakers. It is only maintained that tho orii^inal dialects of mankind wero numerous and diverse, it heinj;; an ossontial part of tho theory that but comparatively few of the early stock now survive, the rest having? been rliiniuated in tho strujrglo for existence. It should also be romcnibcred that, so far as wo can judge, tho primitive Aryans and Semites must have comprised only a relatively small ])ortion of tho earth's inhabitants, and that it was their inherent intellectual and moral superiority that secured their gradual progress, and their survl.al of tho vast civilizations that preceded them. Ilenco wo see that no real advantage would be lost if the theory of tho original multiplicity of language could bo proved. Still, as it might seem to justify a presumption that each present great division of human speech had a separate beginning, it may bo proper to say a few words upon tho subject of its pretensions. Those who maintain this polygenetic theory of language are usually disbelievers in the doctrine of the common origin of mankind. But we do not need to assume that they are pre- judiced to any extent, by their views upon the latter question, formed upon other grounds than tho results of linguistic research. Some eminent linguistic scholars think that tho final decision of the question as to the original unity or diversity of language rests with physical science.^ Others maintain that ethnology and the science of language should not be mixed up together.' However this may be, we have 1 DeTorigino der langago (4th cd.). Paris, 1864, pp. 204, 208. ' E.g. Bcnfcy, Gcschichto d. Sprachwissenschaft in Deutschland, p. 789 f. ' E.g. Max Miillcr, Science of Language, i. p. 326 f. 40 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. now to consider simply the worth of the linguistic proof which the advocates of the theory of a primitive diversity of dialects have to offer.* The argument upon which reliance is chiefly placed may be stated as follows : — Although it is natural to the human mind to seek for and to expect unity of origin in all forms of existing things, the facts of linguistic history point us to an opposite conclusion with regard to the development of language. It is a fact that widely-spread idioms owe their predominance to the influence of civilization ; that if we turn to savage tribes (among whom are certainly to be sought traces of the earliest modes of Nature's workings), we find an endless diversity of dialects, each village, sometimes, having an idiom of its own ; that if we go back to the ear- liest records of written speech, we see the same conditions exemplified, as in ancient compared with modern Greece ; and that a number of subordinate considerations (which we cannot here adduce) strengthen and illustrate the position thus assumed. Since, therefore, as far back as we can go in the history of language we meet the same diversity as at present, or even a greater, it is only in accordance with the methods of science to conclude that it was always so.^ But surely it is only scientific to draw like inferences from like conditions. It is surely a perilous assumption to regard the conditions of the formative periods of language as analogous to those of its historical progress in the latest ages of the earth. Apart from the peculiar physical and psychological factors that must have entered into the forma- tion of early speech for a long period, there is one possible 1 The theory is maintained elaborately by Sayce, Principles of Comp. Phil- ology, chap, iii., " Idolum of primeval centres of Language " ; Ilcnan, Orig. da Lang. chap. viii. ; Hist, gdne'rale des langues S^mitiques, p.93ff.;cf.Pott,Un- gleichhcit menschlichcr Rassen vom sprachwiss. Standpunkte, p.SOlf. Fr. Miiller, Grundriss dor Sprachwisscnschaft, p. 50 ff. A neat statement of the general position is given by Schleicher, Compendium d. vergleich. Grammtik d. indogermanischcn Sprachcn, 1866, p. 2f. 2 On the origin and growth of dialectical differences in contravention of the above general theory, see Whitney, Language and the Study of Language, p. 177 ff., and in American Journal of Philology, 1880, p. 341. RELATIONS OF THE ARTAN AND SEMITIC LANGUAGES. 41 difference of vital importance which is assumed not to have existed. It is regarded as an unquestionable fact that lan- guage could only have arisen when mankind had become very numerous and scattered. Passages might be cited from bordinato arguments employed by these scholars involve the same fallacy. Thus Rcnan (Orig. du langage, p. 1 77 ff. ), lays great stress upon tho fact that the terms employed by early tribes to designate their neighbors were usually dcrivrd from some notion implying the unintelligibleness of their language, they being usually styled " stammerers," " dummies," or some other Bueh unsocial designations. Ho cites in confirmation such words as the Ger- man Wdh (Welsh), the Sanskrit Mlecelia (supposed to be cognate with the former), the Greek Aglossoi and Barbaroi, the Abyssinian Timtim. He then pro- ceeds to argue that language must have been originally divided no less imi assa- bly. On this it is obvious to remark that we do not know whether these terms in all languages did not arise after the diverging dialects had become mutually unintelligible from familiar causes. Further, many of tho cases are taken from within tho Aryan family ; and it is now certain that there was once a time when all those who used that idiom could make themselves mutually understood. To this opinion Renan himself elsewhere (op. cit. p. 49 ff.), professes his adherence. RELATIONS OF THE ARTAN AND SEMITIC LANGUAGES. 43 of the parts of speech, and which attains a perfectly adequate capacity of expression, merely through the relative position of the words, and the use of a small number of particles. But there are some who would forbid us to assume such a hypothetical Aryo-Semitic type of language, and who main- tain strenuously that it is both improbable and unexampled ; that it has no ground in linguistic philosophy, and no anal- ogy in the history of speech. It is maintained by them that no language has ever passed from an isolating stage (as above described) into an agglutinative or combinatory, and none from either of these into an inflectional. Probably the strongest assertion of this dogma has been made by E. Kenan and A. H. Sayce, in their works already cited. The question is so vitally important to our discussion, that it demands a serious, though necessarily a brief, consideration. We shall therefore present the best evidence we can in favor of the theory of the development of each of the families from a more primitive type, considering the opinions and objections of opposing theorists as they may occur to us in connection, with different points in our argument. Our theory as to the divarication of the two families rescs upon the doctrine that every inflectional language must have passed through a simpler combinatory stage (of longer or shorter duration), which itself arose from an original iso- lating type. In our grammatical comparison of the two systems we did not think it necessary to discriminate be- tween the first two stages, both because in these languages the combinatory period appears to have been comparatively brief, and because the structural divergences seemed so radical as to exclude the probability of a common form of speech after the process of combination had once begun.^ The evidence for this may be gathered from what lias been said of the modes in which the formative elements of full- grown words are attached in each group, as well as of the differences in their internal structure. We have to go right back to the most simple and priiuitive type of language, ^ Comp. Max MUlIer, Bede Lecture on the Stratification of Language, Chips from a German Workshop (Eng. ed.), ir. p. 102. i' : 44 RELATIONS OF THE ART AN AND SEMITIO LANGUAGES. and we tliiiik the step may be justified demonstrably by proof that each system has been developed from a more rudimentary condition. As to the psychological causes which led to the adoption of the more complex forms of expression, we admit that they are to a large extent mysterious, but claim that they are not without historical exemplification. As to the occasions which led to the perpetuation of each system, after its origin, we hold that they at-c easily discoverable, and are being constantly repeated in the history of human speech. We would remark, first, that we have an exhibition of tendencies in many languages which clearly reveal the possi- bility of such development. It is said, however, that tl^ere is no instance of a clear transition from one state to another. Certainly there is not ; nor have we any right to expect that, after the forms of a language have been hardened through the course of ages, they could be changed easily and speedily. We do not claim, however, that any language has made this decisive transition under conditions similar to those with which we are now familiar. But it is manifest that in the early state of every form of speech, the possibili- ties of such a serious change were immeasurably greater. In those times men were seeking after suitable forms of ex- pression, not having at hand any that had been gradually worked up into a familiar and adequate instrument of thought. One class of them would attempt, by various de- vices, to perfect, without radical change, the primitive rudimentary type, a task in which they succeeded admirably, as we learn from the adaptability of the Chinese to an un- limited range of ideas. Others would adopt the expedient of combining their roots; and this idea was carried out apparently in two main directions. Among the founders of the so-called agglutinative languages, predicative roots were modified (so far as we can determine) generally by other nominal and verbal forms ; while the pioneers of inflectional speech made as decided a choice of demonstrative or pro- nominal roots to accomplish a similar end. In the former case, since both elements of the new compound stood out RELATIONS OP THE ARYAN AND SEMITIC LANGUAGES. 45 |: with equal prominence, they would naturally retain their former importance, and oj^pose persistently the inevitable tendency to phonetic corruption ; while in the latter tho comparative unimportance of the determinative elements would subject them to the predominance of the radical por- tion, their individuality would, after a time, become lost in the consciousness of the speakers, and phonetic decay having one begun, the process would soon extend itself to the whole body of the word. So much for the general process by which these complex systems were educed from the primitive condition of sim- plicity. The force which operated in each system to produce uniformity of structural type throughout its whole extent must have been chiefly the powerful influence of analogy. How potent this was in early times we may infer from its power even within historical periods, as we learn from the development of varied forms in such idioms as the Romanic languages, and most conspicuously, perhaps, in the dialects of France. And we maintain that the possibility of a tran- sition from the isolating to a combinatory stage in early ages, oug^t not to be more difficult of conception than tho change which has actually taken place in the development of the modern analytic out of the ancient synthetic languages. We must remember that men were groping after more com- plete and satisfactory modes of expression. They had not yet lost the spontaneity of primeval speech, and with an inherent, almost creative, facility they could achieve without reflection that which, to us, would seem to involve a radical intellectual change. When the superior fitness of the new principle of formation was once percei\ed, the whole family in which the change began would assimilate its speech with equal readiness to the forms of the more deserving system. The condition of things was very different after these ag- gressive principles became dominant. Each family, having moulded for itself a suitable instrument of thought, then possessed it. It did not seek any other, since it did not feel the need of it. Hence, we do not find in the acces- iH I 40 HKLATIONS OF THE ARYAN AND SEMITIC LANGUAGES. Bible forms of language, the very earliest of which is much later than the period we are describing as essential to the development of each family of speech,^ any instance of a comi)lclc transition from one type to another; nor should we expect it. The faculty of language is drawn upon only at need. It does not even furnish new words, unless these are required for the expression of new ideas ; much lesa should we look for the creation of new grammatical cate- gories without necessity. Yet we do find languages, some of whose features seem inexplicable on any other theory than the one we are advocating. We have such idioms as the Finnish, which are almost as much inflectional as aggluti- nativc.2 We have that most puzzling of languages, the an- cient Egyptian, about which scholars hesitate to say whether it should be called isolating, agglutinative, or inflectional.' But of more importance than these facts are the peculiaritiea of some of the languages classed as isolating, such as those of Thibet and Siam, which partake largely of the com- binatory character, while the Chinese itself, in some of its forms, exhibits a marked tendency in the same direction. If such mutability is manifested in languages checked in growth and fixed in general type through age, tradition, and usage, what must have been the capacity of radical change inherent in the earliest forms of speech, with all their sim- plicity and vagueness ! Our next argument is based upon the fact that an exami- 1 It will be seen from what has been said that we consider all languages, from isolating to inflectional, to have undergone this, so to speak, subjective process of development. We must not make the mistake of assuming that all languages have started from just such a state as that now represented by the Chinese. This language itself must have passed through important changes in modes of expression before assuming its present condition. It is not a primeval lan- guage, but only a more primitive type of language than those familiar to us. A study of its system would show that it presents the result of a considerable psychological development. ^ The approximation of agglutinative to inflectional idioms is of secondary though considerable importance. The psychological interval between these conditions is not nearly so great as that between the isolating and the com binatory stages. ' Comp. Whitney, p. 342 f. ; Benan, Eistoire g^ntfrale, p. 83 £F. RELATIONS OP THE ARYAN AND SEMITIC LANGtJAGES. 47 nation of fully-formed words in Aryan and Seuiitic speech attests the doctrine that they are ultimately duo to the accretion of originally independent forms. Tlie determina- tive elements added to the roots have been ascertained in a vast number of cases, and shown to possess a significance of their own. The natural assumption is, that the same is true of all the original compounds. In the Semitic family, where the process of analysis is peculiarly easy, this conclusion may almost be taken for granted. But the advocates of the opposite theory prefer to consider the Aryan languages, where, confessedly, there is much more that is obscure in the ultimate constitution of some of the more primitive forms. Even with regard to these, however, the same pre- sumption is probable. We are told,^ indeed, that as far back as wo can trace the Aryan languages they are inflec- tional, and, beyond that, they must be remitted to the prov- ince of physical science, which, as we are told with great confidence, could only prove that the brain of the earliest Aryan was capable of originating no other type of language. But surely this is claiming too much. Inductive reasoning has surely something to offer on the opposite side. While explanations of forms hitherto obscure are continually being made, we feel a strong presumption that if we could only penetrate the misc through which the opening dawn of Aryan speech is faintly discernible, all that remains myste- rious would yet be brought to light. If these elements are always significant, it would be certain to the ordinary mind that they were once used independently — a conclusion which would establish our theory. Such a conclusion, it may be said, is only an inference from a partial analysis, and not a demonstration based upon the working of a universal principle. Even if this were to be conceded, there is another way of considering, the general question which leads to the same result. It may be shown that the opposite theory is psychologically inconceivable. The formative elements were originally significant, or they > Sayce, Principles of Comparative Philology, p. 158. |1!1 vf 48 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. were not. If they were significant, they were previously independent vocables. If they were not significant, how account for their employment as determinative symbols in the earliest attempts of the race to achieve an intelligible method of oral communication ? Now, it is maintained (by Prof. Sayce) that although (as proved) later forms in these languages arose through the attachment of sitrnificant terms, or fragments of these, yet the example of inflection in the earliest i)criods was set in the creation of forms which conveyed in one single word both the fundamental and the modifying idea, the latter being expressed by " unmeaning terminations." ^ Thereafter, as the needs of the languages demanded, the p* ogress would be easy to the attachment of significant terms. Which of these two theories has the greater inherent probability may appear from a candid pre- sentation of the assumptions demanded by each. According to the one theory, at the very birth of these languages, when, as we are bound to assume, men were just accomplishing the task of giving forth in sound intelligible signs for the objects of nature and the simplest qualities and actions, we are to believe that they expressed the various relations of these by attaching to the phonetic expression of the root-idea (which must itself have been held on precarious probation) any one of a number of mere grammatical symbols, these having no existence save in such combination. It is natural to suppose that the earliest efforts of speech were, at best, not very easily understood, and that at least the relations between various objects would at first have to be indicated by various contrivances, such as gestures or other outward signs. But to attempt to express such relations by drawing, on occasion, upon a number of arbitrary (since not signifi- cant) sounds, would have tended very much to discourage incipient vocal communication. The other theory assumes » Op. cit, p. 151. The words are evidently equivalen*, to "suffixes of little meaning " (p. 145, note). The use r ' the latter phrase may show how difficult it is to conceive of the growth of inflection by the attachment of unmeaning sounds to the root. In Prof. Sayce's Introduction to the Science of Language, 1880, 1, p. 85, cf. p. 119, a similar theory of Ludwig (Agglutination oder Adap- II RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES 49 thai at an early period, though not the earliest, of a given inflectional language, terms which had already grown faniilinr to the 8{K;aker8, gradually came to have their various rela- tions expressed by the eomljination with them of other words which were already accepted vocables ; that at first those of early origin and of most frequent usage, such as demonstra- tive particles, were employed ; that thereafter, as the circle of ideas widened, more special expressions came into use ; and that in course of time, the sense of the independence of the two elements being lost, the word became one indivisible form in the popular consciousness. The choice lies between these two hypotheses, and onlythese ; and hesitation between them does not, antecedently, seem possible. But a very plausible argument is presented, to the effect that the farther back we go in the history of inflectional languages, the greater complexity of structure is to be found, while their tendency always has been, and still is, to greater simplicity, and wo are therefore to assume that the primary types of expression were synthetic. Here again there is a fallacy, due to the failure to pass from the ob- served facts of accessible forms of language to the necessary conditions of its early development. The assertion that in- flectional languages are continually becoming more analytic in their structure is based upon the phenomena of idioms that have received a literary cultivation, analysis being the necessary accompaniment of reflection, and the result of a self-conscious endeavor to attain greater simplicity and clear- ness of expression. Yet it may readily be conceded that back to a very remote period in the history of any such lan- guage the assumed conditions did exist. But the argument is valid only against any who might claim that throughout the progress of such an idiom a tendency to greater com- plexity prevailed. This, however, is not the position main- tained here at all ; for a multiplicity of complex forms is tation, 1873), is cited and supported by the researches of M. Bergaigne into the nature of the Aryan case suffixes. In this instance he acknowledges more fully the difficulties attending both theories. 50 RELATIONS OF THE ARYAN AND SEMITIC LAN0UAQE8. ju8t what we would expect to have happened after the combi- natory iuipulso l)egan to manifest itself, in accordance with what wo know of the general diversity and confusion of early efforts at lungiiage-malcing. Afterwards when any language i)ecame fixed in its structural type, and was much employed in the expression of manifold thought, the simplifying process was equally inevitable. A more particular form of the same general objection to the root-theory has yet to be considered, and in it an extreme seems to have been reached in the way of crude philosophizing. We are told that language begins with sentences, not with words ; that an idea .annot bo communicated by the use of single words, and that even in the most primitive utterances of men such single terms had to be eked out by gestures or other signs so as to convey the ideas intended to be ex- pressed ; that the form in which such utterances were made characterized "jach linguistic type, and was perpetuated un- changeably in the development of the language ; that the sentence is the unit of significant speech, and it is therefore evident that all individual words must once have been sen- tences ; that the student of language therefore cannot deal with words apart from sentences.^ Many considerations oppose this reasoning, any one of which is fatal to its sweeping con- clusions. In the first place, even if it is admitted that spoken language can never consist of the use of a mere word without some form of predication concerning it, it does not follow that such a form is permanent from the first, and becomes crystallized about the word with its earliest utter- ance. On the contrary, since we know that the first means whereby men conveyed their ideas about objects, or the qualities of objects, must have been . the employment of some kind of outward sign apart from the words that ex- pressed those objects or qualities, these mechanical symbols of gesture, tone, and so forth, must necessarily have varied with the habits and genius of each community, while the names of the objects or qualities, once settled upon, would 1 Sayce, Introdiiction to the Science of Language, Vol. i. p. II1-II6. RELATIONS OF THE ARYAN AND SEMITIC LAN0UAQE8. 51 ( l>ocoino more permanently held in their essential phonetio representation. Such words, being conceptual, were jKsr- |)otuated, their permanence being derived from the intel- lectual judgment that established them. The supplementary elements in the primitive utterance varied with each group of Bpcukers or each community that helped to popularize and extend the much needed vocables ; these demonstrative ex- pressions being spontaneous, natural, and easily understood, were not jxjrmanent just because they were variable. It is unnecessary to point out how the reasoning employed is out of harmony with what is observed of all organization either in nature or in human history. The elements of the assumed " sentences " are all before us, each of them a separate entity ; but the theory denies that there was any synthesis in their combination. It is as unphilosophical to assert that words could never have had an independent origin and history because in actual speech they are always found organized into sentences, as it would be to maintain that oxygen or nitrogen never had a separate existence because they are regularly found in definite combinations. The main fal- lacy, however, lies in the abuse of the term " sentence," as a grammatical category, in its application to the simple utter- ances of the makers of language. The stereotyped forms of fully developed speech could not possibly have been repre- sented in such primitive expressions. If it is said that every utterance implies a sentence ; we deny the statement, if the implication is that every utterance is capable of a formal grammatical analysis ; for an intelligible expression can be made by the use of but a single word. When it is said that every such word would need to be accompanied by signs to indicate its bearing or special use, we reply that such signs as gesture, tone, and facial expression are not language at all, that is, not human speech ; and with anything beyond this science has nothing to do.^ * Even taking these theorists on their own ground we can find much that proves the root-theory as against the sentence-theory. Thus we know that in a simple lentence the copula is of late origin in all langaages, being osoally an 52 RELATIONS OP THE ARYAN AND SEMITIC LANGUAOKS. If' I 1 1 These olmcrvations, whicli ore all we have Hpaco for here, will 8how how little reason thorn is for accc))tii)^ the dicta of Ueiian ' that '' lan^agCH isHiio ready made from the mould of the huniuii mind," and that linguistlo** familieo appear as C8tahliHhed types once for all." ailnptation of the late r.iotaphyxical conception, to be or exiNt ; and that in luch wiilvly Hupnrutud idiunm hm Ilehruw and Baniiirit, as well as In many othon, in btu'h cxprcMHiuna as " this is a troo," the copula was primarily not employed, the form l>cinK> " this a tree." Such an example shows how near even highly de« velopcd tonguut) still lay to the source of their individual life ; and when we add to thcHo couHidurations the fact of the ambi||;uity in the use of demonstrative pronouns in the curly literary stages of such languages, the same example points us almost directly to a " sentence " and a word in one. * Origino du langage, pp. tt9, 116. CHAPTER III. COMPARATIVE PHONOLOGY. In dealing with Aryan and Semitic sounds as they come up for comparison, three questions present themselves in the following order. The first is : Does a marked difference in the current phonetic stock of the two families properly pre- clude all discussion of their ultimate identity ? The second is : Will a fair examination of the sounds of the two idioms result in showing that the dissimilar elements have arisen in their respective systems from more primary sounds ? The third, which is entirely distinct from the second, is : How do the Aryan and Semitic sounds represent one another in the accessible forms of hypothetical Aryo-Semitic speech ? The first of these questions was answered in the last chapter, where it was shown that sounds are not a primary criterion of relationship. The answer to the other two questions will be given in the present chapter. Our first task will accordingly be to take up the contents of the Aryan and Semitic alphabets, eliminate the sounds which may be proved to be secondary, and thus reduce them inde- pendently to their primary limits. Two practical results will thus be gained : we shall be able to determine what were the Proto-Aryan and Proto-Semitic sounds in which their earli- est vocables were clothed ; and we shall be able to reduce to its primary form any root that may come up for comparison containing sounds that are proved to be secondary. Of course the present discussion has nothing to do with bring- u f: 54 RELATIONS OF THE ARfAN AND SEMITIC LANGUAGES. ing together the sounds of the two systems, and determining whether they corresponded to one another in actual Proto- Aryan and Proto-Semitic speech. That is a question to be settled apart. At present we have to take up the phonetic repertory of each family and reduce it to its primary limits irrespectively of other considerations. In this endeavor the chief work will have to be done in the Semitic department. Aryan phonology has progressed so rapidly and surely, in keeping with Aryan etymology, that although there is still dispute on some points of minor importance there will be no great difficulty in presenting a correct working scheme of ultimate Aryan sounds. It is hoped that the attempt will be equally successful with the Semitic alphabet. The first class of sounds to claim our attention is the gut- turals. The development of these in the Semitic languages especially is remarkable, particularly in Arabic and Ethiopic. That these were net all employed from the very earliest stages of Semitic speech, but were gradually proc.uced in later times, can be made to appear at least very probable from the following considerations. In the first place we have the notorious fact that when we compare together roots which were undoubtedly Proto-Semitic, agreeing in other sounds but differing in their possessing different gutturals, an agreement or resemblance of meaning is shown in an im- mense number of cases. This seems to point to the conclu- sion that many of these forms were modif.cations of these synonymes through a variation of the guttural elements, a process which throws light on the production of such sounds in earliest Semitic times. Again we have the analogies prp- sented by other languages. Thus within the Aryan family, which started with no true guttural, these sounds have been variously and sometimes strongly developed, notably in the Keltic and Armenian branches. So also in some of the American dialects.^ In the next plpce, we must remember that in the growth of Semitic speech with its peculiar structure, » Prof. FnMeman in Proccediogs of the American Oriental Society, Oct. 1874 (Journal, Y(A. x. p. ciii^. '1i RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 55 it was inevitable in attempting to express the great variety of notions bred in the minds of an intellectual people, that they shou'd employ a greater variety of sounds than those with which they at first started. There was a two-fold inner necessity for ihis. First, the vowels could not be used in forming now roots among the Semites, but only in form- ing derivatives, or in expressing different aspects of the root-idea. Secondly, there was no compounding of words with prepositions or other modifying terms to express new relations or kindred notions. When the need for various expression was felt, resort must have been had unconsciously to the stock of consonants, from whose fundamentally distinct sounds there gradually arose variations, at first, perhaps, slightly, and finally quite strongly marked. Other causes no doubt conspired with these in each case of differentiation, and we think it probable that the strongest gutturals, such as are met with both within and without the Semitic family, were produced by those general influences, such as food, climate, and mode of life, which led to their development in the Armenian and both of the great Keltic dialects. But we think that these finer distinctions, peculiar to the Semitic, such as the Arabic and £, as well as some of the non- guttursl variants, were due not only to such occasions, but to those others which are peculiar to Semitic speech. Hen 3, as it appears to us, the immense range of consonantal expres- sion shown in the Semitic idiom, exceeding anything in the pure Aryan languages, even the Sanskrit,* some of whose sounds (the " cerebrals ") are possibly borrowed, and others mere euphonic variants. But, in the third place, however we may account for the variety of consonants, the fact of the gradual development of the different sounds does not rest entirely upon theory. We can trace the process of develop- ment in the later stages of development. The Aiabic p, is not found as a fixed independent sound in the other 1 Max Mailer's Science of Language (Am. ed.), p. 180, gives the number of current Sanskrit consonants as thirty-seven. J '! ^ : 1 ! 1 1 , '; .1; ■ % m % • : 3 •'Si 1,i n JC RELATIONS OF THE ARYAN AND SEMITIC LANGUAGE3. dialects, not oven in Ethiopic, which went hand in hand with it so long after the other dialects left the parent stock. We can see a tendency to its use in Hebrew, or rather a pi'onun- ciation of the 9 somewhat resembling it, since we find the s sometimes represented by the Greek 7 in proper names in the Septuagint, even in the middle of a word when it is usually not represented at all (e.g. piyfia for rrasi, Gen. x. 7). But this only shows how it was possible for the Arabs to develop an occasional into a fixed sound, ^ and so throws light upon the subject of the origin of the Semitic gutturals gen- erally. In Hebrew one character stood for both sounds, ard therefore we must assume that the divergence was of later origin than the invention of their alphabet. So with the n in Hebrew and its representatives in the northern Semitic dialects. The Arabic and Ethiopic made of this letter, which had a fluctuating, uncertain character in Hebrew, two dis- tinct unvarying sounds, for which they devised special char- acters, . ; ff\, H^. Looking at this tendency to multipli- cation of guttural sounds, which is so unmistakable in those languages which had the best scope for the development of their inherent capabilities — a tendency whose operations can be so easily traced ; and looking, on the other hand, at the liability to the reduction of those gutturals to the simple smooth and rough breathings which we find essentially in all languages, we naturally conclude that they were all gradually developed out of those primary sounds. That this is so is reduced almost to a certainty when we attempt to utter those sounds, and find that they are all distinctly related in two orders which have as close a relation to one another as d bears to t. The Arabic c and £ (=*) are developed from 1 Ayin, the most peculiar of the gutturals, seems to have had a tendency in two opposite directions after its origination, more marked than in any other of its class. The tendency to greater strength and variety wo see exemplified best in Arabic. The inclination to weakness and assimilation we see in the later history of all the other dialects, while in Assyrian it is only and always a mere vowel. In the Samaritan, Galilean, and Talmudic dialects, and in the later Phenician s took the place of 9. Later Ethiopic and Mandaite retained only the smooth ami rough breathings. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 57 I (=»), while and • are developed from s (=") ; the former order being just the sonants of the latter respectively. For the sounds in each order essentially the same organs are employed. The possible modifications in position may be illustrated by the use of the German c/t, or better still by the Welsh ch, as compared with the ordinary h. The [Kjculiarity of the Semitic pronunciation is, that it has brought out the ewith its surd more distinctly than any other language ; though; as Dr. Merkel tells us,^ an approach to the c or s is heard in German speech under certain circumstances. A more minute physiological analysis of these sounds than we can give here ^ would only confirm what we have said of the easy gradations of the Semitic gutturals, and of their development from the simple breathings. From all this it appears not only that the variety and peculiarity of these Semitic sounds offer no bar to a com- parison with other linguistic systems, but also that we have arrived at the same phonological level as that upon which the primitive Aryan breaths are found to stand. Let us look at the Aryan side of the equation for a moment. We find here that, so far as we can determine, they had only the spiritus lenis,^ not the spiritus asper. Th-s, however, does not prevent a final equalization of the sounds in question ; for the history of speech shows how soon the h was developed, as phonology shows how easily it arises and falls into disuse.^ Jf is really the surd of - (=**). If the organs remain in t'le position which they assume upon the pronunciation of ai V vowel at the beginning of a word, and if then we blow 1 C. L. Merkel's Physiologie der menschlichen Sprache (Leipzig, 1866), p. 74. * The reader is referred to Max Muller'a Science of Language (Am. ed.), Vol. ii. p. 148, and to the works alluded to in that chapter, particHlnrly to those of Lepsiua, Brucke, and Czermak ; also to the thorough and very able work of Merkel above cited. ^ Schleicher's Compendium d. vergl. Grammatik d. indogerm. Sprachen, 4. Auflagc (Weimar, 1876), p. 11. ♦ See the sounds of Zend, Old-Italian, Greek, Old-Irish, etc., in Schleicher's Compendium ; and compare the phenomena of the so-called Cockney speech, as well as the use or disuse of h in Modern French. I > I ■ f r n 58 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. instead of breathing, or, which is the same tiling, make a surd instead of a sonant sound, we shall have a light apiritus asper instead of the spiritus lenis.^ Wo have no doubt that the same thing was done by the Semites as by the Aryans, and that from the fundamental smooth breathing they also differentiated their h sound. From these, as we have seen, the surd and sonant orders of gutturals were thereafter developed. Hence we see nothing to prevent us from re- garding all the Semitic gutturals as comparable with the spiritus lenis of the Aryans, which the Greeks alone ex- pressed by a definite sign, since they borrowed their alphabet directly from a Semitic people. Of course this can be proi^ed only by adequate comparison ; but we are concerned now to show that the formidable list of Semitic gutturals ought not to divert us from the attempt to institute such a comparison. From what has been said it is clear that we are not justified in receiving, with Dr. Delitzsch* the Aryan gh as the analogue of what we may call the surd or h order of Semitic gutturals. In the first place i|; is moat probable that the Aryan sonant aspirates, gh^ dh, bh, arose, during the remoter history of the family, from the earlier g*, d, and b, just as in Sanskrit the surd aspirates kh^ th, and ph arose after its separation from the main linguistic stem, In the second place, remembering that we have to compare with the spiritus asper, or the simple h, we find that its origin in the Aryan languages is not due exclusively, or even in any large degree, to an original gh. In the old Aryan tongues there were apparently two types of guttural sound ; the one being conveniently represented by the Greek x ^^nd the other by the Greek -. The latter sound is of various origin. It either arises independently, as often in Greek and Latin, and other idioms, or represents an original 5, y, or v, as frequently in Greek, or is due to the dropping of the g, rf, 1 The physiological conditions of the utterance of each spiritus are given by Merkel, op. cit., pp. 72-74, who also shows in the same connection how natnral the transition is from one kind of guttural to any other. * Cited on p. 20. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 59 or b from the original aspirates, as occurs irregularly in all the Aryan tongues, especially in the Keltic. It is never due directly, in our opinion, to an original g*/i. Gh, it is true, is represented in Latin at the beginning of a root by A, as it is in Greek by x > b'-^t this h was originally a rough guttural^ like the Greek, and the sound was heard along with the ordinary h in common speech, as it was in Anglo-Saxon,^ and other old Teutonic languages, until the latter sound took its place entirely. Further, the rough Roman A, as well as the Greek ')(•> must, we think, have passed through a stage in which it had the kh sound.^ But it may be asked how tlie Sanskrit h arose. It represents mostly an original g-A, and is manifestly a corruption of it. It is a sonant, and is the only h in the Aryan tongues that is not surd. It was evidently, therefore, not primarily formed from the other aspirates through the dropping of their first element ; if so it would have been a surd, as the h so arising became in the other Aryan languages. Its pronunciation probably somewhat resembled that of the German ff in Toffe, though it is not safe to speak with authority on such an obscure matter.* This theory would best agree with its development from g-h. Here, then, we admit, is a guttural breath derived from gh. May it not have been so also with the Semitic family, if we allow it to have had at one time the gh sound ? Certainly not ; for its modifications would have brought it into range with the sonant or m order of gutturals, whereas Dr. Delitzsch makes the gh the Aryan representative of the n, or surd order. Moreover, it stands most nearly related phonologi- 15 1 Corssen : TJ'^bfr Aussprache Betonnng n. Vocalismus d. Lateinischen Sprache (Leipzig, 1868), Vol. i. pp. 96, 97. 3 March's Anglo-Saxon Grammar, p. 17. ' The Keltic (Old Irish) ch is corrupted from c {k), occasionally from g', Zeuss, Grammatica Celtica, 2d ed. (Berlin, 1871), pp. 63-71, comp. pp. 74, 78; Schleicher's Compendium (4th ed.), pp. 273-279. The Aryan gh becomes g in Keltic. ♦ Schleicher (Compendium, 4th ed., p. 17), gives it the sound of the German h, made sonant. Bopp (Eritische Gramm. d. Sanskrita-Sprache, 4th ed., p. 20 f.), makes it equal to the Greek x softened. This agrees more nearly with our own view, and harmonizes better with oar theory as to the genesis of each sound. 60 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. < it cally to ^, the Semitic guttural ;' which, as wo have seen, was not the first, but the very last of the Semitic gutturals in the order of development. We have dealt thus at length with the guttural sounds, because they are so numerous and so peculiarly Semitic, and seem to present obstacles in the way of a comparison with the Aryan family which the other classes of sounds do not. The conclusion at which we arrive is, that all of these gutturals in our comparisons ought to be disregarded, as they are of purely Semitic development. The spiritus lenis, - or K, is all that was common to Aryan and Semitic at the time of their separation, if they ever spoke a common idiom at all. It is impossible for us to write, in this connection, at the same lengtli of the other classes of comparable Aryan and Semitic sounds. The same principles which were maintained with regard to the development of the variant gutturals will hold with regard to the differentiation of other sounds within the bounds of their own generic classes. We shall therefore proceed more rapidly to an examination of the remaining contents of the Aryan and Semitic alphabets. Next to be considered are the other weak sounds v (t^) and y. As far as can be made out at the present stage of linguistic science, these were radical sounds in the two great families, though their history has been strikingly different in many respects. As to the Aryan v, the fact admits of no question ; as to the y, though it does not occur in many Aryan roots, yet these are very ancient, and its use both in the pronouns and in inflective elements shows that it could not very well have been developed from an original i, from which it often arises in both Aryanand non- Aryan linguistic forms. It is to be noted, however, that in roots, not in formative elements, the use of v preponderates largely over that of 1/. The same holds true in the Semitic family. Y is much more rarely found in the triliteral roots than is v. 1 It has been observed by Sweet that the German g in Tage, or the Modem Greek y sometimes passes into a uvular r; this is the vanishing sound of the Arabic letter. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. Gl Wliat is most remarkable, however, about these sounds, from a comparative point of view, is that they are vastly more numerous in roots of the Semitic family than in those of the Aryan. This is certainly a most instructive fact, as it is one that cannot be ignored in any just investigation of the general question of Aryo-Semitic relations. It may be ac- counted for in this way : Over and above the normal repre- sentatives of the Aryan y and w in Semitic, there would be two occasions of a large addition. First, it is natural to assume that these primary vowels of the Aryo-Scmitic stock would often harden into semi-vowels ; i and ^^■ would thus bccouio y and y, in a consonantal system like the Semitic. Again, in many of the originally biliteral roots in Semitic, these would become triliteral through the use of weak letters such as y and v. Hence a Semitic w or y would in com- parisons have to be regarded sometimes as having no repre- sentative in the Aryan speech, sometimes as representing an Aryan u or i, and sometimes their own phonetic equivalent. It is scarcely necessary to say that the Semitic forms in which either of these sounds occurs require great delicacy and caution in treatment; for we must not only ascertain to what class each belongs as regards its origin, but also to discriminate carefully between the two letters, inasmuch as they so frequently interchange, especially in some of the dialects. On these sounds we have nothing further to remark, except to say that, according to our present light, Dr. De- litzsch appears to be fully justified in excluding the Aryan y from his table of phonetic correspondences. The sounds I and r come up next for discussion. Dr. Delitzsch, in his table above cited, makes the Aryan r repre- sentative of the' Semitic r and /. We have no objection to this statement ; but it requires to be properly ex[)l'iined, from a consideration of the true relations of the two sounds to one another. First, as to the Aryan sounds. It is usually held, mostly through the influence of Schleicher ^ and Fick,'' ^ Compendium, etc., pp. 11, 162. ' This is assumed throughout bis Vergl. Worterbuch d. indogcrm. Spracheo. I'"H 62 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. that the jjiimitivo Aryan had no I, and that in all the cases of the appoaraiico of that sound in the diverging languages, it arose from the r. This is one of the most interesting questions in Aryan phonology, though one which cannot be discussed here. Wo only remark upon it that the contrary opinion, wliich has been defended by Heymann,^ seems to bo entitled to at least as luuch support.^ With regard to the Semitic / and r the sources of evidence are still fewer and more doubtful. But as to both families we would maintain that both sounds once existed, though vaguely and even interchangcaldy pronounced. In behalf of this wo would cite the history of the sounds in all families that possess them. There are no sounds in human speech more liable to confusion and interconversion.^ Even in the Aryan tongues, where as a rule / is developed from r, the change from r to / is not infrequent.* In the Dravidian family of languages, the Tamil, Tclugu, Canarese, etc., r also changes i'.ito /, though the reverse is very often the case.* In some of the dialects of Polynesia, of South Africa, and of the Indians of North America the confusion is almost universal.^ In some words the speaker is heard to pronounce I, and in other words r, wl en the sound is radically the same. In some languages the / is wanting, as in Zend, as also in old Persian,'^ in Ar- * Das I der indogermanischen Sprachcn gehort der indogerm. Grundsprache (Gottingen, 1873). " A full review of the controversy and of the state of the question is given in Pt'zzi's Glottologia aria rccentissima, pp. 17-24. The author himself holds to to the hclit'f tiiat / was a primitive Aryan sound. ' Even cultivated persons sjieakiiig highly developed languages arc liable to this infirmity, e.g. Alcibiades who was ridiculed by Aristophanes for his use of I for r, Vespae 44. Cf. Plutarch, Vit. Ale. 1. This was probably not affectation. * See some cxariples in M. Miillcr's Science of Language (Am. cd ), ii. p. 184. * Rev. Dr. Caldwell, Comp. Grammar of the Dravidian Languages, p. 120, cited by M. Miilier, ii. p. 185. 8 Even among the dialects whicli are generally supposed to have no r sound at ail, and whose speakers arc thought to use I in place of it in trying to utter a foreign word, cases are not unknown of the utterance of the r. The writer has had as a guide on angling excursions a Micmac Indian, — a tribe usually thought incapable of the r, — who actually changed a foreign / into an r more frequently than the reverse, saying richer, for examj)le, instead of liquor, » Zeitschrift d. deutschen morgenl. Gesellschaft, Vols. xiii. p. 379 ; xTi.p.ll. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 63 menian,^ and in several dialects of Japan, of Africa, and America.* B, again, is wanting in Chinese, in many dialects of America and Polynesia, and in the Kafir language.^ Some languages, again, have two r's, as the dialects of Australia ; * while others have two /'s, as some of the Siberian idioms.* One tribe, at least, of the last-named family, the Tchuktches, have two r's and two Va.^ It is only necessary to add that in the literary period of the Semitic languages r sometimes becomes l,^ though the reverse is not yet proved. Prom all this it seems clear that in all languages both sounds were originally one, and that, in most cases, a sound viljrating between the two. In most languages as they advanced in age the two were clearly discriminated. In the Aryan, for some time before the divergence of its dialects, they were prob- ably not yet perfectly distinct.^ In Semitic they must have been divaricated veiy early in its separate history. It follows, accordingly, that for purposes of comparison r and / in both families may be regarded as representing the same primi- tive sound. To the hypothetical Aryo-Semitic speech one might then justly apply the remark made by Dr. Blcek of the Setchuana dialects : " One is justified to consider r in these dialects as a sort of floating letter, and rather inter- mediate between I and r than a decided r sound." * M and n do not require much discussion for the settlement of their relations in the two systems. Unlike the last two ^ Zcitschrift d. deutschcn Morgcnl. GcscUtichaft, Vol. xiii p. 380. 2 Ibid., Vol. xii. p. 453. * See the references in Max Miiller's Science of Language, ii. pp. 179, 180. ♦ rriedrich MulIer,Grundri8S d. Sprachwissenschaft, ii. Band, 1. Abth.(Wien, 1879), pp. 1, 81, etc. » r. Mullcr, op. c. p. 100. • F. Miiller, op. c. p. 134. ' So r.i:cbx Ezek. xix. 7 ; Isa. xiii. 22, for pil'sanx , palaces : b'*12Cn Ps. cv. 15 for I'^rilSh, to make to shine (eomp. Ewald, Ausf. hebr. Lehrbuch, 8th ed., 1870, § 51 c). In Assyrian even a sibilant generally becomes / before a dental (Sayce, Comp. Assyr. Grammar, p. 30), but it must first have become r; hence the name Chaldaeans, aa compared with o'^'nilJS . " Comp. Pezzi, Glottologia aria recentissima, p. 24. » The Library of His Excellency Sir George Grey ; Philology (Capetown, 1858), Vol. i. p. 135, quoted by M. Miiller, Science of Language, ii. p. 184. 1 ■■>■■ * , " 1 ' ' i ' i . ^ G4 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. Rounds, they arc totally distinct in their origin in all lan- guages. As nasals they are lialilo to occasional interchange in both luniilies, l»ut are not regularly inter-convertiWo. In the Seniitic roots care must be taken to distinguish hotween the undoubtedly radical n and the samo sound where it seems, upon evidence which wo bhuU adduce horeufter, to have been used as a mere determinative, as it apjxjars to be one of the letters most fre(iuently cmjUoyed for the purpose. MoreovcM', being next in weakness to y and v, it is I'ablo to take the place of other liquid letters, as well as to inter;iiange with I/, a matter of very fre(i[uent occurrence. M, on the other liand, is nmch more stal)lo than n. It i)a8scs into n much less frequently than the reverse occurs, and very rarely takes the place of the other liquids.^ Of course, in Semitic the rn is liable to interchange with the other lahials — a phe- nomenon apjtcaring ih 1 languages possessing these sounds.* The Semitic m and n may be provisionally taken to represent the corresponding Aryan sounds, with important restrictions which may oixiratc in consideration of the foregoing cautions. We pass now to the sibilants of the respective systems. At first sight, a comparison seems very difficult, if not im- possible. In the primitive Aryan there was only one s, the ordinary fundamental sibilant. In the Semitic idiom there are several, and it will be necessary to examine them, to classify tliem, and to reduce them, if j)0S8ible, to their funda- mental primitive sounds, so that we may get a proper basis of comparison with the Aryan *•. A careful comparison of tlie Semitic sibilants leads us to the conclusion that before the breaking up of the family there were developed four distinct sounds, answering respectively to 6, u:, t, y. These sounds emerge on comparing all the dialects, — Arabic, Ethiopic, Aramaic, Hebrew, and Assyrian, 1 A rare instance of m arising from / is shown in the Arabic » r'<^ . >^ skull, answering to the Hebrew rbabft . " That is, nearly all known languages. In a few they arc wanting altogether, as in those of the Six Nations and the Uurons in North America; in others some of them arc absent, as in a few of the dialects of Africa, and throughout Australia. RELATIONS OF THE ARYAN AND SEMITIC LANGUAQKS. C5 with tlicir subordinate varieties, — and to them all tlie otlicr sibilant modifications may ho rodncod. The z sound (pro- nounced a.s in EngliHh) is* tlie Honanl of the surd s, and arises from it normally in all languages which possess it, though also occasionally springing from other sounds. Hence wo have to account for the sounds e , o , and y. These con- clusioiiB wo shall try to make clear. In the first place, the * and sh sounds (Heb. o, to, and cJ, Syr. jj and taA) Arabic ^j,^ and Ji,, Ethiopic |*| and QJ, Assyrian s' and s *) sprang from the same source. This might be argued from the history of the sounds in languages generally, in which sh is developed from *. But we have other evi- dence, drawn from the phenomena exhibited by these sounds in the history of the different Semitic idioms. The distinction between the to and uJ sounds, by which the former approximated to the sound of e^ was made in Hebrew alone sufficiently important to be represented by a special sign. Leaving these aside, as of clearly late orijjin, we find that the s and sh sounds have fluctuated and varied greatly from the time of the separation of the different branches of the family. If these dialects be divided roughly into Northern and Southern Semitic, — the former including Hebrew, Aramaic, and As- syrian, the latter, Arabic and Ethiopic, — it will be found that the sh sound of the northern division is represented mostly by the s sound in the southern, while the s of the former corresponds radically for the most part to the sh of the latter. Yet the correspondence is not sufficiently regular to make this a fixed principle of sound-shifting ; nor can the division given above be regarded as anything more than a very general classification. A multitude of facts could be adduced, in addition to the above, if space permitted, to show how these 1 ^ ss Hcb. C, and 8 = V3. This is the usual methud of transcription. It is one of our misfortunes that sh seems to represent a double instead of a single sound. * That D and to were originally distinct in Hebrew is proved by the fact that the t is represented in Arabic by ^j^ more frequently than by ^i, while the to >» represented only by jA. See Ewald, Ausf. hcbr. Lehrbuch (8th ed., 1870), 1 50 a. In later times and tt) were much interchanged. a . f; 3' M RELATIONS OF TIIR ARYAN AND SEMITIC LANG C AUKS. niljilant.s varied nnd iiitcrclmn^'cd from tlio oftrlicst known Semitic times, and tliut too accordinj^ to no Htublo law of permutation, l)ut according to local and tril)al pcidiaritiert, Buch an has made the s/i sound diiricult to Ihi; J<]|iliraimil(>H,^ and to many others tlirou^d\out tho world, the same .{)osition of a development of s and sh from two fundamen- tally distinct sounds, — a notion improl)ai»le on general prin- ciples, — is nntenalile also on historical grounds. What tho original sound was, it is impossihle to deternuue with exact- ness. Most pr()l)al)ly, however, as wo shall sec |)resently, it was that of the Ilelirew o, or the ordinary s, with a slight tendency to palatization which would account for the fre- quency of tho s/i sound in tho southern dialects, and its pre- ponderance in the northern, where other influences wore also hrought to bear, tending the same way. The z sound (Hol>. t) arose sometimes from 5 and some- times from the sound represented by tho Hebrew y. In either case, as wo shall see, the primary source was probably tho same. It clearly was not an original Semitic, as also it was not an Aryan, sound. It remains, then, to account for y and its representatives in the other dialects. This is peculiarly Semitic, running through all the branches of tho family. Yet is only peculiarly Semitic as a constant letter ; for the sound itself is probabiy heard in every language possessing sibilants at all. In English, for example, we come near it in saying costj as distinguished from the s in cast. It is due there to the vowel sound with which it is connected ; but in the Semitic languages its sound is the same no matter what may be the accompanying vowels. In tho northern Semitic there seems to havo been a slight hardening of the first part of the utterance, with almost a 1 Jud^. xii. 6. ^ Sec Dillmann, Aethiop. Grammatik, p. 51. His whole discussion of the Ethiopic sibilants is very instructive, and contirms very strongly the view hero Bdvocatcd. * Schleicher, op. cit., p. 275. Zeuss. op. cit., p, 119 f. et al. RELATIONS OF THE ARYAN AND SEMITIC LAN0UA0E8. 07 coinploto cloMiiif? of the organs, givinj^ tlio cfToct of a uli^ht t Noimd hcforo tho Hibiluiit. Ihitevcn this hoiiiuI wiih usuully tmnscriWcd by « in tho Septnaj^int, as SafScuoO lor pi-m^x, and vory HcMom by {^. Tho pronunciation primarily was evi- dently that of a Htronj? s, made with tho tongue turned back against tho roof of tho mouth. It stood related to tho ordi- nary s as tlio llel)rcw o to p. Vnrious lines of evidence ])oint to tho conclusion that it was not an original sound, lait one developed from tho primitive s. First, its organic asso- ciation with the latter. Tho original s was )>robably, as we have soon, pronounced indistinctly, and perhaps soniowlmt variously. There was that tendency among the Semites to nmltiply consonantal sounds which wo have already dis- cussed. What more natural than to take the occasional sound of s, just described as existing in English and elsewhere, and make it a fixed one, without regard to the vowels accom- panying, especially when it is considered that tho vowels ])laycd a secondary part, and were necessarily varied con- tinually within tho samo invarialjle consonantal formula? The Semites, in developing their roots, necessarily had the sense of consonantal stability dcvcloj)ed and continually exercised ; while the Aryans have regarded the preservation of the vowels as essentially bound up in vital union with the consonants. The Semites, then, would be inclined to hold fast to each distinct consonantal sound when once made familiar to tbeir ears. The Aryans could not ; for the same vowels being retained in each utterance would prevent the discrimination of the consonantal variations, just as we are still ordinarily unconscious that the s in cast and the s in cost are different sounds. Secondly, the same thing is illustrated in the history of the most fully developed Semitic dialects, — the Ethiopic and especially the Arabic, — where the tendency, having once fairly set in, was carried so far that not only the simple s and t, but also the d and the z, had their secondary sounds. It is fair to argue, within cer tain well-considered limits, from the living facts of a lan- guage to its inherent tendencies, and in these later develop- fi 68 RELATI0:i8 OF THE ARYAN AND SEMITIC LANGlJAtES. ments of the Semitic idiom we see exemplified the principlca of its primitive Tvorking. In the third place, the y sound seems not to have been originally a distinct sibilant, for it intcrchari2es with the s and z sounds so frequently in kindred roots that we can hardly attribute the coincidences to the confusion of the sounds. Une must have developed from another. For proof of this statement we must appeal to the lexicons, since we cannot afford the space needed for an adequate exhibition. It is now proper to show how all the Semitic sibilants may be classified as to their immediate origin. To the Hebrew y and its equivalent in Aramaic and Assyrian answer the Ethiopic ^ and ^ and the Arabic ^^, ^, and Jsj. These were all developed at first from the y sound, though on account of their similarity to other sounds, such as those of n and o , they were often interchanged with the latter, and more rarely with other sounds. According to the modern Arabic pronunciation, which may be taken as sufficiently near the ancient for our purpose, there were thus two orders of sounds ; the one uttered with the tongue close to the teeth : o, « (to) , t, or their equivalents ; and the other with the tongue turned back against the roof of the mouth : ^ (y), ^ ^, and io.-^ Tn these groups b and y represent the primary sounds of their resj)ective ranks. The historical development is p.-obably to be represented by the order of the letters as tliey here stand, except that T, in all likelihood, arose later than y. While all of these were thus primarily developed from one sound, it ought to be observed that sometimes we hud a sibilant degenerated from a mute, as «} from n, t from "t. In comparisons these must, of course, be carefully distinguished from those which are unquestionably of sibilant origin. The last group of consonants to be considered are the 1 Pronounccil is d would be in the emphatic English syllable odd. The ori- ginal sibilation was gradually lost. In Ethiopic it was resumed again. Sea Dillmann's Grammar, p. 52. ^ Pronounced as 7 in the cooibination ose. ' .1 RELATIONS OF THE AYRAN AND SEMITIC LANGUAGES. 69 so-called mutes or explosives, represented in English by k^ g" ; t, d ; /J, b. Looking first at the Aryan alphabet, we find that it had, besides the primary, also a secondary ^ series of gut- turals, most probably developed from the former. The sonants were also aspirated so as to yield three additional letters, g-A, , represented throughout the whole family. It had its origin some time before the sepa- ration of the Semitic trilies, as is proved by its individuality and vitality throughout the history of Semitic speech. It was doubtless developed from the ordinary k sound through the same tendency that led, in the same family, to the pro- duction of the deep gutturals. Yet it had also strong affinities with the g; sound, as is shown by the great number of cases in which they interchange, as well as by the fact that in later Semitic times it has shown a tendency to a sonant utterance, as in Ijal)ylonian ^ and in some dialects of Modern Arabic.^ So it also interchanges, though less fre- quently, with the gutturals; and, in all cases of its citation in comparisons, its true relations will have to be ascertained. Next, as to / and d. Tlio latter sound has a variant oidy in Arabic and Ethiopic. This has been developed from the Y sound, as already seen ; but from its resemblance to the primary d sound, the latter was often interchanged with it. 1 Sayce, Comp. Assyrian Grammar, p. 3G. 2 Ibi, though sometimes also from p itself. These labials in Ethiojiic are the most fluctuating sounds in tlie language, and have rendered their compt.rison with labials in the other dialects somewhat uncertain in many cases. The Semitic b had virtually the same pronunciation throughout the whole system. It should be added that in all the dialects not only does b sometimes take the place of p, but the other labial m takes the place of either. Naturally, however, this did not take place in the earliest Semitic times, and a careful examination ought to enable us, as with other cases of permutation, to determine the primary forms. This ha^-^^y survey brings us to the simple sounds k, I, », ^, d- b, as the original Semitic mutes. We have thus reduced both the Aryan and the Semitic consonants to their primary limits. We have found that the original Aryan stock consisted of the following sounds : A", t, w, g-, d, b, s, r (/), w, n. i/, v, with the spiritus lenis. The original Semitic stock har been reduced to precisely the same sounds. No root, therefore, can be found in either 1 If the tongue bo very slightly moved from the roof of ihc mouth while the organs are in the 13 position, and an emphatic hissing sound be made, the result will be V. 72 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. family which contains a consonant not reducible ultimately to some sound in the above catalogue. A discussion of the vowels will not be of much importance for actual comparison, since the Semitic subordination of the vowel to the consonant precludes it generally from admission into the field of inquiry about to be entered upon. At the same time it is worth noticing that the two families show that at one time they had in common simply the primary vowels a, i, u. The recent diiicovery, which shows that the reputed Proto-Aryan a must be differentiated into a^, a^, a^, does not invalidate the conclusion ; while on the Semitic side, the fact that the Arabic and Assyrian possess just the simple scale a, i, u, proves the case for that family.^ As has been already suggested in connection with several of the consonants treated of above, it is not maintained that the secondary sounds were developed from the primary in every case of their occurrence in actual roots. In the fore- going discussion I have only cited the fact of such normal representation as affording evidence of the real relationship of the sounds in question. In quite a number of cases, even in Proto-Semitic, these secondary sounds, since they had been firmly established in the language, arose by degenera- tion from other sounds than their primary originals. The frequency with which n appears in roots as coming from an original j? is one of many obvious illustrations of the general fact. So much for the Proto-Arvan and Proto-Seraitic sounds considered separately. The next question, as to how the sounds of the two systems represented one another in actual roots, which is one of more practical importance in the general subject, can, of course, be fully answered only when we shall have presented a comparison of the roots that may seem to claim relationship. It will* be necessary, however, 1 Arabic and Assyrian did not simplify the original system of vowels ; the other dialects amplified it. See Schradcr, in Zeitschrift d. d. morg. Geselischaft. Vol. xxvii. p. 408. i-;y^;\v-rl, ii . iiV a- J*S: RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 73 to insert here provisionally the results of our observations as embodied in the following scheme : Pro to-Semitic.^ D A;, p &, Proto-Aryan. ' = spiritus lenis. k, k t 1 d d t> P a b P bh 10 S, 0, S'y X S s •^ r, b I r(0 •a m m a n n 1 V V The following remarks should be made with reference to this table. 1. In the roots that have been examined for phonetic rep- resentations, no account has been made of those which show degeneration of sound from earlier forms. The earliest historical expression of the root-idea has been taken in each instance. A fuller exhibition of later secondary sounds would otherwise have been made. 2. The absence from the above list of certain sounds that existed in one family or the other or in both is to be well noted. No root common to the two idioms containing the y sound originally has been discovered. It was noticed in the foregoing discussion that roots with i/ as one of their ele- ments were not numerous in either family. Among the Proto-Semitic sounds it is to be observed that n and n are absent from the gutturals and t from the sibilants. Of Proto- Aryan sounds dh is unrepresented in the hypothetical Aryo- Semitic roots. These facts lead us to suggest here that if the 1 The Proto-Semitic sounds are represented by Hebrew characters ; ttJ is, of course, unpointed. No satisfactory transcription has yet been devised for y. I have adopted that most employed in Germany, which is not to be confounded with the rough breathing. I) An I ^vmfWWr^W' 1. If 1^ 74 RELATIONS OF THE ARYAN AND SEMITk T ANGUAGES. theory of the development of the secondary sounds is true, the correctncHs of the comparison of roots is confirmed in a remarkablo manner ; while, if the comparison of roots should prove to he successful, the truth of the theory of secondary sounds will have been demonstrated. 8. It will be observed tliat these phonetic representatives are also in most cases approximate phonetic equivalents. In this connection it will be necessary to notice arguments that have been made against the likelihood of hyiwthetical Aryo-Semitic roots having preserved the same sounds until an accessible period. The form which such objections are apt to assume in the minds of scientists may be exemplified by the following citations from Professor Max Miiller.^ Tiie remarks quoted are made in the way of caution, since, as we have already said. Professor Miiller admits the possibility of an Aryo-Scmitic affinity, and holds earnestly to the scientific legitimacy of the widest comparisons among the various fam- ilies of speech. After speaking of the vagueness of current Semitic roots as an obstacle to just comparison, he says : " I have by no means exhausted all the influences that would naturally, nay, necessarily, have contributed towards produc- ing the differences between the radical elements of Aryan and Semitic speech, always supposing that the two sprang originally from the same source. Even if we excluded the ravages of phonetic decay from that early period of speech, we should have to make amj>le allowance for the influence of dialectic variety. We know in the Aryan languages the constant play between gutturals, dentals, and labials (^(/uin- que, Skr. panka, irivre, Acol. irefiTre, Goth. pimp). We know thedialecticinterchangeof aspirate, media, and tenuis, which from the very beginning has imi)arted to the princijjal chan- nels of Aryan speech their individual character (T/^ets^ Goth. threis, High German drei). If this, or much more, could happen within the dialectic limits of one more or less settled body of speech, what must have been the chances beyond these limits ? " And again ; " We know that words which 1 See Chips from a German Workshop (London, 1875), Vol. iv. p. 99-lOS. I '•! RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 75 have identically the same sound in Sanskrit, Greek, Latin, and German, cannot be the same words, because they would contravene those phonetic laws that made these languages to differ from each other The same applies, only with a hundred-fold greater force, to words in Hebrew and Sanskrit. If any triliteral root in Hebrew were to agree with a triliteral word in Sanskrit, we should feel certain at once timt they a''0 not the same, or ihat their similarity is purely accidental. Pronouns, numerals, and a few imitative, rather than pre- dicative, names for father, mother, etc., may have been pre- served from the earliest stage by the Aiyan and Semitic speakers; but if scholars go beyond, and compare such words as Hebrew barak, to bless, and Latin precari ; Hebrew lab, heart, and the English liver ; Hebrew melech, king, and the Latin vmlcere, to smooth, to quiet, to subdue, tlkcy are in great danger, I believe, of proving too much." It may be said, in general, with reference to such strict- ures that they are invalid, because the question is not one of antecedent probability, but of direct evidence. The comparer is not bound to assume that Aryo-Semitic roots will appear with the same sounds. He makes his investigations among roots having the same primary meaning in the two families, and if he finds that a large number of such forms have the same or similar sounds tlien it becomes probable that they were originally the same. \i it further appears that the es- sential part of what must have been the primitive working stock of ideas of the two systems are expressed in the same or similar sounds then the probability amounts almost to moral certainty. As was said at the opening of the first cliapter, the evidence is of precisely the same kind as that which obtains in linguistic comparison generally. It is undeniable, however, that there is some plausibility in the arguments above cited based on the analogies of dia- lectic changes within other spheres ; and those arguments therefore require some examination. The first remark to be made is, that there seems to be a misconception of tho con- ditions of a proper inquiry. We have nothing to do, in the Pf wn 76 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. actual comparison, with Hebrew roota and Sanskrit words. We only use those, along with the other Semitic and Aryan dialects, for the puii)080 of finding out Proto-Aryan and Proto-Scmitic roots. That these are accessible, if strict sci- entific methods are pursued, there is no doubt. Now, when these are obtained we have, of course, still two separate lan- guages. Supposing them, however, to have come from the same source, we cannot tell how long a time had then ela[)sed since they had emerged from the radical stage. A glance at the difference in flcctional characteristics, and at the deveio})ement of secondary sounds would seem to show that it was considerable. That it was not necessarily very long, is probable from the consideration that the formative principle must have been very busily at work in those early days of language, and must have evolved new phenomena with great rapidity. These are the conditions with which we have to do, and not those assumed by the critic. In the second place, the inferences from Aryan phonology are sonicwhao overdrawn. Still remembering that we are dealing with roots primarily and not with current words, we do not seem to see the same prevailing variation and inter- changing of sounds that Professor Miiller speaks of. Let any one take a comparative phonological table of the Aryan languages, — such, for example, as that given in Curtius's Grundziige, — and he will probably be struck with the gen- eral correspondences rather than with the variations. There does not appear to be a " constant play " between gutturals, dentals, and labials, such as the somewhat exceptional and still puzzling words for five would seem to indicate. More- over, it does not seem quite just to include the Teutonic languages with the others as the best representative of Aryan phonology. Taking Sanskrit, Greek, and Latin, and with them any form of Aryan speech that is not Teutonic, it is certain that roots (if not words) occurring among them which have identically the same sound and meaning must be the same roots. Sound-shifting in the mutes is not a regular, but rather an exceptional, sort of phonetic change in the m RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 77 Aryan tongues, and in the languages of tlio world at large. The vast development of the Teutonic family and its in- fluence on liistory and civilization have given its dialects greater prominence among the Aryan idioms. Grimm's law has also assumed a large space in linguistic discussion, on account of its intrinsic interest and importance in the science of phonology. But its relative scope within the Aryan sphere does not entitle us to assign it the same importance in general questions of comparative etymology. The languages which have remained nearest to the Proto- Aryan type are free from the regular operition of this prin- ciple ; and there is no evidence whatever for the assumption that the Semitic family was ever subject to its influence. On the whole, therefore, and apart from the evidence of actual comparison of roots, there is no reason for believing that the Aryans and Semites had varied their fundamental sounds radically at the period represented by the earliest accessible forms. We have seen that secondary sounds were developed and have shown how they may be reduced to their primaries. In the third place, it is noticeable that a concession is made in the above criticisms which annuls their whole force. When it is said that" pronouns, numerals, and a few imita- tive rather than predicative names for father, mother, etc. have been preserved from the earliest stage," the question arises, How do we know this ? What is the test ? Obviously only phonetic correspondence. But how is phonetic corre- spondence possible among these groups of words if radical changes of sound had inevitably occurred throughout the two systems of speech ? There is no escape from the force of this counter-criticism, which is applicable not only to Pro- fessor Miiller, but to a large number of other glottologists, some of whom are much less favorable than he to the legiti- macy of such investigations as the present. The writer may say here that, for his own part, he not only regards as ridicu- lous the comparisons cited at the close of the foregoing ex- tract, but also does not think very highly even of the evidence drawn from pronouns, numerals, etc. 78 RKLATIONS OF THE ARYAN AND SEMITIC LANQUA0E9. 1., 'f! In connection with the tabic of phonetic rcprcHentation a))ovc fx'cscntcd, tins subject will he clo»ed with the following recapitnlation : At the time of the breaking up of the Aryan family it j)088es«cd tlie following Htock of consonants : ' (light guttural breath), k, k, if, ^'•/i, g^, /j^/i^, t, r/, f/A, p, bh^ s, r (/), wj, w, y, v. Of these the following Bounds are represented in hypothetical Aryo-Semitic remains : ', Ar, ky g', gh^ g^,^ t, d, p, b/i, «, r (/), w, W, IK At the time of the breaking of the Semitic family, it pos- sessed the following stock of consonants : N, n,», n, 3, p, A, n, i3,T,i!,a,w, o.is, To,b,a, 3,% i (=',/*,', A, A:, /fc,^,<,^,y each of the two systems before its breaknig up into dialects was given at tlic close of the last chapter. Any verbal forms in these languages that are to be comjiared must first be reduced to these simijlc phonetic elements. I have also stated that there were two principles which must determine the choice of com[>arable forms : first, the »>rimar) signification of each must be shown to be the same ; secondly, each term to be com{)ared must be reduced to the form it possessed l)efore the system of speech containing it (Proto-Semitic or Proto-Aryan) became broken up into different dialects. Keeping these principles in view, we have to proceed to an analysis and comparison of the words in the two systems that seem worthy hypothetically of such treatment. It will be necessary, however, to begin the investigation by showing how we are to deal with the living elements of language, whose seemingly endless diversity would appear to forbid any attempt to harmonize them. In both districts of speech, and especially in the Semitic, we seem to be wandering about in a vast wilderness, through which the explorer moves in a hopeless entanglement of bewilderment and confusion, never reaching a meeting-place for the paths that either lead no-whither, or cross one another perpetually, without beginning and without end. It will be needful to show that some central eler ation may be gained from which we may look down upon this " mighty 79 w IMAGE EVALUATION TEST TARGET (MT-3) A % >* ^£^ 1.0 12.8 122 1^ ■2.5 I.I £ |££ 12.0 i 1^ IM 1^ / ^? ^.** ^ ^>1 Fbotographic Sciences Comoralion 23 WEST MAIN STREET WEBSTER, N.Y. MS80 (716) 872-4503 ^■n ^n^HF 4- 0^ ^ jjWPaij.i^iWHi'ilPj^,. w ,, ! 111 , ; 1 ' • ' 1 80 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. maze,'* and see that it is " not without a plan " ; from which we shall be able to see that the paths which are interrupted by so many obstacles, interposed by the careless ages, still keep on their course, whether converging or diverging, and run from side to side of the great wilderness. In plainer language, it will be incumbent on us, knowing how the current terms of each idiom may be referred to their proper stems, and further to their conventional so-called roots, to show according to what laws of formation the " roots" them- selves may be analyzed into Iheir simplest expressions. A root has been well defined by Curtius as " the significant combination of sounds which remains when everything form- ative and accidental has been stripped away from a given word."^ In inflectional languages, at least, such so-called roots do not appear clearly at the first showing ; and the only way of arriving at them is obviously to make sure that the forms to be examined are primary and not derivative, and then by a thorough analysis of them, with a careful application, if need be, of the known phonetic laws of the language in question, to eliminate in each case the invariable significant term from the variable and unessential suffix, prefix, or infix. When this is done, however, we find that in many cases the process of analysis is not fairly complete. In both great families of speech are still left multitudes of similar roots, with similar meanings, whose relations to one another it is the duty of students to determine. In har- mony with what we would naturally suspect with regard to the growth of living speech, it is found that the primitive stock of roots at the command of the earliest speakers was enlarged according to need by internal changes or external additions. The modifying or formative 'elements are seen to be attached with equal freedom and regularity to all these variant similar forms, showing that these forms are inde- pendent of one another. This is not the proper place for an extended exhibition of the evidence in favor of such a 1 Grundziige d. griechischen Etymologie (5tb ed., 1879), p. 45 ; cf. p. 43 f., or In the English translation (4th ed. London, 1875, 1876), Vol. i. p. 58; cf. 55 f. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 81 "Til I :! doctrine. Wc shall presently have to cite groups of words in each family that will illustrate the position licrc assumed. Meanwhile, it will be enough to say that a twofold distinction has to be made with regard to the forms under discussion; and that by the common consent, if not always by Iho verbal agreement, of leading etymologists. First, we must dis- tinguish secondary from primary roots, or discriniiuatc forma that seem to have been developed out of earlier ones from those which we cannot reduce to prior conditions. Hccondly, we must note a difference between absolute and relative roots ^ ; remembering that in many cases analysis brings us at last to forms which it is impossible to regard as the exact ultimate expression of the radical idea ; since, for example, the combinations arrived at are sometimes unpronounceable, and sometimes appear in a slightly different form in different dialects of the same family. This latter distinction, liowever, is evidently not to be made use of practically, and must only be kept in mind as a constant warning against the temptation to fancy that we can always succeed in harmonizing the form and substance of language according to their original iden- tity. But the principle of the existence of both primary and secondary roots is of vital importance in glottological re- search, and much of what we have yet to say will be simply an attempt to trace its manifestations in Aryan and Semitic speech. We shall first deal with the current roots of the Aryan family. The discussion of this subject will be necessarily short ; and the reader is referred for a full presentation of all sides of the question to what has been written by such eminent etymologists as Pott,^ Curtius,^ and Pick.* We I i!;' "\t : ' This distinction, adopted by Curtius, was first made in these terms by Pott. Etymologische Forschungen (2d ed.), VoL ii. p. 246. '^ Etymologische Forschungen (2d ed.), Vol. ii. p. 225 ff. ' Op. cit , pp. 31-70, English translation, pp. 40-90. * Verglcichendes Wtirterbuch d. indogermanischen Sprachen(3iied., 1874-76), Vol. \\\ pp. 1-120, This acute and ingenious etymologist attempts to show at length that Indo-European ultimate roots fall under three classes: 1. those which consist of a more vowel (a, i, u) ; 2. those formed of the vowel a -|- a 82 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. m u • i J shall give here the principles whicli seem to be most surely established with regard to the verbal or predicative roots. Those who are familiar with the late ingenious theorizing on the subject will see that we hold a position as conserva- tive as is possible to any one not belonging to that obstructive sect of glottologists who refuse to analyze the current roots of any system of speech on the ground that there wao no development within that sphere of language. In analyzing the Indo-European roots we must have regard tf> a distinction which divides them into two great classes. We must distinguish between those forms in which new elements have been added to the old, and those in which the old have been simply modified. Both of these processes of change or development were energetically carried on, after the biCaking up of the Aryan household, in every branch of the family ; but their operation may also be traced more or less clearly within that stock of root-forms which was the linguistic property of all in common. First, as to the development of new roots through modifi- cation of the old, without addition. Here we have inde- pendent Indo-European roots arising, (1) Through the modification of a vowel in the original form. Thus the vowel a interchanges with i, as in the root dik, to shov,', as compared with dak (represented in 8i8dcrK(i) and Lat. doceo') ; in di, to divide, and da ; pi, to drink, and pa. consonant (as ad, ap, as) ; 3. those made up of a consonant or double consonant + the vowel a (da, jm, sa, sta, sjm, sua). We have space lor only two or three brief criticisms of this theory. First, to be formally accurate, classes one and three ought to be brought together. No root, and, in fact, no independent artic- ulate sound can consist of a vowel alone ; the spiritns lenis preceding the vowel bound is a consonant. Second, the universal elimination of i and u from classes two and three does not seem justitied by the examples given. There are some roots in which these sounds cannot be shown to bo secondary ; e.g. in di to hasten, pri to love, di to shine, the t cannot easily be reduced to a ; nor can a like origin be found for the m in m to beget, hhu to be, ru (lu) to separate, or yu to join. Third, there are many cases in which a vowel cannot be shown to have been the original closing sound ; thus, mar to rub, grind, in which the notion of jjhysical action is inherent, is probably not developed, as Fick claims, from ma, to diminish ; nor can an earlier vowel-ending root be well found for vas {us) tc bum, spak to see, hhar to bear, rtdf to know, yag to honor. m ■ Li.u^j iiu^yimniJiMi.M!'*.^""'' ■'"' RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 83 Less frequently, but as clearly, a is obscured into w, as in mud, to be lively, comijared with mad ; bhat;, to enjoy, share in Cfunffor^, as related to bhii-aiva>) , to obtain, has changed into rabh (Xaft/S-avco), to take hold of. In class first we cannot appeal with confidence to estab- lished laws of phonetic change, if we wish to determine, in any given case, which of the double or multiple forms is the earliest. Secondly, we must consider those roots which differ from similar ones by the possession of additional elements. (1) We find the additional factor at the beginning of the form. The only soured that seems to play this part in the Indo-European is s. Its occurrence there is limited to a few cases ; though in the subsequent divided life of its several dialects such a use or disuse of s became much more common. The root nu, to float, is clearly Proto-Aryan ; but so also is the kindred snu. The root stan, to sound, was also heard along with the related tan, to stretch, just as o-tovo? is found in Greek in coupany with t6i/o?. There seems to be no good reason to suppose that new Indo-European roots were ever developed by the infixing of a new sound in the old. The only sound for which such a function can be claimed plausibly is n. But if we examine all the forms in which this additional sound occurs, it will be found that the two hypothetical roots are not used inde- pendently of one another to form separate verbal and nominal stems, but occur side by side as the basis of derivatives that evidently spring from the same source. They are thus shown :i .1 n ;U 1*1 mmm 84 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. J] J ;!■ Ill.i ■ If'r " ■ to be variations of one another, rather than distinct roots with a separate range of development and an appreciable difference of meaning. Thus the root ag^h, to press, com- press, is evidently the same as ang-h; for while the former ap- pears in the nearest Sanskrit derivative, ag'ha, oppressing, evil, or as substantive, affliction, sin, as well as in the hometymous ^ words a;^09, grief; e;)^*? (= constrictor, the Sanskrit a/n'), serpent, and the Sanskrit ahu, narrow, the latter is as evi- dent in the corresponding Sanskrit, anhas, affliction, sin ; the Latin anffuis, serpent, as well as in ang-ustus, narrow, and the German eng; angor^ anxius, and the Germ, angst. This wo give as a fair specimen of the whole class, and accordingly assume for the Indo-European system, that the insertion of an n sound is nothing more than the nasalization of the pre- ceding vowel, rather accidental than essential to the autonomy of the root. It is, perhaps, a phenomenon similar in origin to the epithetic v in Greek QeXeyev < eXeye), the nunnation in Arabic, and the mimmation in Assyrian, and does not corres- pond to an additional etymological element. On the other hand, it is probable that in many cases the n was heard in the original root, and the form containing it would have to be regarded as the earlier one, from which the other arose through the weakaning of the sound by denasalization, till ifc disappeared entirely in some of the forms ; though within the Indo-European sphere this process gave rise to no new roots, in the strict sense of this term.'-^ 1 This much needed tenn, with the corresponding "hometymon," the writer owes to the invention of Mr. S. R. Winans of Princeton College, his friend and companion in philological studies. '^ The lately developed theory of nasal vowels casts some light upon the ultimate origin of such cases as those cited above. It was first suggested by Brugman in Curtius's Studien, Vol. ix., '^.nd has sin'ie been rectified and extended by the same scholar, and by others. See the admirable statement by Maurice Bloom- field in the American Journal of Philology, Sept. 1880, p. 292 fF. The main position is that in the Aryan system there is a full set of nasal vowels, answering to n, m, as Skr. i', I answer to r, I in the Unguals. In Proto-Aryan these are rep- resented by n, m. The n and wi remain consonantal before vowels, but before consonants they take the vocalized sound which is heard under like conditions in English and other languages, as in heavenly, handsomely (i= hevpli, hansmli). In Sanskrit p becomes a and an, ifi becomes a and am. In Greek, p is a and av, m is a and an- In Latin they regularly appear as en and em ; in Gothic and High German as un and urn. In Greek and Sanskrit, therefore, an original an or am may appear as a mere vowel a in certain inflections. jiiiilip5IPiipi,fiiWiii|i. I RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 85 (2) We have the most important class of root-distinctioiia in those forms which differ from similar ones in having the additional sound at the end. These sounds, which are quite various and usually distinguishable with clearness, have been named bj Curtius^ root-determinatives. This term, which would properly indicate a radical siirnijicant clement, we shall adopt throughout this discussion as applying to any additional sound in either family, under the guise of a i)relix, infix, or suffix, which is not a mere expansion or strengthening of the root, or a mere unessential variation of a previously existing element through ordinary laws of phonetic change. The justness uf this comprehensive distinction we shall show by-and-by. Here it is in order to enumerate the letters that seem to play this part at the end of Indo-European roots. The only vowel that appears as a post-determinative in undoubted Indo-European roots is a, which is found in a few secondary forms, as dhya, to see, from did ; ffna, to know, from g-an, (Eng., ken). As to the determinative consonants, taking them in the order of the Sanskrit alphabet, we have first ^ k, k, which ap- pears to us as certain only in the roots mark, to touch, stroke, (jnulc-ere^^ as compared with war, to rub ; dark^ to see, as related to dar, (Sanskrit and Lithuanian) ; dale., to bite, as compared with da^ to divide, tear (whence dornt^ tooth,') ; bhark, to shine, (^of)«09, bright)^ as related to bhar, itself a very 3arly development from bha. It appears, moreover, at the end of many lengthened onomatopoetic roots, whose etymological relations are, of course, not so clearly definable, g appears as a determinative in yng, to join, as compared with yu ; marg, stroke, wipe (o-fwpy-vvfu, milk}, as related with mar ; bharg, to shine (^Xcto), flag-ro, bleach), in con- nection with bhar, and a few others. Pick, in his discussion of these points,^ calls attention to the existence of so many 1 In Kuhti's Zeitschrift fur vergL Sprachforschung, VoL iv. 211 ff. See his Grundziige (5th ed., 1879), p. 69 ; English translation (of 4th ed.), p. 89. 2 Fick, op. cit., iv. p. 51 ff., cites a large number of supposed cases for a deter- minative k, but most of these seem to rest on no sure etymological foundatioB. • Op. cit, iv. p. 58 ff. 8G RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. ; i 1 : u ! !f ll 'ji ^ w> 4 J r r,' roots that differ from similar forms only in having ff instead of A" at tho end, and assumes that g" in such cases is only a weakening of k. This is hardly probable. We find no such regular concurrence of p and b in secondary forms, nor of d and t ; and it is not likely that fc alone of tho hard mutes would thus be softened, g" is also an independent Indo- European sound, of at least as much radical importance as k. It seems best, therefore, to assume that the affinity of the ideas to be expressed, was conveyed to tho ear by the employ- ment of similar sounds. Out of the many cases cited by Fick * in which gh is sup- posed to be a determinative, we can regard as well established only dhargh (Eng. drag^, as related with d/iar, to bear. t is plainly a determinative in kart, to cleave, as compared with kar (= skar, shear') ; in pat, to rule, as related with pa, to protect ; and, perhaps, in pat, to attain to (^peto,Jind'), as connected with pa, to obtain. d seems to appear certainly as a determina'. e only in a few roots. One clear case is that of mard, to crush, related to mar. For sad, to sit, there appears evidence of a primary sa, in Sanskrit ava-si-ta, literally, situated, and Latin si-tus, po'si-tus, placed ; mad, to measure, as compared with ma, is also probably Proto-Aryan. dh is found as a determinative in a few well-proven cases : levdh, to conceal, may be compared with ku (shu), yudh, fight (join battle), with i/u, to join. Final n in roots appears to be often a mere nasalized vowel. We may compare gan, to beget, with ga, (as in ye-^ov-a, 7e-7a-a>9) ; tan, to stretch, with ta (as in 7a-T09, ra-a-t?) ; man, to measure (as in mensus}, with ma. p is one of the most common Proto-Aryan determinatives, and easily recognized in most cases. We may bring together karp (kalp) to procure, help, and kar, to make ; dap, to divide out, and da, to divide ; rip (Jip, ar\eiit , V 88 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. i I kinds of wordu are also BupjKJScd occasionally to i)crf()riu the same olTico, as the adverb su, well, in svad, to taste (^dvb-uvw, iJ8uv, sivect}, made up of su and ad, to eat. Some of the alleged instances of such combinations arc very i)lausible, and many are not so. For full discussion of tho whole huI>- joct, the reader is referred to Curtius' Grundziigc,^ where tho theory is, we think, shown to bo untenable. The results of the investigation am brielly these : 1. Of those forms which differ from others in showing an additional element, there is only one group that has this at the begin- ning, namely, those in which s apf)cars as the added factor. 2. There is good reason to hold that no root is modified by the insertion of any letter : the infix n we may call a stem- determinative, rather than a root-determinative. 3. Wo have found the vowel a used as a post-determinative, and also nearly every one of the original Indo-European conso- nants. If we compare the various forms in which tho additional letter occurs, it will be seen that these added sounds are of different degrees of significant value, and that the same sounds are not always of equal importance in this res|X5ct. Thus the vowel a seems to have usually little modifying power; but mna (== mana), to think upon, remember, is clearly discriminated by it from the more general man. Again, the added nasals seem sometimes, like the inserted w, to modify stems, rather than roots ; but in dam., to subdue, tame, we have an obvious specializing of da^ to bind.* Again, the initial s (as in snu, to float, compared with wm), gives or takes away no apparent force, in most cases, from the shorter form ; and for this reason, as well as on account of the general uncertain tenure of the s in various languages of the family, Fick and others choose to regard the longer form as the earlier, and so do not consider s as a determina- ^ See in the fifth German edition, p. 31 fF. English translation (of 4th cd), p. 38 ff. "^ Fick, in his classification, to which we have been very much indebted, gives m and n a. place by themselves as being of less importance than the other dete^ minaiives. i ji i- ::i / ! ^ RELATIONS OF THE AUYAN AND SEMITIC LANGUAGES 89 tivo at all in such cases. It is impossible to prove, however, that the s was really dropjKjd from the hcgitmiiig of any Proto-Aryan root ; and it would seem to be more in accordunco with analoiry in root-formation that the shorter form should have preceded. But wo think we can show, in one case at least, that the s is a true determinative, and tlie shorter tho more primitive form. Tiio root tan, already alluded to, means to stretch. But it yields derivatives which, along with this sense, also express tho notion of sounding. Thus Skr. tana and Gr. t6vor than any one of the northern languages. But there should be as little doubt that the members of the northern division, taken togetiier, must count for more than the southern alone, in the comparison of roots. Cf. Schrader, Zeitschrifk d. deutschen morg. Ges., Vol. xxvii. p. 401 flf. ■'I.' ! 1! :!' ii i i :i «' ■,!' 1 1,: I ! 94 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. may seem to havo disregarded tlie principle in some cases, which wo shall notice presently. The question at onco arises : Must we hold that all these roots were tri-consonantal from the beginning, and that the apparent exceptions are only degenerated, shortened forms ; or do any of the roots show peculiarities that would lead us to infer that they have developed from more elementary conditions ? An affirma- tive answer seems due to the second alternative; and, though this is not the place for a full discussion of the matter, we shall adduce a few of the considerations that seem to point clearly to that conclusion. First, we have the co-existence of a large number of roots of similar sound and related meanings, which differ from one another only in one of tb.e radicals. Thus (a) the first two consonants of each member of the group are the same, the third being different throughout the list ; or (6) the last two radicals of some roots may contain the constant formula, the first being the variant; or (ir) the second letter may appear as additional, the first and third representing the essential significant combination. This would seem to show that the forms with the variant letters were developed from earlier roots represented in the present stage of the language by the two constant letters in each hometymous group. Further, we have still more conclusive evidence from those hypothetical forms in which the third radical is the same as the second. Comparing with class (a), mentioned above, we find that in nearly all those groups of roots which agree with one another in the first two consonants and differ in the last, there appear forms in which the last letter is not a variant, but merely the second repeated. Moreover, such forms (giving rise to the so-called ^^9 stems) are generally more comprehensive in meaning than the related roots with variant letters, containing the generic idea whose specific modifications are expressed by the divergent forms. These facts indicate that they represent an earlier expression of thought thiin the longer roots, and this is naturally obtained by dropping the repeated consonant. In other words, we »• '.- RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 95 infer that the early speakers developed these aHsuined tri- literals from earlier biliterals by simply repeating the second sound. The production of the hometyraous forms is thus more easily accounted for, upon any theory of phoiiolo), as compared with VP. which, though only found in Hebrew, is almost certainly Proto-Semitic. "We may also compare '«'', the root of the Semitic word for the right hand, with ''sk, to be firm, found in all the divisions of the family ; and "I'S-^, to be right, prosperous, with the kindred 10X, both Proto-Semitic, as being found in all the dialects. •• was not employed in this way by the early Semites nearly so often as i. n is probably a determinative in the Proto-Semitic psn, to press, choke, make narrow, found in all the dialects, either in noun or verb stems. This may be connected with the equally ancient p5s, to put round the neck, if the primary notion of the latter is of close binding ; while tho Syr. ■ »t *-- Chald. p|«j , Arab. j££^, to strangle, is clearly a kindred causative. Another case is perhaps the h in Proto-Semitic iin , to let go, cease, etc., as connected with V*i, to be loose, which is devel- oped in various forms throughout the family, onn, to close, seal, may possibly furnish another example, but the proof would be precarious. We must acknowledge that tho evi- dence is not conclusive fov any other instance of the use of n as a predeterminative. The persistence and independent 1 According to the law discovered and established by Oppert (see his Qrftm- maire Assyrienne, 2d ed., 1868, p. 9 f.), the Hebrew ^''B forms nsually become *KD in Assyrian, if they correspond to ^lo in Arabic ; but when the Arabic preserves the Hebrew i the Assyrian does bo also. The Hebrew forms require no explanation. ' .1 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 101 force of this sound from tlio earliest Semitic times, is one of the most important facts in the phonology of the system. c is a predetcrminative in one or two roots with a causa- tive force. Thus biso, to extend, lengthen (Ueb., Arab., and Targ., either in noun or verb stems), may be compared with bio , to be long (as in Arabic ; the Hob. b^urj means to throw = send along). Such developments were common enough in the several dialects in their seimrate history. In Ethiopic they became quite fashionable. In the primitive speech they were very rare — a fact which may perhaps go to show that a as a servile letter was of later origin than some of the others, being a nominal, not a verbal formative. 3 was a very common Proto - Semitic predetcrminative. Thus, ins, to give (Heb., Chald., Samar. ; the Assyr. pa shows a customary softening of t to ar.s likowiso in all tho divisionH of the funiily. rio also •nn, to revolve, keep going (in Assyr., Aram., Iloh., and Araltic, citlicr in nonn or veil) Htonis), deveh)ped from tho ancient common root "n (cf. "iin). In this use n is nearly as common as ». ^, more frequently than any other letter, represents an in- ternal dovolopmeiit of tho root. It is, of course, demonstraldo that this shows a secondary form only when wo can compare with the simi)ler so-called 'ss roots. Such cases, iiowevor, are quite numerous. Thus wo have "iia, to turn aside, sojourn, found in all tho dialects, as compared with 13, to turn, to twist, to roll, equally Protcj-Semitic ; inn, to revolve, as related with *n, which expresses various kinds of irregular motion in the different dialects. We may compare also 113 and 13, l)oth primitive roots expressing rapid motion and flight ; *is and "iis, both Proto-Semitic, of which the former means, to arrange in a series, to number, and the latter, to re- peat. 3llany other cases might be cited ; and it may be slated as a general fact, that when we have an '"y and an Vj root, side by side, with tho first and last letters the same in both, the radical notions in both may be easily connected. Objection might be brought on the score of the want of association between a few of such cases. The only exceptions we know of in Proto-Semitic arc the roots from which spring ni*', day, and c, sea (but wo have not any verb-stems from these roots, and therefore can say nothing as to the primary meanings), and bin, to whirl, twist, which does not seem connected with bn, to pierce, to open, aio, to return, may be explained as connected with 30, to turn around ; at least, that is the only primitive root with which it can be compared. It is now proper to give what seems to us to be the true view of the origin of these forms. It being quite certain that inflection had begun long before the roots had been universally raised to the tri-consonantal type, the matter of ' 'ii I I I 11 . 1 * rl I- i \ . I 104 RELATIONS OF THR ARYAN AND SEMITIC LANGUAGES. asHimilating the HJiorter furms to that Htandurd wuh nuconi* pliHhed apparently in this way. Wliilo tho 'ys roots rcauljcd this level by having the 8eeond radical emphasized or douhled (and afterwards, in certain inllection^, repeated), the*'r roots entered upon the same stage by having the characteristic vowel of each Ktcin lengthened. Thus fcam in indcctiou would heconio kum : and knm, kum} Not till a nnich later period did the more highly developed of the Semitic dialects, Arabic and Ethiopic, make of these stems distinct roots. From this it follows that a medial i represents merely a lengthened inflective vowel in Proto-Semitic, and not a radical sound. n as an ancient indeterminativo can bo held to he i)robabl6 in only one instance that we can adduce. A plausible case is •Tw, the root of I'^rro, price, which it would seem jjrofxjr to connect with "iha, to sell, and *>"«, to exchange.*'' JJut it is not Proto-Semitic in that sense, only Hebrew ; the Assyr. mahirn, offering, tribute, which Lenormant^ connects with ^flra, being derived from the native root "ina, to bo in front, and in causative forms to bring before, or present.* A surer instance is found in *no, to go round, traverse (Heb., Aram., and Assyr.), as compared with *ino, to be round (Ileb. and 0^ Aram. ; cf. Arab, .^g a , moon, with Heb. Tiinig, and Syr. l5oup)) hoth of which may bo connected with Heb. ^«id, to turn aside, from the primary notion of bending. Of course, it may be suspected that "ino may be merely a strengthened form of "ino, especially as in Assyrian the former root has the intransitive meaning attaching in the other dir.lects to the latter. In general, we may say of medial n what has 1 See a brief but instructive discussion of this question by Prof. A. Miiller in Zeitschrift d. d. morg. GescUschaft for 1879, p. 698 ff. ^ Not with "^5^ to sell, which is probably a secondary, derived from f^^ to buy (cf. the use of q as a prcdeterminativo discussed above). " Etudo Bur quelques parties des syllabaires cundiformes (Paris, 1876), p. 247. * The conjecture of Fricdr. Dclitzsch (Assyr. Studien, Part i. p. 125), that the Hebrew and Assyrian roots are connected, is probably wrong. : ! B£LATI0N8 OF THE ARYAN AND SEMITIC LANGUAGES. 105 been snid aliTiuly of initial n, tlint it i« normally an intlc|)cn. dent Htiihlo Hounil. Modiiil -> uppoars to ivproHont an cxpan«ion of tlio root in several cusch. n^, to j)lare, lay down (in noun or vcrli 8tcni8 in Hen., Aram., and Awsyr.), nuiHt 1)C compart'd witli nttJ, which in also poRwihly j)riniitive, iieinn found in l»oth Ilcb. and Aram, in the Kamo sense. So also apparently with Yp, to fasiiion, forge, as compared with yp, to set right, prepare. In thcHo •• appears to bo Proto-Somitic ; and ye* here, as well as in tlio many cvlhca where '•» and '^jr forms exist nido hy side in tlio wamo Honse, it is very doubtful whether the ^ is pri- mary. It seems more i)rol)ublo that it took the place of 1 in these instances ; it having perhaps been shortened from the causative form of the verl)-atem in each case, since such Ny stems are mostly transitive. If this view is correct, wo cannot maintain that ■» represents a Proto-Semitic indetor- minativo, but are obliged to hold that medial "^ stands with medial i for that very early lengthening of the inflective vowel by which the primary roots were made to assume a triliteral guise. 9 is an indeterminative in *i5a, to be separated from (repre- sented in Ileb., Arab., and Ethiopic) as compared with the universal root *ia, to divide ; also in nra, to cut off, ( nsume (appearing in Heb., Aram., and Arabic) as related with the primitive root *a, to divide ; so too evidently in *i5::, to be small (in Heb., Aram., and Arabic) as developed from ix, to press together, contract, also Proto-Semitic ; and in several other cases, amounting to about one half of the whole number of roots in which » appears as the middle radical. In nearly all the remainder with medial s the first letter is a determinative: thus, it would seem, » was not liked as the second letter of primitive biliterals, while, as we have seen, it was frequently employed as the first — an instructive fact in Semitic phonology and morphology. These are the only letters we can regard as undoubted Proto-Semitic indeterminatives. Others (as 3,-i,b,p) were I! I' If i 106 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. ;.,,: i; ill if' I : ) 1 4 .1 ,i. r n used raoiG or less freely in the different dialects during their separate history, especially in the formation of quadriliterala, which arc all secondary roots.^ Lastly, we have to take the final determinative letters in Proto-Scmitic. These are much more numerous than cither of the other two classes ; the true place of the additional sounds in secondary roots being at the end, as in the Aryan family. X represents a post-determinative very frequently. So in xna, to hew out, fashion, create (in Heb., Aram., and Arabic), from -13, to cut, which is variously represented in all the dialects. So also in ttba, to shut out, to obstruct (Heb., Aram., and Arabic), as compared with ba, to shut, close, finish (found in noun or verb stems in all the dialects). It appears in many other examples that might be cited ; and we are inclined to set it down as a principle that wherever M appears as the last letter of a root, it is of secondary origin, unless the first letter is a determinative. This might be inferred from the character of the sound itself, which only exists for the sake of its vowel ; but it may be proved in nearly every case by actuf.i comparison with kindred forms. The only instances in which this is not practicable are prob- ably xbo, to fill ; K5p, to be mov>.d with passion ; and kcx, to thirst ; and here it is better to assume that the kindred roots are lost or their connection obscure, than to maintain that the K stands so exceptionally foi an independent consonant. a is apparently a post-determinative in ana, to be scabby, leprous (Heb., Aram., and Arabic in noun or verb stems), from the widespread root *ia, to scrape ; in aisn, to hew wood (Heb., Arab., and Ethiopic) from the common root an. to cut ; in ab:c, to hang up = make incline (Aram., Arab., and 1 If the Troto-Semitic root IPS to prepare, could be regarded as having a similar origin to that of Aj conj. VIII. in Arabic, an instance would be at band of the use of n senrile as an indeterminative ; but this we cannot regard as probable. V .i . M RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 107 :a Ethiopic), as compared with »bs, *-o incline, also Proto- Semitic ; and perhaps in a few other cases. a is a post-determinative in abu, to divide (in various noun or verb stems in Aram., Heb., Arabic, and Ethiopic), from the root is, to cleave, burst asunder, variously represented in all the dialects ; and perhaps in a"i*i, to go, proceed by steps (Aram., Arabic, with a Heb. noun-stem), as compared with "p"i, and the primary root *in, which seems to express lively motion in general. We cannot adduce any other probable instances from Proto-Semitic. T is a post-determinative in *ron, to be ardent (with related meaning in all the dialects^), as compared with en, to be warm ; also in "iib, to separate (in Heb., Aram., and Arabic), from no, to rend asunder ; and in several other cases. jn is an post-determinative apparently in pjba, to be stupid, embarrassed, timid (cf. the Hob., 4.ram., and Arabic mean- ings), from ba, to be confounded, confused ; probaldy in >nix, the root^ of a Proto-Semitic name for God (Heb., Aram., and Arabic, which), as we prefer to think, is a denominative from the shorter bx, also proved to be Proto-Semitic by the Assyr. il-u ; and, in general, wherever it occurs as the third radical, as it does but rarely in the primitive speech. 1 , or rather the vowel u, was used as a post-determinative in the primitive speech.^ So apparently in 'ba , to draw off, lay bare, reveal (in Heb., Aram., and Arabic ; in Ethiopic, to draw on, cover), as compared with a root ia, evident in aba, "nba, nba, of kindred meanings, all Proto-Semitic. So too 1 The Hebrew and Chaldee forms mean to desire ardently ; the Arabic has one meaning, to be angry (or "warm") ; another to deem worthy of praise, i.e. desirable ; the Assyrian means to hasten, or pursue ardently. " What the specific meaning of this root was, or whether it ever had more than a theoretical potential significance, is doubtful. The Arabic meaning, to adore, is probably secondary, = regard as God. ' It is not easy to say in all cases whether u or i was the original determina- tive vowel. It is only in Arabic and Ethiopic that the distinction bctweoi. the two has been regularly preserved. Moreover, in these languages so many new roots were developed in later times with these as f.nal sounds, that the question of priority is still further obscured. It is only where the two idioms agree in important roots, that we can infer surely as to the real state of the case. m' * ' I'.' ^ lu ! fs| i !i . *'-l ??■! 108 RELATIONS OF THE ART AN AND SEMITIC LANQUAQES. 'i i \i ^'"1 li 'fJ u \ ■ 1 ! f! in ibi, to let down, suspend, weigh (cf. the various related meanings in Assyr., Ethiop., Heb., Aram., and Arabic, which has also •'H), from the root bi, to hang loose, no less widely represented through the system ; and in other cases that might be adduced. t is a post-determinative in na, to pierce. This root i» found only in Heb. and Aram. ; but it is proved to be Proto- Semitic by the word for iron, bna (bne), which is found in all the dialects, and is evidently developed from it, as we shall see later. The ultimate root is *ia, to divide open, already frequently cited, nt, to separate, branch out, is also Proto-Semitic, from the common root *«, related to -a. T, however, is rarely used for this purpose, as we would natu- rally expect from the fact that it is a secondary sound arising from s : cf . in Hebrew I'lB, gib, pa ; tbs, Dbs, ybs. n is a frequent post-determinative. So in n*a, to pass through, to pass out, escape (cf. Jie Heb., Arabic, and Ethiopic stems), as related with la. So also in nia, to make bare, smooth, bald (Heb., Aram., and Arabic), as compared with hba, etc., cited above. It is found, besides, irf a few other cases ; but was employed far more frequently in each, dialect after the dispersion of the family. ts is perhaps a post-determinative in obs, to break away, escape (cf. the related senses in Heb., Aram., and Arabic ; the Assyr. oba, to live = to be preserved, is the same root), from ba, to cleave or break open. Possibly, also, in o*in (Aram., Arabic, and Hebrew in noun or verb stems, and perhaps Assyrian) to cut into, grave, engrave, as compared with a root 'vt, represented in Arab. ^ , to cut open, pierce, divide ; in oin, etc. The Heb. and Arab. »"io of like meaning, we may compare with a root *no, represented in the Heb. *iit) and into, to saw, and elsewhere. t», however, was not a very common determinative. •',1 or rather the vowel t, was apparently the most common of all the post-determinatives. The following are a few of its examples : •'ss, to smite, injure (Arab., Ethiop., Heb., and 1 S«e the remarks just made on i as a post-determinative. ;l: RELATIONS OF THE ArYAN AND SEMITIC LANGUAGES. 109 Aram.), from a common root^s, cognate with 53; •'pj, to l)e separate, pure (with interesting derived meanings in Arab., Aram., Heb., and Assyrian), as compared with ps, a widely represented primitive root, meaning to strike asunder ; np, to erect, to establish, acquire, possess (in noun or verb stems in all the dialects), from "jp, to be erect. 3 is a probable post-determinative in yn, to tread (with various associated meanings in Heb., Aram., and Arabo- Ethiopic), as compared with a-n and the primary -n cited above. Also in "p^) to break in pieces, crush, oppress (cf. the noun and verb ■ t.eras with related meanings in Syr., Arab., Heb., and Assyrian), from the familiar root *\t, to rend asunder ; and in a few other instances. ^ is a post-determinative in b*ia, to twist together, make strong or great (cf. the various meanings in Aram., Heb., Arab., and Ethiopic), as related with the root ia, to bind, which appears in "lax and "j'^a, both Proto-Semitic. It is also found in bia, to tear off, drag off, as related with *ia, already cited (both of which are found in Heb., Aram., Arabic) ; and in a few other cases. s is a post-determinative in d», to be firm, strong, great (cf . the noun and verb stems in Arab., Heb., and Assyrian '), as related with y9^ to be strong, as found in nxs, y'\s, etc. Also in D-'S, to be naked, bare, as compared with nis and m», of a similar meaning, all of them being Proto-Semitic. A few other cases might be adduced. 3 as a Proto-Semitic post-determinative can hardly be proved. The only plausible instance we can adduce is pa, the root of the Proto-Semitic word for threshing-floor (Heb., Arabic, and Ethiopic), which seems to be developed from a root "ia, of manifold expressiveness, but having clearly the general sense of dragging along, rubbing, crushing, so that pa may perhaps be = the place of threshing grain.^ -jaa, to 1 The Assyr. asmu means material, analogous with Heb. d:{|; bone, in the Inscription of Ehorsabad, line 164 (see Oppert's Commentaire philologiquc). " There does not seem to be any verb-stem pa cleariy Proto-?emitic, which woTild give a stiitable intermediary sense. The Arabic ' I^, however, means il :t.| I' I -f. Is ?rt 1 10 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. be curved or arched, if it is Proto-Semitic in that sense, might be connected with aa of kindred meaning ; but it is difficult to comprehend all the divergent meanings of the former root under one general satisfactory notion, a was used more freely for this purpose in each dialect after the family separation. e^ is a post-determi native in eiB, to cleave asunder, break up (cf. the noun and verb stems in Heb., Aram., Arab., and Ethiopic) as related with the familiar root "id ; and in a few other cases. » is a post-determinative in »is, to hew off (Heb. and Arabic) from the root "»a, variously represented in the sense of cutting; in 5-it, to scatter, to sow (represented in Heb., Aram., Arab., Eth., and Assyrian), from the root it, to spread, scatter, shown in lit and several other kindred forms ; and in many other cases. It is clearly a determina- tive in nearly every instance of its use as the last radical. Those few cases are of course excepted when the first letter is a determinative, as in »iti, to place ; stsD, to set in or set out. It is probable that no ultimate triliteral ended in 9. sate, to be full, satisfied, is probably no exception.'' Those who hold to a common origin of yao ^ an J the Indo-European word for seven will have no hesitation in considering the 9 as secondary in the former word. D is a post-d|terminative in cpa, to carry away, sweep away (in noun or verb-stems in Heb., Aram., Arab., Ethiop., and Assyrian), from the root -la, to drag along, already to grind corn, thus furnishing a notion kindred to the one required. Its other meaning of smoothing, wiping clean, does not throw satisfactorj light on the word for threshing-floor, though it is usually assumed as explaining it. i The distinction between this and O was, as we have seen, obscured in some of the dialects. The Hebrew b appears to have preserved the sound best, though not in all cases. With it agrees in general, the Arabic i>m , the Ethiopic M the Aramaic UO and Q , and the Assyrian a, as it is conventionally represented ; though the disagreements are frequent, except in Assyrian. * See Gesenius' Thesaurus, p. 1319, for kindred forms. ' Not sat). That the other is the Proto-Semitic form, a comparisoa of Auyrian iUw with the Arabic and Ethiopic shows plainlj. RELATIONS OF THE ARYAN AND SEMITIC IJINOUAGES. 1 1 1 alluded to; in tisj (Heb., Aram., and Ethiopic), to smite, from the widespread root as, to strike ; and in several otlier plain cases. X is a post-determinative in -pt, to cleave or break open (Heb., Assyr., Arab., and Aramaic), from the common root IB, to divide ; and in a few other instances eqa^Uy clear. p is a post -determinative in pn^ to scatter, sprinkle (in noun or verb stems in Hob., Aasyr., Aram., and Arabic), from the root *it already referred to ; and in several other r? 'TIS. ■« is a post^eterminative in lot, to open (with various associated meanings in Assyr., Heb., Aram., and Arabic), from the common root xit (pb) of kindred meaning ; and in many other forma that might be cited. to is a post-determinative in td-ib, to separate, scatter, disperse (Heb., tons; Aram., ua^s and ti^jB ; Arab., jiji; Assyr., «ib in Niphal, to flee away), from the familiar root *iB. It appears besides in only a few other cases ; but, like e, was more commonly employed as a secondary formative in each dialect after the Semitic dispersion. n also is an infrequent post-determinative. It appears in pes, to be silent and bring to silonce (cf. the associated mean- ings in Heb., Syriac, Arabic, and Ethiopic), as related with the root as, with the primary notion of binding, shutting up, which is extended in the different roots so as to express the divergent ideas of fasting, deafness, dumbness. It is found also in a few other cases, and in some instances of its occur- rence the root is perhaps a denominative, formed from a feminine abstract. We must now put together the results of this investigation into the structure of Semitic secondary roots, and try to classify those sounds used in forming them. First, as to predeterminatives, we found that k, n, i, n (probably), % a, a, 5, w, and r were thus used. Of these « represents only a prefixed vowel ; for though it is a true consonant it is only used in the interest of the vowel sound that conditions it. u ! li I' Uf I is- m ' '■■ ' h i ■■*■ 1 I. i. 112 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. With regard to i and ■', it might seem doubtful whether they were originally prefixed as consonants, or as the corre- sponding vowels u and i. On the whole, we incline to the belief that they were at first vowels, and then in course of inflection hardened into semi-vowels. For this the following arguments may be offered: (1) the analogy of the post- determinatives *> and "^ ; (2) the frequent interchange ol)- served in every Semitic period of 'lu or '•'B with *«s forms developed from the same primary root, — a phenomenon easy of explanation upon this theory, but more difficult upon the other ; the '"19 stems being, as we have seen, merely vowel expansions of 's» forms ; (3) the fact that consonants are not normally liked as predeterminatives : o, 3, ©, and n are used because they are inflective formatives ; the other con- sonants are breathings, and of them n and t are rare, and n doubtful. In all probability we may set down » as repre- senting fl, 1 and "^ as representing u and i respectively, when used as predeterminatives. n, n, and 9, used as predeterminatives, probably arose in this way. n is the surd breathing corresponding to the sonant k, and arose from it through the process of dialectic Tariation familiar in all languages. Its rarity as a radical prefix is a proof of its late employment for this purpose. "Prom it n arose by strengthening, and was employed still more rarely. s> is the deep guttural development of » ; and as n is rarer than n, so 3» is rarer than k as a predeterminative. The true consonants used as radical prefixes, o, s, o, n, are among the rarest used as post-determinatives ; while other con- sonants, some of which are very common at the end of roots, are not used at all as predeterminatives. The solution of this enigma 3an only be gained from the consideration that these are letters used frequently as prefixes in the formation of verb or noun stems. And it is remarkable that the fre- quency of their occurrence, respectively, varies according to the priority of their introduction as stem-formatives, as the phenomena of the Semitic idiom seem to indicate : 9 is most commonly employed, then to, n coming next, and finally », I* RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 1 13 which seems to have been used for only a short time before the family dispersion. We liavc, then, as Proto-Seraitic predcterrainatives the vowels o, t, u (wliich were displaced by the corresponding n, % and N under the later consonantal sys*^era),the breathings n, n, and s, and the consonants v, 5, o, and p, originally inilectivo forma- tives, themselves relics of old independent stems or words. All of these, save the vowels o, i, m, were introduced in the consonantal period. As indeterminatives we found the breathings k, n, and s to be used, and n, whicii is rarest as a radical prefix, does not appear here at all, being too much like a true consonant. These all belong to the consonantal stage of Semitism, as also does the vowel expansion, already treated of, expressed currently by i and "^.^ As to post-determinatives, we found that all of the conso- nants, with the possible exception of a, were so employed. M, 1, and •'j however, represent vowels that were used as radical affixes before the establishment of the consonantal rdgime. As in Proto-Aryan, so in Proto-Semitic, the regular place for determinatives is the last part of the root. A study of the character of the prefixed and inserted radical letters, as compared with the post-determinatives, makes it probable that they would not have been used at all, except in the interest of a manifold development of roots ; since the need of various expression, as ideas multiplied, could be met in no other way ; the genius of Semitism, unlike that of Aryanism, being averse to the use of compound words. There are a great many Proto-Semitic roots which, so far OS can be seen, show no determinative letter ; and there is, of course, every reason to suppose that many of these, as well as many of the Aryan ro3ts, possessed three consonants from the beginning. Of quadriliterals there are no sure examples in verb-stems. In noun-stems there are a few whose triliteral origin is apparent. ^ Of conrae there is no inconsistency in making M at the beginning represent a primary vowel, and in the middle a consonant ; for a vowel must have heea heard already in all vocal expressions beginning and ending with a consonant. \H r lU RELATIONS OF THE ARYAN AND SEMITIO LANGUAGES. i I i 1 ■ If I iff. ' 1 -I I: Two classes of cases yet remain to be considered. First, wo have those triliterals in which the third radical is the same as the first. This form, which seems so inconsistent with the ordinary types of Semitic root-structure, is accounted for by an analysis of the roots in question, from which it appears that they are developed from shorter forms by the repetition of tlie first radical.* These also occur in noun- stems in Proto-Semitic, not in verb-stems, except, perhaps, in denominatives. They are common enough in the several dialects as developed later, where their origin can be clearly traced. Another and very important class of secondary roots are those *is roots that end in M, as the Proto-Semitic Kia, to go in. With regard to such cases we claim, without hesitation, in accordance with the principles already established, that the root originally consisted of a consonant and a vowel. The root was raised later to the triliteral standard only graphically, and not in actual speech, just as the Hebrew vb, not, is sometimes written Kib, though it was never anything in sound but Id. The fact is, that the Semitic roots, before the consonantal period, had as great variety of form as the Proto-Aryan. It is an error to maintain that all the Semitic roots are ultimately tri-consonantal ; but it is also an error to hold either that all were developed from biliterals, or that in general the bi-consonantal form is their shortest or ultimate type. Guided by the principles above set forth, we shall now attempt to draw up a scheme of the possible and actual root- forms in the two systems of speech. I. A Proto-Aryan root may consist : (1) Of a consonant and a vowel, as 'i,^ to go ; A;t, to lie down ; da, to give. p hi \ [hi ' l-iU • i, . K-'H h 1 This throws light on the origin of a nnmber of obscare words ; for example, hdl) the Proto-Semitic word for gate is, a» we conjecture, from the root jta, to go in, enter. '■^ The Greek ' is here used to represent the breathing, corresponding to », which precedes every vowel-sound at the beginning of a word or sellable. ii'i ^ ! RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 1 IS ». (2) Of a consonant, a vowel, and a consonant, as W, to eat ; pat, to fall ; tar, to go through. (8) Of two consonants and a vowel, as A/u, to shut ; pri, to lovo ; pru (^plu), to swim. (4) Of two consonants, a vowel, and a consonant, as dram, to run ; prak, to ask ; praty to spread out. (5) Of a consonant, a vowel, and two consonants, as kart, to cut ; b/iarg; to shine ; mard, to bruise. (6) Of two consonants, a vowel, and two consonants, aa sparg-h, to strive after ; smard, to gnaw at. A root in any of these classes but the first maj be secondary. In class (6) probably all, in class (G) certainly all, are secondary. II. A Proto-Semitio root (taking in both the preconso- nantal and the consonantal period), miglit consist (1) Of a consonant and a vowel,^ as »ia < «a, to go, or go in (Ileb., Ethiop., Arabic, and Assyrian) ; ''»n <^ »i, to see (Heb., Arab., and Ethiopic). (2) Of two consonants, as *i3, to separate (represented ia all the dialects) ; 15, to be strong (in all the dialects). (8) Of two consonants with internal vowel expansion, as i^K, to be strong, superior (Heb., Arab., and Assyr. in noua or verb stems) ; lis, to be set up, or established, exist (in all the dialects). (4) Of a consonant, a consonant, and a vowel, as Kba, to shut up or out (in all the dialects) ; ^Vn, to let down, sus- pend (represented in all the dialects) ; •'pa, to be separated, pure (represented in Arab., Aram., Heb., and Assyrian). (5) Of a vowel, a consonant, and a consonant, as *iax, to be lost, perish (Heb., Aram., Ethiopic) ; iai , to contain, be capable (Heb., Arab., and Assyrian) ; noi, to be right, pros- perous (in all the dialects). (6) Of three consonants, as "^na, to kneel, bless (in all the dialects) ; lo^p, to be pure, sacred (in all the dialects) ; isbo, io be strong, to rule (in all the dialects). ^ In this classification a vowel is cited as an integral part of the root, only when it is original and determinate. :il i 01 i ■ . •; i I w 1 1 6 RELATIONS OF THE AYRAN AND SEMITIC LAN0UA0E8. 1 .1 •vuj w ■'• >) (7) Of four consonants. Noun-stems, as bna, iron (rep- resented througliout the system), presuppose a true root ; and nuriB, to 8})read out <; una, is certainly Proto-Seraitic, being represented in Hebrew, Arabic, and Assyrian. A root in any of these classes but the first and second may be secondary. In classes (4) and (6) probably all, and in class (7) certainly all, are secondary. In the next Article we shall consider whether the morpho logical differences between the two systems of roots may be reconciled, and enter upon a comparison of the roots that may seem to inyite such treatmeul. 1 .1: f iii {■ - I'' 'I la • ■; CHAPTER V. COMPARISON OP ROOTS. ■t . Ha VINO in the last Chapter taken up the most important questions relating to the formation of the predicative roots^ considered as primary and secondary, in the two systems of speech, and having presented a scheme of the typical forma under which these roots are expressed, it remains for us to determine how wo may reconcile the seemingly discordant principles according to which they are formed. The main difficulty presented arises from this fact, that while in the Aryan system the vowel is a significant part of the root, in the Semitic, on the other hand, — at least in the inflectional period of that idiom, — the vowel is not essential to the expression of the radical idea. The difficulty is great, hut perhaps not insurmountable. The following considerations are o£fered as tending to show that a reconciliation is possible : (1) The Semitic principle of root structure bears evidence of a secondary and, so to speak, artificial origin. In the language as it is presented to us, the vowel is not co-ordinate with, but subordinate to, the consonant. Now, we do not claim that the vowel once heid an equally important place with the consonant. If language is a growth, and not an institution, the two elements cannot have' been originally co-ordinate, even in those systems of speech where we find them currently of equal value. The consonants, as the harder and more stable elements of speech, must have secured their independent recognition and employment before the vowels, in all early forms of human language. But it may be said that the Semitic is an exception to other systems in 117 I II' 11 JW w % 118 RELATIONS OF THE ARYAN AND SEMITIC LAN0UAOE8. thiw, tlmt the vowels never Hccurcd complete autonomy for tlicmHclvoB. TluH m true ; l»ut it is not true that tlicy nlwuys filled that subordinate function which we see iiHHi^ned to them in die full-blown inflectional fteriod. It huH been Hliown already that voweU even formed a constituent part of distinct, inde|)endent roots; wo have not only an internal vowel ex- pansion, but also a development of secondary roots by the use of any one of the three original vowels a, t, m, each of which has maintained a distinct and clearly r(H;ognizal)lo influence until the latest Semitic times. We have even found that some roots coimisted o' a consonant and a vowel ; and if it cannot l)e clearl) shown in each instai^^e what that vowel was, it still remains true that, though it is there sub- ordinate to the consonant, its subordination is of an essen- tially difTerent kind from that which is seen in the function of vowels in the "strong" stems of the inflectional period ; it is, in fact, due merely to that indefini*eness which we have shown to be necessary to the vowel in all primordial Bpeech. It would, of course, be absurd to maintain that in the earliest Semitic the vowel was of equal importance with the consonant for the expression of radical ideas. But it would be just as absurd to hold that it counted for nothing. If there is anything which can be maintained with certainty as a necessary feature of primitive language in general, and of the constitution of its roots, it is this, — that in both the vowel played an independent part. On the other hand, the only sure inductioh <^rom the jihenomena of root develop- ment, as we have siiuied the subject, is, that the vowel was subordinate and fluctuating.^ (2) The P»*oto-Aryan roots also give evidence of a previous 1 Here, R8 well as in related discussions, it makes no difference what theory is held as to the nature of " roots," whether we regard them as having once been Rctual words, or as being mere abstractions — forms theoretically assumed as the basis of actual words. Unless the distinction between primary and secondary roots, to whose elucidation the last chapter was devoted, is an utter delusion, we shall have, upon either theory, to go back of the current triliterals, if we wish to determine these ultimate forms to which the name of " root" is applied ; and in the last analysis the indefiniteness as well as the originality of the vowel io •nch forms, will be equally apparent under either view. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES 111) Btago in their hls.ory when the vowel did not pos«oHH the curtain and Htnldo character nmnif(«Ht in their current fornm. At loaHt, it i» allowaldo tu infer as niucli us this from tiio fact that 80 many formH are found cxpnjssiujuf Uk! suuic or kindred ideas which agree in tlicir consonants and dilTcr in their vowels. Thus wo have Z*/m^-, to eat ; h/utij;-, to enjoy ; mand and m and, to decorate; mad, to ho exidted ; tni(d, to bo gay, joyful ; skad and skid, to Hplit ; as and t.v, to throw ; di and du > div, to Hhine ; pa and pi, to drinic ; bhad and bhid, to pierce, cleave ; si, to bind, and sn, to sew ; ska and skit, to cover; and a multitude of other divergent associated forms.* Those cannot very well 1)0 regarded as primary and secondary roots re8f)ectively, because there is no developniont of meaning and no addition or degeneration of form.'' At this point the two great systems of siKJCch seem to meet We find Semitic roots in which the vowel is indeter- minate, and yet an independent constituent ; and wo find Aryan roots with fixed consonants, but varying vowels. Doth phenomena are just what would be expected in the necessary development of early language ; and the subsequent diver- gence of the two idioms in root formation can also be explained. In both systems definiteness of expression was aimed at equally and necessarily. In the Aryan system tliis was secured by giving greater precision to the vowel elements in each utterance, till at last they were made co-ordinate with the consonants in every respect. In Semitic, on the other hand, the original vagueness of the vowel remained, and definiteness as well as variety of expression was sought through the multiplication of consonants, either with or without the use of determinative letters. Hence wo are prepared to find that while the bulk of the current Aryan roots have two consonants, and are monosyllabic, the bulk of the Semitic have three, and were perhaps originally dissyllabic, ^ Such forniB may be collected and collated from Pott's Wurzel-Lexicon, or more readily from Fick's Vergl. Worterbuch d. indogerm. Sprachen, Vol. i. ^ This extensive group must be distinguished from that small class of forms with vowel variationB which we cited in th« last Chapter as consisting of secondary root*. m '[■ i 120 RELATIONS OF THE ABYAN AND SEMITIC LANGUAGES. ;l i; ' ^•i.i |:' : Hence also it happened that in Semitic the vowel elements had less precision and importance in each utterance, till at last they lost their independence entirely, and became sub- ordinate to the consonants in every respect.* From this it follows that whatever roots in the two idioms are to be adduced for comparison must be represented by their consonants alor.e. This, of course, need not be any bar to an association of such roots, if they are eligible in other respects. For even within the Aryan range alone a consonantal formula might often be chosen as comprehending the same idea under various vowel variations. Thus, in accordance witli examples of roots just cited, MD might convey the general notion of highly wrought feeling, and S* ( 5 -|- an indeterminate vowel) might stand for the idea of fastening together ; just as in the Semitic sphere ni means to be high, and ks means to go. We thus see how the Proto-Aryan and Proto-Semitic roots may be brought together, so far as the forms are concerned. It remains for us to determine what kinds of roots are to be compared as regards their signification. (1) First, it is evident that we must exclude those roots which are clearly onomatopoetic. In many languages through- out the world we find the same or like forms occasionally used to express the same ideas, when the sound seems to be a sort of echo of the sense, as when words seem to be * J. Grill, in an elaborate Ebsaj in the Zeitschrift d. dentschen morgendl. Gesellschaft, Vol. xxvii. pp. 425-s^60, attempts to show that the roots of the two systems may be unified in structure by reducing them to a hypothetical stage of development in which the vowM a alone was heard in them all (p. 449). Under those circumstances he thinks the vowels would not count for anything m determining the specific expression of thenroot-idea, since they would be the same in all the forms. Tho validity of this conclusion depends upon the correctness of the assumption of such a form of speech, an " Alpha-Sprache " as he terms it. But there is no strong evidence of it. The preponderance of the vowel a in Aryan roots may be accounted for on the principle that it is the most com- mon of sounds in general, not necessarily the only primary vowel. The reader is referred also to the criticisms upon the similar, but not so far-reaching, theory of Fick, mude in our last Article. On the other hand, we have abundant evidence of the original vagueness and variations of the vowel-«onnds in tbe roots of both systexBs. m RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 121 simply ■ litative of the movements of the objects of nature, or of the utterances of men or lower animals. Some writers have made undue use of this fact, and applied it to the explanation of many cases in which onomatopeia has b id no part. It offers an easy solution of innumerable difficulties, and can often be plausibly appealed to when no etymon is at hand to wl ich a given form may be referred. Thus the comparer of obscure roots runs a double risk. On the one hand, he is liable to cite forms as being of kindred derivation whose likeness is due to their origin in the imitative tendencies of early speakers ; and, on the other hand, he is in danger of being accused of citing cases which are all " more or less onomatopoetic," and therefore not necessarily of common origin. Now, while it is true that such a charge has often been made unjustly against etymologists, it is not to be denied that it has always been made with some justice against those who have attempted to compare Aryan and Semitic roots. It will be our aim to avoid occasion for such an accusation, except as it may come from those who see in onomatopoeia the universal solvent of etymological diffi- culties, and would therefore give no credit to any comparison whatever made within our present sphere. (2) It is also evident that we ought to include only those forms which express common and elementary notions. This must be insisted upon rigorously ; and the principle is adopted not only for our guidance, but also as our defence against the opponents of all attempts at comparison in this obscure region. It is clear, in the first place, that if the two families were originally one they must have separated at a time when only the most rudimentary arts of life were practised, and the most primitive conceptions of the world without and within the mind were attained. Hence a combination of forms conveying conceptions peculiar to a more advanced state of thought must be regarded with suspicion. Coinci- dences between forms expressing such notions are, indeed, not common; but they have been used too freely by com- parers, and discredit has thus been cast upon such investiga* tions in general. i . u p m '■(■•i:i rili r:;i : " 122 UELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. It is manifest, in the second place, that if a large number of notions clearly elementary are found to be expressed in the two idioms by like sounds, in whose production onoma- topoeia has had no share, the evidence in favor of previous unity is very strong. We have not only the fact of a coin- cidence of such words as we should expect to find agreeing, but also the consideration that the occurrence of such coin- cidences ought, if we judge from the analogy of languages in general, to argue the existence at one time of many more similar phenomena which are now lost to view. For if we regard any great family of tongues, — the Aryan, for ex- ample, — it is surprising, as well as instructive for our present purpose, to note how many of the most elementary notions are expressed differently in the different dialpo,ts, and how many expressions once common to the wholo faj'i 1} have been dropped in one or several of them in the course of ages. We must not, and ought not, from the very nature of the question, to look for many agreements ; and if, after all, the number is found to be considerable, the evidence in favor of an original unity, which rises with cumulative force with every additional case, becomes well-nigh irresistible. These, then, are the conditions under which forms may be cited for comparison. If it is urged that it is not always easy to determine what notions are primary or elementary, and what are secondary, the answer is that we are not left to a priori judgments alone in the matter ; for the science of etymology has pushed its researches into various lan- guages so far and so successfully that we can appeal to the analogy of similar developments outside our proper sphere ; and this is the surest resource for those who seek to have light thrown upon the workings of the human mind as they are revealed in language. WOBDS IN COMMON IfELATINO TO PiRB. If the Aryans and Semites came from a common stock we should expect to find some trace of their early civilization in their common possession of one or more words for burning. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 123 Fire \7aa one of the earliest discoveries of mankind, and plays an important part in the legendary and mythical systems of most primitive communities. The fact is that we find no less than four words belonging to both systems, comprising most of the Proto-Aryan terms relating to that subject, and a large part of the Proto-Semitic. 1. Proto-Aryan kav (ku) ; Proto-Semitic is , to burn. The Proto-Aryan character of the root is proved by the fol- lowing forms : Gr. Kaiay for KaF-uo, to burn ; Skr. <;ona (for pri- mary A;awwa) flaming red, and as a noun, fire (see the Peters- burg Diet., and cf. Curtius,5.ed., p. 145 ; Fick i. p. 61). That it was developed from an earlier ku appears further from the occurrence of secondary roots, meaning to shine, most of which arc found only in Sanskrit; one, however, kvid (wlience Eng. white} being Proto-Aryan. — For the Proto-Semitic root we may compare Heb. n;ii, Assyr. kav-u,^ Arab. 'V Syr. ]ao, to burn. The root la here inherent was probably devel- oped from an earlier H5 like tlie Proto-Aryan, though this is not essential to the validity of the comparison. 2. Proto-Aryan kad (kand') ; Proto-Semitic *ip, to burn. This is one of the most wide-spread of Porto-Aryan roots. In Sanskrit it appears in some of its senses with a prothetic s (cf. tan and stow, to sound), in the sense of glowing, for ac- cording to the Petersburg Diet, the root cand^ to shine, is from ocand. But kand-u^ a fire-pan, shows no trace of it. Nor do any of the hometymous forms outside the Sanskrit, unless the Gr. ^av6-6''* - ; ; T r i;j •I: M ! 124 RELATIONS OF THE ARTAN AND SEMITIC LANGUAGES. i. p. 241 ; Curtius, p. 622) is due to an over-deference to the Sanskrit. The primary form is lead ; the principle of nasali- zation resulting in hand, and the use of a prothetic s in cases similar to the present, were discussed in our last Article. The Proto-Semitic ip is illustrated fcy the Heb. ''1?7 , Arab. ji" » Syr. ^r' I to burn, in which •« is a predeterminative. Also by the Heb. nnjs, to kindle fire, Arab, 'j^j , and Syr. Mi»|-e, of similar meaning, in which the n is a post-determinative. 3. Proto-Aryan kar (kaV)^ to heat, to cook ; Proto-Semitio Vp , to roast, to fry. Icar (kaV) is represented by Skr. qrd^ to boil, cook, from i^ar (=kar) as mnd from man ; Lat. cal-eo, cal-or, cre-marcj and several other Aryan forms — ^p appears with a post- determinative vowel in Heb. nbp , Arab. '^ and Ij; , and Ethiop. 'PA©, to fry» Chald. «ip, to roast, to burn, Assyr. kalrv,, to burn. This is perhaps the most striking combination of all the group ; for we see here that a term used by both families in the sense of burning was also specialized in both so as to apply to the preparation of food by fire. 4. Proto-Aryan ««, to burn ; Proto-Semitic v», fire (prob- ably = the burning thing). Skr. ush, to burn, scorch ; Gr. av-a> for ai^(7-a>, to kindle, eiJ-6) for €v, to singe ; Lat. ur-o for ms-o, to burn ; Old Norse us-liy fire ; A. S. ys-elf 0. H. Germ, us-el, ashes. — Cf. Heb. »x , Chald. kvm , Syr. jiLJ) , Eth. AH^ , Assyr. 'w-w, fire. There is also an Aryan by-form vas, to enlighten, which is commonly thought to be the earlier root. Whether the Semitic words have arisen from wi , through the dropping of the original v or w,^ or whether they themselves represent the earlier form, must remain undecided. This combination is highly probable, though not so certain as the other three. In accounting for the common possession of these similar 1 According to the usage which became nniversal in Assyrian. — Fick (ii. p. 27) combines the Teutonic word for ashes, aa-gan, with the Lat. ar-eo for as-eo and ard-eo for aad-eo, pointing to a root a«, to be hot. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 125 forms, it is apparent that onoiuuLupoeia must be excluded, as well as the theory of a chance coincidence. The only refuge left to doubters is the assumption tliat one language borrowed the sounds from the other. But why there should have been any borrowing at all of such primitive essential matters, or why it should have been done on so large a scale, is not easy to imagine. Words for Shining. 6. Proto-Aryan bhar (bhaV) ; Proto-Semitic •ina (ia), to shine. The Proto-Aryan form points, according to what was said on comparative phonology, to an earlier bar. It is represented in Skr. bhdl-a, star and brightness, bhdl-u, sun (also in bhalla^ etc. a bear, from its sleekness?) Gr. ^aK-7jp6opK-6,to shine, burn ; L&t. Jlag--ro, to burn ; A. S. blic-an, to shine (cf. Eng. bleach, and Germ, bleicfi). — In Semitic we have the Heb. aba (in Hiph.), to be bright, cheerful, Arab. ^J^ , to shine forth, e^ be clear. This Proto-Semitic root has no associations with any forms with medial b, and in consideration of the essen- tial character of the / sound, we may without presumption assign it to the root *a exemplified in the foregoing cases. Accepting number 6. as a highly probable combination, we have in Proto-Aryan bhd > bhar > bhark and bharg. The last three forms are the principal ones developed from bhdy and with them we find in Semitic exact correspondences in form and sense, which seem to preclude the possibility of merely accidental resemblance. ^ See this with other forms in Assyrian established by Lenormant, £tade sur qnelques parties des syllabaires cun^iformes, p. 231. Most of the Semitic words mean both to be bright and to lighten, and though the latter predom- inate., the former is the primary sense. The resemblance of p^3 to many words meaning to cleave, split, might suggest that the word for lightning arose from this notion, and that the sense of shining was secondary. But the natural order of the ideas, as well as the analogy of other languages, shows that the name for lightning was drawn from the idea of its brightness. So with onr word itself, with the German Blitx, the French idair, the Latin ^jjiur, and evea fidvrien. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 127 9. Proto-Aryan bhas; Proto-Seraitic p, ©a, to shine. The ProtoAryan character of bhas is pretty safely estab- lished by Fick, i. p. 153. Cf. Skr. bhds, to shine, bhds, bhds-^y splendor; Zend banh^ light (nh for a primary s ; see Schleicher, Compendium d. vergl. Gramm. 4. ed. p. 190), with Slavo-Teutonic bas-a, bare, manifest = Eng. bare. — In Semitic we have the form ya clearly presented in Arab. J^' , to shine, probably appearing also in Heb. pr. nomi V5PJ; cf. ya^,to be white, shining, \\^ and '(J, Heb. yia > niijia egg^ with hometymous noun-stems in Aramaic and Arabic. The root «:a seems to convey the same idea, for we find Ja^' along with J^' with a like meaning ; cf. A and " ^ , to be joyful. The last named root suggests the Proto-Semitic name for flesh, which we may represent by Heb. itoa. It was probably so called from its bright color. Perhaps bca, a Semitic word for cooking, came from the same source, as Lat. frig-o, Gr. tppvy-to, to roast, are con- nected with the root bharg (No. 8). 10. Proto-Aryan ark (rak) ; Proto-Semitic p"i , to shine. The root ark is proved from the Skr. arc , to shine forth, arc -is, splendor, and especially ark-as, the sun, as compared with Gr. rj-XiK-imp, the sun, or sun-god. See Curtius, 5 ed., p. 187. Pick, i. p. 22, cites a number of Keltic words point- ing to the root lak <^rak as the Gr. ^\€K-Tcop as well as ^\eic-rpov, amber, point to a root alk <^ark. With rak we may connect as a by-form the common Proto-Aryan root ruk (Juh), to shine, and with ark the root arg of the same meaning, whence Skr., Zend, Gr., Lat., and Oscan words for silver, ark : arg = rak : rag, to color, a wide-spread Proto- Aryan root. The root rdg, to shine forth, is a further devel- ^ Miihlan and Volck in their edition (the eighth) of Gesenius Handworterbnch (Leipzig, 1878), make the notion of whiteness, shining, to be secondary, and derived from the words for egg in the different dialects. But our citation of verb- stems shows this to be impossible. Cf. the derivation of albumen. f m ,i.j: w N ', ) '. i 1 '> ; '1 » ''1 T '4 ;l i|<:R 'I Bi Kl; vi \ 1 i k 1 128 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. opment, whence the Skr. rdg, to shine, and the Proto-Aryan word for king. — The existence of the corresponding Semitic root p*i is not so evident at first, but is easily established. It appears most usually represented with a predeternainative i as in pi% wliose sense of shining is attested by its derivatives \i\ all the dialects. The predominant meaning is to be yellow, whence a name for gold: Eth. ^C^ t Arab. J", coined money ; cf . Heb. PjI??"; , as applied to gold, Ps. Ixviii. 14, Assyr. rakrakku, yellow,^ also arhu and araku, yellow, green ; Heb. PT,^ green; p*:;, Syr. iijl, green herbs. Cf. also Heb. VP7?» paleness, yellowness, which like Arab. .ti!j> ^^^^ .? I , denotes a disease in men, and a blight in grain, produc- ing a yellow complexion. These several meanings can only be explained from the comprehensive sense of shining in- herent in the root.'' But we have the root in a simpler form, which puts this meaning beyond doubt. From some of the Assyrian and Arabic forms above cited, it appears that the i is not primary. Now we cite further, Arab. T, med. Waw, to be bright, clear (used of wine and the eyes) ; .'^f^, med. Ye, ^ ^e^ oV to shine brightly (used of the mirage) ; '..%' , to shimmer. Still further, the Arab, j^jj , to shine, and j(jj| y splendor, show that here as well as in Greek and Keltic the primary r was sometimes replaced by an / ; and a comparison of all the Semitic words shows clearly that the primary form was pi , which is thus assimilated perfectly to the Proto-Aryan ark or rak. 1 See Friedr. Delitzsch, Assyr. Studien. i. p. 105. 3 The most instructive analogy that we know of is the Proto-Aryan root ghar. Meaning primarily to shine or glow, a large number of its deriva- tives show the signification of being yellow or golden, and green. So the Skr. harita, green and yellow, hirxaia, gold, Gr. xpva' gold, for x/>vi-t6s, and Goth. gulUi, Eng. gold. Derivatives are even found in Zend and Slavonic (see Fick, i. p. 81), having the sense of green shoots of plants, as with pnx relations of the aryan and semitic languages. 129 Words for Cutting and Separating. 11. Proto-Aryan bhar ; Proto-Semitic "la, to cut, to pierce. The value of these roots in the present discussion is their agreement not simply in the general sense, but in two allied meanings. For bhar^ cf. Zend bar, to cut, to bore ; Gr. ^fhot, a plough, dp-arf^, a cleft, ravine, ^dp-vy^, opening, gullet ; L&t fur-are, Eng. bore. — la is illustrated by the Ileb. rri^ , to cut ; Arab. J[\ lH > to hew, hew out ; Assyr. fra,^ to cut into, grave ; also by Kia , to cut out, form, create, rep- resented in most of the dialects. It shows also in forms with consonantal postdeterminatives, as ^"^a, to pierce, the root of the Proto-Semitic itia , iron, nia , to pass through, seems to have had the same origin, if we may judge from the Assyr. buruhi, spear.* Arab. ^^''^^ to cut, appears in yy*J , cutting, yy *j , an axe ; cf . Eth* 'AC't' » bronze, from the same root, as ^na, iron, meaning to pass through, perforate. The idea of boring, however, is most distinctively conveyed by the -«^ form with indeterminative m , ixa (as in the Arab, ij , to pierce), whence the word for a well in Heb., Syr., Arab., and Assyrian. Again the Arab, 'ij, to explore, investigate = Heb. *ia (Eccl. ix. 1), points clearly to the same origin with a figurative application. With a stronger indeterminative, isa means to cut ofT, consume (with various associated senses in most of the dialects) ; and with a predeterminative, lan means to divide up, in Hebrew and Arabic. 12. Proto-Aryan bhad (bhicT) ; Proto-Semitic *ia , to divide, split open. Cf. Skr. bhid, to split ; Lat. find-o, fid-i ; Goth. beit-aUf A. S. bit-an = Eng. bite. The Lat. fod-io, to dig ; cf . Gr. 1 A very probable root ; see Friedr. Delitzsch, Assyrische Studien, i. p. 9. 3 See Schrader, Keilinschriften u. d. Alte Testament, p. 106. 4 r i u < !:: ■•1 ino KELATI0N8 OF THE ARYAN AND SEMITIC LANGUAGES. fiod-po^, a pit, seemH to point to an old by-form hhad. — TIio root ^a I1Q8 a conHidcrabIc development. In Hob. n^ meann to divide, and keep apart = Arab, j^ the same root having derivatives in Aramaic also ; with b as a post-determinative, bn^ means to divide, and with p the primary meaning of splitting comes out in pn^, to cleave. With » as an indeter- minativo, we have isa, to separate from, represented by noun or verb stems, in Heb., Arab., and Ethiopic. The physical notion of cutting asunder is better preserved in the kindred root ra, which has a wide representation throughout the Semitic system. 13. Proto-Aryan pat; Proto-Semitic n», bb, to separate, open. These roots apparently stand remotely connected with No. 12. We fi.id pat represented by the common consent of leading etymologists (see Fick, i. p. 13') ; Curtius, 5 ed., p. 211 ; Pott, W. Wb., iv. p. 154), in the Gr. irlT-vij/jLi, irer-uv- vvfii, to spread out, open out, and Trer-aXo?, spread out ; Lat. pat'CO, to open, and pat-ulus =7r€T-aXoi? ; A. S. fath-m, the out-spread arms = Eng. /fl^Aow. We should also add, with Fick, the Zend path-ana, wide. — The Semitic ne has the fun- damental notion of separating. So the Heb. no, with the cor- responding Arabic and Ethiopic, means to break off ; hence various noun-stems in these dialects, meaning a fragment or morsel, or, as we say, a bit (see No. 12). But the simplest modifications of the root have precisely the sense that pre- dominates in Proto-Aryan. Thus the Heb. p-* , as illustrated by the Arab. y^\j and its own derivative rb , means to spread out, while hPB,* in Heb., Aram., and Assyr., signifies to spread out and open. In Heb. and Syr., Arab, and Eth., nr& means also to open, while in Heb. rrn means to open ; and iPB , to interpret, has developed its meaning obviously from the same primary notion. Cf. ibb , to cleave, open, in Heb., Assyr., and Arab., from a kindred root, tsB. . 1 The name D^^, Japhet, of the ancestor of the Aryan race, from PPB, is an historical, if not a linguistic, connecting link between the two families. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 131 14. Proto-Aiyan park; Proto-Seuiitio pie und -jib, to cleave. The root park does not appear in any Aryan vcr1>-Htotn, but wo assume it to l>e represented in the Skr. para<;-^i (of. parr€Xe/c-w> TreXf/e-tfw, to hew off. Curtius, (5 ed., p. 104), refers these forms to a root trkaK, to heat, from which ifKarf in ifKrjaau) and Lat. pla>ig-o arise tlirongh softening. That this is wrong scema to us clear, hccause (1) the Sanskrit forms show clearly that the original root was woi prak but parfe, and (2) all the Greek and Sanskrit words contain only the idea of hewing or cleaving, and not of beat- ing (wood-cutting is the most common notion in both lan- guages). The root is park, and it can be explained only in the sense of cutting or cleaving. — In Semitic the root pne is much more widely extended. In Heb., and Aram., and Ethiopi", its general secondary sense is tliat of separating and loosening; but the primary physical notion of cleaving is apparent also in Heb. as well as in Arabic. The kindred •pB has the prevailing signification of breaking up, but in Assyrian it takes the place also of P'^b , meaning to separate, as well as to break in pieces. In all these dialects the root is represented largely in noun, as well as in verb stems. A very remarkable coincidence with the Proto-Aryan word is found in the Syr. }jaLs) Assyr. pilaJ^i, hatchet.^ The root pbfe , found besides in Arabic, and perhaps in Ethiopic, in the same sense, stands for the primary pie , as the root \t , liav- ing the same general meaning of cleaving, is from *« , both of these latter being widely represented throughout the Semitic family with various determinatives. It is not claimed here that the Syrian and Assyrian word for hatchet is the same as the Proto-Aryan above cited. But both are appar^ ently from the same root, and they show that this root in Aryo-Semitic expressed the special sense of cleaving or hew- ing wood. 16. Proto-Aryan kar ; Proto-Semitic la, "ip, to cut, divide. 1 See Friedr. Delitzsch, Assyr. Studien, i. p. 132 f. •t'i. ! : n ;;i « 132 RELATIONS OF THE ARYAN AND SEMFTIC LAN0UA0E8. The root kar is discuBBcd ftiUy by Pott, WurzelwJJrtorhiicli, ii. p. 140 fT. It iA aUo dealt with by Pick, i. p. 2.^8 f., atid Curtiiifi, p. 147 f. The form skar appeors in some of the dinlccts, but kar predominates, and is rightly token by Pott an the proper root. It is found *'ot only in Skr. kar (kr-^dmi and Arr-fiomt,), to wound, but also in kar^ kar-omi, to make, (cf. Eng. shdfif; and shave, Heb. w-a, to how out, and create). It nlHo ap[)can4 in Zend kar, to cut, and kar-eta, a knife, in Or. Kelpo) for xep-Uo, to shear, as well as in several noun- stems. The Latin has cer-no, to divide, as well as cur-tusj short (= cut off), and in the secondary sense, cre-o, caer- imonia. Tlie CJoth. hair-us, sword, and the A. S. hrudder, sieve, Eng. riddle, also belong here, the occurrence of which in the Teutonic family shows that the skar represented in Eng. shear, scar, and score, is a secondary root. — The exist- ence of the "IS in this sense is proved from the Heb. rria Arab. \j;', Eth. h^P , Chald. ^-J^, to pierce, to dig. The root nia had probably the same sense in Heb., ar 'ab. "ip again apf)ears with a like meaning in Heb. niip , ^o aig out ; Arab, 'jj , to cut out ; also with various determinatives in special modifications of the general notion of cutting. 16. Proto-Aryan hart ; Proto-Semitic onp, rina, to cut off. The root kar (No. 15), is developed into kart by the deter- minative t (cf. Pott, Wurzelworterbuch, iv. p. 115). It is found in Skr. kart, krint-ati, to cut, split ; Lith. kert-H, to hew, kirt-ikas, a hewer, and various other Slavonic words cited by Fick (i. p. 46). The Latin cutter, knife, is adjudged to be- long here by Pott (ii. p. 152) being for cult-ter ; cf. Skr. kart-tri, shears, and kart-ari, hunting-knife. — The occurrence of the root in Proto-Semitic seems clear. The Heb. r.'O , to cut off, has no direct representative in the other dialects ; but jSS'^ short, ^^Is'i a rock, t^j^* an axe, show that it once existed in Arabic ; and ^jf , to cut up, with the Amhario 4'^n\> of ^he same meaning, are matched by the Syr. ^j^. I >' V wt ^ BELATI0N8 OF THE AUYAN AND HEMITIC LANUt'AUlCS. 10.') All of thc«e cannot Imve l>een devi'loju'd intaltle in Prulirit, but proved to bo primitive from the derivatives), krjhdnay a sword, kalp-aka, a barber, krjhdni^ shears ; cf. Lith. kcr/Hi, kirjhti: to cut off, clip, with other Slavonic words cited by Pott. Probably Latin carp-o^ to pluck off, belongs hero ; cf. dis-cerpo. And, as Pott suggests, the Teutonic word haff (A. S. healf, 0. H. Gernmu halb), probably meant originally an equal division, and is liius naturally to be connected with this root. — On the Semitic side of the equation we find Arab. ^jj^, Eth. *tAc^, also Syr. ^.ai, Chald. qbp, to tear off, peck off ; cf . Arab, ^j j , and Eth. ^(^ , of the same meaning. We might be tempted to bring in here qbs, which is the root of the Heb. rinb''? , axes of a certain sort (Ps. Ixxiv. 6), a word to which there are similar terms in Syriac and Chaldee, but as these forms may be onomatopoctic they must be excluded. 18. Proto-Aryan kars ; Proto-Semitic y^p, tt)np, to cleave, tear asunder, drag off. The root kars has mostly the sense of dragging away, a meaning which it is not difficult to connect with that of separat- ing. So the Skr. karsh, karsh-ati, means to drag, but also to tear,i and karsh, krsh-ati, means to plough, that is, to tear or divide the land, to make, not to draw,^ furrows. Hence, 1 Cf. the German xerren, to drag, also to tear, the latter being the primary sense = Engl. tear. How this can indicate violent motion is shown by our col- loquialism " he tore along." " Ploughing, in this expression, is usually explained (see Petersburg and .i fl 134 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. li !| '' l^ 1 1:- if I ' ^ II -;^ii J-1 the derivative karsh-4 means a furrow, but also an incision in general. The sense of dragging is therefore secondary, though as the root evidently implied originally a violent separating, that meaning arose very early, and is exhibited in those EiH'opean forms which seem to represent the Skr. harsh. The root probably appears in the Gr. Kop-ea, to sweep out or away, if this is for Kopa-ito, as the Lat. verr-o for vers-o, and this for cvers-o would seem to imply. This combination which seems bold, has the high authority of Corssen in its favor. It certainly is the best tliat has yet been attempted. The root may be regarded almost certainly as Proto-Aryan, especially as all its meanings in Sanskrit appear also in Zend with corresponding forms. Perhaps a trace of the original sense of cutting off remains in Gr. Kopa-oto, to cut the hair, and Kopa-ij, the temples (as being shorn ; but cf. Pott, W.Wb. ii. p. 157). — Of the corresponding Semitic roots the radical idea is also that of violent separation. So in Heb. y^p , to cut off, also tear away (Job xxxiii. 6). Cf. Arab. "^S-, ^^ cut off, break off; ^li-» a morsel = Chald. V^p, Syr. yl^L', Eth. ^^f\ , to cut into, engrave ; also Arab, "j^^ , to cut off, gnaw off ; Eth. <|>^9 , to cut off, tear off, shear. In these roots the fundamental notion of the Proto-Aryan kars is fully represented. Its secondary sense of dragging comes out in the Arab, "a^^ , which, like the Heb. o"?!; > means first to cut off, but also, and more characteristically, to draw to one's self, to acquire. We also venture to add here the root 'pn , to cut, cleave, open, represented in Heb., Arab., Aram., and Assyr- ian ; and especially the root vm , which, having the general sense of cutting open, furnished also the Proto-Seniitic word for ploughing, Heb. «nn (cf. Arab, ^^^y Syr. A^k), Eth. ihj^fl . Cf. Assyr. hirs^, a ploughed furrow (Lenormant, Benfey's Dictionaries) as the drawing of furrows. But the notion of drawing does not naturally yield that of ploughing, which is expressed by words for cutting or separating in all the cases that we can recall in both Aryan and Semitic. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 135 to op. cit. pp. 155, 202). This brings the Semitic word com- pletely into accord with the Aryan Tears in ail its meanings. In this instance we do not hesitate to regard the roots as by- forms, the p being weakened into n, a change of frequent occurrence. That these letters are here of the same origin is as !''ood as proved by the following correspondences, run- ning uirough all the forms we have cited : B"in (rin) = oip (prp) ; C|"in = q-.p (qbp) ; win, fm = v^p, pp . The agree- ment in meaning between each of these pairs is complete. 19. Proto-Aryan sak ; Proto-Semitic "^lo , p\o , to cut. The root sak appears in Lat. sec-o^ to cut ; sec-uris, an axe ; in sec-tor and seg-mentuni as well as in sic-a, a dagger, and sec- ula, a sickle ; also in various Slavonic words cited by Fick (i. p. 790), and Pott (iii. p. 322). It is also the basis of many Teutonic words ; among them, that from which the Eng. see ^ (A. S. se-on, for seh-wan) is formed. With this the Teutonic word for a saw {saga) is allied, but not homet- ymous. The root is not found in Sanskrit or Zend, but, as Fick says, it is the basis of the Proto-Aryan ska ( > Skr. kshan, to wound, and Gr. KTei-v-fo^KTu-iMevai), and there is no doubt that it belonged to the primitive stock. — -;ia is repre- sented by Heb. -^'w, thorns, and nsto, a sharp weapon; cf. Arab. w< a^ Eth. l|JY\., a thorn, ^»)L& , armed with sharp weapons ; also ^ij[^, to be in doubt (i.e. divided in mind), and SLCm' weapons, po appears in Arab. , «^, to cleave, with many derivatives ; cf . Syr. wAQa > to cleave, > }"n* , a fis- sure. Both "^10 and p» are also found as secondary roots with various determinatives. 20. Proto-Aryan tak ; Proto-Seraitic ^n , to cut, divide. The root tak has the sense of forming, producing (as in Gr. tIk'To), e-reK-ov, to beget), along with other meanings easily con- nected with it (see Fick, i. p. 86 ; Pott, W. Wb., ii. 2. 401 ff. ; 1 For the development of meaning, cf. the Lat. cemo and Germ, mterscheiden, meaning first to separate ; Heb ntn , and Arab. "^^ , to see, primarily to cut. ■I 1* : * m il ■I ^ , i . r i ; ' \ ■' 186 RELATIONS OF THE ARYAN AND SEMITIC LANGUAQES. Curtius, p. 219 f.). What the primary meaning was, may perhaps be inferred from t!ie secondary taks, which in San- skrit means to hew out, to prepare, to make, and gives the noun taksh-an, a carpenter, a wood-cutter, taksh-ana, an axe. The Zend also has tash, to cut (from taks = 0. Pers. tak/ish, to build), and tash-a, an axe. From the same root comes Gr. r€icr- a shoestring, in Etluopic ; and especially by the word for large serpent or sea-monster : Heb. and Chald. ^in; Arab, ^j-yt**' which is derived from •jsn, just as the Lat. re^ulus is from rego, to stretch. With a predeterminative i the idea of extension denoted by the simple root is transferred to time; hence the Arab. %' to be perpetual, and the obsolete Heb. ir^, which is to be presupposed for the noun in*'*, perpetuity. With the pre- determinative 5 the idea of stretching becomes that of giving, or reaching forth.^ So we have the Heb. in?, to give, which appears also in Ghaldee and Samaritan, and of which the Syr. \^ is probably a corruption. The Assyr. pa is the same word with t softened to d, according to a common change. In the Eth. ^'t' J , however, the primary notion has apparently been transferred to the mental sphere, and the word means, in conj. iv. 1, to be busily engaged, assidu- ously occupied,* or, as we say, to have the mind on the strain, to be in-tent. The same root, in, with a vowel postdeter- minative, appears in Heb. njFi , as well as in several of the Aramaic idioms, with the proper sense of rewarding.** As corresponding with the Aryan to we may possibly have a relic of a Semitic ttn or kk in the Arab, reduplicated form 'C ft. 1^1^ > to incline downwards. 28. Proto-Aryan nat (nit) ; Proto-Semitic rs , o5 , to stretch forward, incline. ^ This transference of meaning is very common in language. It is manifest in the origin of the words offer and proffer, Lat. praebeo { = prae-habeo, to hold out), and even in the word give which is probably identical with the Lat. habeo, to hold. So also in the Skr. prayaccMmi, I offer, give, from the root yam, prop- erly to stretch. 3 See Dillman, Lex. Aethiop. col. 660, who, however, with apparent impro- priety, connects the meaning with the idea of giving, and compares the Lat. ex- pr^ssion : m deder^. ' Cf. the Lat. Cono from do, or, as a still oetter illustration, the Qerm. dar» •when, to reach forth, present. li Ji I 1 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 143 All Indo-European combination is given by Fiuk (i. 125). From the aclduciblc examples there would seem to have been not only the root nat, but a degenerated form nit. The Skr. ndtfi means to seek for help.^ Com[)aring this with the Goth. nathy nithan (Teutonic ndlha), to support, help, and the Lat. nit-or, to strive after, to seek or gain support, it is evident that the primary meaning of the root is, to reach after, or stretch forwards. — On the Semitic side the Arab. ^, and with a vowel determinative [Cs\ , to stretch out, lengthen, preserves the primary signification of the root ; but the cor- responding Heb. Mtss, while yielding the same sense, means more generally to stretch or lean forwards, to incline. Again, Eth. ^'t'V i with the post-determinative », means, primarily, to extend, stretch out, as the noun-stem 5'^0 > ^ ^snt, im- plies, which is formed from it as Lat. tentorium, L. Lat. tenta, tent, came from tendo. But ^'{'\} also meant to stretch forward or incline, for its current sense is to flee or to be put to flight.* The proof is complete when we refer to the identical root in Syriac, '^AJ • to Incline, used specially of a scale of the balances. 29. Proto-Aryan mad ; Proto-Semitic to , to extend, to measure. The root ma yields the common Indo-European words for measuring. In its undeveloped form it is found in Skr. md, t'^ measure ; Zend md, to measure, to produce ; Gr. fie-rpov, a measure ; Lat. me-tior, to measure ; Eccl. Slav, me-ra, a measure. The secondary root mad is also Proto-Aryan. It appears in Lat. mod-^s measure, and mod-eror, to keep in > Pott's attempt (Wurzelworterbuch, i. 576), to connect n&tli with n(, to lead, fails, because it begins at the wrong end of the train of ideas. The Slcr. ndtha, means, a " leader," only because it first meant a protector, i e. one who is sought for help or support. As a neuter noun, natha means help or support. ■ Just as the hatfugio is from the root bhug, to incline, bend, which also yields onr English bow. The Arabic ^ , just cited, m«ana also to flee ; cf. Heb. rXff , 1 Sam. xiv. 7. ■ "^V If: 'A- i: H. ti t> t' :i 1! .i '■-. . i 11 1 Jf 144 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. measure, inod-ivs and Gr. /liB-ifivo^, a bushel mcaHurc, and it takes tlic place of ma entirely in the Teutonic mat (Goth. mitan, Engl. mete'). In the figurative sense of considering (cf. Gcrra. ermessen) we have it in Gr. fiijB-ofiai, to think on, fieB-ofiai, to care for ; while it is fonnd also in the same sense in Keltic. The sense of measuring, then, is the prevailing notion attaching to these roots. That the primary idea was that of extension can, we think, be pretty clearly shown. In the first place the idea of measuring is not primitive ; it is essentially a secondary and complex notion, implying a fac- titious comparison with an accepted standard : it must be expressed by the new application of a previously existing term. What, then, is it to .neasure ? It is just to take the length, or rather the extent, of anything. Hence, when we come to examine in various languages the words for measur- ing whose etymology is accessible, we find that the radical notion is that of extending, in nearly every case.^ In the second place, we have apparent secondary forms of the root ma which imply the notion of extending. There are in Indo-European apparently three roots, mak, mag", and ma^h (see Gurtius, 5 ed. p. 328, No. 462), which had the sense referred to. These have given rise respectively to such rep- resentative words as the Gr. fiaK-p6^, long; Lat. mag^us^ great, and Skr. mah-ant^ great. These are most naturally to be connected with a root ma, having the general sense of extending." In the third place, there is more direct evidence from the usage of the root ma itself. In Zend, it means to make, produce, and a similar sense is given by it in Sanskrit, when it is compounded with the prefix nis. But it is more significant still that the Proto-Aryan word for mother, matar^ is from ma, and as it obviously means the producer, it shows how very early this meaning was attached to the root. Now, * The Arab, jj^ia an exception. Like the equivalent Heb. b^S it primarilj meant to hold or contain, and was thus applied to dry and liquid measure. This, of course, belongs to a later order of things. 3 It is noteworthy that mA is the stem of the Latin comparative md-jor, and that there is no flnal consonant in the stem of the Gaelic mdr and Welsh motor, great, which are undoubtedly hometymoua. ti ; ( ii • ! ^ r, and tiMtvr, RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 145 WO cnnnot very rcndily get the idea of producing from that of mcaHuring, Imt wo can very easily awHociatc it, an well as the notion of measuring, with the idea of extending (cf. tho hat. prthduro). — Tho root »« ia preserved in tho Aral). 11^ and ^t^, to extend, spread out, and though it does not appear in otlier idioms in verlvstems without connonantal determinatives, it is prohable that tho Semitic word for hun- dred (Hel). mm ) is derived from it. However this may ho, there is no doultt that this fundamental expression occurs in many other forms. Tho most notable is the root "ro, which appears as Proto-Semitic, not only in tho 8imi)lo form, but also with various determinatives, as nto, iia, *ik«,^ all having the notion of extending. The simple root to had also, from early Semitic times, tho sense of measuring, as ap{)ears from tho Heb. iTo , to lengthen, to measure, as compared with tho Arab, a^ , the name of a certain dry measure, from the root jjo'y ^^" ^^ I'^o meaning. In the same way, as wo have seen, the root mad yields the Lat. mod-4us and Gr. fiiS- ifMva, and thus tho analogy is completed with the root ia. 30. Proto-Aryan rak; Proto-Semitic, ^n, to extend. In the Indo-European sphere the two roots rak, raff lie side by side ; each of them means, properly, to stretch, ex- tend. Whether the form rag" has been weakened from rak, according to the analogy of a multitude of roots in Greek (Curtius, p. 633 if.), and occasional examples elsewhere, or whether they are equally autonomous, we do not need to attempt to determine. The root rak, in tho sense of extend- ing, seems to survive in the Zend ra to dispose, arrange. For the Indo-European root, see Fick i. 188 f., and cf. Pott III. 216 ff. (Nos. 1024, 1025). It is allowable to compare the Skr. rac^ to arrange, compose, set right ; Goth, rak^yan^ ^ It is a fancy of the Hindu grammarians that this is erroneously written for r&ai. But no root r&a or nu yields the proper sense. > Cf. Cornsen : Ausspracho u. s. w. d. lat. Spracho, i. 500 f. He assumes a root riky which he finds represented in many other words. Most of the combinations seem hazardous. The most plausible is that with O. High Germ, rih-an (cf. Eng. row), to place in line. 8 This root in Arabic also means to stand still. For the sense, may we not compare ^VS , to stand, with the root IS , already discussed ? * For a full discussion of the Assyrian words, see Lenormant : Etude snr quelques parties des syllabaires cun^iformea (Paris, 1876), p. 143 ff. * This must be carefully distinguished from Eng. rec^-on, A.S. rec'tum, which is from the root rag, to extend, direct. RELATIONS OF THE ARYAN AND SEMITIC LANOUAOES. 147 to reckon, dotormino ; Lith. renk-ii, to collect. — The Semitic root, like the preceding, is fonn'l with a liglit predeterniiiia- tive: Hob. Tp^, to sot in lino, arrange, adjuHt; Eth. {JJ^^p and VjiJ^ t ^^ adjust, reconcile.* Words for Bending or Curving. 82. Proto-Aryan kap^ kiip; Proto-Soiuitic 5)3, to bend, to curve. Kap is represented in the Gr. Kafiir-Ta), to bend, Kafiir-iiXow curved, and probably in Lat. cap-erare^ to wrinkle. The Skr. kamp^ which is undoul)tcdly the same root, to tremble, the expression being suggested by the curvature of trembling objects ; ccf/xi, a bow, from the primitive form kap, preserves the earlier notion. The same notion is apparent in kap-and; Gr. Kdfiir-f), a worm (cited by Fick, i. 89). — The Semitic tp has a very wide representation, and in its simplest form it appears in Hob. C)»^. Syr. ^asj Chald. ri?, to bend, to be curved ; Arab. ,"^< ; to turn away or aside ; while the Assyr. has it as a noun-stem in kap-u? a hollow place. The ap- parent derivative C)?, the palm, or hollow hand, is found throughout the system. With closely related meanings the root is also found with various determinatives in verb and noun stems that are surely Proto-Semitic. 83. Proto-Aryan kmar; Proto-Semitic nap, to bend around. The researches of Pictet^ and of Pott (W. Wb. i. 503) have made highly probable the existence of a primitive root kam, with the sense of bending (comp. also Fick, i. 40). More certain, however, is the occurrence of a root kmar, with three 1 There can be no doubt that the last two pairs of roots (Nos. SO, 31 ) were originally the same. The idea of arranginfr is a secondary one, and, according to a multitude of analogies, it is usually expressed by words that mean extend, etc., to put in line. 80 with our word ar-range, the Lat. or^o (of. or-ior), dit- poiio, rec-4us, our word right, reck-on, and a great number of hometymous words from the root rag. Indeed, the root *pK (No. 30), has also the sense of fitting, adjusting, in Hebrew, Talmudic, and Arabic. > For examples of this word, see Norris, Assyrian Dictionary, p. 592 f. ; cf. 516. * Les originet indo-europ^nnea (2d ed. 1377), ii. p. 277. iH m* ' If V lU - i:i it!. I 148 RELATIONS OF THE ARYAN AND SEMITIC LANQU." ^ES. consonants. The Sanskrit has a root kmar, kmarati^ to bo curved, and although the verl>stem does not emerge else- wliere, we find in Zend the noun kamar-a, a vr\ult, and a girdle ; of. Gr. Kafidp-a, Lat. camer-a, a vault, and Lat. cam- vr, bent inwards (used of horns). It is possible, as Fick suggests, that the same stem appears in 0. S. himil (Germ. himmeV) as the vault of heaven. — The Semitic root is devel- oped in precisely the same way. Cf. Eth. ^<^Z^ , to vault over, to make round, ^f*^*^, a vault, and an orb, with Chald. "lojp, to gird; "TittJ?, a girdle, Syr. X^ioa? 34. Proto-Aryan ank (oA;) ; Proto-Semitic p», to bend, curve. The Indo-European has mostly the nasalized ank in stems from this root ; but ok appears in some forms, and according to what was said on the subject of nasalization in Indo-Euro- pean roots in Chap, iv., the primary sound may be repre- 8 silted by ak. Cf. Skr. ac, ancy to bend, afc-a, tl.e curved bosom, and a kock , Zend ck-a, a clasp ; Gr. cfy/ic-09, a clasp, hook, ar/K-fov, the bent arm, ar/K-vXo^, bent, curved ; Lat. unc-us, bent, and a hook ; 0. Irish Sc-ad, a hook ; Engl, ang-le in its two senses. — The Semitic root is not found in its simplest representation ; but appears with a variety of determinatives, all of which reveal its primary force. Thus bps (in verb or noun stems in Heb., Chald., Syr., and Arab.), to bend or twist ; Eth. Q^^^; and Arab, l^t to bend, restrain, shut up; Syr. >Q.a^> Chald. dj??,, to twist, to turn ; sp'^, to bend, to arch, yps and cps, to twist, all of which also are Proto- Semitic. Forms with other determinati , es are found besides in the separate dialects. Moreover, the ancient roots pi» and nps give the idea of restraining, already adduced. 1 This root, though not quotable iu the literaiy language, is attested by the Dhi\t to go swiftly ; Syr. '^i^, to slip down. The agreement between the Aryan and Sem- itic roots in both general and special meanings should be well noted. 37. Proto-Aryan rag-h; Proto-Semitic a*i, to move quickly. For the various representations of the root rag-h Fick, i. p. 190, may be compared with Curtius, p. 192 (No. 168). "We shall cite a few cases in which it undoubtedly appears : Skr. rank (== rah = rag-h^ and rangh (= ragh}, to run, to hasten ; langh, to spring up or over ; rangh-as, ranh-as, rah-as, swiftness, haste ; lagh-^ and ragh-u, quick, svnall ; Gr. i-\ax-v^^, to lead, ]Ai^*9, a torrent; thence also a Proto- Semitic word for foot (found in Heb., Syr., Chald., Arabic, and some minor dialects), Heb. ban.* 38. Proto-Aryan di (da) ; Proto-Semitic Kn (in), to move swiftly, to fly. The root di shows itself in Skr. di and di, to hasten, to fly ; Gr. hi-(o, to flee, to hasten, Bi-€fj,ai, to speed away, Sirvo's, a 1 For the connection 6f ideas, cf. Lat. ctipio, which is hometymous with Skr. kup, to move quickly, to be angry ; see Pott, W.Wb. v. 91. Our word to long for and the Germ. er4ang-en, are from the root under discussion. ^ So our word yoot, representing the Proto-Aryan term, is from the root pad, to go. i-r. !.i! RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 151 pad, whirling, Uvm, etc., to whirl ; 0. Irish di-an, swift. That there was another, perhaps earlier, form da, as Fiek (iv. 106) suggests, seems probable enough from the Gr. Bo-vio), to shake, to drive about. — The root K"t is seen in ITeb. i-x'n, to fly swiftly (see especially Deut. xxviii. 49 ; Ps. xviii. 11) ; cf. Arab. |jfj, to run swiftly, also to roll about. Hence, or from a cognate in , we have the Heb. n*^ ; Chald. stjij ; Syr. |2u9 , the name of a bird of prey, so called from its swift flight. 89. Proto-Aryan tal; Proto-Semitic in, to raise, to weigh. The root tal has a very wide distribution. For a very satisfactory discussion of the history and mode of its develop- ment, see Curtius, p. 220 f. (No. 23G) ; cf. Fick, i. 94 ; Pott, II. 304-314 (No. 442). In Greek the fundamental form has been retained, though it also appears as tel and tol. Thus we have, with other forms, rX-aco, for raX-ao), to bear, Tai\raat-yan\ Engl, set^, and corresponding terms in Slavonic and Celtic. The Gr. eS, for o-eS, is transitive ; cf. el-aa, for e-aeB-aa, I set, e^-ofiai, for (re^ofiai, I sit = Germ. ich setze mich. The causative form sad-aya is also Proto- Aryan, and a large number of primary noun-stems in all the dialects preserve the ancient root. The force of the caus- ative verbs throughout shows that the word meant first not to sit, but to be situated or placed. — The Semitic no appears mostly as causative or transitive with the predeterminate '* j so Heb. "10^, to place, to lay a foundation, to set in order =: Chald. "lO"; ; Arab, j^" , with a specialized meaning, to set a pillow ; Assyr. isid,^ a foundation ; cf . Heb. niD"^, , etc. That ^ So in Greek rJiX-apos, a basket, and rt\-an^v, a supporting strap, from the root tal. These a? well as the words for weighing, above cited, hp.ve their mean- ing from the sense of suspending. 2 Cf. the Germ, an-liangm, to cling, adhere. 8 It should be mentioned that in Hebrew, Chaldce, and Syriac the same root means to raise, and to be heavy ; the additional meaning in Assyrian well lUas- trates the Greek and Sanskrit usage. * See Norris, Assyr. Diet, (ii.), p. 495, for sufficient examples. m ni RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 153 the root ">d was primarily intransitive is clear from the Arab. j^j to be placed, to be in the way, to obstruct ; cf. Heb. i?; Chald. Kpo; Syr. ]^i a block ; while the Heb. "'"o, an as* sembly (cf. Lat. consessvs') has as its most probable etymon an obsolete verb I'l^ or lio, meaning to sit. With a post- determinative "» we find -lie (Heb., Chald., and Syriac), mean- ing to set in order, like the Heb. *io^ in one of its applica- tions. 41. Proto-Aryan as, ds ; Proto-Semitic ck, to sit, to re- main. Cf. No. 53. For discussion of the root ds see Pott, W. Wb. II. 2. 299-302 (No. 683) ; Curtius, p. 379 f. (No. 568). The following forms clearly represent it : Skr. ds, to sit, dwell, remain ; Zend dh, to sit, to remain ; Gr. ^imi„ for ^a-ixac, I sit. Very probable derivations are, Lat. d^us, for as-nus, the fundament, and Lith. as-fd, floor, ground. — Tlie Semitic OK does not seem to be retained as a verb-stem, except in denominatives, but its existence in the sense indicated is shown in many noun-stems. Cf. Arab. ^ "^^ and " | , a foun- U-' LK" dation, also anything that remains or abides ; '''t , the foundation of a house = Assyr. asas-u, uss-u, foundation ; ^ ef- Heb. la'^iax . Hence Arab, '^'j , Assyr. asas-u, to lay a foun- dation. The root also comes out in nox with similar mean- ings in Heb. and Arabic. From these instances it is clear that, as in the Proto-Aryan ds, the root t'K meant originally to be placed, to remain. 42. Proto-Aryan man; Proto-Semitic io, to stay, to be fixed. For a full exhibition of the words that spring from the root man see Pott, W. Wb. n. 2, 118 ff. (No. 607). The discussion of Curtius, p. 311 ff. (No. 429), is complicated by the identification of this root with man, to think. This com- bination, which is maintained by leading Indo-European ety- mologists, is of no significance for our present business, inasmuch as maw, to remain, is an independent Proto-Aryan .1 i ,1 '* .1 • "1 ' I 1 , ' ' ! i 1 .'i:B H| 1 o'H HH |';?| ^9 1. '^ - f *';■■ I I tj 1^ r 154 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. root.^ We cite Zend and Old Persian man, to remain ; Mod. Persian mdn, to remain, also abiding, eternal ; Gr. fiiv-to, 1. to stand fast, to endure; 2. transitivelv, await, expect; fi{-fiv-a), to remain, await ; Lat. man-eo, to remain, also to wait for. Such noun-stems as Gr. fiov^ and Lat. marirsio show well the inherent notion of the root. — Precisely the same primary sense appears in the various representations of the Semitic "JO. With the lightest predeterminate M the root yan , widely represented in verb and noun stems in all the dialects, means to be fixed, firm, enduring, and in caus- ative uses and forms, to make firm, establish. The figura- tive sense of enduring, abiding, comes out in all the dialects as clearly and fully as it appears in the root man. Thus the simplest abstract expression of the root is Heb. p»» , for pyo^ . Assyr. amat^u; Arab. j^| J Eth. A^^'t', truth, fidelity, religion, i.e. what is fixed and abiding. This figurative use is almost the exclusive one in some of the dialects ; but the primary physical notion is exhibited in all. With the pre- determinative 9 the Arab. vJ^ means to stand still, to re- main in a place. This last form, though Hot certainly Proto- Semitic, shows the presence and force of the ancient root, with its meaning as above given. Words for Shutting or Enclosing. 43. Proto-Aryan klu; Proto-Seraitic Kba, to shut, enclose. The Indo-European root is not found in the Indo-Eranian division, but it appears in every other branch of the family, and must have a Proto-Aryan origin. For its manifestations see Pott, W. Wb. i. 684 ff. (No. 227) ; Curtius, p. 149 f . (No. 1 The identity of these two roots is nothing more than a brilliant hypothesis. No apt analogy for the etymological association of the ideas is at hand. Some- thing more is needed than a mere plausible connection of the notions expressed. And the association is nothing more than plausible. The intermediary idea is given by Pott, for example, as that of expecting or waiting in meditation. But it will be found that in all the cases where the root shows the two meanings of expecting and remaining, the latter is primary, the former secondary. So with vianere, fiivw, filfivu. In any case man, to remain, and man, to think, should be treated as separate roots. RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. Ifto 69) ; Fick, i. 641. The most significant representations are found in Gr. /cXt;-/?, AcXet?, for k\€F-i<{, a key, K\eia>, for K\eF- idD, to shut, /cXot-09, a collar, KXel-dpop, a bolt or bar ; Lat. cldv-is, a key, cldv^s, a nail (as a fastener), clavrdo, to shut ; 0. Irish cld-i, nails; Lith. kliuv-h, to fasten on, attach. Whether the Old High Germ, stiu-zan, for skliu^z-an, to shut (whence Germ. scfUiessen, sc/iloss, etc. ; cf . Engl, sluice, slat^ slot), belongs here is doubtful ; but its affinity would not prove, as Curtius imagines, that the root was primarily sklu. See our remarks on the prothetic s in the discussion of the morphology of Aryan roots. — The Semitic Kbs is represented by Heb. k^s, to shut, enclose, k^d, a prison; Chald. k^s; Syr. )lo, to shut, ) °V*» t a bolt; Eth. "HAA* to shut out, prohibit ; Arab, ^j^, to guard, watch ; Assyr. Kba ,^ to hold back, to refuse. The root has also the secondary sense of shutting out, separating,^ as appears from the Heb. o':?;'?, different species, with hometymous words in Ethiopic and Arabic. A great number of Semitic forms point to a siin[)ler root, is, represented in all the dialects, with the general sense of including, holding, containing. It should also be observed that the Aryan root klu has not the physiognomy of an ultimate root. Words for Guarding against or Fearing. 44. Proto-Aryan var; Proto-Semitic ii, to guard against, to fear. The root var may be traced through its various manifes- tations in its treatment by Pott, W. Wb. ii. 1. 552-597 (No. 612) ; Fick, l 211 ; Curtius, p. 346 f. (No. 501), and p. 550 (No. 660). We shall cite only a few of the many cases in which the root appears, according to the judgment of these and other leading etymologists. These instances will be found to be the most truly representative : Skr. var, to cover, protect, ward off; vdr-a, var-Htha, defence ; Zend apa-var, to » E.g. ik-lu-u, Inscr. of Khorsabad (ed. Oppert), lines 28, 69, 113, and ik-lcHJi, lines 79, 122. ' Cf. ex-dudo, dia^luth, and Sia-KAf/o. .1, m ma RELATIONS OF THE AYRAN AND SEMITIC LANGUAGES. ward off, hold back, var-atha^ defence ; Or. 8p-ofiai, for fop* ofxat, to keep watch, ojJ-popov'pd, (onrpo-Fop-a^ a guard, &p-a, care, apprehension, top-d(o, to see ; Lat. ver-eufj to fear, ver-ccundus^ modest, i.e. diffident, apprehensive ; Goth, var-ian, to keep off, var-as^ careful ; 0. High Germ. wdr-a, care, regard (cf. Engl, nmr-y, ware, ward, a-itmre). — The Semitic root unites in the most signal manner the two meanings of guarding and fearing, indicated by the Aryan var. We first call attention to the Arab, j "', to repel, hin- der. Comparing this with the Eth. ^^/\, an apron, from the corresponding obsolete root (D^A , it is clear that the primary meaning was to keep off, to guard against. Now the same root in Hebrew is k'^7', meaning to fear, which completes the parallel. If further assurance is needed, we may cite the Arab. cC. » e«« , and c^. , which is the same root "nn with post-determinative 9, and means to be afraid of, to keep away from, c ». , pious. God-fearing (cf. Lat. re-ver- ens). Its equivalent, the Heb. stC* means to tremble, i.e. to quake with fear (Isa. xv. 4). No two related words in dif- ferent branches of the Indo-European family show more striking correspondences in meaning than do the root var and "II. 1 i i 1 ^ 1 ; i \ ■{ ^HK 1 ■' It ,1 ; i ■ Words for Binding together. 45. Proto-Aryan sar ; Proto-Semitic ie, *is, to bind to- gether. For the root sar see especially Curtius, p. 353 f. (No. 518), and the references to Kuhn's Zeitschrift there given. We cite the following forms : Skr. sar-at,^ a thread ; Gr. op-fio/i09, a fastening, eip-epof, bondage ; Lat. ser-o, to string, to tie, ser-a, a bolt (fastener), ser-ies, a 1 See the Petersburg Dictionary, s.v. The word is not cited there from current literature, but from a native lexicon. RELATIONS OF THE ARYAN AND SEmTIC LANGUAGES. 157 series, ser-tum, a garland ; 0. Norse s6r-vi, a collar ; Lith. ser-is, a thread. — The Semitic 10 has projierly the sense of holding together firmly. With prcdcterminative M it yields the Hel). ne»; Syr. j^j; Arab, '^^'j Eth. /^^^ and AUU-*^, to bind, with many hometymous noun-stems ; for the Assyrian we may compare ««V-w,^ a band. With another predetermina- tive, the Heb. 10^, to punish, chasten, obviously meant at first to bind." The root^x, with a like primary force, a|>- pears in Heb. T»; Arab. °^j Syr. i!^, Chald. "^x, all mean- ing to bind together. The same root, is, reveals the same meaning in many developed forms ; the examples just given will, however, suffice for our purpose. Words for Pressing and Crushing. . 46. Proto-Aryan mnk; Proto-Semitic yo, to press, to crush. Certain of the ideas expressed by this pair of roots agree with some conveyed by the group meaning to rub, to bruise (Nos. 21-26), though the fundamental notions are different. For the root male, cf. Skr. mac,^ with the bye-form mane, to crush ; Or. root fiay, for fioK, in fidaao) (= ^ay-tcD), to kneed, uay-€vm da, to bind ; Lat itringo in "Virgil, Aen. 9. 294; Germ, h&ndigen " Attested by Hindu lexicographers ; see the Petersburg Dictionary. iililjl -!■ :r^'? I ; > . ■ l' i j; 11 1^ i j ■ i ; ■ '" i t,l; 158 UELATIONS OF THE ARYAN AND SEMITIC LANQUAGES. to consume, and in conj. v. to oppress a debtor. With in- dcternii native s the Heh. 'n?^ means to press and to crush (cf. 1 Sam. xxvi. 7 and Ezek. xxv. 8 with Lev. zxii. 24) ; and Arab. ^)^ means to rub and, as the dcrivotions show, to crush small. The Ghald. ^?v has a meaning similor to tliot of the Hebrew ; and as the root "^ts, with tlie secondary -|y«, run a perfectly parallel course through Hebrew, Aramaic, and Arabic, they are plainly Proto-Semitio in the senso indicated. Words for Carving or Graving. 47. Proto-Aryan grap, glup ; Proto-Semitic tiba, to carve, to grave. For these Aryan roots cf. Curtius, p. 178, 180 (Nos. 134, 188), with Pick, i. 674. The root grap is seen in the Gr. ypd^, for ypdir-w, to cut into (as in Iliad 17, 599), to write ;^ A. S. ceorf-an; Sweu karf-va; Engl, carve. The root glup appears in Gr. 7\v<^-o), for 7\u7r-«, to grave, 7\iJ^. avo'it a graving tool, 7\v<^, carved work, 7\if7r-T?J?, a sculp- tor ; A. S. cleof-an^ to hew ; Engl, cleave. The / in the primary Teutonic forms shows that the final letter was origi- nally/?. The A. S. graf-an; Engl, grave, may possibly be from the root grap, with g exceptionally retained ; but this is by .no means certain. We cannot agree with Curtius in comparing the Lat. glulho, to peel off, with y\v-m a root allied to grap with prothetie i and just as sculpo is related to glup. RELATIONS OF TUE AUYAN AND SEMITIC I-ANdKAOKS. 1.0;) regard to tlio roots horo combined it Hhould l»c (dwc'ivcil that neither of them is secondary in its oriKiu ; the evideiico of their primary identity is strengthened from the considcrution that to all appearance they are ultimate roots. Words for Piercinq, Infixing. 48.? Proto-Aryan $mar ; Proto-Semitic ie», to pierce, infix. All leading etymologists hold to the originality of the s in the root smar. For the forms cf . Pott, W.Wb. v. 7 13 ff . (No. 650) ; Fick, i. 254 ; Curtius, p. 830 (No. 40(3). The follow- ing forms will show that the current Indo-European sense of the root is to hold in mind ; Skr. smar^ to remember, keep in mind ; Zend wmr, of like meaning ; Lat. vie-mor, mindful, etc. ; Or. fiep-ifiva, anxiety, fiep-fiep-o to cut, to pierce, 1 Fick assigns here the Lat. mord-to, to bite ; but see No. 34. m m l¥r M \m ml •If 160 RELATIONS OF THE ARYAN AND SEMITIC LANOUAORS. to commit to memory; ^J^t to cut, conj. V. to keep iti momory. Tlio root smar, thou, according to tlio host liglitH, meant first to picrco. — That i«o and naio moan to pierce, to infix, in apparent from the following examplew : Hel». "^xstf"?' Ghald. M7if9^; Arab.^UM^) a nail; Chald. "ivQ; Arab. ..m*" conj. jr., to fasten with nails. Now the Heb. "insd means a o»^ thorn > and Arab. ^^, thorns, especially " spina Egyptiaca" ; Heb. ■I'^pi^ and Assyr. semir-u also meaning a diamond. The Heb. '^«^, and Chald. "1^9 mean to keep in mind, to watch, i.e. obviously, to pierce, or fix in the mind. The analogy is thus completed with the root smar. Words for Wetting or Pouring out. 49. Proto-Aryan sak (sik) ; Proto-Semitic pio, to moisten, pour out. For the Indo-European forms see Pott, W. Wb. v. 831-884 (No. 1069) ; Curtius, p. 187 (No. 24 ft) ; Fick, i. 229. The following forms from sik are representative : Skr. sic, to moisten, sprinkle, pour out, sek-a, sec-ana,a, sprinkling, etc. ; Or. lK-fid to water. In Hebrew, Aram., and Assyrian the corresponding verbs mean to be moist, to drink in, and in the causal forms, to water, give to drink. The notion of drinking is, of course, secondary. It is not found at all in Ethiopic, and is subordinate in Arabic, as it does not appear in any of the sixteen derivative nouns. \\\ RELATIONS OF THE ARYAN AND SEMITIC LANOUAOES. 1 01 IVTi WoUUa DKNOTINO CoLD. 60. ProtoAryan kar ; rroto-Somitic ip, to l)€ cold, to freeze. The root kar is OHtahliHhod liy Fick, i. ^>1. Cf. Skr. qt-c.wV-a, cold (as noun and adjective); Zend in conj. iii. 2, means (1) to de- vise means, in general, and (2) to devise cunningly, fraudu- lently. The first meaning is, of course, the j)rimary one. The corresponding Arab, 'l^, mid. Ye, retains the second- ary sense of the Ethiopic, and means, to use deceit, to lie (cf. the use of Lat. mentior^ ; but with mid. Waw it corres- ponds to the primary sense of the Ethiopic and to the sense of '(^, above cited, meaning to care for, provide for. But tuo same root exists in Heb. rtMrn, likeness, image, form, and l^o, a species, and is then evidently used to express the idea of a mental conception or image transferred to sensible objects^ (cf. the various uses of the Gr. ISed). The notion of thinking is thus shown to be Proto-Semitic. If the pri- 'iiary notion of the root is sought for, it seems more than probable that it is to be found in those common Semitic words from the root •)» which convey the fundamental idea of measuring. For example, the Heb. nsa ; Arab. _2^ means to measure out, allot (cf. Germ, ermesseri), and tlie same root in all the dialects means to number, while the Arab. [J^ means a deunite measure or weight. Derivations and kin- dred roots illustrate the same general signification. The Aryan and Semitic roots are thus shown to be completely in accOid. Words for Knowing. 52. Proto- Aryan vid; Proto-Semitic ni, to know. The root vid is one of the most familiar of the whole Indo- European stock. The citation of the following forms will suffice : Skr. vid, perf. ved-a, 1 know, vid, to find ; Gr. 18-elv, for Fi8-€iP, to see, oi-Ba, for Folha, I know = Skr. veda, 1 Hence, in Job iv. 16, flMrri is expressively employed for a form appearing in visions of the niyht. Gescnius' association of these words with the Arabic sense of deceiving, is as though one should derive specie* from specious, or Jingo from feign. m'' RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 160 m tS-€o, a conception, etc. ; Lat. vid-ere, to see, etc. ; Goth, vait, I know = Ski*, ved-a ; cf . Engl, ivit^ wot, wil-ness ; Eccl. Slav, vid-eti, to see, vhl-eti, to know ; Old Prussian imid-imai, we know. The idea of knowing predominates in the system as a whole, but in some of the dialects the notion of seeing prevails ; and it may be true, as Curtius says (p. 101, Engl. transl. of 4. ed., p. 124), that the fundamental expression was that of a seeing whicli apprehended and discovered. This fact, however, has no direct bearing upon the validity of our combination ; for the sense of knowing evidently goes back to early Proto-Aryan times, — The Semitic root is no less ancient, as it is ijund in all the great divisions of the family. It sometimes expresses the idea of observing, though the physical notion of seeing is not found. We cite the following verbal forms: ITeb. y^!?; Chalu. s't*); Syr. '?!,j^' Assyr. id-^, to know; Eth. P^U> conj. ii. 1, to make known, etc. That the first radical was originally i appears from the Heb. S'n^rn in tlie Hithpael, and the Assyrian forms ^ are rightly assigned to the Assyr. 'nb, or original 'ib class. by leading authorities. The Ethiopic P in the place of the first radical is probably an early dialectic variation. That the third radical, y, is merely a determinative is made plain from the fact that the fundamental notion is expressed also by the Proto-Semitic root mi. This in the causative forms, Heb. m-ih; Syr. J^o]; Chald. r)^«. cf. Arab, y, conj. x., means both to celebrate and to confess,^ i.e. to make known. 1 See Lcnormant, Etude sur quelques parties des syllabaires cun^iformes, p. 171 ; Schradcr, Keilinschriften u.d. alte Test. p. 223. 2 These ineanin},'8 can be bist cxj)lained on the hypothesis of a conne "tion between nil and S"11 . The common way of treating them is to make them causatives of the homophonous root n*l1 , to throw. But this docs not explain them at all suitably. Nor is the attempt more successful (Gcsenius's Hebrew Handworterbuch, 8th ed. by Muhlau and Vokk), to associate 511 with tiiu Arab. " "i' to place. The connection is not obvious ; and since the root in the sense of knowing is absent from the Arabic only of all the dialects, and m the 6cn,se of placing is found only in Arabic, the combination shows bad etymologizing. yPM,lllflHii^flBPI««WfW)|.ff,RV]ll,UDI!.' i.Uii'WUyv.W!flJ.JjJi^^,^|^J»!),J^^y f' 11 164 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. The root 11 is thus shown to be as old and independent as the root vid^ and it is worthy of attention th?.t the mean- ings coincide precisely. The apj)lication ci both roots is almost exclusively to mental, not to physical apprehension. They do not signify to be acquainted with, but to know within the strict sphere of self-consciousness. These two roots seem thus to claim a common origin through their individuality, ntiquity, and commanding influence in the fulfilment of a common destiny. Words for Being or Existing. 53. Froto-Aryan as ; Proto-Semitic lax , to be, exist. For the root as cf . Skr. as, to be = Gr. €«? in etr-r/ ; Lat. cs, es-t; Lith. es-mi, I am ; Goth, im, is, ist; Engl. is. It is generally agreed that it rests upon the root ds, to be fixed, to sit (No. 41). — The Semitic root is represented by the Heb. a;: and lax • ttJ? and m , there is = the Arab, "a \ Syr. tJ\] Assyv. is^. The "* in Heb. la."; is plainly secondary, ttJH representing the fundamental Semitic sound, which is revealed in all the other forms. With regard to its origin, it should be remarked that several independent observers have already suspected its affinity with the root tos, to be fixed, to remain (No. 41).^ Is not this remarkable double parallel with Proto-Aryan forms very strong evidence of the identity of the two pairs of roots here involved ? I ,■ 1 ■ • 11^ 1 il*", » I have thus taken up the predicative roots of the two sys- tems of speech which seem to justify an attempt to identify them. Something should be said now of those nominal forms which show a mutual resemblance. It should be re- marked that, as a general thing, such forms cannot furnish nearly such strong evidence of relationship as do tlie verbal roots. The reason is plain. The general conceptions con- 2 See what is said by Miihlan and Volck in their edition (the eighth) of Gesenius' Heb. Hondworterbuch. Even Gesenius, who wrongly assigned the Heb. d|^ directly to a root hlO^ , did not fail to perceive the connectioa will* Uriast, etc. (Thesaurus, p. 636). I4i !■ RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES 165 veyed by such predicative roots as we have been discussing are necessarily expressed by a corujiaratively limited number of words in any language. If in a large number of these the primary forms and notions correspond to a certainty, the proof of ancient unity is overwhelming. But derivatives are numerous, and are based upon secondary applications of the roots, and not only upon their radical meaning. The chances of coincidence are therefore greater in this region. It should be noticed, again, that the chances of one family borrowing from another the names for sensible objects are immeasur- ably greater than the chances of appropriating signs for fun- damental and generic conceptions, just as it is easier to appropriate a formula than a system of thought, or a maxim than an idea. Very much stress should, therefore, not be laid upon most of the examples of homophonous and synony- mous words that might easily be brought forward. We shall, however, discuss two or three that seem worthy of special consideration from the character of the notions they express. Words for Horn. 54. Proto-Aryan karna ; Proto-Semitic "jip, a horn. The Indo-European forms are Lat. cornru; Irish, Welsh, and Cornish corn; Teutonic horn-a (Goth, havrn; Engl., etc. hoTTi). The Greek may possiljly have had the same word ; see Curtius, p. 147 (No. 50). In Skr. it is probably repre- sented in qrn-g-a, horn. There is another Proto-Aryan word for horn, A;ar-?;a (Fick, i. 58), which seems connected with words for head, such as Skr. qir-as ; Gr. Kcip-a, etc. ; liut no satisfactory root has been found. — For Semitic forms cf. Heb. i:;?; Chald. »rp.> Syr. |jjIo; Arab. ^"J; Eth. tC'^J; Assyr. karviru. No plausil)le roots can be found for these forms. If karn-a and W. are not the same, the identity of the forms might be accounted for either on the assumption that the two were developed quite separately from distinct roots, or on the supposition that in very early times one family borrowed the term from the other. Ojnsidering the i ! ! i I ■[ f; r 166 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. ; ' 1 i-i I it IE 1 apparent priority of Proto-Aryan related words it would seem as if, oil the latter theory, the Semites must have borrowed from tlie Aryans.^ Neither of these hypotheses seems prob- able, but of the two the second is the less improbable. 55. Proto-Aryan ag-ra; Proto-Semitic lax, a field. For ag-ra of. Skr. ajra, a plain, open country ; Gr. wypo^ ; Lat. ag-er ; Teutonic akra A. S. acer; Engl, acre, cf. Germ. acker), tilled land. The Gr. adj. ayp-io^: agrees with the identical Skr. ajr-ya in its sense of belonging to the country, rustic, wild. It is a plausible, though not certain, conjecture of Kuhn (Zeitschrift III. 334), who is followed by Pictet (Origines indo-europ^ennes, 2. ed., ii. 108), t^t the word means properly pasture ground, from off, to drive (Lat. ag-o; Gr. dy-Q}, etc.), or the place to which flocks are driven.^ But, as Pictet remarks, the use of the Latin and German words shows that it was very early employed to denote cultivated land. — The Semitic term is found in Assyr. ag-ar, a field,^ in Eth. [J')Ct (1) cultivated, inhabited land, a region, (2) a village, (3) a town or city.* In the Himyaritic dialect of Arabic ^]^1^ means a district, a town. The Ethiopic form appears in Amharic as AOC? ^^^ this is probably a degen- eration.^ These forms are not susceptible of explanation from any Semitic source. The same alternatives are pre- 1 Prof. Sayce says, in arguing against Aryo-Se:nitic relationship (Assyrian Grammar for comparative purposes, p. 14): "Words like 'j'np compared with Ktp-ai are borkowed." This implies the lielief that such resemblances are not due to mere chance or " onomatopoeia." If they are not borrowed, therefore, they must point to a primary identity. A fortiori, then, the conceptual roots compared above, which cannot have been borrowed, point to an ancient on'jness of origin. But who would compare directly "pp with the simpler Kt(t-asl 2 Cf. Heb. "laip , wilderness, from *ai, to drive, and the hometymons Syriac and Ethiopic words (see Gesenius, Thesaurus, p. 318). ' For examples of this word, see Norris, Assyr. Diet. i. p. 15. * See Dillmann, Lexicon, col. 20. ' Ewald (Ausfiirhliches hebr. Lehrbuch, 8th ed. p. 402), who i followed br Dillmann (I.e.), combines these words with Heb. '^S)S<, a tiller, husbandman, and its hometyma in Syriac and Arabic, at the same time connecting all of them with Lat. ager, etc. But ISX is probably from "OM, to dig, found in conj. v. in Arabic. 11 RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. 107 Bented as in No. 54. In the present case the chaucos of the words being borrowed seem very slight, and the chances of fortuitous coincidence no sti'onger. Words for Wine. 56. ? Proto-Aryan vain; Proto-Semitic T'^, wine. Leading etymologists are at variance upon nil possible questions connected with this most common Indo-European word for wine. The ascertainable forms are Gr. olf-o? ; Lat. vinrvm, anciently vain-om; Goth, vem; Armenian g-iriri, ior gwinri (= Georgian ^wifb-o}, for win-i. Similar words in the Keltic seem to have been borrowed from the Latin. For a full discussion of the possible origin, as well as the history, of these words the reader is referred to Pictet.^ It is difficult to find a suitable etymology in the Indo-European fiiuiily, though several notable attempts have been made. — The Semitic forms are Heb. 1^? , for T.1 > wine ; Aral). ' ^ dark- colored grapes ; Eth. 0^7 > wine and a vineyard. No satis- factory etymon has been found for these words. It should be remarked that some eminent Indo-European etymologists, after Friedrich Miiller, hold to the Semitic origin of the non- Semitic forms. It is probable that both the primitive Sem- ites and primitive Aryans cultivated or were acquainted with the grape-vine. The evidence for the theory of the ancient identity of the terms involved is of the same general charac- ter as that adducible for Nos. 54 and 55, though borrowing on one pide or other is perhaps more probable in the present instance. Although many other cases more or less plausible could be cited, these are the only nouns which seem worthy of serious discussion in a treatise like the present. I think they are worthy of attention from impartial students ; +he agreement between the first two especially seems hard to account for on any other theory than that of oneness in origin. Another class of words should be mentioned, though not 1 Op. cit. ii. p. 311 ff. ; cf. Hintner in Fick's Vergl. Worterbuc^, ii. 795. ll w 168 RELATIONS OF THE ARYAN AND SEMITIC L/ ^GDAGES. if- i I. I: ■i > '4' S/* ■ -( . w discussed. These are pronominal and demontjuiitive roots which are surprisingly alike in the two systems. But for two reasons the treatment of them here would be unprofitable : (1) In most cases only a single consonant is found in each one of a pair of similar roots, and the identification is not 80 conclusive as when two or three consonants are the same. At all events, such combinations would meet with thot objec- tion. (2) Such roots are found to be (though in less measure) alike in most of the languages of the world ; and it is easy to put aside all these resemblances on the assump- tion that demonstrative roots, being interjectional in their character, are apt to be alike everywhere, since men, in a state of nature, are held to express similar feelings by similar sounds. The following table will exhibit in one view the comparable forms which have just been expounded. Some of the forms have a twofold representation which is not exhibited here in every case. Proto- Proto- 1 Aryan. Scmitio. 1. ku V\ 2. kad np 3. kar Vp V to bum. 4. ? tis ox J 5. hha ra > 6. hhar 13 7. bhark pn= > to shine. 8. hJiarg ana 9. hhas U93 10. ark p-n J 11. hhar 13 " 12. bhid "l^l. 13. pat rB 14. park pIB 15. kar^ 13 to cut or 16. kart M3 ^ separate. 17. karp tjip 18. kars taip 19. sak -|«5 20. tak in J 21. Proto. 4ryBu. mar Proto- Semitio. ia >, 22. 23. 24. mark marg mard pi« aia iia to rub, or ' bruise. 25. mars \aia ) 26. 27. 28. gam tan nat oa to unite, to stretch, 29. mad ia [ extend. 30. 31. 32. 33. rak rak hap kmar T i«p to arrange. > to bend. 34 35. 36. ok sad sar IS ^ IB ^ to go. 37. ragh ai to move 38. di «i quickly, to fly. I i RELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. ICO Proto. Aryan. 89. tal Proto- Semltio. to raise, 48. Proto- Aryan. ? smar Prot> Bcmltlo. ISO to pierce. weigh. 49. sik pio to moisten. 40. sad 41. as to sit. 60. 61. kar man to be cold, to think. 42. man 1« to be fixed. 62. vid 11 to know. 43. klu Nba to sliut. 63. as UM to be. 44. var 11 to keep off. 45. sar 10 to bind. 54. karna TP horn. 46. mak •p to press. 55. agra lax field. 47. grap glup ) tiba to carve. 66. ? vain ri wine. With regard to these forms, taken in connection with the ideas they express, it is necessary to make some closing remarks : (1) It should be observed that no form has been admitted against which the ol)jection might fairly be made that it is onomatopoetio in its origin. The list might have been largely increased if such terms had been included.^ On the other hand, it is impossible, in the case of most of the terms compared, to see how onomatopoeia could have had to do with their origin. The only ones in which this might be suspected are those which express the ideas of cutting or separating and rubbing or bruising. But these notions might be expressed in a hundred different ways ; and here the coincidences are so numerous and striking, in both primary and secondary forms, that we must, in reason, either maintain that the onomatopoeia acted in primitive Aryo-Semitic speech, or reject that theory altogether for those classes of roots. (2) The close phonetic correspondence between the forms compared should be well considered. If it is admitted, as I think it will be, that in these discussions there has been no straining after an imaginary identity of primary meaning 1 It is to be noted, however, that ideas which are usually held to be expressed most fMii|iiently by onomatopoeia are rarely convnyed by similar terms in the two systems of speech. For example, no two terms for breathing are alike, and only one pair of words for calling resemble one another. The onomatopoetic theory is a very easy one to employ, but it is apt to be overworked. 170 KELATIONS OF THE ARYAN AND SEMITIC LANGUAGES. P I i ' I u ;^ . : rt' ■ i \, in the roots and no false phanologizing in the harmonizing of the forms, the results are well worth serious attention from this standpoint. The main fact in the question is simply this : leaving out the cases in which an interrogative mark lias been used, we have over fifty pairs of roots whicli agree exactly in their primary notions and ultimate forms. The value of this fact, as bearing upon the issue involved, may be estimated from the attempt to conceive what the chances would be against such an agreement, if the two linguistic systems did not spring from a common source. Tlmt two peoples, not having a common origin or a common early history, should have separately framed a primitive speech from precisely the same elements would seem to be a phono- logical and psychological miracle after which such difficulties as are presented by the confusion of Babel would become problems only fit for the kindergarten. The chances would have been just as good for a merely partial agreement in any one of an infinite variety of ways. In bi-consonantal forms the first radical and the second in each pair might have been the same and the ether two have differed from one another by the whole range of phonetic expression. Or in the dis- similar letters the divergence might have been slight, involv- ing only cases of possible sound-shifting.* Of the tri-conso- nantal roots, of which a goodly number have been cited, a much more various and bewildering series of combinations than even these might have been presented, if the theory of a chance coincidence were valid. And the proved conditions of the question must shut us up to that theory of a purely fortuitous resemblance, unless we assume that the two sys- tems were originally one. (3) The ideas which are found to be expressed by the ^ In a few cases, but only in a very few, there are bye-forms in one f.4mily or the other, which differ from the forms above compared, by merely this slight measure. The Proto-Aryan root rag, to extend, along with the form rak, has been already alluded to (No. 30). In Proto-Semitic, the only ones are "iB and pa , to separate, along with *ia , 13 , and nB (Nos. 1 1 , 12, 13) ; B« , to extend, along with *IB (No. 29) ; i*i, to raise, along with in (No. 39), and perhaps 3S , to be round, along with C)3 , to bend (No. 32). RELATIONS OF THB', ARYAN AND SEMITIC LANGUAGES. 171 same fi»rms in the two Hystems arc just those which wc should naturally expect to have been enij)loyctl liy a prinutive people. The notions are simple and primary. The action of the forces of nature ; the most spontaneous works and ways of men and animals ; the efforts and movements required in the most essential acts and arts of life, are what we find repre- sented in this brief, but rich vocabulary. There are only three ideas expressed here which do not relate to the world of sense ; but these are the most essential of all metaphysical conceptions : to think, to know, to be. Only one term is absent which we might seem to have a right to expect : there is no word in our list relating to human speech. But even this accords with what our observation of language would lead us to look for. Words for speaking are notoriously different, for example, in the different branches of the Indo- European family. They arc mostly secondary and originally figurative.^ The same remark holds equally good of such terras within the Semitic family.'' From all that has been said it seems to be a just and nec- essary conclusion that the primitive Aryans and primitive Semites possessed in common a good working vocabulary. 1 Proto- Aryan words for speaking are but few, and most of them are but sparsely represented. Only one, the root vak has been at all persistent. Pictct has no treatment of this subject in his " Origines indo-eurcpe'ennes." ^ In fact, it is doubtful whether any Froto-Semitic word for speaking haa BurviTed. w m. m INDEXES. [Comptre the Table of Contenta and the LUt of Comparable Booti, pp. 168, IW. large flguroi refer to the upeolal ooroparlioni]. The I. PROTO-ARYAN. 11 1 n^i PAoa rAoi 'ok i'ank), to bend, curre. 148 ^1, to lie. 114 'agra, field, 166 ^M(^ai>), tobnm. 188 'agh ('angh), to press, 84 (cvid, to be white. 123 'ad, to eat, 115 'ark (rak), to shine. 127 ga, to beget, 16 'arg (rag), to shine, 83,87, 127 ga (gam), to gO, 87 'arbh, to obtain. 83 gan, to know, 85 'oB, to throw. 119 gan, to beget. 86 'as [is), to sit, remain, 168 gam, to unite, 140 '. to bend, 99 bK 'i7-«, God, 107 nbt* 'i/aA-«, God, 107 yom '-m-n, to remain, be firm, 100, 164 qOK '-»'-i'> to accumulate, 99 *10» '-^-r, to bind, 99, 167 ■i*iS< '-r-k, to stretch out, 101, 146 UJK '-«, to be placed, remain, 168 ^{jK '-*. to be, 164 ast 'is-u, fire, 124 ^itjJK '-«-r, to go right, 100 "l»a b-'-r, to dig, 102, 129 as bdb-u, agate, 114 na 6-(f, to cut ofi; 99, 105, 115, iso ina b-h-r, to shine, 126 «na b-{w)-', to go in, 114, 115 ■J^ia 6-(m')-?, to be white, 127 *l a 6-{i«)-r, to explore, 129 ia W, to be confused, 107 aba b-l-g, to be brigkt, 126 Wba W-^, to be abashed, 107 15a b-'-d, to be separated, 105, 130 *isa ^'-'■. to cut off, consume, 129 na *-»■, to cut, divide, 99, 102, 105, 108, 129 6<"ia f^r-', to hew out, create, 106, 129 t"ia b-r-z, to pierce, 108, 129 b^"ia (bflB) barzilM (parzil-u), iron, 108, 116,129 nia b-r-h, to pass through, 108, 129 •'"la b-r-if, to cut, 129 •pa b-r-k, to kneel, bless, p"ia b-r-k, to lighten, bttSa .W-/, to cook, ■ittJa &*•»", to be joyful, na ^. to separate, 115 126 127 127 180 ",aa g-b-n, to be curved (7), 109 *1J gr-d, to cut, 110 b"ia g-d4, to twist ♦"gether, make great, 109 sna 5'-rf-', to hew off, 110 *iia g-{w)-r, to turn aside, sojourn, 103 lia fif-Cy)-*!', to bind, 109 aba 9-i-f>, to drag off, 107 nba i7-W, to tear off, 107 iba to carve, grave, 168 BJ g-m, to bring together, 143 "loa g-^-r, to complete, 141 *ia 9^> to drag, scrape, roll, 101, 106 109, 110 a"ia g-r-^i to be scabby, 106 bia g-r-l, to drag off, 109 pa gum-u, threshing-floor, 109 tjia 5'-^-/'. sweep away. 110 K*i d-', to move swiftly, fly, 161 "im d-A-r, to re/olve, 103 inn d-(u>)-r, to tnrii round, 103 •^ d-y, to fly, 161 bl d-/, to hang loose, 100, 108 lb*l d-l-w, io suspend, 108, 115 •11 rf-r, to move round, 103, 109 am d-r-^, to go by steps, 107, 109 •pi d-r-*, to tread, 107, 109 lan h-b-r, to divide np, •^bh h4-k, to go away, fi"lh A-r-m, to be high, bai «>-W, to flow, 102 ni«i uj-d-y (in cauB. forms), to make known, 163 5)*I1 v-d-', to know, 163 ^1 wain-u, wine, 167 bsi w-i-/, to contain, 100, 115 101 vs'-d, to set, place, 162 Kl") v-r-', to guard against, fear, 166 *111 v-r-d, to go down, 99 jj^ll v-r-', to fear, 166 p11 w-r-ifc, to be green or yellow, l&iS pni w-r-k, to be behind, 146 •jni lo-t-n, to be perpetual, 142 INDEX II. 175 09, 129 13 k-n, to be fixed, 102 99 tp Ic-p, to bend, 99, 147 99 113 k-r-w, to pierce, dig, 132 ma ^-r-', to cut off, 132 lit z-r-w, to scatter, 110 ajit «-»■-'> to sow, 110 iin W-/, to let go. 100 un M' to cut, 106 2Bn A-.'-6, to hew wood, 106 on A-m, to be warm. 107 Ian A-m-s, to cut open, 134 ttJin *.-'■■*» to plough. 134 Dnn h-t-tn, to close, seal, 100 bllJ t-(w)-l, to be long, 101 B11 yum-u, day, 103 fii yam-u, sea, 103 l^ai y-wi-n, to be strong (?), 100 SS"^ y-s-', to place, 110 1pi ^-/t-rf, to burn, 124 ypi y-^*-?' to awake, 100 IW y-s-r, to be right, 100, 115 aa k-b, to bend, ^13 k-w-y, to bum, ^13 i-(w)-/, to contain, ■jlS k~(w)-n, to be fixed, *l^S i-(w)-r, to dig, is ^-/, to enclose, {t^B J(:-^-') to shut, 102 123 144 102, 115 132 100, 166 106, 115,166 •^Kb l-'-k, to send, 09 102, 146 145 102 102,146 104 1K53 m-'-d, to be largo, •'Dta m-'-y, to extend, ONtt m-'-sf, to fiow, 11B ff'-'/, to extend, IPia m-A-r, to sell, ■)1ia m-( «■»)-)», to tiiink, conceive, 162 IIB w-(i«)-r, to exchange, 104 ■iriia »n-A-r, to exchange {?), 104 btstt '«-'.-^ to extend, 101 -[13 m-i, to press down, 157 Xba m-Z-', to fill, 106 ijB m-ny, to measure out, 162 Da mi', to be liquid, 102 •^ya m-'-A;, to press, 158 5^a "J-r-^;, to rub hard, to press, 137 inr m-r-rf, to bruise, 138 nm m-r-ft, to rub, 137 1113 m-r-y, to rub against, 137 yna m-r-s> to press upon, 139 pISl m-r-^, to rub off, 187 ISi n-?-'** to move along, •is n-tS3 M"'> to set in, ■f J n-k, to strike, has n-A;-!/, to smite, IDJ ''-s'-^i to weave together, p3 n-k, to strike apart, np5 n4-y. to be separate, pure, 109,115 inj "■'""» to Rive, 101, 142 y^^ n-t-', to stretch forwards, 143 101 103 103 103 103 148 110 109 108 101 109 2f s'-ft, to turn round, 530 s'ab-'-u, seven, '10 »'-d, to set down, IhO ^-h-r, to be round, •llO s'-(w)-r, to turn aside, inO s'-i^-r, to traverse, "TO sf-k, to weave, cover. 103 110 163 104 104 104 101 176 IKDEX n. m\ II tp »'-p, to scrape, •lO s'-r, to bind, ^9 '-J, to Hrranp;e, number, 103 *11S '-(v>]-d, to repeat, 103 plj '-(w)-}/:, to restrain, 148 •»■*> '-(wj-r, to be bare, 109 T5 '-'3^, to be strong, 115 pi9 '-n-fc, to put a jund the neck, (denom.), 100 »SS '-s-w, to be firm, 109 u:t9 '-^-m, to be strong, great, 109 I apS '-If-b, to arch, 148 bpS '-H. to bind, twist, 148 Bp9 '■fp-'n, to twist, restrain, 148 yps '-it-?, to twist, 148 ■ipS '-fc-»'> to cut out, 101 ttJps '-^-Si to twist, 148 1-15 '-r-w, to be bare, 109 B"iS '^-m, to be bare, 109 ■^ny V-i, to arrange, 101, 147 *jr5 '-'-t'j to prepare, 106 r\1B />-(«>)-«, to spread out, 130 IBB p-t-r, to open. 111, 180 bo p-^, to cleave, 107 jbfi Wi7» to divide, 107 tjba /'-^-fi to break away, 108 pba p-^-^i to cleave, 131 nfi p-r, to cleave, 107, 108, 109, 1 1 1 1"iB p-r-d, to separate, 107 tna p-r-z, to branch out, 108 "t-iB p-r-k, to crush, 109, 131 0-iB p-r-s', to break up, 110 y-iB />-r-f , to break open. 111 p-iB p-r-k, to separate, 131 TSna p-rs, to disperse. 111, 116 *1\31B p-r-8- to break off, 130 nrs P-*-?. to open, 130 nx f-'', to go aside, 149 p*iU f-«f-^, to go right, 149 •T'S »-(«;)- tear off, "iBp ifc-m-r, to make round, •^p k-n, to be fixed, 102, 105, 108 KSp If-^-'i to be moved with passion, 106 •ijp k-n-y, to found, acquire, 108 np k-r, to cut, dig, 101, 182 ■ip k-r, to be cold, 161 ^3-ip fc-r-<, to cut up, 182, 186 wp icam-u, horn, 166 y-ip k-r-s, to cut off, 184 W"ip 't-r-s, to cut off. 184 *iK*l r-'-y, to see, J1 r-^, to move quickly, ta"i r-^-2, to tremble, bj"i r-g-l, to move about, run, •n"! r-flf, to move, push, baiB 8-W, to flow, 5310 «-*-', to be full, aniXJ 8-(w)-b, to return, "Ilia «-(M;)-r, to move quickly, r\'ccial attention is ;:ivcn to the occasions and circumstances in which the I'snlms were written, as liavinjr nn im- portant l)earin>j upon their pr()])hotie character, as well as aitiin;; in their elucida- tion." — TUc Ni'W York Ohscnur. " It comprises in itself more excellences than any other commentary on tlio Psalms in our lnnfrua'je, and wc know of no single commentary in the (Jcrman language which, all things considered, is preterahlc to it." — BaiUinl Q,uarlerly. STUART. Critical and Exegetical Commentaries, with translations of the Text, by Moses Stuart, late Professor of Sacred Literature in Andover Theological Seminary. 12mo. Romans, $1.75. Hebrews, $1.75. Ecclesiastos, $1.25. Proverbs, $1.60. The Commentnries on the Romans, Hebrews, and Ecclcsiastes arc edited and revised hy Prof. li. 1). C. Kohhins, " The Commcntiirics of Professor Stuart abide the test of time. Thoujrh some- what diflfusive in style, they contain so much thorou^;h discussion of doctrinal points, so much valuable criticism on pre<.''iiant words, and such an earnest reli- pious spirit, that they must live for <;enerations as a part of the apparatus for tho biblical student." — Tlw. Indiiwudcnl. " His Commentary on the Romans is the most elaborate of all his works. It has elicited more discussion than any of his other exe<;et;eal volumes. It is the result of long-continued, patient thonj:ht, It expresses in clear style his maturest con- clasions. Re-rarding it in all its relations, its antecedents and eonscquuT'ts, wo pronounce it the most important Commentary which has appeared in this country on this Kpistle The Commentary on Proverbs is the last work from the pen of Prof. Sttiart. Both diis Conmientary and the one preceding it, on Ecclcsiastes, exhibit a mellowness of s|)irit which savors of the good nmn ripening for heaven. In learning and critical acumen they are equal to his former works." — BiUiotheca Sacra. Miscellaneous Works of Professor Moses Stuart. Critical History and Defence of the Old Testament Canon. l2mo. Revised Edition. $1.50 " The work covers ground nnoceupied, in the same form and extent, by any other English or American work. And yet it is ground with which not only every biblical scholar, but every well-trained minister, ought to be familiar." — Methodist Qnarterli/. ' Miscellanies. Consisting of, 1. Letters to Dr. Channing on the Trinity; 2. Two Sermons on t!ie Atonement ; 3. Sacramental Sermon on the Lamb of God ; 4. Dedication Sermon — Christianity a distinct Religion ; 5. Letter to Dr. Channing on Religious Liberty ; 6. Supplemental Notes and Postcripts of additional matter. 12mo. pp. xii and 369. $1.25 A Grammar [Greek] of the New Testament Dialect, Second Edition, cor- rected, and mostly written anew. 8vo. Half cloth. $1.26 4-81 P Booh-K Pithlb/icd by W. F. Draper. ('nry. An Iittnulitvtion to the (itvvh of the New Testament. By Goorgo L. Oury, of the Moadvillo Theological Somiuaiy. 12mo. pp. 72. 75 conta " 'I'liis siiiiill voliinic liiis Iiccn ])rc|inrt'tl for ^icrsons, cither stiuleiits of tlicohiLjv, or otlurs who have not IimcI tins ndvuiitnirc of ii knowlcd^re of the (Jrcck. imd who iievcrtliclcs.s would lie ^diid to ri'iid tlif New 'I'cstaiiu'nt in its ori;;iiial t<)n;.'iu'. It (:ives till! rudiments of tlio lan;;ua;:i' so far as iil)S()liiti'iy ueecssary for tlu; uudor- Mtandiii.u- of New 'IVsianicnt (inriv. It is well adapted to its |iiir|)ose, and will provo of cnnit service to siieli us have not had the a(lvaiita;;e of ivehissieal edueution." — LnllicidH QiKiiicrli/. " Tlie siii)]iiicity of its method, its concispness and j)crspicuity ndniirahly luhipt it to the use of such per.-ous. With ii moderate decree of appiication, any ouu wiih an ordinary capacity lor acipiirinji' ian^^niiue may in a short time master ihosu elementary lessons, and liavin:: dotie so will h(^ aide to ' )>roceed at once to tlio readiriLr of llie easier jiortions of tlu^ New Testament.' " — T/i(uloi/inil mid llomiliiic iMiiiii/ili/. " 'I'liis is sulistantially n primary (Jreek (Iramir'ar of the New Testament, in- tended for tliOM; who have iiahica| stand-])oint. This is i'ollowed hy a hrief discussion of the Icadin;: jioints in the History of tno Canon, and then hy a resiling of the suhject of 'l>\tual ('riticism — Ili>tory of the Text, In this ])art of the work, and the fourteen Tahles which accompany it, care has heen taken to coinliine hrevity with the greatest jiossihle accuracy of statement. The hestricjnt authorities have heen consulted, and the author iias received vahiahle aid from eminent scholars in Kn^^land and America, especially from l)r I'/.ra Ahhot, of ("'anilirid;;e, who lias bestowed u|>on the whole of I'art III. and the nceompanyini; '{"allies much ])atient thought, sul)iiisim's (liSiiiiiis, besides incorporatiii}; into tla; liody of the work all the grainmuticul (orms con- tained in IlohiUson's Analytical Appendix. " I re^jard it ; s a very vnhuihle addition to our TL-hrew text-hooks. It is con- cise, neeurate, sulhi.iently full in dcfniitions, and admirably athipted (or the use of took or manual a(hiptud to the thought of our own day, marking; out tho line of the (.'liriHtian delcnccH which his jiarty are now holding and mean to stand by. Moreover, what is wanted, and, Indectl, IH nccesHary to secure attention, over and aliovo ability and learning — of which our author seems to have a fair share — is candor, and a disposition to rest witliin the lines of greatest stren^jth ; and in these respects our author appears to advantai;o. His book is throughout sensible and considerate, therefore inviting; and with promise of nscfulness. It is not often tlnit a parish clergyman is found M) well fitted as ho shows himself to be — oy a knowlcdi^o of what natural science is, and what its methods and rif^htful claims are — for dcalinfj on the one haiiii with the ' op))08ition9 .n dix un the I'riiiolpltiH of Textual Criticism, with n List of'ull tlio Icnuwn Cirvcic L'nuiuU, and a Table re))ri>M>ntinK griiphicully tlio I'nrta of tlio Text uf tno Kcw TuMtuniont contained In each, liy Fredurio Uardiuur D.D., rrofoHMor in the Uerkolcy Ulviiiitjr Sohuol 8vo. ta.OO. Tbk Pbimoifles or Tkxtual Criticism. Taper ooveni, 60 ocnta; cloth, flex., 76 cents. "A ytrj Important mutter in tiie praparntlon of the Harmony li, of eoune, tlie choice of a text. The one •hoeen by I'rofbiiur Oanllnrr U that of Titchvndorf'i elglilli edition of the New Tcitainpnt. 'I'hla li'xt wui ohoeen bccnim Mt emliodled tho latcit rotulle oferitlciim, having hud the udvantnifo throughout of llie Cudva BInailicue ond of a more uloio collullon of the Codex Vallcanui.' It ia an obvloui nicrll in thli Harmony, that the itudent can ice at a glance whether or not tlie text of Tiichendorf ogreeaor confllete with tliut ol Orleabach, lAChmano, and Trt-gillt'i in placee where there la a difference of opinion. It ■• annthur excel- lence oi the woric that llie Greek text !• to accurate, evincing the moit icrupuloui care and thorough echoiar- ■hip on the part of the editor."— llilMotheca mncra, " The note* of the author arc murlied by echolarthip and good eenee. The itudent will find It ■ conven- ient manual for the etudy of the Uoepria, becauee he eeee upon oni- and the lame page Ihu rcadiiiga of the principal edition! and manuicrlpta, together with the quotationa made by the cvangcliata from tiie Uld Tea- tamenl."— I'rinceton Hf.mew. " Dr. Oardlner'a woric Ima been well done, and he haa given ua a Harmony ot great value."— (Jwirltrli Smnew Kvaag. Lulh. CHureh. " By thia achoiarly work Ur. Oardiner haa rendered all dlillgent aludvnta of the Ooapel narrative an in. ▼aluable aervice. The book IXirniihca the beat resulta of the ahlcat and mnal laburioua Invcatigotion of ail known aourcea of knowledge regiirding the original aacred text." — Jt^bniied Church Mmit/ili), " Thia book, the reeuit of great reaearch and utmoat palnataking, la well worthy the couaideratioo of all Bible acholara."— Watchman and Rotator. GARDINER'S ENQLISH HARMONY. A Uarmony of the Four Goiipelfi in Kn^- lisn, Bccording to the Authorized Version ; corrected by the bent Critical b^ditions of the UrlKlnal. By Frederic Ciardiuer, O.D., Professor in the Berlteiey Divinity School. 8vo. Cloth, •a.oo. *< Th* Harmony In Engliah, the title of which if given above, la a reprodncHon of the Harmony in Greek) •o other changea being made than each aa were required to lit the work for the uf« of the Kngiiah reader wh« dealrea to ieam aome of tha Impiovementa which modem critlctou haa made la the authurixed Engliah test* — BiUwUuea Saera. ** We gladly commend thia Harmony to avary Intelligent reader of the Seriptnrea. Tha need of auch a gvila la Ifeit by every thoughtful Churchman at leaat once a year — In Holy Week — when he deairea to r«a4 Hm eventa of each day in tha order In which they happened ao many yeara ago. We do not think that out laymen know how much they will be helped to tha undentanding of the Moapela by a aimpie Uarmoay, f« hapa read aa wa auggeatcd above. In connection with aome aUndard Life of our Lord." — The C'AiircAinaii. IjU'B of CTIBIST The Life of our Lord in the Words of the Gospels. By Frederic Oardiner, D.D., Trotlsflsor In the Berkeley Divinity School. Itimo. pp. 250. $1.00. 'II le wall adapted to the convenience of putora, to the needa of teachera In the Bible-claaa and Sabbath- iehool, to the Teligioua Inatruction of famlllea. It bida fair to Inlroduce improvementa Into the atyie of Itach. lag the Bible to the young." - UMiolheea »Kra. ** Thia little volume will not only anawer aa a Harmony cf the Ooapela for the uae of thoee who only can to have rceulii, but 11 will be tn ezceUent book to read at flunlly prayera, or (o atud^ with a Ulble-claaa."- CAfiatMH Cuaan. 1-«T0. Books I'ultlixhid hy W. F. Braptr. [ WINER'S N. T. GRAMMAR. A OrammBr of tbn Idiom of tha New rMtkircnti |)ri-pari>(l an a Kulld lluith for tliii lntorpr)*t«tlutli Kn. Kcvlnod iinU author* uud IratiHlHtlun. 8vo. pp, 744, Cloth. IM.OU; rhiwp, $.10U) halfKuot. •O.Te. " I'rur. Tli*7*r (ililblt* th« inoit lehoUrly mil pilnt-Uklni ■csurmcy In 4II hli work, ti|ivri«l tllanlioa bflnn itlvan to r«l«ranc« ■ml Inilrie>, on wliluh (li« vitlut nf iiuvh ■ wnrk M much dvpvnili. 'Iht ImUm ■iuiia nil ilghly-ili page!. Tha publliher'a wnrk li hitiulioiiitly KnuUmhr, " Froraaaor Thayer haa Intniduccd numeruiii and linporlaiit ciirrvullona of Maaaon'a tranalalinn, and ha* mnda Iha praaant adltloli uf tha Uraniinur ducblvdiy auparlnr l« any ul tha prrctdlnn tranalatlnna. lie hat made It eapeolally oonvenlent (br the uiea of an KnglUh •tuilrnl by noting on IIil outer margin uf the piigri the paging of the alxth and leventh Orrinan edlllnni, and alio ul'I'mC. Miiaaun'a tianalatlim. I'huc the readi*r of a vominentary which mfvre to the pagea of cither of thiiae viilumra, may catily find tha reliirence by con- iultiiiglhe margin of IhliTolume."— /Ji/i/i«Mr«j .Siicra. "The whole appearance of the work aa II now itanila Indlcatea a earel^l and thorough acholarthip, A critical comparlaon of aaveral pagea with the ori,flnal conlirma (he ImpraMlon made by m general examination Of the book, in ita preaant form, Ihit tnuiilutlun may now be rerommended ua worthy of a place In the library of every mlnlater whodeiireato atudy ihe New 'X'eitamenI with the al4 of the beat critical helpa."— J'hruiug- ical Kekelic, " Ureat paina alao have l>een taken to aecure typographical accuracy, an extremely diflleull thing In a work Ofthla kind. We rejoice that io invaluable a work haa thui been mada ai nearly perfect aa wo can Impe i-ver to haTe it. It If a vork that can hardly Ikll to fauilitat* and Increaaa tlia rararaut and accurata atudy of tha Word of Uod."— ./tmert'can Fretbyttrian hevteie.. l\ i ' ' i ■ Is tl: BUTTMANN'S N. T. GRAMMAR. A (Jrammar of the Npw Testament Oropk. Uy Alkxanuku UuTi'MAMt. Autliurizfd Traiislutlon, by J. llKNUY 'i'iiAYKR. With numerous additioni) aud corructiuiiB by tho Author. 8vo. pp. 494. Cloth, $2.75. " Thia Urammar la acknowledged to be Ihe moat Important work which hue appeared on N. T. Orammar •Ince Wlner'a. Ita uae haa been hindered by tlie fact that In the original It haa the torm of an Appendix to tlie C'lauic Greek Urammar by Ihe Autlinr'a tkther. The inconvenience arlalng from Ihia peculiarity haa b«en obviated '.n thia tranalatlon by Inlrodni'lng in every caae enough IVom that Urammar to render the atate- menta eaally inteiligible to reudera unacquainted with that work ; at the aame time, the Author'a general iCiieme of conatantly comparing New TeaUinent and Ciaaaic naage haa been Ikcilitated Ibr every Htudcnt, by giving running retlBreneea throughout the book to live or aix of the moat current grammatical worka, among them the Urammara of liadley, Croaby, Unnaldion, and Jelf. Additlona and corrrcliona in more than two hundred and Htty placea have been lurniahed fbr thia edition by Ihe Author. " The N, T Index haa been enlarged eo aa tu Include all tlie paaaagea fVom tha N. T. ralerred to In tha Orammar 1 and a aeparate Index haa been added, compriaing all the paaaagea rlleil ftom the Septuagint. T):« other Indexea have been materiiiliy augmented) the cro'a-relerencea hare been multiplied 1 chapter aal ▼erae added to many of the fVaginentary quotatiuna from the N. T. 1 the pagination of the Oermau iiriginai hna been given In the margin 1 and at the end of the book a gloaaary of technical terma encountered more or leaa frequently in commentariea and grammatical worka haa been added for the canvenienca of atudenla." — Tnuulator't Pr^faet. " Profeaaor Thayer haa performed hia taak — which haa lieen i^great deal more than that of a mere tranalator — with remarkable lldelity. It la doubtleaa the beat work extant on thia aubject, and a book which every •rholarly paatur will detire to poaaeaa. Ita uaableoew la greatly enhanced by ita compute aet of Indexea,"— T>ie Aiiveuice. "It ia a thoroughly aclentlflc ^eatiie, and one which will ba helpftil to atudenta, both in connection with Wlner'a and aa diacueaing many pointa ttom a diOhreut or oppoaita point of view. Prof. Tluyer haa added much to the value of the book — aa one to be readily and conveniently uiad — by anlaifing and perfecUni the Indexea," eto. — Jlew Englander. BTIJART'S N. T. GRAMMAR, A Grammar of the New . v,toment Dialect. Bj M. KrUART, Profoasor of Hacred Literature io the Theological Seminary at Ando/ev 8to. Boards, 91.26. 2-8va