THI. 
 
 EASTERN ORIGIN 
 
 OK THE 
 
 CELTIC NATIONS 
 
 PROVED BY A COMPARISON OF 
 
 tljetr Dialects 
 
 WITH THE 
 
 SANSKRIT, GREEK, LATIN, AND TEUTONIC 
 
 FORMING A SUPPLEMENT TO RESEARCHES INTO THE 
 PHYSICAL HISTORY OF MANKIND. 
 
 BY 
 
 JAMES COWLES PRICHARD, M. D. F.R.S.&c. 
 
 OXFORD, 
 
 PRINTED BY S. COLLINGWOOD, PRINTER TO THE UNIVERSITY, 
 
 FOR J. AND A. ARCH, CORNHILL, LONDON. 
 
 MDCCCXXXI. 
 
i\Q 
 
 
 
 I %*>l-\ 
 
 v; 
 
TO 
 THE REVEREND 
 
 WILLIAM DANIEL CONYBEARE, A.M. F.R.S. &c. 
 
 RECTOR OF SULLY, 
 AND TO 
 
 PROFESSOR JACOB GRIMM 
 
 OF 
 THE UNIVERSITY OF GOETTINGEN, 
 
 THIS WORK IS INSCRIBED, 
 
 IN TESTIMONY OF 
 
 THE HIGH RESPECT AND REGARD 
 
 OF 
 
 THE AUTHOR. 
 
ADVERTISEMENT. 
 
 A HE treatise now laid before the public forms 
 a Supplement to my " Researches into the Physical 
 w History of Mankind," and was announced in the 
 first edition of that work, which was printed in 
 1813. Of the motives which induced me so long 
 to withhold it, and of those which have at length 
 determined me to the publication, a sufficient ac- 
 count will be found in the Introduction ; and I have 
 only a few words to premise on the circumstances 
 and designation under which the work now ap- 
 pears. 
 
 It is termed, a Supplement to Researches into the 
 Physical History of Mankind, because it was under- 
 taken with the view of furnishing proofs of a series 
 of facts, of which little more could be introduced into 
 that work than general statements, containing the 
 results of inquiries which had been sufficient for my 
 own conviction. It forms, however, a distinct trea- 
 tise, in exclusion of its reference to the history of 
 nations or races of men; and it may be proper to re- 
 mark, that some of the philological researches which 
 it contains have been pursued into greater extent 
 than the primary object of the work may seem to 
 have required. If this is in one respect a fault, it 
 may be hoped that contingent advantages in another 
 
vi ADVERTISEMENT. 
 
 point of view will be found to atone for it. The exa- 
 mination of cognate languages, while it points out 
 their resemblances and proves the affinity of the races 
 of men of which they formed the vernacular speech, 
 seldom fails at the same time to elucidate, in a greater 
 or less degree, the structure of the respective idioms 
 themselves; and it will appear, if I am not mis- 
 taken, that the relation of the Celtic dialects to the 
 other languages brought into comparison with them, 
 furnishes the means of throwing some light on the 
 European idioms in general. I have followed the 
 investigation which thus suggested itself, and have 
 stated the results. If the latter are well establish- 
 ed, they will be found both interesting by them- 
 selves to the philologist, and will at the same time 
 strongly confirm the principal inferences obtained 
 in respect to the origin and mutual affinity of the 
 European nations. 
 
 As I have had occasion in several parts of this 
 treatise to allude to the grammatical forms of some 
 languages, with which I am but imperfectly ac- 
 quainted, I have endeavoured to cite correctly the au- 
 thorities on which I have depended for information. 
 The names of various grammarians and other writers 
 on philological subjects, with the designations of 
 their works, will be found in the marginal references 
 scattered through the following pages, and need not 
 be mentioned in this place. But there are four 
 living authors to whom in a more especial manner 
 I am indebted, and am anxious to acknowledge 
 
ADVERTISEMENT. vii 
 
 my obligation. These are Mr. H. H. Wilson, the 
 learned secretary of the Asiatic Society, author 
 of the Sanskrit dictionary, and Professors Bopp, 
 Rosen, and Grimm, to whose well known works I 
 have made throughout this essay frequent refer- 
 ences. 
 
CONTENTS. 
 
CONTENTS 
 
 INTRODUCTION. 
 
 Sect. I. Different opinions respecting the population of 
 the world — Autochthones — Hypothesis of the ancients — 
 Modern opinion — Way of investigating the subject — 
 Physical evidence — Comparison of languages — How far 
 this inquiry has tended to elucidate the history of nations 
 — General relations and value of philological inquiries. 
 
 Sect. II. Nations of Europe at the earliest periods of 
 history — Eastern origin of several of them proved by 
 their languages — Indo-European languages — Is the Celtic 
 allied to them ? — Denied by several writers — Motives 
 for the discussion of this question. 
 
 Sect. III. Of the Celtic dialects extant — Modes of ortho- 
 graphy — Authorities. 
 
 CHAPTER I. 
 
 Preliminary survey of the forms of words and the permuta- 
 tions of letters. 
 
 Sect. I. Introductory Remarks. 
 
 Sect. II. Of the permutation of letters in composition 
 and construction — Of Sandhi and Samasa in Sanskrit — 
 Of the same principles as discovered in the Celtic dialects 
 — in the Welsh — in the Erse — Of the digamma and sibi- 
 lant in Greek. 
 
 Sect. III. Of the interchange of particular letters between 
 different languages — Table of numerals — Observations 
 deduced from it. 
 
 CHAPTER II. 
 
 Further proofs and extension of the observations laid down 
 
 in the preceding chapter. 
 Sect. I. Introductory Remarks. 
 
 b 
 
x CONTENTS. 
 
 Sect. II. Of the interchange of palatine or guttural conso- 
 nants with labials in the different languages. 
 
 Sect. III. Of the interchange of sibilant and soft palatine 
 consonants with gutturals or hard palatines. 
 
 Sect. IV. Of the relations of the aspirate — Of the substi- 
 tution of the aspirate in several languages for S and for 
 F — Of the aspirate as a guttural or hard palatine. 
 
 Sect. V. Of the interchange of dental and sibilant letters. 
 
 Sect. VI. Of the substitution of R for S. 
 
 Sect. VII. Of the relation of different vowels and diph- 
 thongs to each other in different languages — Synoptical 
 table of letters interchangeable between different lan- 
 guages. 
 
 CHAPTER III. 
 
 Proofs of common origin in the vocabulary of the Celtic and 
 other Indo-European languages. 
 
 Sect. I. Names of persons and relations. 
 
 Sect. II. Names of the principal elements of nature, and of 
 the visible objects of the universe. 
 
 Sect. III. Names of animals. 
 
 Sect. IV. Verbal roots traced in the Celtic and other Indo- 
 European languages. 
 
 Sect. V. Adjectives, Pronouns, and Particles. 
 
 CHAPTER IV. 
 
 Proofs of a common origin derived from the grammatical 
 structure of the Celtic and other Indo-European lan- 
 guages. 
 
 Sect. I. Review of the preceding facts and inferences. In- 
 troductory remarks on the personal inflections of verbs. 
 
 Sect. II. Personal endings of the Sanskrit verbs. 
 
 Sect. III. Terminations characteristic of the persons of 
 the Greek verb. 
 
 Sect. IV. Personal endings of the Latin verbs. 
 
 Sect. V. Terminations which distinguish the persons of 
 verbs in the Teutonic dialects. 
 
 Sect. VI. Personal endings of verbs in the Sclavonian dia- 
 lects and in the Persian language. 
 
CONTENTS. xi 
 
 Sect. VII. Terminations characterising the persons and 
 numbers of verbs in the Celtic languages. 
 
 CHAPTER V. 
 
 Of the personal pronouns in the Indo-European languages, 
 and of the derivation of the personal endings of verbs. 
 
 Sect. I. Personal pronoun of the first person in the San- 
 skrit, Greek, Latin, Russian, Mceso-Gothic, and Old 
 High German languages. 
 
 Sect. II. Pronoun of the second person. 
 
 Sect. III. Pronoun of the third person. 
 
 Sect. IV. General observations on the preceding facts. 
 
 Sect. V. Of the Celtic pronouns. 
 
 Par. 1. Of the entire personal pronouns in the Erse. 
 
 Par. 2. Of the entire pronouns in the Welsh. 
 
 Par. 3. Of the pronouns in a contracted state, or as used 
 
 in regimen. 
 Par. 4. Comparison of the personal endings of verbs 
 
 with the contracted forms of the pronouns. 
 Par. 5. General result of the foregoing analysis in re- 
 spect to the personal inflections of verbs in the Celtic 
 language. 
 
 Sect. VI. Conclusions respecting the personal terminations 
 of verbs in the other- Indo-European languages. 
 
 CHAPTER VI. 
 
 Of the inflections of Verbs through tenses and moods. 
 
 Sect. I. General view of the subject. 
 
 Sect. II. Modifications of verbs common to the Sanskrit 
 and the Greek languages. 
 
 Sect. III. Forms common to the Greek, Latin, and San- 
 skrit. 
 
 Sect. IV. Formation of the preterperfect tense in the 
 Greek, Latin, Sanskrit, and Teutonic verbs. 
 Par. 1. Professor Jacob Grimm's analysis of the Teu- 
 tonic verbs. 
 Par. 2. Analysis of the preterperfect in the Greek and 
 Sanskrit verbs. 
 
xii CONTENTS. 
 
 Par. 3. Analysis of the preterperfect in Latin verbs. 
 Sect. V. Of the remaining forms of the verb — Poten- 
 tial, Optative, and Conjunctive moods — Future tenses — 
 Middle and Passive voices. 
 
 Par. 1. Potential moods — Professor Bopp's opinion. 
 Par. 2. Future tenses — Formation of these tenses in 
 Sanskrit, in Greek, in Latin — General remarks on Fu- 
 ture and Past tenses. 
 
 CHAPTER VII. 
 
 Illustration of the principles developed in the preceding- 
 chapter — Conjugations of the verb substantive and of at- 
 tributive verbs, both in other Indo-European languages 
 and in the Celtic dialects. 
 
 Sect. I. General Remarks — Analysis of the verb substan- 
 tive in several languages. 
 Par. 1. Verb ^i \ ™ — asmi and its cognates. 
 Par. 2. Verb M^i 1 1 *H — bhavami and its cognates — 
 Conjugation of this verb through its various forms. 
 
 Sect. II. Analysis of the Celtic verb substantive. 
 
 Par. 1. Conjugation of bod or bydh, and comparison of 
 its forms with those of the verb substantive in the San- 
 skrit, Persian, and Sclavonic. 
 Par. 2. Of defective verbs used as verbs substantive in 
 the Celtic dialects. 
 
 Sect. III. Inflection of regular verbs in Welsh analysed. 
 
 Sect. IV. Conjugation of regular verbs in Erse. 
 
 Sect. V. Concluding observations on the Celtic verbs with 
 general remarks on the grammatical peculiarities of the 
 Celtic languages. 
 
 Sect. VI. General Inference, 
 
INTRODUCTION 
 
 SECTION I. 
 
 Different opinions respecting the population of the world — Au- 
 tochthones — Hypothesis of the ancients — Modern opinion — 
 Way of investigating the subject — Physical evidence — Com- 
 parison of languages — How far this inquiry has tended to elu- 
 cidate the history of nations — General relations and value 
 of philological inquiries. 
 
 IMaNY writers on natural history and geography 
 have maintained the opinion that each particular 
 region of the earth must have been supplied from 
 the beginning, by a separate and distinct creation, 
 with its peculiar stock of indigenous or native inha- 
 bitants. Among the ancients this notion prevailed 
 almost universally. There existed, indeed, in the 
 pagan world an obscure tradition of a primitive pair 
 fashioned out of clay by the hand of Prometheus or 
 of Jupiter; but this belonged to mythology; which, 
 in its literal sense, at least, was of little authority 
 with the best informed, and the frequent occurrence 
 of such terms as autochthones, indigence, or abori- 
 ginal inhabitants, whenever reference is made to the 
 population of different countries, indicates a general 
 prevalence of the ideas which such expressions are 
 fitted to suggest. The prevailing opinion in modern 
 times has referred all the nations of the earth to a 
 common parentage ; and this it has done chiefly, as 
 it would appear, on the authority of our Sacred His- 
 tory, the testimony of which seems hardly to be re- 
 
 B 
 
2 EASTERN ORIGIN OF 
 
 conciled with a different hypothesis. Of late, how- 
 ever, many learned men, chiefly on the continent, 
 have been strongly inclined to adopt an opinion si- 
 milar to that of the ancients ; and this seems now to 
 be gaining proselytes among the French naturalists 
 and physiologists, and among writers on history and 
 antiquities in Germany. Some of the former speak 
 of the Adamic race as of one among many distinct 
 tribes. Von Humboldt, who has collected so many 
 evidences of intercourse between the inhabitants of 
 the eastern and western continents, yet seems to 
 have regarded the primitive population of America as 
 a distinct and peculiar stock. The celebrated geo- 
 grapher Malte Brun has plainly taken it for granted 
 that each part of the earth had indigenous inhabit- 
 ants from the earliest times, into whose origin it is 
 vain to make inquiries ; and even the accomplished 
 Niebuhr, who is not more distinguished by the great 
 extent of his learning than by the novelty and in- 
 genuity of his critical speculations, has adopted a si- 
 milar opinion in connexion with his researches into 
 the early history of Italy a . 
 
 It would be no difficult matter to cite names of 
 equal celebrity on the other side of this question b , 
 but it is not by the authority of opinions that it can 
 ever be decided. The most learned men, and those 
 of the most profound research, are equally liable 
 with ordinary individuals to adopt erroneous notions 
 on subjects which lie beyond a particular sphere ; 
 they are perhaps even more disposed to prejudices 
 of certain kinds. It is only by examining the evi- 
 
 - a Romische Geschichte von N. G. Niebuhr. 1. Ausgab. Vor- 
 rede, p. 38. b Sir W. Jones. 
 
THE CELTIC NATIONS. 8 
 
 dence which may be drawn from a variety of dif- 
 ferent sources, that those persons who feel interested 
 in this inquiry can hope to arrive at a satisfactory 
 conclusion. 
 
 Perhaps those arguments which bear with the 
 greatest weight upon this question, and on which 
 the ultimate opinion of philosophers respecting it is 
 to be determined, are considerations resulting from 
 a survey of the natural history of the globe, and 
 facts connected with physical geography, and with 
 the multiplication and dispersion of species both of 
 animals and plants. On the evidence which is to be 
 deduced from these sources, I shall say nothing at 
 present. I have endeavoured to take a comprehen- 
 sive view of the whole of this subject in my Re- 
 searches into the Physical History of Mankind. 
 
 Among the investigations which belong exclu- 
 sively to the history of our own species, an analysis 
 of languages, affording the means of comparing their 
 component materials and ascertaining their affinities 
 and diversities, is one of the most important. 
 
 It must be a matter of regret to those who are 
 aware of the real value of this resource, that it has 
 been applied with so little judgment, and that many 
 writers who have devoted themselves to the study 
 of what is termed pJiilology have mixed up so much 
 that is extravagant and chimerical with the results 
 of their researches, as not only to throw a shade of 
 doubt and uncertainty over them, but even to bring 
 ridicule and contempt upon the pursuits in which 
 they have been engaged. A fondness for wild con- 
 jecture and for building up systems upon the most 
 inadequate and precarious foundations has been sup- 
 posed to belong to the whole class of writers on the 
 
 B 2 
 
4 EASTERN ORIGIN OF 
 
 history and affinities of languages, and it has cer- 
 tainly prevailed in no ordinary degree among them. 
 Even some of the latest works on these subjects, 
 though abounding with curious and valuable infor- 
 mation, are in a particular manner liable to this 
 censure. The treatise of Professor Murray on the 
 European languages, though it displays extensive 
 knowledge and diligent research, is scarcely men- 
 tioned without ridicule ; and in the Asia Polyglotta 
 of M. Julius Klaproth, which has added very con- 
 siderably to our acquaintance with the dialects and 
 genealogy of the Asiatic races, we find the results of 
 accurate investigation mixed up and blended with 
 too much that is uncertain and hypothetical. It 
 must, however, be allowed, that there are not a few 
 writers, in both earlier and later times, who are 
 scarcely, if in any degree, chargeable with the same 
 faults, and whose acuteness and soundness of dis- 
 cernment are equal to their extensive and profound 
 erudition. This may be truly said of Vossius and 
 Edward Lhuyd among the philologists of former 
 ages, and in more recent times of Professor Vater, 
 the Schlegels, Bopp, and Professor Jacob Grimm. 
 
 The comparison of languages is perhaps incapable 
 of affording all the results which some persons have 
 anticipated from it. It would be too much to expect 
 from this quarter to demonstrate the unity of race, 
 or an original sameness of idiom in the whole hu- 
 man species. But this resource, if properly applied, 
 will furnish great and indispensable assistance in 
 many particular inquiries relating to the history 
 and affinity of nations. 
 
 It would be easy to point out instances in which 
 the examination of languages has rendered substan- 
 
THE CELTIC NATIONS. 5 
 
 tial and undoubted services to the historian. The 
 history of the Goths, who conquered the Roman 
 empire, will furnish an example. The real origin of 
 this people could not have been known with cer- 
 tainty, if we had not come into possession of an 
 ample specimen of their language in the version of 
 Ulphilas. By this we learn that they were not 
 Getae or Thracians, as most of the writers who lived 
 near to the era of the Gothic invasion supposed 
 them to be, and as some modern historians have 
 maintained, but, in conformity with their own tra- 
 ditions, nearly allied in kindred to the northern 
 tribes of the German family. 
 
 The origin of the Polynesian races has been 
 illustrated by an investigation in one respect simi- 
 lar. Some of these tribes are found in islands so 
 distant from all other inhabited regions, as to furnish 
 an argument in favour of the opinion, that they had 
 the beginning of their existence in their present 
 abodes. But a comparison of their languages has 
 furnished proof that all the most remote insular na- 
 tions of the Great Ocean derived their origin from 
 the same quarter, and are nearly related to some 
 tribes of people inhabiting a part of the Indian con- 
 tinent and the isles of the Indian archipelago. 
 
 Even the history of the African and American 
 tribes has been in many particulars elucidated by 
 an inquiry into the relations of their languages, 
 though the results which have been obtained have 
 not proved to be precisely those which were hypo- 
 thetically anticipated, and with hope of arriving at 
 which these researches were in part undertaken. 
 
 Philologists have sought in vain in the old con- 
 tinent for a nation, from whose speech the diversified 
 
 B 3 
 
G EASTERN ORIGIN OF 
 
 idioms of America may with any degree of proba- 
 bility be derived ; but an examination of the Ame- 
 rican languages themselves has led to some interest- 
 ing results. The native races of North America are 
 referred by a classification of their dialects to a few 
 great divisions, several of which extend as radii 
 issuing from a common centre in the north-western 
 part of the continent, where it is divided from 
 Asia by Beh ring's Strait. The traditions prevalent 
 among the ancient Mexicans seem to have derived 
 credit from the discovery of a chain of nations ex- 
 tending almost from New Mexico to Mount St. 
 Elias, in the neighbourhood of the Esquimaux 
 Tschugazzi ; their languages, particularly those of 
 the Ugalyachmutzi and Koluschians, bearing a cu- 
 rious analogy to that of the Aztecs and Tlaxcallans. 
 Another series of nations, the Karalit, or Esqui- 
 maux, connected by affinities of dialect, has been 
 traced from the settlements of the Tschuktschi in 
 Asia, along the polar zone to Acadia and Green- 
 land. Light has also been thrown in a similar 
 manner on the history of the Lenni Lenape, and 
 the great kindred family of Algonquin nations, on 
 that of the Iroquois, and likewise of the Floridian 
 and other races of North America, by a comparison 
 of their national traditions with the indications dis- 
 covered in their dialects. One circumstance, which 
 is perhaps of more importance than all the pre- 
 ceding, is the singular congruity in structure be- 
 tween all the American languages, from the northern 
 to the southern extremity of the continent. To this 
 I only allude at present, having already in another 
 place surveyed the facts on which the observation is 
 founded, as they have been developed by the re- 
 
THE CELTIC NATIONS. 7 
 
 searches of Barton, Hervas, Von Humboldt, Hecke- 
 welder, and Duponceau. In Africa a remarkable 
 and interesting fact was the discovery of a nation 
 occupying nearly the whole northern region of that 
 continent, to which the Kabyles of Mauritania and 
 the Tuarik of the Great Desert belong, and whose 
 branches extend from the Oasis of Siwah on the 
 eastern, to the mountains of Atlas, and even to the 
 Canary islands, on the western side ; the Guanches, 
 the old inhabitants of those islands, whose remains 
 are said to lie embalmed in the mummy caves of 
 TenerifFe, spoke, as it appears, a dialect of the same 
 language as the Kabyles and Berbers. The Fela- 
 tahs, who have spread themselves over the interior 
 countries of Nigritia, have been traced by a similar 
 investigation to the mountainous districts above 
 the Senegal, where the Foulahs, who speak the 
 same language, have been long known to Europeans 
 as a people in many respects distinguishable from 
 the Negroes. To the southward of the equator a 
 connexion still more extended has been discovered 
 among the native tribes across the whole of the same 
 continent from Caffraria and the Mosambique coast, 
 on the Indian ocean, to the countries which border 
 on the Atlantic, and form a part of the region 
 termed the empire of Congo. 
 
 I have thus pointed out some of the most striking 
 instances, well known to those who have made phi- 
 lological subjects their pursuit, in which researches 
 of this kind have thrown some light on the origin 
 and affinities of nations, when all other historical re- 
 sources have failed. I shall presently consider the 
 application of this inquiry to the European nations, 
 as this is my principal object in the present work. 
 
 B 4 
 
8 EASTERN ORIGIN OF 
 
 It is requisite, however, before I proceed so far, to 
 make some general remarks on the evidence which 
 languages appear to furnish in proof of the affinity 
 of nations. 
 
 The use of languages really cognate must be 
 allowed to furnish a proof, or at least a strong pre- 
 sumption, of kindred race. Exceptions may indeed, 
 under very peculiar circumstances, occur to the in- 
 ference founded on this ground. For example, the 
 French language is likely to be the permanent idiom 
 of the negro people of St. Domingo, though the 
 latter are principally of African descent. Slaves im- 
 ported from various districts in Africa, having no 
 common idiom, have adopted that of their masters. 
 But conquest, or even captivity, under different cir- 
 cumstances, has scarcely ever exterminated the na- 
 tive idiom of any people, unless after many ages of 
 subjection, and even then vestiges have perhaps 
 always remained of its existence. In Britain the 
 native idiom was nowhere superseded by the Ro- 
 man, though the island was held in subjection up- 
 wards of three centuries. In Spain and in Gaul 
 several centuries of Latin domination, and fifteen 
 under German and other modern dynasties, have 
 proved insufficient entirely to obliterate the ancient 
 dialects, which were spoken by the native people be- 
 fore the Roman conquest . Even the Gypsies, who 
 have wandered in small companies over Europe for 
 some ages, still preserve their original language in 
 a form that can be everywhere recognised. 
 
 But the question is here naturally suggested, 
 
 c Without adverting to the Bas Breton, the Basque in Aqui- 
 taine and the Biscayan in Spain afford proofs of the fact above 
 asserted. 
 
 - 
 
THE CELTIC NATIONS. 9 
 
 what degrees and species of resemblance must be 
 considered as indicating any given languages to be 
 cognate, or as constituting their affinity? In advert- 
 ing to this inquiry I shall be allowed to repeat some 
 remarks which I have made on a former occasion. 
 
 A comparison of various languages displays four 
 different relations between them. 
 
 1. In comparing some languages we discover little 
 or no analogy in their grammatical structure, but 
 we trace, nevertheless, a resemblance more or less 
 extensive in their vocabularies, or in the terms for 
 particular objects, actions, and relations. If this 
 correspondence is the result of commercial inter- 
 course, or conquest, or the introduction of a new 
 system of religion, literature, and manners, it will 
 extend only to such words as belong to the new 
 stock of ideas thus introduced, and will leave un- 
 affected the great proportion of terms which are ex- 
 pressive of more simple ideas and universal objects. 
 Of the description now alluded to is the influence 
 which the Arabic has exerted upon the idioms of 
 the Persians and the Turks, and the Latin upon 
 some of the dialects of Europe. But if the corre- 
 spondence traced in the vocabularies of any two 
 languages is so extensive as to involve words of the 
 most simple and apparently primitive class, it ob- 
 viously indicates a much more ancient and intimate 
 connexion. There may be instances in which this 
 sort of affinity is so near as to render it probable, 
 that the dialects thus connected had a common 
 origin, and owe the diversities of their grammatical 
 forms to subsequent changes and difference of cul- 
 ture. 
 
 2. There are certain languages which have very 
 
10 EASTERN ORIGIN OF 
 
 few words in common, and which yet display, when 
 carefully examined, a remarkable analogy in their 
 laws of grammatical construction. 
 
 The most striking instances of this relation are 
 the poly synthetic idioms, as they are denominated 
 by Mr. Duponceau, of the American tribes, and the 
 monosyllabic languages of the Chinese and Indo- 
 Chinese nations. 
 
 3. A third relation is discovered between lan- 
 guages which are shewn to be connected by both of 
 the circumstances already pointed out. These are 
 the languages which I venture to term cognate. 
 The epithet is applied to all those dialects which 
 are connected by analogy in grammatical forms, and 
 by a considerable number of primitive words or 
 roots common to all, or in all resembling, and ma- 
 nifestly of the same origin. 
 
 4. A fourth relation exists between languages in 
 which neither of the connecting characters above 
 described can be discerned ; when there is neither 
 analogy of grammatical structure, nor any corre- 
 spondence in words sufficient to indicate a particular 
 affinity. Such languages are not of the same family, 
 and they generally belong to nations remote from 
 each other in descent, and often m physical cha- 
 racters. But even among languages thus dis#£vered, 
 a few common or resembling words may often be 
 found. These resemblances are sometimes casual, 
 or the result of mere accident : in other instances 
 they are perhaps too striking and too numerous to 
 be ascribed to chance or coincidence. Such are the 
 phenomena of connexion which M. Klaproth hypo- 
 thetically terms antediluvian, and those which Mr. 
 Sharon Turner has lately pointed out between the 
 
THE CELTIC NATIONS. 11 
 
 idioms of nations very remote from each other. 
 More strongly marked are the traces of approxima- 
 tion observed by Professors Barton and Vater be- 
 tween the vocabularies of tribes in North and even 
 in South America, and the dialects of the Samoi- 
 edes, Yukagers, and other races in North-eastern 
 Asia. Such facts are sometimes difficult of explana- 
 tion ; in other instances they may lead to interesting 
 results. Whatever may be thought of them, the va- 
 riety of languages, nearly or wholly unconnected, is 
 on a general survey so great, that it seems difficult 
 to avoid being led to one of two conclusions : either 
 that there existed from the beginning divers idioms, 
 or that the languages of mankind were rendered va- 
 rious by a miraculous change, according to the most 
 obvious import of a well known passage in the book 
 of Genesis. It would be foreign to my present de- 
 sign to consider these opinions more fully, and I shall 
 pass them by with a single remark on each. The 
 former, besides other objections, involves one which 
 has scarcely been adverted to. It implies that the 
 world contained from the beginning, not three or 
 four, as some writers are willing to believe, but some 
 hundreds and perhaps thousands of different human 
 races d . With respect to the latter, it seems incum- 
 bent on those who reject this passage of Sacred His- 
 tory on the ground of its making a reference to a 
 
 - d The languages of the African nations, according to Seetzen, 
 who has made the most extensive and original researches into 
 this subject, amount to ioo or 150. In America, there are said 
 to be 1500 idioms " notabilmente diversi." Such was the opin- 
 ion of Lopez, a missionary of great knowledge in the languages 
 both of South and North America. (See Seetzen's letters in 
 Von Zach's monathliche correspondenz. 1810. p. 3 28. and Her- 
 vas's Catalogo delle Lingue, p. 11.) 
 
12 EASTERN ORIGIN OF 
 
 supernatural, and, as it may be termed, an unknown 
 agency, to furnish us with some account of the first 
 existence of our species which, does not imply events, 
 at least equally miraculous. Unless the events which 
 certainly took place can be understood in a different 
 way from that in which the Sacred Scriptures repre- 
 sent them, we may rationally adhere to the whole 
 of the same testimony, as involving the operation of 
 no other causes, than such as are both proved and 
 are sufficient to account for the phenomena. 
 
 In the inquiry on which I have now to enter, I 
 must confine my view within a narrower sphere, and 
 advert to the relations of languages which, though 
 displaying great variety in their vocabulary, yet ap- 
 proximate in their most essential constituents and are 
 nearly connected in their grammatical formation. 
 Such phenomena can only be explained on the suppo- 
 sition that a different superstructure has been raised 
 by different nations on a basis originally common. 
 Tribes having a common idiom scanty in its stock of 
 words, appear separately to have added to their speech, 
 partly by new invention, and partly by borrowing 
 from their neighbours, such terms as the progress of 
 knowledge among them required. The accessory 
 parts of languages may have come at length to 
 bear a considerable proportion to the primitive one, 
 or even to exceed it, and the grammatical structure 
 may have been diversified under different modes of 
 cultivation. Hence arise in the first place varieties of 
 dialect ; but when the deviation is greater in degree, 
 it constitutes diversity of language. The German 
 and French are never termed dialects of one speech ; 
 and yet all who compare their respective sources, the 
 old Teutonic and the Latin languages, are aware that 
 
THE CELTIC NATIONS. 13 
 
 between these, a near and deeply rooted affinity sub- 
 sists. 
 
 Those who will duly weigh the facts which asso- 
 ciate themselves with this last consideration, will, I 
 believe, experience no difficulty in admitting all such 
 languages to be cognate, which have in common, 
 together with analogy in grammatical forms, a 
 large number of undoubtedly original and primi- 
 tive words. Such words are simple vocables, ex- 
 pressive of the most natural and universal objects 
 and ideas, terms for family relations and for the 
 most striking objects of visible nature, as likewise 
 verbal roots of the most frequent and general occur- 
 rence. These are elements of language which must 
 have belonged to every tribe of men in their origi- 
 nal dispersion over the world, and which must have 
 been the most tenaciously retained, and scarcely in- 
 terchanged between different nations. When such 
 elementary parts of speech are common to several 
 languages, and when their grammatical structure 
 displays likewise undoubted marks of a real and fun- 
 damental affinity, we may be allowed to regard these 
 languages as cognate, though the number of words 
 peculiar to each may be very considerable. 
 
 I have dwelt the more fully on this last consider- 
 ation, because on it will depend the validity of the 
 conclusions which I shall endeavour to draw in the 
 course of the following treatise. I shall now advert 
 particularly to the population of Europe and the 
 history of the races of which it consists. 
 
14 EASTERN ORIGIN OF 
 
 SECTION II. 
 
 Nations of Europe at the earliest periods of history — Eastern 
 origin of several proved by their languages — Indo-European 
 languages — Is the Celtic allied to them ? — Denied by several 
 writers — Motives for the discussion of this question. 
 
 At that era when the earliest dawning of history 
 begins to dispel the mists which had hovered over 
 the first ages of the world, we find the different 
 races of people in Europe nearly in the same rela- 
 tive situations which they now occupy, and we can 
 discern scarcely a trace, even in the oldest memo- 
 rials, of those wanderings of tribes which may be 
 supposed to have filled this region of the world with 
 inhabitants. In the remotest quarters of Europe, 
 towards the setting sun, we are told by Herodotus, 
 that the Celtae and Cynetae dwelt about the sources 
 of the Ister and the city — perhaps rather the moun- 
 tains — of Pyrene, and it is unknown during how 
 many ages they had occupied the region thus de- 
 scribed, before the father of history obtained these 
 earliest notices of them. It would seem, however, 
 that before the Trojan war even Britain must have 
 had inhabitants, since tin was at that time in 
 use, which was brought from Britain by Phoenician 
 traders a . We know likewise that the Teutonic na- 
 tions inhabited the northern countries of Europe at 
 a period not long subsequent to the age of Hero- 
 dotus. Pytheas, the navigator of Marseilles, who 
 was nearly contemporary with Aristotle, is well 
 known to have made a voyage of discovery towards 
 the north beyond the pillars of Hercules, by far the 
 
 a This at least would appear from the account given by He- 
 rodotus of the Phoenician commerce. 
 
THE CELTIC NATIONS. 15 
 
 most ancient that is recorded in that direction. In 
 the course of this voyage he visited Britain, and 
 even obtained some knowledge of Thule, or Iceland, 
 and of the coast of the Baltic sea. Pytheas men- 
 tions the Gnttones, who inhabited the shores of an 
 estuary which must have been the mouth of the 
 Vistula, and who carried on with their neighbours 
 the Teutones a traffick in amber, a native production 
 of their country b . The Teutones are well known 
 under that name ; the Guttones are probably the 
 Goths ; and thus we already discern in the north of 
 Europe two of the most celebrated nations belong- 
 ing to the Germanic family, in an age when even 
 the name of Rome had scarcely become known to 
 the Greeks. The Finns and the Sclavonians are 
 generally supposed to have been the latest among 
 the great nations who formed the population of Eu- 
 rope. But Finningia and the Fenni are mentioned 
 by Tacitus and Pliny, who place them beyond Ger- 
 many and towards the Vistula. In the age of these 
 writers the Finns were situated near the eastern 
 parts of the Baltic, and had probably extended 
 themselves already as far as those districts, where 
 their descendants were known under the name of 
 Beormahs or Biarmiers, in the times of Ohthere and 
 St. Olaf. The Sclavonians, indeed, are not early 
 
 b " Pytheas Guttonibus Germanise genti accoli aBStuarium 
 " oceani Mentonomon nomine spatio stadiorum : ab hoc diei 
 " navigatione insulam abesse Abalum : illo vere fluctibus ad- 
 " vehi, et esse concreti maris purgamentum : incolas pro ligno 
 " ad ignem uti eo proximisque Teutonis vend ere. Huic et Ti- 
 " maeus credidit, sed insulam Baltiam vocavit." Plin. Hist. Nat. 
 lib. xxxvii. cap. 2. The island of Abalus, or Baltia, may be 
 Abo. 
 
16 EASTERN ORIGIN OF 
 
 distinguished in Europe under that name, but by 
 the appellation of Wends, given to the Sclavonic 
 race by the Germans, we recognise them in the geo- 
 graphical descriptions of Pliny and Tacitus, who 
 mention the Venedi, and place them near the Finns, 
 and on the borders of Finningia. There the Ovev&at, 
 or Winidae, are stationed by Ptolemy and Jornan- 
 des, and the last of these writers appropriates ex- 
 pressly the name of Winidae to the Sclavonic nations. 
 It is besides highly probable that the Russians were 
 known to Herodotus, and that they are mentioned 
 by him under a term little varying from that which 
 is now applied to the same people by their Finnish 
 neighbours ; for the Finns distinguish the Muscovites 
 by the name of Rosso-lainen, or Russian people, and 
 call themselves and nations of their own kindred 
 Suoma-lainen. The word Rosso-lainen heard and 
 written by a Greek would be Rhoxolani. The 
 Rhoxolani, who are first described by Herodotus, 
 are said in the age of Strabo to have inhabited the 
 plains near the sources of the Tanais and the Bo- 
 rysthenes. 
 
 It appears, then, that the European races, in the 
 earliest periods in which we have any information 
 respecting them, held nearly the same relative situa- 
 tions as the tribes of people who are chiefly de- 
 scended from them still continue to occupy. Thus 
 far the facts which history developes afford no evi- 
 dence against the hypothesis, that different parts of 
 the world were originally filled with indigenous in- 
 habitants. It would be vain to attempt, merely from 
 traits of resemblance in some customs or supersti- 
 tions, or even from the doctrines of druidism and 
 the mythology of the sagas, to ascribe a common 
 
THE CELTIC NATIONS. 17 
 
 origin to the nations of Europe and those of the 
 East. By a similar mode of reasoning we might 
 perhaps as well deduce the Turks and the Tartars 
 from Arabia, and the Buddhists of northern Asia 
 from India or Ceylon. Nor can historical traditions 
 fill up the void. We can only hope by an analysis 
 of the European languages to obtain a proof, that 
 these races of people, having preserved common ele- 
 ments of speech, were connected in origin with the 
 nations of Asia. 
 
 The languages of the Finnish nations, the Lap- 
 landers, the Hungarians, the Ostiaks, and other 
 Siberian Tschudes, have been compared and care- 
 fully analysed by several German and other northern 
 writers, particularly by Gyarmathi, Adelung, Gat- 
 terer, and Julius Klaproth. The result that ap- 
 pears to have been sufficiently established is, as I 
 have elsewhere remarked, that all these nations 
 sprang from one original. The primitive seat of 
 this great race of men, or rather the earliest station 
 in which we can discover them by historical inqui- 
 ries, is the country which lies between the chain of 
 Caucasus and the southern extremities of the Ura- 
 lian mountains. 
 
 But our chief concern at present is with the Indo- 
 European tribes. That term was designed to in- 
 clude a class of nations, many of them inhabitants 
 of Europe, whose dialects are more or less nearly 
 related to the ancient language of India. This dis- 
 covery was originally made by comparing the San- 
 skrit with the Greek and Latin. A very consider- 
 able number of words were found to be common to 
 these languages, and a still more striking affinity 
 
 c 
 
18 EASTERN ORIGIN OF 
 
 was proved to exist between the grammatical forms 
 respectively belonging to them. It is difficult to de- 
 termine which idiom, the Latin or the Greek, ap- 
 proaches most nearly to the Sanskrit, but they 
 are all evidently branches of one stem. 
 
 It was easily proved, that the Teutonic as well as 
 the Sclavonian dialects, and the Lettish or Lithua- 
 nian which are in some respects intermediate be- 
 tween the former, stand nearly in the same rela- 
 tion to the ancient language of India. 
 
 Several intermediate languages, as the Zend and 
 other Persian dialects, the Armenian and the Ossete, 
 which is one of the various idioms spoken by the 
 nations of Caucasus, have been supposed by writers 
 who have examined their structure and etymology 
 to belong to the same stock . 
 
 Thus a near relation was proved to subsist be- 
 tween a considerable number of dialects spoken 
 by nations who are spread over a great part of Eu- 
 rope and Asia. It may be remarked, that the more 
 accurate the examination of these languages has 
 been, the more extensive and deeply rooted their 
 affinity has been discovered to be. Those who are ac- 
 quainted with Professor Jacob Grimm's able and lu- 
 cid Analysis of the Teutonic idioms, will fully admit 
 the truth of this remark. The historical inference 
 hence deduced is, that the European nations, who 
 speak dialects referrible to this class of languages, 
 are of the same race with the Indians and other 
 Asiatics to whom the same observation may be ap- 
 plied ; and this conclusion seems to have been ad- 
 mitted by writers who in general have displayed 
 
 c Klaproth, Asia Polyglotta. 
 
THE CELTIC NATIONS. 19 
 
 little indulgence towards the visionary speculations 
 of philologists' 1 . 
 
 The inquiry has frequently been made, whether 
 the Celtic dialects belong to the class of languages 
 thus allied, for which the term Indo-European is the 
 most suitable designation. The question is an in- 
 teresting one, because it has a particular bearing on 
 the origin of the nations of western Europe, includ- 
 ing the British isles, as well as a more extensive one 
 on the physical history of mankind. We have to 
 
 d The Edinburgh Reviewers, in a late critique, to which the ob- 
 servation in the text may particularly be applied, have remarked: 
 " We are free to confess that the result of our inquiries has 
 " been to produce a conviction in our minds that the affinities 
 " known to subsist between the Sanskrit, Greek, Latin, and 
 " German languages are perfectly irreconcilable with any 
 " other supposition than that of their having all been derived 
 " from a common source or primitive language spoken by a 
 " people of whom the Indians, Greeks, Latins, and Germans 
 " were equally the descendants." Ed. Rev. No. 102. p. 562. 
 Baron Cuvier has admitted the same inference as far as it re- 
 lates to the Indians and the Greeks, which is equivalent to its 
 general admission. He says, "The Pelasgi were originally from 
 " India, of which the Sanskrit roots that occur abundantly in 
 " their language do not permit us to doubt. It is probable that by 
 " crossing the mountains of Persia they penetrated as far as the 
 " Caucasus ; and that from this point, instead of continuing 
 " their route by land, they embarked on the Black Sea, and 
 " made a descent upon the coasts of Greece." In another pas- 
 sage of the same lecture, M. Cuvier observes, "that the San- 
 " skrit language is the most regular that is known, and that it 
 " is especially remarkable for the circumstance that it contains 
 " the roots of the various languages of Europe, of the Greek, 
 " Latin, German, and Sclavonic." (Baron Cuvier's Lectures on 
 the Natural Sciences.) He has omitted the. Celtic nations, the 
 earliest inhabitants of Western Europe, and perhaps regards 
 them as Aborigines. 
 
 c 2 
 
80 EASTERN ORIGIN OF 
 
 inquire whether the same arguments which prove 
 most of the other nations in this quarter of the 
 world to have sprung from an eastern origin, may 
 also be applied to that stock whose branches at the 
 earliest period of history were spread over Gaul and 
 Britain, and a part of Spain. Writers on the his- 
 tory of languages and the antiquity of nations have 
 been divided with respect to this question. Ade- 
 lung and Murray have regarded the Celtic as a 
 branch* of the Indo-European stock. But the latter 
 of these writers has passed over the subject in a 
 very cursory manner, or rather, he has left that 
 part of his work which relates to the Celtic dialects 
 in an incomplete state. And Adelung, who has 
 been followed in this particular by many foreign 
 writers, has committed the error of supposing the 
 Welsh tongue to be a descendant from the language 
 of the Belgae, and not from that of the Celtae, who 
 inhabited the central parts of Gaul, and, as it is gene- 
 rally supposed, of Britain. A want of access to in- 
 formation respecting the Celtic dialects has prevented 
 the learned men of Germany from forming correct 
 opinions on their relations to each other, and hence 
 it has arisen, that this department in the history of 
 languages — a subject which has been principally in- 
 vestigated by German writers — still remains but 
 imperfectly elucidated. Many of the continental 
 writers, among whom may be mentioned Frederick 
 Schlegel and Malte Brun, seem to have believed the 
 Celtic to be a language of a distinct class, entirely 
 unconnected with the other idioms of Europe ; and 
 in England the same opinion has been expressed by 
 several well-known authors. Mr. Pinkerton has de- 
 clared in the most positive terms that the Celtae 
 
THE CELTIC NATIONS. <>1 
 
 were a people entirely distinct from the rest of man- 
 kind. He says that their language, "the real Celtic, 
 Y is as remote from the Greek as the Hottentot from 
 •* the Lapponic." " The mythology of the Celtae," 
 adds Mr. Pinkerton, " resembled, in all probability, 
 •* that of the Hottentots, or others the rudest sa- 
 •• rages, as the Celtae anciently were, and are little 
 *• better at present, being incapable of any progress 
 *• in society." A late writer, in a work of extensive 
 research, at the conclusion of a chapter, in which he 
 has refuted some of the opinions of Pelloutier and 
 Bullet with respect to the Celtae and their language, 
 thus sums up the general result of his inquiries 6 . 
 " With regard," he says, " to the languages of Asia, 
 " I may adopt the words of Davis in the preface to 
 " his Dictionary, after substituting the word nullam 
 " for mamfestam. ' Ausim affirmare linguam Bri- 
 " tannicam (Celticam) turn vocibus, turn phrasibus 
 " et orationis contextu, turn literarum pronuncia- 
 " tione, nullam cum orientalibus habere congruen- 
 " tiam et affinitatem f .' The Celtic, therefore," con- 
 tinues the same writer, " when divested of all words 
 u which have been introduced into it by conquest 
 " and religion, is a perfectly original language : but 
 " this originality incontrovertibly proves that nei- 
 " ther Greek, Latin, or the Teutonic dialects, nor 
 " Arabic, Persian, or Sanskrit, were derived from 
 
 e Researches into the Origin and Affinity of the principal Lan- 
 guages of Asia and Europe, by Lieut. Col. Vans Kennedy, &c. 
 London, 1828. p. 85. 
 
 f " I dare to affirm that the British or Celtic language has no 
 " connection or affinity with the languages of the East, either in 
 " words, or phrases, or the construction of sentences, or the pro- 
 " nunciation of letters." 
 
 c 3 
 
 
22 EASTERN ORIGIN OF 
 
 " the Celtic, since these languages have not any af- 
 " finity whatever with that tongue." 
 
 In the first edition of my Researches into the Phy- 
 sical History of Mankind, which was published in 
 1813, fifteen years before the work from which the 
 preceding extract has been taken, I ventured to 
 make the following statement on this subject, the 
 result of what appeared to myself an adequate ex- 
 amination. 
 
 " We have remarked above that there is historical 
 " proof of the connexion of the Sclavonian, German, 
 " and Pelasgian races with the ancient Asiatic na- 
 " tions. Now the languages of these races and the 
 " Celtic, although differing much from each other, 
 " and constituting the four principal departments of 
 " dialects which prevail in Europe, are yet so far 
 " allied in their radical elements, that we may with 
 " certainty pronounce them to be branches of the 
 " same original stock. The resemblance is remark- 
 " able in the general structure of speech, and in 
 " those parts of the vocabulary which must be sup- 
 " posed to be the most ancient, as in words descrip- 
 " tive of common objects and feelings, for which ex- 
 " pressive terms existed in the primitive ages of so- 
 " ciety. We must therefore infer, that the nations 
 " to whom these languages belonged emigrated from 
 " the same quarter £." 
 
 g Researches &c. p. 534. The following note was appended 
 to this passage : 
 
 " The author of the review of Wilkins's Sanskrit Grammar, 
 " in the thirteenth volume of the Edinburgh Review, has given 
 " a comparative vocabulary of the Sanskrit, Persic, Latin, and 
 " German languages, which completely evinces the truth of the 
 " position here affirmed, as far as the above languages are con- 
 
THE CELTIC NATIONS. £3 
 
 The extent which my work necessarily assumed, 
 and the apparent incongruity of filling up any con- 
 siderate part of a physiological essay with glos- 
 saries or remarks on grammatical forms, combined 
 with other reasons in preventing me from entering 
 at full into the proof of these assertions, and the 
 same circumstances operated likewise at the publi- 
 cation of the second edition. I have, however, had 
 the subject occasionally in view during the interval, 
 and have collected from time to time materials for a 
 treatise upon it, which many circumstances have at 
 length determined me to lay before the public. 
 Among these may be mentioned the decided opinion 
 advanced in the work from which I have above 
 cited a passage, proving, unless I am mistaken, that 
 there is not as yet sufficient information before the 
 public on a subject of considerable moment in re- 
 spect to the history of the human race, and the re- 
 lation of its various branches to each other. An- 
 other motive to this determination has been the 
 advice of some learned friends with whom I have 
 conversed on the subject of the following treatise, 
 and particularly of the two highly distinguished 
 men, to whom it is dedicated. 
 
 The main object which I have had in view in the 
 composition of this work has been, to institute such 
 a comparison of the Celtic dialects with the lan- 
 guages allowed to belong to the Indo-European 
 stock, as may tend to illustrate the relation of the 
 Celtic people to the rest of mankind. In the course, 
 
 " cerned. But the proof would have been more striking, if he 
 " had added the Celtic dialects and the Greek. I have made 
 " an attempt to supply this deficiency, which I intend to make 
 " public." 
 
 c 4 
 
24 EASTERN ORIGIN OF 
 
 however, of this inquiry, I have incidentally disco- 
 vered that the relations between the languages 
 above mentioned and the Celtic, is such as not 
 merely to establish the affinity of the respective na- 
 tions, but likewise to throw light upon the struc- 
 ture of the Indo-European languages in general, 
 and particularly to illustrate some points of ob- 
 scurity, to which many writers on grammar and 
 etymology have adverted without fully elucidating 
 them. The following pages will contain such re- 
 marks as I have thought requisite in this point of 
 view. 
 
 SECTION III. 
 
 Of the Celtic dialects extant — Modes of orthography — Au- 
 thorities. 
 
 It may be doubted whether the term Celtic lan- 
 guages is the most proper epithet for the class of 
 idioms generally designated, and which I shall con- 
 tinue, in compliance with custom, to designate by 
 that name. The Celtse, properly so called, were a 
 people of Gaul. Of their language we have no 
 undoubted specimen. There are, indeed, strong- 
 grounds for believing that it was a kindred tongue 
 with the dialects of the British isles ; but it would 
 be better to take the general name of a whole class 
 of languages from something that actually remains. 
 
 There are six dialects of the language termed 
 Celtic which may be said to survive, as five are still 
 spoken, and one of them, viz. the Cornish, is suffi- 
 ciently preserved in books. These six dialects are, 
 the Welsh, the Cornish, the Armorican, the Irish 
 or Erse, the Gaelic or Highland-Scottish, and the 
 
THE CELTIC NATIONS. 25 
 
 Blanks. The three former are relics of the idiom 
 of the ancient Britons ; the three latter, of that 
 spoken by the inhabitants of Ireland. We have 
 historical evidence 8 that the Britons of Armorica, 
 the Britanni of Gregory of Tours, emigrated from 
 Britain, through the whole extent of which, with 
 the exception of some parts of the southern coast, 
 where the Belgae from Gaul had settled, it is pro- 
 bable that one language prevailed at the era of the 
 Roman conquest b . Of this language the three dia- 
 lects of Wales, Cornwall, and Lower Brittany are 
 descendants. Of the Irish language, the Scottish 
 Gaelic is a slight modification : the Manks differs 
 more considerably, and it is probable that the Isle 
 of Man had inhabitants from this branch of the 
 Celtic stock long before the emigration of the Scots 
 from Ireland to the coast of Argyle. 
 
 I shall in general take the Welsh as a specimen 
 of the Britannic dialects, and the Erse, or old Irish, 
 as an example of the other class ; but I shall add 
 occasionally words or forms which exist in the sub- 
 ordinate dialects, and are lost, or have become less 
 distinct, in either of the principal ones. 
 
 I have experienced some difficulty in adopting a 
 regular method in the orthography of Celtic words. 
 The modern system of representing consonants in 
 the Welsh and the Erse languages is so remote from 
 the usage of other tongues, that I have thought it 
 advisable to deviate from it in some instances. In 
 
 a Chiefly in the works of Gregory of Tours and Eginhardt — 
 I have surveyed the evidence on this subject in my Researches 
 into the Physical History of Mankind. 
 
 b That the Caledonians had this language has been proved by 
 Chalmers and Ritson. 
 
 
26 EASTERN ORIGIN OF 
 
 the former I have occasionally followed the ortho- 
 graphy of Edward Lhuyd, in preference to that 
 which is sanctioned by the authority of the Welsh 
 translators of the Bible. The grounds for this pre- 
 ference will appear obviously in the particular in- 
 stances in which it has been made, to those who are 
 acquainted with the Welsh language and its pro- 
 nunciation, and other readers will have no reason to 
 complain of a method which will guide them to the 
 proper utterance of words, when it would otherwise 
 have escaped them. In the Irish orthography, which 
 can scarcely be said to have any fixed standard, I 
 have followed the best authorities within my reach. 
 In the orthography of Sanskrit words I have de- 
 viated but little from the system proposed by sir 
 William Jones. In some few instances, however, 
 which will be obvious to those who are acquainted 
 with that method, I have endeavoured to approach 
 more nearly to the habit of our own language . 
 
 c I have followed Mr. Yates in substituting for the four San- 
 skrit diphthongs, v^,V^, 3f|, *3\ |, the following, ai> ox, b, au 
 There being some uncertainty as to the exact pronunciation of 
 vowels in ancient languages, it seems allowable to use those 
 vowels as representatives of each other, which in fact generally 
 are found in corresponding words, provided this method is not 
 used in such a manner as to produce an appearance of resem- 
 blance in words which are not in reality cognate. 
 
 O cms* 
 
THE CELTIC NATIONS. 27 
 
 CHAPTER I. 
 
 Preliminary survey of the forms of words and the permutations 
 of letters. 
 Section I. Introductory Remarks. 
 IN order to display, in its real extent, the affinity 
 which subsists between the Celtic dialects and other 
 languages, it will be necessary to compare them in 
 two different points of view, and to examine, in the 
 first place, the relations between their respective vo- 
 cabularies or stocks of primitive words or roots, and 
 secondly, the peculiarities and coincidences in their 
 grammatical structure. But before we enter into 
 details which properly belong to either of these sub- 
 jects, we must consider some general principles of 
 inflection, which have an important influence on the 
 structure of words and sentences in several of the 
 languages to be examined. 
 
 SECTION II. 
 
 Of the permutation of letters in composition and construction — 
 Of Sandhi and Samasa in Sanskrit — Of the same principles 
 as discovered in the Celtic dialects — in the Welsh — in the 
 Erse — Of the digamma and sibilant in Greek. 
 
 It is a habit common to many of the Indo-Euro- 
 pean languages to interchange certain letters accord- 
 ing to rules founded originally on euphony, or on 
 the facility of utterance ; and from this circum- 
 stance arises the great capability, which these lan- 
 guages possess, of composition, or the formation of 
 compound words. The substitution of consonants of 
 particular orders for their cognates, which takes 
 
 t 
 
28 EASTERN ORIGIN OF 
 
 place in Greek in the composition of words, and in 
 some other instances, is an example of this pecu- 
 liarity. 
 
 In Greek, in Latin, and in the German dialects, 
 the mutation of consonants is confined to words 
 brought together under very peculiar circumstances, 
 as chiefly when they enter into the formation of 
 compound terms, and it is scarcely observed in 
 words which still remain distinct, and are merely 
 constituent parts of sentences. Either the attention 
 to euphony and the ease of utterance has not ex- 
 tended so far, or the purpose was attained by a 
 choice of collocation, the words themselves remain- 
 ing unaltered. But in the Sanskrit language words 
 merely in sequence have an influence upon each 
 other in the change of terminations, and sometimes 
 of initial letters, on the principle above alluded to. 
 Thus, instead of atishtat manajak, stabat homo, the 
 man stood, we find the words written atishtan ma- 
 nujak, the final t of the verb atishtat, stabat, being 
 altered into n on account of the liquid consonant 
 with which the next word begins. This change in 
 distinct words is termed by the Sanskrit gramma- 
 rians WtV? Sandhi, conjunction ; but the laws ac- 
 cording to which compound words are formed, and 
 which have a similar reference to euphony, are de- 
 signated *^*i m, Samasa, coalition. This last pro- 
 cess is to be observed in most, if not in all the Eu- 
 ropean languages, and the rules which govern it in 
 all instances are very similar ; but the alteration of 
 consonants in entire words, according to the rules 
 of Sandhi, have been considered as in a great mea- 
 sure peculiar to the Sanskrit. It is, however, a re- 
 markable fact, that in the Celtic dialects, and more 
 
TI-IK CELTIC NATIONS. 
 
 29 
 
 especially in the Welsh, permutations in many re- 
 spects analogous to those of Sandhi are constant and 
 indispensable in the formation of sentences. It is 
 impossible to bring three or four words together in 
 the Celtic languages, without modifications similar 
 in their principle to those of Sandhi. 
 
 The general principle of these changes in Sanskrit 
 may be understood by the following table of conso- 
 nants, arranged according to the organs by means 
 of which they are pronounced, and likewise accord- 
 ing to the intensity and mode of utterance. The 
 former arrangement is analogous in some respects 
 to that of the Greek mutes, but more numerous and 
 comprehensive. It consists of five classes, termed 
 Guttural, Palatine, Lingual, Dental, and Labial. 
 To these is added a sixth, consisting of semivowels ; 
 and a seventh, containing Sibilants and an Aspirate, 
 which is associated with the Sibilants. The second 
 division is into two orders termed Surds and So- 
 nants. 
 
 
 
 Surds. 
 
 
 SONANTS. 
 
 
 Gutturals 
 
 K 
 
 
 K'h 
 
 G 
 
 G'h 
 
 Ng. 
 
 Palatines 
 
 Ch 
 
 
 Ch'h 
 
 J 
 
 J'h 
 
 Gn. 
 
 Linguals 
 
 T 
 
 
 T'h 
 
 D 
 
 P'h 
 
 N. 
 
 Dentals 
 
 T 
 
 
 T'h 
 
 D 
 
 D'h 
 
 N. 
 
 Labials 
 
 P 
 
 
 P'h 
 
 B 
 
 B'h 
 
 M. 
 
 Semivowels 
 
 
 
 
 Y- 
 
 _R_L- 
 
 -V. 
 
 Sibilants 
 
 S 
 
 Sh 
 
 S H 
 
 
 
 
 The vowels are included 
 among the Sonants. 
 
 The laws of Sandhi forbid the meeting of conso- 
 nants of different orders. Hence a Surd consonant 
 at the end of a word is changed with the corre- 
 sponding sonant, if the next word begins with a so- 
 
30 
 
 EASTERN ORIGIN OF 
 
 riant ; and sonants are changed into surds if the fol- 
 lowing words begin with surds. 
 
 Nearly of the same description are the mutations 
 of consonants in the Celtic language ; but in order 
 to obtain a view of the whole system of these 
 changes, it is necessary to compare several dialects, 
 as there is not one which preserves them all in an 
 entire state. The Welsh alphabet has them, how- 
 ever, in greater variety than the others. In this 
 all mute consonants of the order termed above surds 
 have four forms, and those which correspond with 
 the sonants have three. The semivowels or liquids 
 have two. The sibilant letter had probably its mu- 
 tation into the aspirate, but this is lost in Welsh, 
 though preserved, as we shall see, in the Erse. 
 
 First order, the primitive letters being surds. 
 
 
 First form, 
 Sharp. 
 
 Second form, 
 Obtuse. 
 
 Third form, 
 Aspirate. 
 
 Fourth form, 
 Liquid. 
 
 Gutturals 
 
 Dentals 
 
 Labials 
 
 C 
 t 
 
 P 
 
 g 
 
 d 
 b 
 
 ch 
 
 th 
 P h 
 
 ngh 
 
 nh 
 
 mh 
 
 Second order, in which the primitives are sonants. 
 These have two changes. 
 
 
 Primitive. 
 
 Obtuse. 
 
 
 Liquid. 
 
 Gutturals 
 
 Dentals 
 
 Labials 
 
 g 
 
 d 
 b 
 
 initial omitted 
 dh or Saxon p 
 
 V 
 
 ng 
 
 n 
 
 m 
 
 lird order, Liquids. These have one chan 
 
 lh (corresponding with the surd 
 lh or lr of the Vedas.) 
 
 1 
 
 m 
 rh 
 
 
 
 
 V 
 
 r 
 
THE CELTIC NATIONS. 
 
 31 
 
 Of the mutations of consonants in the Erse or 
 Gaelic. 
 In the Erse dialect of the Celtic language the 
 imitations of consonants are not so varied. Each 
 consonant appears in two forms only, which are 
 termed the plain and the aspirated. But the aspi- 
 rated forms in the Erse are often the obtuse forms 
 in Welsh, the aspiration being deceptive, and arising 
 from the imperfect orthography adopted in this lan- 
 guage. The addition of h to the primitive conso- 
 nant serves only to render it obtuse, or in other in- 
 stances to obliterate it. On this account I shall set 
 down the table of consonants, with one column for 
 the obtuse letters as usually spelled, and another in- 
 dicating their pronunciation, which is in general si- 
 milar to that of the obtuse forms in Welsh. 
 
 Gutturals 
 
 Dentals 
 Labials 
 
 Liquids 
 Sibilants 
 
 It f is to be observed that H never stands as the 
 initial of a word in Erse in the primitive form, or 
 
 "Plain or primitive 
 
 form. 
 
 Secondary form 
 as spelled. 
 
 Secondary form 
 as articulated. 
 
 Cor K 
 
 
 Ch 
 
 X aspirate] 
 or Kh \ 
 
 G hard 
 
 
 Gh 
 
 
 T 
 
 
 Th 
 
 H 
 
 D 
 
 
 Dh 
 
 
 P 
 
 
 Ph 
 
 F 
 
 B 
 
 
 Bh 
 
 V 
 
 M 
 
 
 Mh 
 
 V 
 
 F 
 
 
 Fh or H 
 
 H 
 
 L (like Welsh 
 
 Lh) 
 
 L 
 
 L plain 
 
 N 
 
 
 N 
 
 
 R (like Rh) 
 
 
 R 
 
 
 S 
 
 
 Shi „ 
 . > or H 
 S ) 
 
 H 
 
32 EASTERN ORIGIN OF 
 
 is never in fact an independent radical letter. It is 
 merely a secondary form or representative of some 
 other initial, viz. F or S. It must likewise be no- 
 ticed, that the same words which begin with S or F 
 as their primitive initial in the Erse, taking H in 
 their secondary form, have in Welsh H as their pri- 
 mitive initial. This fact affords an instance exactly 
 parallel to the substitution in Greek of the rough 
 and soft breathings for the iEolic digamma, and in 
 other words for the sigma. O/W, as it is well 
 known, stands for Foiva, "Eo-nepos for Fev-rrepos, and 
 en™ probably replaced a more ancient form of the 
 same word, viz. aeTrra ; ef stands for o-ef, lg and epvco 
 for <Tv$ and a-epirw. These instances might lead us 
 to suppose, as Edward Lhuyd has long ago ob- 
 served, that the Greek language had originally a re- 
 gular mutation of initial consonants, similar to that 
 of the Celtic, though it was lost, except in these in- 
 stances, or rather, as pointed out by these vestiges, 
 previously to the invention of letters. 
 
 It is necessary to explain somewhat more fully 
 the nature of these mutations of consonants in the 
 Celtic language, and this can only be done by point- 
 ing out the circumstances under which they take 
 place. The following examples are from the Welsh. 
 
 1 . Words of four initials. 
 
 Gutturals. 
 Car, a kinsman. 
 
 1. form Car agos, a near kinsman. ** 
 
 2. form Ei gar, his kinsman. 
 
 3. form Ei char, her kinsman. 
 
 4. form Vy nghar, my kinsman. 
 
THE CELTIC NATIONS. 
 
 33 
 
 tu y 
 
 Dentals. 
 Tad, a father. 
 
 1. form, Tad y plentyn, the child's father. 
 
 2. form, Ei dad, his father. 
 
 3. form, Ei thad, her father. 
 
 4. form, Vy nhad, my father. 
 
 Labials. 
 Pen, a head. 
 
 1. form, Pen gwr, the head of a man. 
 
 2. form, Ei ben, his head. 
 
 3. form, Ei phen, her head. 
 
 4. form, Vy mhen, my head. 
 
 2. Words of three initials. 
 
 Gutturals. 
 Gwas, a servant. 
 
 1. form, Gwas fydhlon, a faithful servant. 
 Ei was, his servant. 
 Vy ngwas, my servant. 
 
 Dentals. 
 
 
 2. 
 
 form, 
 
 3. 
 
 form, 
 
 Duw, a god 
 
 1. 
 
 form, 
 
 2. 
 
 form, 
 
 3. 
 
 form, 
 
 
 
 Duw trugarog, a merciful god. 
 Ei dhuw, his god. 
 Vy nuw, my god. 
 
 Labials. 
 Bara, bread. 
 
 1. form, Bara cann, white bread. 
 
 2. form, Ei vara, his bread. 
 
 3. form, Vy mara, my bread. 
 
 <?2^t^v-^-w <** v 
 
 3. Words of two initials, viz. liquids and sibilants. 
 Lhaw, a hand. 
 
 1. form, Lhaw wenn, a white hand. 
 
 2. form, Ei law, his hand. 
 
 D 
 
 
 
34 EASTERN ORIGIN OF 
 
 Mam, a mother. 
 
 1. form Mam dirion, a tender mother. 
 
 2. form Ei vam, his mother. 
 Rhwyd, a net. 
 
 1. form Rhwyd lawn, a full net. 
 
 2. form Ei rwyd, his net. 
 
 As the sibilant has no similar inflexion in Welsh, 
 I must take an example from the Erse. 
 Suil a , an eye. J* 
 
 1. form Suil. 
 
 2. form a Mil, his eye. 
 
 Slainte, health. — - 
 
 2. form Do hlainte, your health. ~V^ 
 
 N. B. F has the same mutation. 
 
 Words beginning with vowels in Welsh are sub- 
 ject to changes similar to those belonging to the 
 form Guna in Sanskrit. They also take the aspi- 
 rate after words which cause the consonants to be 
 aspirated. 
 
 In Welsh composition these changes in the initial 
 consonants take place more frequently in reference 
 to the sense of words and the rules of grammatical 
 construction, and without any respect to the prin- 
 ciple of euphony which governs the Sandhi in San- 
 skrit. But there are a great many similar changes 
 in Welsh, for which no other reason can be assigned 
 than some real or fancied advantage in respect to 
 
 a In these instances the initial s, though converted into an 
 aspirate in pronunciation, is sometimes retained in orthography, 
 either with a dot over it, or followed by h. But in either case 
 the sibilant is entirely lost. There seems to be no precise rule 
 of orthography in this instance. 
 
THE CELTIC NATIONS. 35 
 
 sound or the facility of utterance. It must likewise 
 he observed, that in some instances changes are in- 
 duced in the terminating consonants of preceding 
 words, as well as in the initials of succeeding ones. 
 
 In the influence which some of the numerals have 
 on other words examples may be found tending to 
 illustrate these remarks. 
 
 Un makes no change in the following noun ; as 
 
 un gwr, one man. 
 Tri and chwech change the initials into the cor- 
 responding aspirates ; as 
 
 tri char, for tri car. 
 chwech char, or chwe char. 
 Deg, ten, before blynedh, years, changes not only 
 the initial of the following word into its correspond- 
 ing liquid, but likewise its own final consonant into 
 the liquid analogous to it. Thus instead of 
 
 deg blynedh, 
 we read 
 
 deng mlynedh, ten years ; 
 and instead of 
 
 pump blynedh, 
 we find 
 
 pum mlynedh, five years. 
 In like manner, when the preposition yn is pre- 
 fixed to a noun, it not only changes the initial of 
 the following noun on the same principle of eu- 
 phony, or ease of pronunciation, but is likewise itself 
 changed. Thus for 
 
 yn canol, we read yng nghanol, 
 
 yn pen ym mhen, 
 
 yn ty yn nhy, 
 
 yn bara ym mara, 
 
 yn gwr yng ngwr. 
 
 D 2 
 
36 EASTERN ORIGIN OF 
 
 The changes above described are in a great mea- 
 sure analogous to those which have been pointed 
 out as taking place in Sanskrit, except that the 
 latter affect principally, though by no means exclu- 
 sively, the terminations of words. 
 
 SECTION III. 
 
 Of the interchange of particular letters between different lan- 
 guages — Table of numerals — Observations deduced from it. 
 
 There is another comparison of corresponding 
 consonants and vowels, or of letters frequently and 
 habitually interchanged, which it is necessary to 
 take into consideration, before we can proceed with 
 advantage in examining the analogies which exist 
 between languages of the same stock. I refer to the 
 phenomena which relate to the interchange of par- 
 ticular letters in the derivation of words from one 
 dialect into another, or in deducing them into both 
 from a common original, and to facts which prove 
 that these changes take place according to certain 
 rules, and not by a merely accidental variation or 
 corruption. 
 
 In order to ascertain the rules which govern this 
 system of changes it is necessary to proceed with 
 great caution. The vague conjectures in which 
 writers upon etymology have too frequently in- 
 dulged, have brought ridicule and contempt upon 
 the legitimate pursuits of the philologist, and upon 
 the philosophical study of languages, and have in- 
 duced some persons to entertain doubt, whether it is 
 possible to deduce from this quarter any historical 
 conclusions of importance, either as to the deriva- 
 
THE CELTIC NATIONS. 37 
 
 tion of languages themselves, or of the tribes of 
 people who are found to use them. As an introduc- 
 tion to what may be stated on this subject, I lay 
 before my readers a comparison of the cardinal 
 numbers in several Indo-European languages. It 
 will be apparent on a very superficial glance, that 
 the words expressing these numerals in all the dia- 
 lects mentioned are derived from one origin, though 
 variously modified; and a survey of these modifica- 
 tions will shew the particular changes which words 
 and the elements of words assume in the respective 
 languages. 
 
 D 3 
 
Si s 
 
 13 
 
 cf 
 
 C 
 
 "8 
 
 CO 
 
 CO 
 
 a 
 
 2 
 
 a 
 
 .o 
 
 
 60 
 
 3 a 
 
 -a 
 
 be 
 
 o 
 
 a 
 a 
 ja 
 
 J- 
 o 
 
 <a 
 
 2 
 
 a 
 
 a § 
 
 I 1 
 
 a +-> 
 
 a 
 So 
 
 
 s 
 
 a 
 
 60 
 
 CO 
 
 a 
 
 a 
 a 
 
 ^3 
 
 
 §■8 s-fe * 
 
 fr2 v $ 
 
 £r £r t- fc= t= i- fc £ <** 
 
 e bo 
 
 s 
 
 ft 
 o 
 
 ft 
 
 rem 
 
 *3 
 
 !> eg a 
 
 t3-o a 
 
 cu a> a 
 
 PhP, A 
 
 •3 
 
 
 -§ 
 
 £ 60 
 
 a <cd 
 a t3 
 
 w 
 
 o a 
 
 CD CD 
 CO CO 
 
 o 
 
 3 
 
 I 
 
 u 
 
 CD 
 
 •T3 
 
 a 
 •5b 
 
 3 
 
 a 
 
 '60 
 
 
 P-i CO 
 
 co O 
 O > 
 
 c3 
 
 I 
 
 cu 
 -T3 
 
 c3 
 > 
 n3 
 
 % 2 
 
 
 ^ 
 
 a 
 
 CO 
 
 I I 
 
 ^ 
 «> 
 
 c8 !FH 
 
 =1 -I 
 
 a 
 
 ctf 
 
 H3 
 
 ■c 
 
 K 
 
 I I J 
 
 pte'&dEr 
 
 ■8 
 
 •8 
 I 
 
 A- 
 
 8 
 
 CD .*■* 
 
 4 
 
 CD 
 
 a cd 
 a ^ 
 
 t i a 
 
 I t 
 
 J-l 
 
 R 
 
 ■9 
 
 -9 
 
 a 
 
 CCJ 
 
 
 a 
 
 a 
 
 XL 
 
 as 
 60 
 
 etf 
 
 s 
 
 03 
 
 Ph 
 
 ^3 
 
 > 
 
 a 
 
 c3 
 
 a 
 
 
 cS 
 Ph 
 
 CO 
 
 a 
 
 3 
 
 *G 
 
 > 
 
 z 
 
 tr fer tF RT 
 
 
 c3 
 60 
 
 a 
 
 I 
 
 1 
 co 
 
 I 
 
 a 
 
 ccj 
 
THE CELTIC NATIONS. 
 
 39 
 
 A very slight inspection of these tables will be 
 sufficient to convince any person that nearly all the 
 words contained in them are derived by each lan- 
 guage from some of its cognates, or by all from a 
 common source. It is therefore allowable to make 
 them a subject of examination, from which the pe- 
 culiarities of each dialect may, so far as such a spe- 
 cimen can extend, be discovered. 
 
 It is easy to observe that certain consonants, or 
 certain classes of consonants in one language, are 
 almost uniformly substitutes for certain others in a 
 different language ; and although this observation 
 can here be made only on a confined scale from so 
 small a specimen of the vocabulary, it may be suffi- 
 cient for furnishing suggestions which will be amply 
 established from other materials. 
 
 One of the most striking facts that appears on 
 comparing these lists of numerals is, that in some of 
 the languages of western Europe guttural or hard 
 palatine consonants abound, and take the place of 
 the sibilants, soft palatines, and dentals, and even of 
 the labial consonants, which are found in the more 
 eastern and in some northern languages. Thus 
 
 S£I — sh ] f c, i. e. k 
 
 *T— s q 
 
 ^ — sn . , • g 
 
 r-r >are converted mto< , . 
 
 H — p r ^ ch, 1. e. x 
 
 D 4 
 
40 
 
 EASTERN ORIGIN OF 
 
 
 > become 
 
 
 The following examples prove this remark: 
 Numeral 4. 
 chatur, Sansk. 
 chetyre, Russ. 
 chehar, Pers. 
 
 T€TTap€$, Gr. 
 TMTVpef, 
 
 pedwar, Welsh 
 petor, Oscan. 
 fidwor, Goth, 
 fiuuar, Teut. 
 
 Numeral 5. 
 
 pancha, Sansk. p & ch 
 penj, Pers. p & j 
 
 7reVT€, Gr. 7T & T 
 
 Gr. 7T & 7T 
 
 Welsh p & p 
 Goth, f & f 
 
 I ch 
 
 }: 
 
 quatuor, Lat. 
 keathair, Erse 
 
 pump, 
 fimf, 
 
 s 
 
 o 
 
 q and q, quinque, Lat. 
 k and g, kuig, Erse 
 
 Numeral 6. 
 shash, Sans, sh & sh ' 
 shesh, Pers. sh & sh 
 sex, Lat. s & x 
 saihs, Goth, s & s 
 
 g f ch & ch ) , «.„•».* 
 , fchwech, Welsh 
 
 
 - g j guttural 
 
 f) and f, ef, Greek 
 
 Numeral 7. 
 
 saptan, Sansk. s & pt " 
 septem, Lat. s & pt 
 saith, Welsh s & th 
 
 ashtan, Sansk. 
 hesht, Pers. 
 wyth, Welsh 
 
 a 
 
 V o 
 w 
 u 
 
 s and cht, Erse 
 
 h and ft, Pers. 
 
 ( f ) and or, Greek 
 
 Numeral 8. 
 
 cht ocht, Erse 
 
 kt oktw, Greek 
 
 ct octo, Lat. 
 
 ht ahtan, Goth. 
 
 shr 
 
 .1. 
 
 sht 
 
 
 U 
 
 th 
 
 0) 
 
 
 
THE CELTIC NATIONS. 
 
 41 
 
 
 Numeral 10. 
 
 
 
 
 
 K §€K(X 9 
 
 Greek 
 
 
 
 , Xfl 
 
 c decern, 
 
 Lat. 
 
 
 dashan, Sansk. 
 
 sh 
 
 s 
 
 ■ o - 
 o 
 
 0) 
 
 ch deich, 
 g deg, 
 
 Erse 
 Welsh 
 
 
 
 . pC 
 
 h tehan, 
 h taihun, 
 
 Teut. 
 Goth. 
 
 
 
 Numeral 20. 
 
 
 
 
 "! GO 
 
 g viginti, 
 
 Lat. 
 
 
 vinshati, Sansk. 
 
 sh 
 
 S 
 
 ■ o " 
 u 
 
 en 
 
 g ugain, 
 
 k eiKocri a , 
 
 Welsh 
 Greek 
 
 
 
 J & 
 
 ch fichid, 
 
 Erse 
 
 
 Numeral 30. 
 
 
 
 trinshat, Sansk. 
 
 GO 
 
 sh } | \ K T ?' lcucov ™> 
 f § (g triginta, 
 
 Numeral 100. 
 
 Greek 
 Lat. 
 
 
 
 
 K €KO,T0V 9 
 
 Greek 
 
 
 shatum, Sansk. 
 sad, Pers. 
 
 sh|| 
 
 c centum, 
 c cant, 
 k kett, 
 
 Lat. 
 
 Welsh 
 
 Erse 
 
 
 
 
 
 h hunt, 
 
 Goth. 
 
 
 The preceding facts suggest the following obser- 
 vations. 
 
 The Sanskrit and some other languages holding 
 a near relation to it in the forms of words abound 
 in sibilants and soft palatine consonants. They 
 have these letters in several instances, in which cog- 
 nate words in other languages have in the place of 
 them gutturals, or hard palatines, or dentals. 
 
 a e'Uocri was probably FeUovn. 
 
42 EASTERN ORIGIN OF 
 
 The Greek substitutes for the sibilants and soft 
 palatines of the Sanskrit, chiefly the tenues of the 
 hard palatine or guttural class and of the dental, 
 viz. k and t. In several instances the Greek, parti- 
 cularly the iEolic, has tt in the place of the Sanskrit 
 soft palatine, or ^ — ch ; as in wcpre for pancha, neavpa 
 (neTvpa?) for chatur. 
 
 The Welsh makes nearly the same substitutions 
 as the iEolic Greek. It puts p for the soft palatine 
 ch in the instances before mentioned. It substitutes 
 more generally hard palatines or gutturals (either c, 
 i. e. k,) or ch for the soft palatines and sibilants of 
 Sanskrit. It has the aspirate guttural ch instead of 
 the aspirate sibilant sh. It has th in the place of ct 
 and pt. 
 
 The Erse substitutes for the sibilants and soft 
 palatines of the Sanskrit, gutturals, as the hard 
 c or k, as also in some instances the guttural as- 
 pirate ch. 
 
 The Latin displays nearly the same phenomena 
 as the Erse. It puts c or q, equivalent to k, in the 
 places of the letters above mentioned. Neither the 
 Erse nor the Latin adopts the p of the Welsh and 
 iEolic Greek, but they have c or q instead of it, as 
 in other instances where the Sanskrit has ch — ^. 
 
 The Gothic and other Teutonic dialects resemble 
 the Welsh and the iEolic Greek, except in the cir- 
 cumstance that they prefer aspirate consonants, as 
 finfe for wfywre or pump, fidwor for pedwar, or vervp, 
 thri for tri. They likewise substitute the simple h 
 in the place of palatines and sibilants in other lan- 
 guages, as may be seen in a variety of instances, as 
 in the numerals, 6, 8, 9, 10, 100. The Persic and 
 
THE CELTIC NATIONS. 43 
 
 the Greek languages use the aspirate in some in- 
 stances in a similar manner. 
 
 We are not yet prepared for entering on a com- 
 parison of the vowels and diphthongs as they are 
 related to each other in these cognate languages. 
 
44 EASTERN ORIGIN OF 
 
 CHAPTER II. 
 
 Further proofs and extension of the observations laid down in 
 the preceding chapter. 
 
 Section I. Introductory Remarks. 
 
 X HE changes which I have pointed out in the 
 preceding section between particular consonants 
 in the derivation of words from one language to 
 another, appear, in some instances, so unlikely, and 
 the analogy, if any, in pronunciation is so remote, 
 that many of my readers may be disposed to regard 
 the examples on which I have founded my remarks 
 as a mere result of accidental coincidence. These 
 changes are, notwithstanding, regular and systema- 
 tic. I shall not attempt to account for them, or to 
 say how they took place, but they are accordant 
 with observations which may be traced to a great 
 extent in the comparison of kindred languages. As 
 I cannot, however, expect that any person should be 
 convinced of this fact on my assertion, I shall here 
 adduce some further evidence. 
 
 SECTION II. 
 
 Of the interchange of palatine or guttural consonants with la- 
 bials in the different languages. 
 
 The interchange of cognate letters, both mutes 
 and liquids a , is a thing familiar to every body, but 
 
 a The cognate mutes are 
 
 t, d, th. 
 
 k, g, ch. 
 
 p, b, ph. 
 
THE CELTIC NATIONS. 45 
 
 the permutation of palatines into labials appears 
 much more improbable. We have observed that 
 this interchange has taken place in several instances 
 in the numerals of Indo-European languages. Great 
 as the difference is between such elements of articu- 
 lation as k and p, we find them to stand as repre- 
 sentatives for each other even in two different dia- 
 lects of the same language. Some dialects of the 
 Greek language afford a well-known exemplification 
 of this remark. The Ionians and iEolians inserted 
 Ka7nra in a variety of words, instead of wi, used in the 
 other Grecian dialects. This remark has been made 
 by many of the scholiasts and old grammarians, and 
 more fully by Vossius b , who says, " Iones in interro- 
 1 gativis et relativis mutant 7r in k. Ita kx$ dicunt 
 1 pro nag ; OK&g pro oncog ; kyj pro 7tyj ; 7zwo§-, Koaog ; 
 1 07roaog 9 oKoaog ; voio$, Kohg ; 07ro?og, OKoiog ; 7tot€, Kore ; 
 ' o7tot€, okot€. Grsecis quoque Kva^og est faba. iEoles 
 ' quoque uti k pro v testatur E ty mologici auctor 
 6 in Kohg. Sic Latini jecur a Gr. tjirap, et scintilla, 
 6 quasi spintilla, a o-irivtyp" 
 
 The same writer has adduced other instances in 
 which this interchange has taken place between the 
 Greek and Latin. 
 
 Lupus. 
 
 kvKOg. 
 
 Sepes. 
 
 <TY]K0g. 
 
 Spolia. 
 
 GKvka. 
 
 Vespas. 
 
 a(f>v]Kag. 
 
 Cognate liquids or semivowels are in many languages the fol- 
 lowing. 
 
 1, r, v. 
 
 a Gerard. Joh. Vossii de Litterarum permutatione Tractatus, 
 Etymol. Ling. Lat. prefix, p, 24. ed. Neap. 1762. 
 
46 EASTERN ORIGIN OF 
 
 " Maxime tamen locum id habet in iis vocibus, 
 " in quibus juxta Ionicae et iEolicae dialecti proprie- 
 " tatem, it transiit in k. 
 
 Equus ab iEolico Ikkos pro iWof. 
 
 Inquio ab iEolico evvUco — Gr. hveirw. 
 
 Linguo ab iEolico Ac/aw — Gr. AeiWw vel a Ae/™, 
 
 \ifA7ravoo. 
 Qua ab Ion. kyj pro Gr. ir*\. 
 
 Quatuor a neTTOpa, KeTTOpa. 
 
 Quinque a nevTe, irefxire, /avK€. 
 
 Quis a tis, Kig. 
 
 Quoties ab Ionice kotc, Gr. irore. 
 Quotus a kotos, pro ttotos. 
 
 Sequor ab eKopoci pro evofmi G . 
 
 The learned Edward Lhuyd has observed that a 
 similar interchange of p and k takes place regularly 
 between the Welsh and Erse dialects of the Celtic 
 language. I shall cite his words and the evidence 
 he adduces for this remark. 
 
 " It is very remarkable that there are scarce any 
 " words in the Irish, besides what are borrowed 
 " from the Latin, or some other language, that be- 
 " gin with p ; insomuch that in an ancient alpha- 
 " betic vocabulary I have by me that letter is 
 " omitted ; and it is no less observable that a consi- 
 " derable number of those words, whose initial letter 
 " in the British language is a p, begin in the Irish 
 " with a k, or, as they constantly write it, with a c. 
 " This partly appears by the following examples : 
 
 Paul, W. a pole or stake, Kual, Ir. 
 
 pA,, fa thing, part, share, \ Kod, Koda, 
 
 ( some, j Kiiyd. | 
 
 c Voss. ubi sup. p. 24. 
 
THE CELTIC NATIONS. 
 
 47 
 
 Pa? 
 
 
 what ? 
 
 Ka? 
 
 Pask. 
 
 
 Easter, 
 
 Kasg. 
 
 Pencas, 
 
 Corn. 
 
 Whitsuntide, 
 
 Kaikis. 
 
 Peiswin, 
 
 W. 
 
 chaff, 
 
 Kaithsloan. 
 
 Pesuch, 
 
 
 a cough, 
 
 Kasachd. 
 
 Pen, 
 
 
 a head, 
 
 Keann. 
 
 Puy, 
 
 
 who? 
 
 Kia? 
 
 Puylh, 
 
 
 sense or meaning, 
 
 KiaL 
 
 Plant, 
 
 
 children, 
 
 ( Klann, and 
 1 Kland. 
 
 Plyv, 
 
 
 feathers, 
 
 Kluvv. 
 
 Peduar, 
 
 
 four, 
 
 Kathair. 
 
 Pymp, 
 
 
 five, 
 
 Kuig. 
 
 Pair, 
 
 
 a furnace or cauldr 
 
 on, Kuir & Koire. 
 
 Pren, 
 
 
 a ton, 
 
 Kran. 
 
 Par, 
 
 
 a couple, 
 
 Koraid. 
 
 Pridh, 
 
 
 earth or clay, 
 
 Kriadh. 
 
 Praidh, 
 
 
 a prey, 
 
 Kreach. 
 
 Pa raid, 
 
 
 wherefore, 
 
 K'red. 
 
 Pryv, 
 
 
 a worm, 
 
 Kruv. 
 
 Pob, 
 
 
 every, 
 
 Ceach or Gach. 
 
 And sometimes in other parts of the words we 
 find the same : as 
 
 Yspydhad, a hawthorn, Skiathach, 
 
 Map or Mab, a son, Mak. 
 
 The preceding examples are quite sufficient to 
 establish the fact asserted in the present section. 
 We shall hereafter find the application of this re- 
 mark. 
 
48 EASTERN ORIGIN OF 
 
 SECTION III. 
 
 Of the interchange of sibilant and soft palatine consonants with 
 gutturals or hard palatines. 
 
 It has been customary in many languages, and in 
 our own among others, to soften the guttural or hard 
 palatine letters, or to interchange them with other 
 elements of pronunciation which are termed sibilants 
 and soft palatines. We substitute the ordinary ch 
 in the place of the hard c, or the k of other cog- 
 nate languages, and say church for kirk or kirche. 
 The Italians pronounce Tschitschero, a name which 
 the Greeks wrote Kucepcov. Secondly, many nations 
 are in the habit of softening the g, and giving it the 
 pronunciation of our j, as we are accustomed to do 
 when this consonant comes before the vowels e and i. 
 Thirdly, we shall find some languages converting 
 the guttural aspirate % or ch into sh, as the Welsh 
 substitute chwech for the Sanskrit shash. 
 
 It will illustrate the two former of these changes to 
 observe that the Sanskrit ^ — ch d is interchangeable 
 in the regular inflexions of that language for ^ — k, 
 and 3T — j likewise for 3J — g. Thus, verbs beginning 
 with k, in the reduplication of the initial, which in 
 Sanskrit as in Greek is a character of the preter- 
 perfect tense, substitute ch for k, and verbs begin- 
 ning with g substitute j for that consonant. The 
 following are examples. 
 
 Root. Present. Preterperfect. 
 
 ^i kri, (to make) ^vUTrTj karoti ^^THC chakara. 
 
 3| goi, (to sing) ^TRm, gayati 3FTT, jagau. 
 
 d Ch, as in cherry. 
 
TIIK CELTIC NATIONS. M) 
 
 We cannot find a parallel fact in the Sanskrit 
 language for the third remark, which respects the 
 interchange of the aspirate sibilant for the aspirate 
 guttural, because the Sanskrit has no consonant ana- 
 logous to the Greek %ror the Welsh ch. 
 
 The preceding remarks will be more perspicuous 
 if we place these changes in a tabular form, as fol- 
 lows ; 
 
 ^i, k, or c, or q — interchanged for ^T, ch. 
 
 sometimes for ST sh v ^T sh, or ?f s. 
 
 ^T g, for 3T j. 
 
 y? or ch aspirate guttural, for ST sh, ^" sh, or ^T s. 
 
 It must be observed that the Greek BT and Z^ra 
 are to be included in many instances among the pa- 
 latine letters, and fall under the same rules of per- 
 mutation. E7 is sometimes represented in Sanskrit 
 by %$ ksh, but frequently by the simple character 
 corresponding with sh. Z^t«, when it is the cha- 
 racteristic of verbs making the future in fey, may 
 properly be considered as a palatine letter, and it 
 will be found represented in Sanskrit by palatine 
 consonants. 
 
 I shall exemplify these remarks by some lists of 
 words in addition to those instances already disco- 
 vered among the numerals, in which the above- 
 mentioned interchanges occur. The first series con- 
 tains examples of soft palatines in one language and 
 hard palatines or gutturals in another ; the second, 
 cases in which j is substituted for hard g, and the 
 third, words in which sibilants appear in the place 
 of gutturals or hard palatines. 
 
 E 
 
50 EASTERN ORIGIN OF 
 
 I. Words in which ^ — ch or ch soft is inter- 
 changed with hard palatine letters. 
 
 Words having of ch or ch soft. Words having hard palatines. 
 
 ^ cha, and, subjoined ) . _, 
 
 ,, r Kai, Gr. que, Lat. 
 
 to the noun, J ^ 
 
 chatur, quatuor. 
 
 locha, look, Eng. 
 
 chatai, \ 
 vacet, ) 
 
 looketh. 
 
 lochayati, lucet, Lat. 
 
 lochan, (an eye) lhygad, i. e. lhugad, W. 
 
 vachas, voces, Lat. 
 
 vachati, orl Paget, i. e. £«*€/, unde 
 
 vakti, j i9a?/f. 
 
 chyotati, yeverou, yjeerat. 
 
 richch'hati, S. ) f opeyerou. 
 
 ) 
 
 reacheth, Eng. j (erreicht, Germ, 
 
 uchcha and 1 , . , ( uch, uchel, W. 
 
 uchchah, ) (hoch, Germ, 
 
 uchchata, (arrogance) uchad, W. (act of rising.) 
 
 church, Kvpioucvj, kirche, &c. 
 
 II. The following are examples of j or 3f in 
 Sanskrit supplying the place of y or g in Greek and 
 other European words. 
 
 Sanskrit. European languages. 
 
 januh, genu, ydw 9 knee. 
 
 janus, (birth,) yovog. 
 
 jani, ywYj. 
 
 jarami, yqpwh I grow old. 
 
 jaran, yepwv. 
 
 jarati, ■ ypavg. 
 
THE CELTIC NATIONS. 
 
 51 
 
 jarjati, 
 jagaras, 
 
 jatus, 
 
 taijatai, he sharpens. 
 
 ajah, 
 
 raj am, 
 
 jurgat. 
 
 eyprjyopos. 
 begotten, yerys. 
 Oyyerai, 
 ouya 9 goat, 
 regem. 
 
 III. Instances of sibilant consonants interchanged 
 for gutturals e . 
 
 Sibilants. 
 
 
 Gutturals. 
 
 dresh, root. 
 
 
 SepKtiv. 
 
 dadarsha, 
 
 
 ^e^opKa. 
 
 dansh, root, 
 
 
 SaKveiv. 
 
 danshati, 
 
 
 SaKvei. 
 
 mishrayatai, 
 
 
 fjLio-y erect. 
 
 mishrum, 
 
 
 mixtum, 
 
 ashwah or eshuus 
 
 ) f equus. 
 
 asb, (Persian) 
 
 1 
 
 (each, (Erse) 
 
 shwashurum, 
 
 
 socerum. 
 
 shwashrus, 
 
 
 socrus. 
 
 pashus, 
 
 
 pecus. 
 
 swasaram, 
 
 ) 
 
 
 sororem, 
 
 sch wester, Germ. 
 
 suir, Erse 
 
 i 
 
 f khauhir, Pers. 
 Ikhwaer, Welsh. 
 
 sister, 
 
 J 
 
 
 ^poaog, 
 
 
 druchd, (Erse.) 
 
 seta, Lat. 
 
 
 kaishah, Sansk. 
 
 suess, Germ. Sweet, 
 
 chwys, W. 
 
 silex, Lat. 
 
 
 X«A/f. 
 
 sch wan, Germ. Swan 
 
 KVKVOg. 
 
 e The words in the left hand columns not otherwise specified, 
 and neither English, nor Latin or Greek, are Sanskrit. 
 
 E 2 
 
52 EASTERN ORIGIN OF 
 
 short, Eng. curtus, court, 
 
 chien, French. canis. 
 
 sus, Lat. ) ( khuk, Pers. 
 
 if, Or. ) jhwch, Welsh. 
 
 SECTION IV. 
 
 Of the relations of the aspirate. — Of the substitution of the as- 
 pirate in several languages for S and for F. — Of the aspirate 
 as a guttural or hard palatine. 
 
 The state of Greek words beginning with the 
 aspirate, or with the digamma, has long been an 
 object of attention among grammarians. Some of 
 the facts connected with this subject are capable of 
 elucidation by a reference to the laws of the Celtic 
 language. 
 
 It was observed by Edward Lhuyd, that H is 
 never the first or proper initial of any word in the 
 Erse language, but that words beginning with F or 
 with S change that initial according to the laws of 
 permutation peculiar to this dialect of the Celtic 
 into H. Hence he infers with probability, that in 
 the primitive form of these words they began with 
 F or S, and that cognate words which begin with 
 H in other languages have lost their proper initial. 
 In like manner some Greek words now beginning 
 with an aspirate have lost an original digamma, 
 while others, as hra and ef, corresponding with 
 septem and sex in Latin, and with sapta and shash 
 in Sanskrit, have in all probability lost an initial S*. 
 
 a Lhuyd remarks with great probability, that such pheno- 
 mena indicate the former existence of a system of permutation 
 in other languages, similar to that which is still preserved in 
 the Celtic dialects. 
 
THE CELTIC NxVTIONS. 
 
 53 
 
 The following words, collected by Lhuyd, are 
 cognate in the Welsh and Erse languages. In the 
 Erse they begin with S in their primitive form, and 
 with H in a secondary form, or in regimine. In 
 Welsh they have only one beginning, with H. I 
 add a third column to shew the correspondences 
 presented by other languages, or merely to point out 
 the meaning. 
 
 ] 
 
 Irse. 
 
 Welsh. 
 
 Other languages 
 
 IMeaiiing. 
 
 saileog < 
 
 3r haileog 
 
 helig 
 
 salix L. 
 
 willow. 
 
 salan or 
 
 halan 
 
 halen 
 
 sal, a\s 
 
 salt. 
 
 sailte or 
 
 hailte 
 
 halht 
 
 salitus 
 
 salted. 
 
 saith or 
 
 haith 
 
 haid 
 
 
 swarm. 
 
 saith or 
 
 haith 
 
 hath 
 
 
 thrust. 
 
 sav or h 
 
 A 
 
 av 
 
 hav 
 
 
 summer. 
 
 savail 01 
 
 • havail 
 
 havail 
 
 similis 
 
 like. 
 
 skoiltea 
 
 
 holht 
 
 
 cleft. 
 
 se 
 
 
 e 
 
 
 he. 
 
 seavak 
 sealv 
 
 
 hebog 
 helva 
 
 
 hawk, 
 herd. 
 
 sealva 
 sealga 
 sealgaire 
 
 > 
 
 helu 
 hela 
 helliwr 
 
 
 possession. 
 
 hunting. 
 
 huntsman. 
 
 sean 
 
 
 hen 
 
 senex 
 
 old. 
 
 seasg 
 seile 
 
 
 hesg 
 haliu 
 
 sedge and 
 saliva. 
 
 hedge. 
 
 seol 
 si 
 
 
 huyl 
 hi 
 
 a sail, 
 sie, she. 
 
 
 sin 
 slth 
 
 
 hyn 
 
 hedh and ) 
 
 hedhwch j 
 
 
 this, 
 peace. 
 
 sil 
 
 
 hil 
 
 
 seed. 
 
 sir 
 
 soinean 
 
 
 hir 
 hinon 
 
 
 long. 
 J fair wea- 
 1 ther. 
 
 suan 
 
 
 hyn 
 
 somnus, vnvos, 
 
 
 
 E 3 
 
54 
 
 EASTERN ORIGIN OF 
 
 Sometimes the H in Welsh is lost, as in 
 segh ych ox 
 
 suas yuch super, virep. 
 
 silastar elestyr flag. 
 
 It would be easy to point out numerous instances 
 of a parallel description, in which words beginning 
 in Greek with the aspirate have in Latin and other 
 languages either S or the F — V b . The following 
 are examples chiefly from Vossius. 
 
 1. Aspirate substituted for S c . 
 sus. 
 
 «\ 
 
 <•/ 
 
 ayio$ 9 
 
 apnea, 
 
 Yj[xiav 9 
 
 ClpfAOS, 
 €p7TvXX0V y 
 
 OfxaKog, 
 
 ICTTCC, l(TTY)fJU, 
 €KUpO$, 
 
 serpo. 
 
 sal. 
 
 salis. 
 
 sacer. 
 
 sarpo. 
 
 satis. 
 
 se. 
 
 secus. 
 
 sedes. 
 
 semis. 
 
 sequi. 
 
 sermo. Scaliger. 
 f serpyllum. Servius in 
 t Eclog. 2. 
 
 sexus. Festus. 
 
 similis. 
 
 sisto. 
 
 socer. 
 
 b Grimm's Deutsche Grammatik, p. 583. 
 
 c Lacones, Argivi, Pamphylii et Eretrienses 2 eximere solent 
 atque aspirationem ejus loco sufficere ; ita fiovaa iis est paa : 
 fiovariKrj, fKOLKa ; nao-a, naa ; fiovcroa, /3ovoa : 7rot^<rat, noirjai, &C. 
 Voss. vid. Prise. L. V. Lhuyd, p. 30. 
 
THE CELTIC NATIONS. 55 
 
 (KVpa, 
 
 socrus. 
 
 o\og, bkov, 
 
 solus, 
 
 V7rap, 
 
 sopor. 
 
 ■jpafj 
 
 sorex. 
 
 i$poo$, and v$(op, 
 
 sudor. 
 
 cii, 
 
 sui. 
 
 6\ko$, 
 
 sulcus. 
 
 r \ 
 V7T0, 
 
 sub. 
 
 V7T€p, 
 
 super. 
 
 €7TTa, 
 
 septem. 
 
 
 sex. 
 
 vTrepfiiog, 
 
 superbus 
 
 V7TT10$, 
 
 supinus. 
 
 
 suus. 
 
 5a^, 
 
 sylva. 
 
 2. Instances of the rough aspirate substituted for 
 F or V d . 
 
 €G7T€pa 9 
 
 
 vespera. 
 
 e/Aw, 
 
 
 volvo. 
 
 f Ev€T0f, 
 
 
 Venetus, 
 
 'EXia, 
 
 
 f Velia. Serv. ad iEneid. 
 t 1. 359. 
 
 
 
 etTTtav, 
 
 
 festum. 
 
 Ofjukia, 
 
 
 familia. 
 
 apfxot, 
 
 
 ferme. Scaliger. 
 
 epfxa and ) 
 elpfxog, ) 
 
 
 firmus. 
 
 
 
 In other instances the Greek language seems even 
 
 to have lost the 
 
 spiritus asper, and pronounces such 
 
 words with the 
 
 gentle 
 
 \ aspiration e , as in the fol- 
 
 lowing. 
 
 
 
 d Chiefly from G. I. Vossius, ubi supra. 
 
 e Grimm, th. i. 
 
 P. 587. 
 
 E 4 
 
56 EASTERN ORIGIN OF 
 
 a\$o$, Dorice, 
 
 V 
 €7T0>, 
 
 > 
 
 pro aXvog, 
 
 saltus. 
 sequo, dico 
 si. 
 
 avev, 
 
 
 sine, 
 sino. 
 
 apio-Tepoc, sinister. 
 
 op(pea), v. po(f>€w, sorbeo. 
 
 In these instances the spiritus lenis stands, where 
 probably the spiritus asper once stood, for an ori- 
 ginal S. In the following, the digamma was ori- 
 ginally the initial letter : 
 
 cap, ver. 
 
 f videmus, Sansk. vidmus, 
 scimus. 
 
 a\6iry%, vulpes. 
 
 haXog, vitulus. 
 
 Perhaps we may trace the effect of a similar dis- 
 position to soften and obliterate the initial S in the 
 following words beginning with consonants. 
 
 yXa<f>u), SCalpo. 
 
 ypa<fiu>, scribo, 
 
 ykvcfxo, SCulpo. 
 
 6fA€V, 
 
 H representing a hard palatine or guttural conso- 
 nant in the Teutonic languages. 
 
 In the foregoing paragraph it has been shewn, 
 that the rough aspirate or H represents in several of 
 the Indo-European languages, a sibilant, or the 
 digamma or vau. Thus the Welsh, as well as the 
 Greek language, drops the S or the F entirely, and 
 substitutes the aspirate in words which originally 
 had either S or F for their initial, or which appear 
 
THE CELTIC NATIONS. 
 
 57 
 
 to have had one of them, as far as can be judged 
 from the cognate languages : while in the Erse the 
 aspirate is still used as a regular inflection of words 
 properly beginning, and yet often retaining the ori- 
 ginal S or F. I shall now shew, that in the Latin 
 as well as in the Teutonic languages, H is the sub- 
 stitute for, or is to be considered as, a radical hard 
 palatine or guttural. It stands for k, g, or ch. 
 
 In the following words H in Latin seems to be a 
 substitute for the Greek X f . 
 
 hiems, 
 
 
 X € ¥*»s- 
 
 halo, 
 
 
 yaXw. 
 
 hara, 
 
 
 X° r P°S' 
 
 heri, olim hesi, 
 
 %ecr/, unde %0€*\ 
 
 hio, hisco, 
 
 
 yaw, ya<JKW. 
 
 hir, 
 
 
 x e <>- 
 
 hirundo, 
 
 
 yzXithv, 
 
 hortus, 
 
 
 yoprog. 
 
 humi, 
 
 
 XafJ-ai. 
 
 humilis, 
 
 
 yapakog. 
 
 humor, 
 
 
 XVfW- 
 
 veho, 
 
 
 FoX$. 
 
 In the following 
 
 instances the Teutonic languages 
 
 substitute H 
 
 for 
 
 a palatine in Greek and Latin 
 
 words?. 
 
 
 
 claudus, 
 
 
 halts, halz, halt, (lame.) 
 
 Kavvaj3i$ 9 
 
 
 hanpr, hanaf, hemp. 
 
 caput, 
 
 
 haubith, houbith, haupt 
 
 Kaptiia, COr, 
 
 
 haerto, herza, heart. 
 
 f G. I. Vossius, ubi supra. 
 
 £ The list is taken from Dr. Jacob Grimm's Deutsche Gram- 
 matik. The Teutonic words are Mceso- Gothic, Old High Ger- 
 man, and English. 
 
58 
 
 EASTERN ORIGIN OF 
 
 kvqm, canis, 
 
 hunths, hund, hound. 
 
 K0l\0$, 
 
 hoi, hole, hollow. 
 
 celare, 
 
 hilan, heln. 
 
 KaXafxoc, calamus, 
 
 halam, halm, 
 
 Kaprog, Kaprepos, 
 
 hardus, hart. 
 
 cornu, 
 
 hatirn, horn. 
 
 collum, 
 
 hals. 
 
 KpVfAO?, 
 
 hrim, rhyme, (old Norse.) 
 
 KXateiv, 
 
 hlahan. (Goth.) 
 
 Kpa%6iV 9 
 
 hrakjan. (Goth.) 
 
 JifAeSTtyft 
 
 hleftus. (Goth.) 
 
 lux, (i. e. luks,) 
 
 liuhad, light, licht, 
 
 oikos, 
 
 veihs, (Goth.) house. 
 
 tacere, 
 
 thahan, dagen. 
 
 socer, 
 
 svaihra. (Goth.) 
 
 In Sanskrit we often find ^ — H, corresponding to 
 
 the r in 
 
 Greek words. 
 
 maha, 
 
 peya. 
 
 ahan, or ehon, 
 
 » / 
 eyccv. 
 
 SECTION V. 
 
 Of the interchange of dental and sibilant letters. 
 
 In a variety of languages, either for the sake of 
 euphony, or from caprice or accident, sibilant letters 
 have been interchanged with dentals. The conver- 
 sion of the Greek sigma into tau is familiar to all 
 classical readers. The use of the double tt instead 
 of the double a a- is said to have been introduced in 
 Athens by Pericles, but it probably preexisted as a 
 custom somewhere, otherwise it would have been 
 too great an innovation. It was probably a Boeo- 
 tian habit, for the Boeotians said avpiTTeiv instead of 
 
THE CELTIC NATIONS. 59 
 
 avpi&iv, or the iEolian avpiacreiv, and 07tXittco for oicklty. 
 The interchange of aa- and f for tt is a well known 
 dialectic variety in the Greek language. The single 
 t was also put for a by the iEolians and Dorians, as 
 (par), "| r <f>a(Ti. 
 
 > for J „ 
 
 €7T€T0V, €7T€(70V. 
 
 norei^av, J V. iroveiftav. 
 
 The late Attics adopted this custom, and said tyj- 
 fj.epov, iA€Tav\o$ 9 for aYjfxepov, ^eaavkog a . 
 
 In the Teutonic languages the frequent use of s 
 and z in the one class, and of t by the other, has 
 always been a characteristic distinction of the idioms 
 which belong to the High and Upper German divi- 
 sion, and of those allied to the Platt-Deutsch or 
 Lower German dialects. For water in English and 
 Holland-Dutch and Platt-Deutsch, the High-Dutch 
 has wasser ; for aut, aus ; for sweat, schweiss ; for 
 foot, fuss ; for sweet, suss ; for let, lass. This fact 
 is so well known, that it is superfluous to dwell 
 upon it. 
 
 SECTION VI. 
 
 Of the substitution of R for S. 
 
 The interchange of s and r is very frequent in 
 many Indo-European languages. 
 
 Among the Greeks it is said that the Lacedaemo- 
 nians substituted p for <r, and probably other Do- 
 rians had the same custom. For 7mro$ 9 novg 9 Qeog, nous, 
 they said "nnrop, nop, aiop, no'ip b . The interchange of 
 
 a Matthise's Grammar, ed. i829 ? . p. 34. 
 *> Ibid. p. 33. 
 
60 EASTERN ORIGIN OF 
 
 pa- for pp was much more frequent ; it is, at least, 
 more commonly seen in books, pa- being peculiar to 
 the Attic dialect. 
 
 We learn from Quintilian, Varro, and Festus, 
 that the Romans substituted r in a great many- 
 words for s, which had been more anciently used. 
 According to the last mentioned writer the ancients 
 wrote majosibus, meliosibus, lasibus, fesiis, for ma- 
 joribus, melioribus, laribus, and feriis. 
 
 It has been observed, that r is the most recent 
 form in all these instances, and s the most ancient c . 
 In the very oldest specimen of Latinity that is ex- 
 tant, and which has been ascribed to the age of Ro- 
 mulus, viz. a hymn of the Fratres Ar vales, engraved 
 on a stone which was discovered A. D. 218, are 
 found these words, " Enos Lases juvate," mean- 
 ing, in all probability, " Nos Lares juvate d ." It is 
 said, indeed, that the letter r was unknown to the 
 older Latins, who used s instead of it, till the time 
 of Appius Claudius Csecus, who introduced the r. 
 
 The following are examples of the substitution of 
 r for s, in which we can trace both forms in the 
 Latin language. 
 
 assus, 
 
 arsus. 
 
 robur, 
 
 robus, unde robu* 
 
 honor, 
 
 honos. 
 
 arbor, 
 
 arbos. 
 
 pignora, 
 
 pignosa. ^ 
 
 plurima, 
 
 plusima. > Festus, 
 
 holera, 
 
 helesa. J 
 
 c Grimm's Deutsche Grammatik. Bopp's Conjugationsystem 
 der Sanskritsprache. 
 
 d Lanzi Saggio di Ling. Etrusca, t. i. p. 142. Adelung's 
 Mithridates, th. ii. p. 460. 
 
THE CELTIC NATIONS. 
 
 M 
 
 ara, 
 
 asa. 
 
 
 arena, 
 
 asena. Varro. 
 
 carmen, 
 
 casmen. 
 
 feriae, 
 lares, 
 
 fesiae.) r . 
 . [-Varro 11 . 
 lases. ) 
 
 erit, 
 
 esit. 
 
 Anrelii, 
 
 Auselii. 
 
 Furii, 
 
 Fusii. Quintilian. 
 
 Papirii, 
 
 Papisii. 
 
 Valerii, 
 
 Valesiir 
 
 
 labor, 
 clamor, 
 
 labos. 
 clamos. 
 
 ^Quintilian 
 
 vapor, 
 
 vapos. J 
 
 
 The same change may be inferred to have taken 
 place in all words which take r in the increment in- 
 stead of s ; as, acus, pecus, foedus, pignus. 
 
 The Latins substituted r for s in other words 
 cognate with the Greek ; as 
 
 celer for *-eA^. 
 cruor — Kpvos. 
 
 The importance of this observation will hereafter 
 be apparent, when we come to trace the relationship 
 of Latin words with those of other languages. We 
 shall find r frequently substituted in the former for 
 an s or some equivalent in the Latin, and the re- 
 semblance is more decided between such words 
 when we restore the original s. Thus sororem, per- 
 haps originally sosorem, is almost identified with 
 the Sanskrit swasaram. The same change of letters 
 has an useful application to the inflections of verbs, 
 as we shall have occasion to observe. 
 
 e Vossius, ubi sup. 
 
62 EASTERN ORIGIN OF 
 
 SECTION VII. 
 
 Of the relation of different vowels and diphthongs to each other 
 in different languages. — Synoptical table of letters inter- 
 changeable between different languages. 
 The vowels are by no means to be discarded in 
 tracing the derivations of words and the relations of 
 languages, as some learned philologists have erro- 
 neously maintained. If any proof is necessary of 
 this remark, a very striking and sufficient one may 
 be found in Dr. Jacob Grimm's analysis of the Teu- 
 tonic verbs. 
 
 The first letter of the Sanskrit alphabet, 3f , or the 
 akara, which we generally represent by a, is a short 
 vowel, and, as it has been already observed by Pro- 
 fessor Bopp, corresponds in different instances with 
 nearly all the short vowels of the Greek and Latin 
 languages. It is easy to point out many examples 
 in which it occupies the place of the short e and of 
 the Greek epsilon, and the short 6 and omicron, as, 
 dashan, S. Ukcl 9 decern, 
 
 ashta, S. oW«, octo. 
 
 It corresponds also very frequently with the Latin 
 u before s or m in the terminations of words. The 
 endings of Sanskrit adjectives and nouns are fre- 
 quently in (3f :) or (3f^J) for the masculine, (3ff) 
 for the feminine, and (3f or 3f *T) for the neuter : 
 these are most correctly represented by ah or us or 
 o$ 9 a, and um or ov. Thus 
 
 shabiis, shuba, shabum, is 
 kol\o$, Ka\yj 9 KaXov. 
 
 Instances may be found in which 3f corresponds 
 with other short vowels, but they are not so fre- 
 quent ; as 
 
 agnis, ignis, fire. 
 
THE CELTIC NATIONS. 6ft 
 
 The Sanskrit 3fT or long a is most frequently 
 found to occupy the place of 6 or co in Latin and 
 Greek words ; as in 
 
 dadami, ^i^fxt. 
 
 The other Sanskrit vowels, i and u, long and 
 short, correspond with those nearly related to them 
 in sounds; viz. the long and short i or Iutx, and the 
 long and short u or tytXov. 
 
 OF DIPHTHONGS. 
 
 The semivowels 2f — ya, and ^T — va, or wa, and 
 the diphthongs T^ — ai, and T£ — oi, correspond with 
 the Greek and Latin vowels; thus, 
 
 2J — ya, and T^— ai, with ai Gr. and e Lat. 
 <T — wa, and"Q^ — oi, with a and 6. 
 Examples of the former kind occur in the ter- 
 minations of verbs in the middle and passive voices 
 in Sanskrit and Greek. They are thus : 
 in Sanskrit. in Greek, 
 
 ai, \kai. 
 
 sai, ai and aou. 
 
 tai, rat. 
 
 2J — ya, is used in Sanskrit where at occurs in 
 Greek and e in Latin, in very many cases. One 
 example occurs in the form of verbs having in San- 
 skrit a future signification, but used in Greek and 
 Latin with a preterite potential sense. Thus from 
 the root Bhu or Fuo, we have 
 
 Bhavishyami — fuissem \ related to ( o-atfxi. 
 
 yasi es >the Greeks at$. 
 
 yati et J form in \.ou. 
 
 ^T — va or wa, ^7 — va or wa. 
 swanum, S. sonum. L. 
 
 shwashurum, socerum. 
 
64 
 
 EASTERN ORIGIN OF 
 
 vacham, vocem. 
 
 vakshatai, av&Tou. 
 
 swasaram, sororem. 
 
 I shall conclude the foregoing remarks on substi- 
 tutions or interchanges of letters in different lan- 
 guages by an attempt at a table of correspondences. 
 
 JO 
 
 '2 
 
 
 
 
 
 
 CO 
 
 
 
 i 
 
 Cm 
 
 «+H 
 
 «m 
 
 42 
 
 4* 
 
 43 
 
 43 cc 
 
 co 43 
 
 42 
 
 o3 
 
 U 
 
 42 
 42 
 
 
 
 
 
 42 43 
 
 X 43 
 
 
 o 
 
 8 
 
 o 
 
 
 
 42 
 
 8 8 
 
 8 8 
 
 
 J* 
 
 4= 
 
 4* 
 
 
 
 U 
 
 CO «+h 
 
 co cm 
 
 
 
 
 
 
 
 
 ,__ 
 
 
 
 
 
 
 
 
 
 e$ 
 
 
 
 
 
 
 
 
 
 m 
 
 
 4 
 1 
 
 a* 
 
 > 
 
 8 
 
 42 
 
 PL 
 
 &U0 
 
 &J0 
 
 6JD 
 
 3 
 
 bo 
 
 ■a x 
 
 J- 
 O 
 
 * 
 J3 bJO 
 
 
 o 
 
 
 
 
 
 
 
 
 c 
 
 42 . 
 
 
 
 
 
 
 
 
 I 
 
 3 a ro 
 
 
 3 
 
 
 
 h 
 
 
 
 
 o*a g 
 
 
 cr 
 
 
 
 8 
 
 O co 
 
 y C+_| 
 
 O Cm 
 
 
 uAO 
 
 6M 
 
 cJ" 
 
 o M 
 
 &J0 
 
 c/T >" 
 
 co" > 
 
 
 
 
 
 
 
 
 ^ 
 
 b ^ 
 
 •% 
 
 
 
 
 t- 
 
 
 
 . 
 
 . 
 
 05 
 
 
 
 
 t» 
 
 
 
 
 
 1 
 
 
 <£L 
 
 
 «\ \/> 
 
 
 X^, 
 
 X 13 
 
 O O 
 
 ^ 
 
 «v 
 
 
 b „ 
 
 
 
 - '-N 
 
 r- s^-v 
 
 
 t? 
 
 -9- <£> 
 
 i- 
 
 b <u 
 
 ^ 
 
 5C *T 
 
 ST w 
 
 X^X^ 
 
 
 
 
 
 
 
 
 43 
 
 
 .2 
 
 9* 
 
 
 42 
 
 
 T 
 
 i 
 
 CO 
 
 1 
 
 
 <J2 
 
 1 
 
 pq 
 
 1 
 
 
 i 
 
 1 
 
 B 
 
 
 •1 
 
 I 
 
 if 
 
 J* 
 
 42 
 4* 
 
 1 
 
 43 
 u 
 
 1 
 
 T 
 
 ce 
 
 1 
 
 ■a 
 
 41 
 
 1 
 
 X) 
 
 ^« 
 
 
 A 43 
 
 _!2 
 
 
 
 42 
 
 4= 
 
 
 'A 
 
 9^ 
 
 5J 
 
 o 
 
 
 "P 
 
 CO 
 
 CO 
 
 CO 
 
 •^ 
 
 
 1 
 
 [T3 ^d 
 
 
 1 ^ 
 
 1 ^ 
 
 1 
 
 
 J- 
 
 tJ- 
 
 
 P9 
 
 E?" 
 
 feT~ 
 
 £ 
 
THE CELTIC NATIONS 65 
 
 CHAPTER III. 
 
 Proofs of common origin in the vocabulary of the Celtic and 
 other Indo-European languages. 
 
 Section i. Names of persons and relations. 
 
 IT is now time to examine how far the Celtic and 
 other Indo-European languages are related to each 
 other in their vocabularies, or by the possession of 
 a considerable number of common roots or primitive 
 words. From the comparison to be instituted for 
 this purpose, it is obviously requisite to exclude all 
 such words as from their nature appear likely to 
 have been introduced at a late period by foreign 
 commerce, by conquest, or with the adoption of a 
 new religion or system of manners. I must confine 
 my observations to the original materials of speech, 
 and to expressions which denote simple and primi- 
 tive ideas. 
 
 On entering on this part of my inquiry, I shall 
 take some of the groupes of words collected in the 
 Amera Cosha, or Sanskrit Vocabulary of Amera 
 Sinha, and try whether the corresponding terms in 
 the Celtic dialects have any resemblance to them a . 
 In general, I shall place the Sanskrit words first, 
 and then the Celtic, subjoining any terms which ap- 
 pear to be of cognate origin in the other European 
 languages. 
 
 a I do not confine myself to the particular vocables given in 
 the Cosha, when other genuine words can be found which are 
 more to my purpose, nor do I think it necessary to follow the 
 exact order of arrangement observed by the author of that vo- 
 cabulary. 
 
66 
 
 EASTERN ORIGIN' OF 
 
 ]Vr^a,-..'.-^w, 
 
 ' 
 
 I. Words denoting persons and family or other 
 relations b . 
 
 3TPT — jani, (a woman.) 
 
 Celtic, gean, Erse. (Lhuyd.) SK^O 
 
 Russian, Jena ; Gr. Tvvyj ; Pers. Zen, Zenne. 
 
 ^ H I — Varna, (a woman.) (Am. Cosh.) 
 
 Celtic, fem, Erse. (Lh.) 
 
 ^ H«lT — vamani, (a woman 
 
 Celtic, femen, Erse. (Lh.) 
 
 Latin, foemina. 
 
 ^ MrlT — vanita, (a woman.) 
 
 Celtic, Bean and Bhean or vean, Erse ; benw and 
 
 benyw, Welsh. In regimen venw and ve- 
 
 nyw. 
 Greek, Byva et Bdva, Boeotice vel Dorice, est mu- 
 
 lier vel filia (Salmasius, p. 402. de Hellenis- 
 
 tica) Bccvyreg, (Boeot. pro ywaTices,) mulieres. 
 
 ^TX! — virah, a hero, warrior ; vir. 
 Celtic, fear, Erse c ; man. 
 
 Gwr and WE, Welsh, man; pi. gwyr and 
 
 wyr; viri. 
 Hence, gwraig and wraig, a woman. Com- 
 pare wr and wraig with vir and vi- 
 rago or virgo. 
 
 N. B. Frag, Erse for gwraig; Germ. frau. Com- 
 pare the Greek Ftjpcos, heros, with fear. 
 
 •1 < I — narah, man, also Lord, applied to the Deity. 
 Celtic, Ner, lord. 
 Greek, awjp. 
 
 3. There are three words in the Amera Cosha 
 corresponding with the word father. These are 
 
 b Amera Cosha, book ii. chap. vi. sect. i. 
 
 
THE CELTIC NATIONS. 67 
 
 r| |rl, ^5l«ich: and IMriT. One of these, janakah, is 
 derived from a root allied to the etymon of genitor 
 in Latin, which will be compared with its cognates 
 hereafter. The other two may be traced as follows: 
 
 r\ |c| : — tatah, plural tatah. 
 
 Celtic, tad, plur. tadan, Welsh; tat, Armoric; 
 taz, Corn. ; taid, a grandfather, Welsh. 
 Cognates : ata, Moeso-Goth. ; aita, Cantab. 
 \ HrJ — pitre, nom. pita, ace. pitaram. 
 Cognates : Persian, pader. 
 Gr. & Lat. Traryp, Trarepx, pater, patrem. 
 Teut. dial, vater, faeder, father. 
 Celtic, athair d , Erse. 
 
 H |rj — matre, nom. mata, ace. mataram. 
 
 Persian, mader ; Russian, mater. 
 
 Celtic, mathair, Erse. 
 
 Gr. & Lat. pvjTvjp, mater. 
 
 Teutonic, meder, mutter, mother, &c. 
 
 ^TJrJ — bhratre, nom. ^j |r1 1, bhrata. 
 
 The nearest cognates are, 
 Russian, brat'. 
 Celtic, brawd, (Welsh, in plur. brodyr ; compare 
 
 the Sanskrit plural bhratarah.) 
 Persian, braudur. 
 
 Teutonic, brothar, (Goth.) bruder, brother, &c. 
 Celtic, brathair, Erse. 
 Latin, frater. Compare fyaTccp. 
 
 d That the word athair is really of cognate origin with pater, 
 and other European words of the same stock, appears probable 
 from a comparison of the following : 
 
 athair pater pitre. S. 
 
 mathair mater matre. S. 
 
 brathair frater bhratre. S. 
 
 F 2 
 

 68 EASTERN ORIGIN OF 
 
 £^1J — swasre, a noun of the same form, making 
 in the accusative swasaram, sororem. 
 This word has two sets of cognates; one, in which 
 the sibilants are preserved as such ; the other in 
 which they are converted into gutturals, according 
 to the mode of permutation illustrated in chap. ii. 
 sect. 3. 
 
 Latin, soror, ace. sororem, probably sosor, soso- 
 rem in an older form e , is nearly allied 
 to swasaram. 
 Some languages alter the middle s by adding t ; as, 
 Russian, sestra; Teutonic dialects, suistar (Moeso- 
 Gothic), sch wester, sister. 
 Others change s for h, and finally omit it, as 
 Celtic, Siur, Erse. 
 
 Secondly, 
 Persic, Khauher. 
 
 Celtic, Chwaer, (i. e. khwaher) Welsh ; chuar, 
 Armor.; hor, huyr, Corn. 
 The Greek has no similar word. 
 
 <5T^rJ — duhitre, nom. duhita, ace. duhitaram, 
 
 (daughter.) 
 Greek, 6vyarY]p f Qvydrepa. 
 
 Pers. dokhter ; Goth, dauhtar ; Germ, tochter. 
 
 Celtic, Dear, (dehar?) Erse. 
 
 The Latin has no cognate word. 
 W^<t — swasurah, a father-in-law; accus. swasu- 
 
 rum. 
 Latin, socer, socerum. 
 
 Russian, svekor'; Goth, svaihra; Germ, sch waeher. 
 Celtic, (converting sibilants as usual into guttu- 
 rals,) Chwegrwn, Welsh; Huigeren, Corn. 
 e Chap. ii. sect. 6. 
 
THE CELTIC NATIONS. ft 
 
 ^PJ; — swasruh or shwashrus, a mother-in-law. 
 
 Latin, socrus. 
 
 Celtic, Chwegyr, pi. chwegrau, Welsh; hueger, Corn. 
 
 «1^, — naptre, a grandson, and naptri, a grand- 
 daughter. 
 
 Latin, nepos and neptis, meaning also in a later 
 sense, nephew and niece. 
 
 Celtic, Nai, Welsh ; noi, Corn. ; ni, Arm. for ne- 
 phew; and Nith, Welsh; noith, Corn.; 
 nises, Arm. for niece. 
 
 N. B. pt is mutable into th in Welsh, according to 
 an observation in chap. i. sect. 2. 
 
 <5^T, — daivre, a brother-in-law, or husband's or 
 wife's brother. 
 This word has no analogue, as far as I know, in 
 any of the Celtic or Teutonic dialects. 
 Greek, tiarjp ; Russian, diever' ; Latin, levir. 
 
 ^XK — vangshah, offspring. 
 
 Persic, pachah, (a child,) pi. pachegaun. 
 
 Celtic, bachgen, Welsh. 
 
 The preceding words are the principal part of 
 those by which family relations and the distinctions 
 of persons are signified in the Indo-European lan- 
 guages. The greater part of them appear evidently 
 to exist in the Celtic dialects under a peculiar form, 
 which is the guarantee of their genuineness. That 
 the Celtic words are really cognates of the Persian, 
 Russian, Greek, Latin, German, and Sanskrit words 
 no doubt can be entertained. 
 
 f 3 
 
70 
 
 EASTERN ORIGIN OF 
 
 SECTION II. 
 
 Names of the principal elements of nature, and of the visible 
 objects of the universe \ 
 
 Light, Flame. 
 
 ZsTIW — Jwala, also Jwalah. 
 
 Celtic, Gwawl, goleu, Welsh; golou, Arm. 
 The Sun. 
 
 ^T^I — hailih, or hailis. 
 
 Celtic, Haul, pronounced hail, Welsh; houl, 
 heul, Corn. ; heol, Arm. 
 
 Greek, 'AeXiog, i. e. haelios ; r/ EA>y, solar heat. 
 
 Note. The spiritus asper being changed for S in 
 conformity with an observation in sect 4. of chap. 2. 
 we have Saul for haul. Compare sol, Lat. ; in Russ. 
 solnste. 
 
 Another Sanskrit word for sun is *J*J I sunuh. 
 Comp. Sunno, M. Goth, and Germ, sonne, sun. 
 The Moon. 
 
 cRS^T — klaida ; also klaidu. 
 Celtic, lheuad, or lhhyad, (pronounce nearly as 
 chleuad,) Welsh. Also 
 glauh, which by Sandhi becomes 
 
 glaur. 
 
 Celtic, lhoer (chloer), Welsh; laor, Arm.; lar, 
 Corn. Compare luan, Erse, with lima, 
 Russ. and lima, Latin. 
 
 5k- 
 
 Star. 
 rTRJ- 
 Greek, 
 
 tara. 
 
 reipeov ; Welsh, seren ; (pi. reipea, Welsh 
 ser and syr.) Armoric, steren. ; Germ, 
 stern. Goth, stairno ; Pers. sitauren ; 
 
 a Amera Cosh, book i. sect, i and 2. 
 
THE CELTIC NATIONS. 71 
 
 Latin, stella (perhaps a diminutive form 
 of stera, as tenella from tenera). Com- 
 pare also aarepa, and astra. 
 
 «TH: — nabhah, sky, atmosphere, aether, heaven. 
 
 Celtic, Nev. Welsh and Corn. ; neav, Erse. 
 
 Russ. nebo. 
 
 ^T^TT — nabha, a cloud, rain : nabholaya, smoke. 
 
 Celtic, Nivwl, (mist, cloud,) Welsh; Neal, Erse. 
 
 Germ, nebel; Gr. vecpekvj, ve<pa$; Lat. nebula, nubes. 
 
 33^T — udum, water; whence 3^ — unda, to 
 wet or moisten. 
 
 Russ. voda, water; Pol. woda. 
 
 Latin, udus, uda, udum, and unda. 
 
 Goth, wato ; A. S. waeter, water. Compare wet, 
 weather. 
 
 Greek, vlcop, i. e. Fvlup, or vudor. 
 
 Celtic, dwvyr, pi. duvrau. . 
 
 ^~_ dyu. (a day.) 
 
 Celtic, Di and dia, Erse ; dydh, Welsh, 
 Latin, dies. 
 
 Goth, dags ; A. S. daeg, day. 
 [•1X11 — msa, night. 
 
 Celtic, Nos, Welsh. Notch', Russian. 
 More remote are the following : 
 
 Nochcl, Erse ; nahts, Goth. ; nacht, night, Germ. ; 
 
 vvf, nox. 
 *TmC' — mirah, ocean, sea. — > ■ 
 Celtic, M or, Welsh; muir, Erse. More, Russian ; 
 
 meer, mere, Germ. D. ; mare, Lat. 
 t|*TT — dhara, earth. 
 Celtic, Daiar, Welsh, (in regim. dhaiar.) 
 
 there is another Welsh word, tir. Compare 
 
 terra, Lat. ; airtha, Goth. ; talamh and tellur, 
 
 Erse, with tellurem, Lat. 
 F 4 
 
T2 EASTERN ORIGIN OF 
 
 <H\y— 
 
 Celtic, 
 Greek, 
 Celtic, 
 
 aashtra, ether, air. 
 athair, Erse. 
 
 aMUvjp — aidpia, avjp. 
 awyr, Welsh ; aer, Lat. 
 
 Latin, 
 Moeso-Goth 
 
 55- 
 
 agnis, or agnih, (fire.) 
 Ignis ; Welsh, tan, i. e. taan. 
 . fon. 
 druh and drus, (a tree ;) derucht, Pers. ; 
 
 Apve, Gr. ; 
 
 Derw, Welsh; dair, Erse (an oak tree.) 
 
 im 
 
 SECTION III. 
 
 Names of animals. 
 Of the terms for different species of animals, it 
 appears that few, comparatively, are common to the 
 Sanskrit and the European languages. Nor is this 
 circumstance difficult of explanation : emigrating 
 tribes in seeking a new climate, and leaving behind 
 them a great part of the stock of animals for which 
 they had previously names, are obviously under the 
 necessity of inventing other significant appellations 
 for those peculiar to their new country. In this re- 
 spect the Celtic dialects are under the same circum- 
 stances as the other European languages ; and it may 
 be clearly shewn that they partake of a common 
 stock of terms with these languages ; for though the 
 European idioms differ from the Sanskrit, they have 
 a common stock of such terms among themselves. 
 There are, however, some instances of agreement 
 with the Sanskrit, and this remark includes nearly 
 all the domestic animals. In all the following in- 
 stances the Celtic terms are cognate with those be- 
 longing to the other European languages, and in some 
 they bear a remarkable resemblance to the Sanskrit. 
 
THE CELTIC NATIONS. 73 
 
 The interchange of sibilant with guttural conso- 
 nants is here to be observed, as in the instances before 
 cited. Refer to numbers 1, 2, and 3. 
 
 Dog. 
 
 Sanskrit, Shunah and shun! ; shiini, bitch. 
 Celtic, Ki, pi. cwn, Welsh ; chana, Erse, 
 Greek, kvuv, pi. Kvveg ; Lat. canis ; Goth, hunths, 
 
 hound. 
 Hog and Sow. 
 
 Sanskrit, Shukarah, (hog;) Pers. khuk, (hog or 
 
 sow.) 
 Celtic, Hwch, Welsh, (sow.) 
 Greek, r T$- ; Lat. sus. 
 
 Horse. 
 
 Sanskrit, ashwah or eshuus. 
 
 Latin, (changing sibilants into gutturals,) equus. 
 
 Greek, wires, Mo\. 7kko$.(?) 
 
 Celtic, Each, Erse ; asb, Pers. Also, 
 
 Greek, Ka/3d\kYjg; Lat. caballus. 
 
 Celtic, keffyl, Welsh ; caual, Arm. ; capul, Erse. 
 Also, Pers. fars; Germ, ross; Eng. horse. 
 Ass. 
 
 Greek, ovog ; Lat. asinus. 
 
 Celtic, Asyn, Welsh ; asal, Erse. 
 Goat. 
 
 Sanskrit, ajah and ch'haga ; Gr. alya. 
 
 Latin, caper. 
 
 Celtic, gavar, Welsh ; gobhar, Erse. 
 Ram. 
 
 Sanskrit, Uranah. 
 
 Celtic, Hwrdh, Welsh; urdh, Arm.; hor and 
 hordh, Corn. 
 
 Latin, aries. 
 
74 EASTERN ORIGIN OF 
 
 Oxen. 
 
 Sansk. Ukshan or uxan, ox or bull. 
 
 Celtic, Ych, Welsh; agh, Erse; ochs, Germ. Also, 
 
 Greek, j3ovg ; Lat. bos, boves. 
 
 Celtic, Buw, Welsh ; bo, Erse. Also, 
 
 Latin, bucula ; Welsh, buwch. 
 
 Bull. 
 
 Greek, ravpog ; Lat. taurus. 
 
 Celtic, tarw, Welsh ; tarbh, Erse. (Compare Tor, 
 Chaldee.) 
 
 Cow. 
 
 Sansk. Go ; Germ, kuh, cow. 
 
 Fish. 
 
 Greek, lyfivg (olim FiyQvg?) 
 Latin, piscis. 
 
 Celtic, Pysg, Welsh ; jasg, Erse. 
 Germ, fisch, fish. 
 
 Swan. 
 
 Latin, olor. 
 
 Celtic, alarch, Welsh ; eala, Erse. 
 
 Pigeon. 
 
 Latin, columba ; golub', Russian. 
 Celtic, colommen, Welsh ; cwlm, Arm. ; colm, 
 columan, Erse. 
 
 Frog. 
 
 Latin, rana. 
 
 Celtic, kranag, Corn. ; ran, Arm. 
 
 Fawn. 
 
 Greek, e\a<po$. 
 . Latin, hinnulus. 
 
 Celtic, elain, Welsh. 
 
THE CELTIC NATIONS. 
 
 75 
 
 Lamb. 
 
 Greek, apog ; Latin, agnus. 
 
 Celtic, oen, W. ; oan, Corn, and Arm. ; an, Erse 
 Crane. 
 
 Greek, yepavos ; Germ, krannich ; Eng. heron. 
 
 Celtic, Garan, W.; and kryr, Erse, korr. 
 Cuckow. 
 
 Greek, kokkv%, Lat. cuculns. 
 
 Celtic, Cog, W. ; chuach, Erse. 
 
 
 PARAG. 2. 
 
 The following list of terms, chiefly for inanimate 
 objects, displays nearly the same degree of affinity as 
 the preceding. 
 
 Welsh. 
 
 Erse. 
 
 Greek. 
 
 Other languages. 
 
 
 
 
 
 Jdruh, Sansk. 
 
 derw 
 
 dair 
 
 
 bpvs 
 
 ( (a tree) 
 
 faw, fawydh 
 
 faidhbhile 
 
 (prjyos 
 
 fagus, beech 
 
 colhen, colh 
 
 
 
 Kopvkos 
 
 cornel tree 
 
 Ihyren and ) 
 Ihyriaid j 
 
 
 
 keipiov 
 
 lilium, lily 
 
 lhech 
 
 leac, 1 
 
 liag 
 
 kCOos 
 
 
 lhaeth 
 
 laith 
 
 
 yakCLKTOL 
 
 lac, lactis 
 
 aradyr 
 
 
 
 aporpov 
 
 aratrum 
 
 cwyr 
 
 ceir or keir 
 
 
 cera 
 
 braich 
 
 
 
 fipayi(t>v 
 
 brachium 
 
 lhwch 
 
 loch 
 
 
 
 lacus 
 
 lhu and ) 
 
 Ihuaws j 
 
 
 
 
 
 
 
 kabs 
 
 lludi, Russ. 
 
 kentar (a ] 
 
 
 
 i 
 
 
 nail) (Lh.) f 
 
 
 
 KtVTpOV 
 
 ( nomen 
 
 enw 
 
 ainm 
 
 
 6vo\xa 
 
 (naman S. 
 
 m g 
 
 
 
 aloka, Sansk. 
 
 lux 
 
 medh (mead) 
 
 
 
 \xiQv (wine) 
 
 
 mel 
 
 mil 
 
 
 fiikt 
 
 mel 
 
 
 
76 
 
 EASTERN ORIGIN OF 
 
 Welsh. 
 
 Erst'. 
 
 Greek. 
 
 Other langiia^cs. 
 
 melin 
 
 meile 
 
 jJLVkoS 
 
 mola 
 
 dant 
 
 
 ohovra 
 
 r dent-em 
 
 1 dantah, Sans. 
 
 awr 
 
 uair 
 
 copa 
 
 hora 
 
 aur 
 
 or 
 
 
 aurum 
 
 corn 
 
 
 Ktpas 
 
 cornu 
 
 co ron 
 
 choroin 
 
 
 corona 
 
 cybhigl 
 
 chuvachail 
 
 
 cubiculum 
 
 rhyn (point) 
 
 sron 
 
 piv 
 
 
 ysgraff J 
 scaff Arm. j 
 
 sgaffa 
 
 aKCL($)7) 
 
 scapha 
 
 wr, wyr 
 
 fear 
 
 
 vir, viri 
 
 wraig 
 
 frag 
 
 
 virago 
 
 gwrach 
 
 
 ypavs 
 
 
 gwin, win 
 
 fin 
 
 Folvos 
 
 vinum 
 
 arriant 
 
 airgidh 
 
 apyvpiov 
 
 argentum 
 
 tervyn 
 
 teor 
 
 rip\ia 
 
 terminus 
 
 einion 
 
 ineoin 
 
 CLK[JL(t)V 
 
 incus 
 
 pen 
 
 cean 
 
 K€(f>akr) 
 
 caput 
 
 colovn 
 
 colbh 
 
 
 column us 
 
 swn 
 
 soin 
 
 TOVOS 
 
 sonus 
 
 byw 
 
 bio 
 
 (3Cos 
 
 vita 
 
 halen 
 
 salen 
 
 h\s 
 
 sal 
 
 cader 
 
 chathair 
 
 KaOihpa 
 
 chair (Eng.) 
 
 sowdwl 
 
 sael 
 
 K€\r] 
 
 ( calcaneus 
 1 heel. Eng. 
 
 croen 
 
 croiccionn 
 
 Xpm 
 
 
 erw 
 
 
 apovpa 
 
 arvum 
 
 dor (W. and | 
 
 
 
 f thur (Germ.) 
 
 Armor. ) 
 
 dorus 
 
 Ovpa 
 
 1 dwar (Sansk.) 
 
 parth 
 
 pairt 
 
 
 parte 
 
 creuan 
 
 
 Kaprjvov 
 
 
 mynydh 
 
 
 
 ( monte 
 1 mountain 
 
 fynnon 
 
 
 
 fonte 
 
 avon 
 
 amhain 
 
 
 arane 
 
 tir 
 
 tir 
 
 
 terra 
 
TIIK CELTIC NATIONS. 
 
 77 
 
 Welsh. 
 
 Erse. 
 
 Greek. 
 
 Other languages 
 
 
 talamh 
 
 
 tellus 
 
 mor 
 
 muir 
 
 
 J mare, meer, 
 (Germ. 
 
 cylha 
 
 
 no i\(a 
 
 
 cylch 
 
 
 KtpKOS 
 
 ( circulus, 
 | circus 
 
 deigryn 
 
 
 haKpVOV 
 
 lachryma 
 
 eigion 
 
 
 J)K€avbv 
 
 oceanum 
 
 hedhwch, pi. ] 
 
 
 
 
 hedhychau > 
 
 
 i](rvyj.a 
 
 
 peace, quiet J 
 
 
 
 
 hwyliau 
 
 
 
 velae 
 
 meidr 
 
 
 ll€TpGV 
 
 
 gwvr and ^ 
 
 
 
 fverum and 
 
 wvr ) 
 
 
 
 ( Veritas 
 
 cariad 
 
 
 Xapt? 
 
 caritas 
 
 cawr, a giant 
 
 
 yavpos 
 
 jghorah S. 
 
 J hrvitvik a 
 
 righ, a king 
 
 | raj a Sansk. 
 j regem Lat. 
 
 SECTION IV. 
 
 Verbal roots traced in the Celtic and other Indo-European 
 languages. 
 
 3T«T, Jan, a verbal root, whence the verbs 3T3Tf^T, 
 jajanti, gignit, and jayatai meaning yiyve- 
 rai, gignitur, he is born ; middle voice, jajana, 
 yeyova. Hence the following nouns in Sanskrit. 
 
 Janah, a man. 
 
 Jani, a mother. 
 
 Janus, birth. 
 
 Janitre, a father. 
 
 Janima, birth, procreation. 
 
78 EASTERN ORIGIN OF 
 
 In Greek, yevvdoo, yivofxai, yevv^fxa, ytvog, k. t. A. 
 Latin, gigno, genius, genero, genitor, &c. 
 Celtic, 
 
 1. Welsh, geni, to be born. 
 
 genedig, brought forth, or born, 
 genedigaeth, birth, nativity, 
 geneth, a girl, 
 genid, birth, 
 genilh, progeny. 
 
 2. Erse, gein, offspring. 
 
 geinim, to beget. 
 
 geineighim, to bring forth, and 
 geintear, gignitur. 
 
 5J" — mre, a verbal root, whence the verb mriyatai, 
 moritur, and the causal verb marayami ; mre- 
 tus, mortuus ; marah, mors. 
 
 The Greek language wants this word, unless the 
 termination [xopog be derivable from it, though attri- 
 buted to another verb. 
 
 Latin, mori, mortuus, &c. 
 
 Celtic, 
 
 1. Welsh, marw, to die. 
 
 marw, and marwawl, adj. dead, 
 and deadly. 
 
 2. Erse, marbh, i. e. marv, dead. 
 
 meath, death ; and meatham, to die. 
 Compare Heb. iltt, meth, dead. 
 Sclavonic, Russian. 
 
 umirat', to die ; umertii, dead, 
 mor', mortality. 
 
 3TT^ — Jiv, a root whence the verb 
 
 jivami, I live, or jivati, vivit. 
 jiva, life, (Am. Cosh.) 
 
THE CELTIC NATIONS. 79 
 
 In Latin, vivo, vita. 
 Greek, /3/W, piow. 
 
 Celtic, 
 
 1. Welsh, BYW, or vyw, verb, to live. 
 
 byw, adj. vivus 
 
 bywyd, vita. 
 
 by wau, to vivify. 
 
 beua, (in Cornish,) to live. Lhuyd. 
 
 2. Erse, beo, to live. 
 
 beatha, life, vita. 
 
 5TT — Jna, a verbal root, whence the verb janami, I 
 know ; janati, he knoweth ; jnatus, 
 jnata, jnatum, adj. notus, nota, notum. 
 Greek, yvocc, and yvSfxt, yivwaKOo, yvoofxy], k. t. A. 
 Latin, nosco, i. e. gnoo, gnotus, &c. 
 German, &c. kennen, know, &c. 
 Welsh, Gwn, I know. 
 
 R|<v — Vid. a verbal root, whence the verb vaida, 
 (oTSa,) in a preterite form, with a present signi- 
 fication, I know, he knows; vaiditum, to know; 
 also 
 
 vidan, wise. 
 
 vidanti, vaida, wisdom. 
 Greek, e/Sew, ?&«, oTba. ol. FctMw, or Vei^eco. 
 
 Latin, video. 
 
 vide, to know; vidende, knowledge, 
 
 Danish, 
 Teutonic, <{ 
 
 weise, wissen, German. 
 
 wit, wot, wise, English. 
 
 Celtic, 
 
 1 Welsh, gwydh and wydh, knowledge. 
 
 gwydhad and wydhad, to learn. 
 
80 EASTERN ORIGIN OF 
 
 gwydhai, gwydhawl, wise, 
 gwybod, (irreg. verb,) whence, 
 gwydhost and wydhost, knowest. 
 
 2. Erse, fis, or fios, knowledge. (Lh.) 
 
 fisc, a seer ; fiosaighim, to know. 
 
 Here the roots are vid, S; /& or €/&; vid. Lat.; vid, 
 wit, Teutonic ; wydh, or gwydh, Welsh. 
 
 ^\f, Budh, a root, whence the verb bodhati, he 
 t knows or understands, 
 budhah, a sage. 
 Celtic, 
 
 Erse, fodh, knowledge, 
 fodhach, wise. 
 
 N. B. The Sanskrit root, buda, to know, or discern, has an 
 equally striking affinity with the Erse, fod, art or skill. 
 
 3£, Shru, a verbal root signifying to hear. 
 Infinitive mood, shrotum. 
 shrutah, heard, adj. and part. 
 In Sclavonic, changing I for r. 
 
 Russian, sluch, hearing. 
 slutat', to hear. 
 In Greek, sibilants changed to gutturals, (ch. 2. 
 sect. 3.) 
 
 kXvco, to hear. 
 k\vto$ 9 adj. 
 In Celtic, 
 
 Welsh, clyw, hearing. 
 
 clust, an ear. 
 Erse, cluinam, I hear. 
 
 clu, hearing, fame. 
 
 cluas, an ear. 
 
 clotha, he heard. — Lh. 
 
THE CELTIC NATIONS. 81 
 
 Here the roots in all the above languages are 
 shru, slu, kXv, clu. 
 
 ^r^[ — loch, a verbal root, to which are referred 
 the two following verbs : 
 
 1. lochatai, he sees ; whence 
 
 lochan, an eye. 
 Greek, AezWe*, he looketh. 
 English, looketh. 
 Celtic, lhygad, W. an eye. 
 
 2. Lochayati, lucet. 
 Latin, lux, luceo. 
 Teutonic, licht, light, &c. 
 
 Celtic, Welsh, f . 
 
 lhwg, light ; lhuched, lightning, 
 lhewychu, to light, 
 lhygu, to brighten, &c. 
 
 ITSCT — Dresh, a verbal root, signifying 
 to see. 
 Pret. dadarsha, I saw. 
 Greek, Sep/too. 
 
 Se^opKa. 
 Celtic, Erse, 
 
 dearc, a verbal root, signifying sight, 
 
 seeing, also an eye. 
 dearcam, to see. 
 dearcadh, seeing, sight. 
 Welsh, 
 drem, sight, &c. 
 Here the roots are dresh, or rather dursh, $epK, 
 and Celtic, dearc. 
 
 K*)^ — lih, a verbal root, signifying to lick, 
 1. person, proper form lihai, lingor. 
 
 / G 
 
 1 
 
82 EASTERN ORIGIN OF 
 
 Greek, Ae/%ey. 
 
 Latin, lingo. 
 
 Goth, laigwan ; A. S. liccan. 
 
 Eng. lick. 
 
 Celtic, Welsh, 
 
 lhyaw, lhyvu, to lick. 
 
 llyviad, licking. 
 Erse, lighim. 
 
 do leigh se, he licked. 
 
 Wf — sht'ha, in inflection t^|, sth'a, a verbal root, 
 whence the verb tishtati, he stands, 
 tishtami, I stand. 
 Greek, lo-rapi or 7<jt^/, i. e. aiarafxi, and the obso- 
 lete araco. 
 Latin, sto, stare, status, sisto, &c. 
 Teut. standan, (Goth.) stehen, stay, stand. 
 Celtic, Erse, 
 
 sta, stand, 
 
 stadam, to stand, 
 
 do stad se, he stood. 
 Welsh, eistedh, to sit. 
 
 , - eistydh, sitting. 
 
 ^58^ — rich'h, a verbal root, whence the verb rich- 
 's 
 
 ch'hati, he moves towards, reacheth, Rich- 
 ch'hami. 
 Greek, opeyeTat, ope%i$. 
 Latin, porrigit. 
 Teut. rakyan, (Goth.) reichen, erreicht, (Germ.) 
 
 reacheth, (Eng.) 
 Celtic, high, a root in Erse ; whence righim, 
 I reach; richeadh, to stretch, reach, &c. 
 
 3f "^ — as, a verbal root, whence the verb substantive, 
 
THE CELTIC NATIONS. 83 
 
 asmi, asi, ASTi, sum, es, est; syam, sim, or 
 siem. 
 Pers. est, he is ; hesten, to be. 
 Greek, e/^/, (ol. eo-fxt,) eavi, cot*. 
 Latin, esum, es, est. 
 Teut. ist, is, &c. 
 Celtic, ys (passive form) and ydis. 
 oes, he is, Welsh. 
 Erse, is, as is me, is tu, i se, I am, thou art, 
 he is. 
 
 N.B. This root is defective in all the above languages, and a 
 great part of the forms of the verb substantive are supplied from 
 the following. 
 
 *T — bhfi, a verbal root, whence the verb BHAVAMI, 
 <V J . . . 
 
 I am ; pret. babhuva, fui ; babhuvima, fui- 
 
 mus. 
 
 Pers. budex, to be; bud, he was; existence, 
 
 being. 
 
 BU, be thou. 
 
 Latin, fuo, fui, fuvimus. 
 
 Teut. beon, to be, A. Sax. bin, bist, be, &c. 
 
 Sclav, buit', to be, Russ. 
 
 Greek, (/>{&, <f>vfjii, <fivvat. 
 
 Celtic, bum, buost, bu, Welsh ; fui, fuisti, 
 
 fuit. 
 
 bydh, erit ; bod, esse. 
 
 Erse, bu mi, I was ; bhith, to be. 
 
 N.B. Compare Byd, the world, from the same root, with Bud 
 in Persian, and Bhuh, the world, in Sanskrit. 
 
 3{«T — an, a verbal root, whence the verb anyatai, 
 
 respirat, vivit, and a?iimi, respiro. The first 
 
 person of anyatai was perhaps, as the analogy 
 
 of the other persons and of the active voice 
 
 G 2 
 
84 EASTERN ORIGIN OF 
 
 would suggest, originally anyamai, instead 
 
 of anijai : whence 
 Latin, animus, anima, animatus. 
 Greek, ave^og. 
 Celtic, anaim, (Erse,) soul, spirit, 
 
 rT^T — tan, a verbal root, whence the verb tanoti, he 
 extends, stretches. 
 Greek, TeiW, ravvcc, Tcivv{j.ou. 
 Latin, tendo. 
 
 Celtic, taen, spreading, extension, 
 taenu, to spread, extend. 
 
 <^T — da, a verbal root ; whence the verb dadami, 
 I give. 
 Pers. daden, to give. 
 Greek, Sftafju, &o'». 
 Latin, do. 
 
 Celtic, daigh, a root in Erse; whence daighim, 
 I give. 
 
 3f^ — ad, a verbal root; whence the verb admi, atsi, 
 atti, edo, edis, edit. 
 Greek, ?&». 
 Latin, edo, esu, &c. 
 
 Celtic, ysu, or esu, edere; ysawl, edax, 
 Welsh. 
 ith, a root in Erse ; whence ithim, I 
 eat ; itheadh, eating. 
 
 2T3T — yuj, a verbal root; whence are derived several 
 verbs meaning to join, and other words, as 
 follows : 
 
 yugum, a couple ; yugah, a yoke. 
 yojami, praet. yuyoja, (conjungere, 
 conjugare). 
 
THE CELTIC NATIONS. 85 
 
 yunajmi, (yuniijmi) prat, yuyoja (jun- 
 gere) and yoksyami. 
 Pers. yoo, a yoke, also yiigh. 
 
 yfighiden, to yoke. 
 Greek, ^evyvvfxi, %vyo$, k. t. A. 
 Latin, jungo, jngnm. 
 Rnss. jgo, a yoke. 
 Tent, joch, Germ, yoke, Eng. 
 Celtic, 
 
 1. Welsh, jau, a yoke ; jeuaw, to yoke. 
 
 jeuad, a yoking : jeuaeth, a yoked 
 state. 
 
 N. B. The Welsh words are nearer to the Sanskrit 
 and Persian than to the European languages. 
 
 2. Erse, cuing, or kuing, a yoke. 
 
 jeugaff, to yoke or couple. (Armoric. 
 Lhuyd, p. 245.) 
 
 <\W- — dansh, a verbal root ; whence the verbs da- 
 shami and dakshyami, (mordeo,) to bite. 
 Noun, danta, a tooth. 
 Greek, &aoa>, n. o^ovra. 
 Latin, dens, dentes. 
 Celtic, 
 
 Welsh, daint, n. aggr. the teeth, 
 dant, pi. dannedh. 
 deintiaw, verb, to bite. 
 Corn, danta, to bite. 
 
 The following verbs, or etymons of verbs, are 
 common to the Celtic and some of the other Euro- 
 pean languages. Where the resemblance is only 
 with the Latin, it may be thought probable that the 
 Britons derived them from the Romans ; but when 
 
 G 3 
 
86 
 
 EASTERN ORIGIN OF 
 
 the coincidence is between the Celtic and Greek, or 
 
 Sanskrit, or other remote branches, 
 
 the fact will 
 
 admit of no similar explanation. 
 
 
 dagru and deigraw W. 
 
 b(XKpV(t> 
 
 
 and 
 
 and 
 
 
 deigryn, W. 
 
 haKpVOV 
 
 lachryma. 
 
 darhunaw 
 
 bapOdva 
 
 
 deu and "l 
 
 
 
 dyvod, W. to come Y 
 
 bva> and bvvco 
 
 
 donet, Armor. J 
 
 
 
 dysgu, W. 
 
 bthdo-Kfo 
 
 disco et doceo 
 
 dylu and ) 
 
 8et and bovkos 
 
 
 dylyaw, to be obliged, W. j 
 
 bovkevco 
 
 
 dyro'i, W. 
 
 bwpia 
 
 
 canu, W. sing ^ 
 
 
 fganum, Sansk 
 
 canam, Erse j 
 
 cano -j 
 
 [singing, song. 
 
 iachau, to heal, from ) 
 iach, sane, whole j 
 
 
 
 idofxai 
 
 
 
 
 r kus and kusya- 
 
 * 
 
 cusau, ^ 
 
 Kvao), Gr. 
 
 -j mi, Sansk. ara- 
 
 cusanu, W. to kiss j 
 
 kiissen, Germ. 
 
 L plector 
 
 elu, W. to go 
 
 ikevdco 
 
 elsynt, they came 
 
 rjkvaav 
 
 
 galw, W. to call 
 
 Kakza 
 
 call 
 
 cleiniaw, W. to lie 
 
 KktVZLV 
 (Kkd(t), Kkd(T(0 
 
 J^et Kkda-Ls 
 
 
 cleisiaw, W. to bruise 
 
 
 cudhiaw, W. to hide, ) 
 kith and kitha, Cornish j 
 
 K€vd(a 
 
 
 curaw, to beat, knock 
 
 Kpoxxa 
 
 
 cyriaw, to limit, border 
 
 KeCpca 
 
 
 chwareu, to gambol, sport 
 
 yaipito 
 
 
 balaii, to spring out, and 
 
 fiakkiaOai 
 
 
 balaw, noun 
 
 fiokrj, e/c/3oAr) 
 
 
 dalw, to catch 
 
 bekct), inesco 
 
 
 eb, to say, as ) 
 eb eve said he | 
 
 €77(0, dico 
 
 
 €(f}7] t said he 
 
 
 elwi, to gain 
 
 eA.eii' 
 
 
THE CELTIC NATIONS. 
 
 87 
 
 ambylu to blunt ) 
 
 ambylus, blunt, adj. J 
 
 degadu 
 
 eichiaw, to sound, from \ 
 
 aich, pi. eichiau j 
 
 gwthiaw and wthiaw, i 
 
 to thrust j 
 
 Iholiaw, to babble 
 lhipau, to droop 
 maclu, to earn wages 
 men, a place 
 medw, the mind 
 meru, to droop ~| 
 
 merwinaw, to benumb, I 
 
 or deaden J 
 
 tormu, to assemble round 
 ystyr and ystyriaw, to} 
 consider, note, reflect i 
 caru, to love 
 cob, cobio, to strike 
 
 menw, mind 
 
 novio, W.) 
 snav, E. j 
 credu, W. ) 
 credeim, E. j 
 eliaw, W. 
 dosparthu 
 
 aixfiXvvv 
 
 | inlaui . 
 
 md 
 
 nila- 
 
 CLfJLj3\vS 
 
 Se/caroco 
 
 Inus, S. 
 
 languid 
 
 r/x^co, n. rj X os, 
 pi. ?}x€a 
 
 
 
 0)6 €0) 
 
 
 
 
 AaAea> 
 
 
 
 
 o\Aei7reti> 
 
 
 
 
 ^eAerdco 
 
 
 
 
 fl€V€(0 
 fJLTjbCti) 
 
 medito 
 
 
 
 ixapaiva) 
 
 turma 
 to-ropeo) 
 car us 
 
 KOTTTG) 
 
 man and manu-1 
 
 tai, S. to know, Vmens, Lat. 
 understand J 
 
 vi<a, no, Lat. 
 credo 
 
 ak€L(f)0) 
 
 dispertior. 
 
 SECTION V. 
 
 Adjectives, Pronouns, and Particles. 
 
 Parag. 1. adjectives. 
 3f^j — aalah, a, m. (ample, vast,) 
 all, alle, whole, Germ. 
 oAof, Gr. holl, oil, Welsh ; idle, Erse. 
 
 G 4 
 
88 EASTERN ORIGIN OF 
 
 3^3"! — uchchah, a. m. (high.) 
 
 uch, higher, upper ; uchel, uchach. (W.) 
 uchchata, pride. (Sanskrit.) uchediad, 
 
 soaring. (Welsh.) 
 vxpov, Greek. Compare v\pY}\o$ and uchel 
 
 in Welsh, 
 hoch, high. Germ. 
 
 *K5J — and H^rJ, maha and mahat, great. 
 Greek, fxeya. Latin, magnus, major. 
 Welsh, mawr. Erse, mor. 
 Germ, mehr, more, &c. 
 
 2J^"*T — yuvan. 
 
 yuvan, Pers. 
 
 jau, jeuant, jeuanc, Welsh. 
 
 juvenes, juvencus, Latin. 
 
 jung, young, Germ, yanuii, Russian. 
 
 >sf]«1 I — jinah, an old man. 
 
 sean, Erse ; hen, Welsh ; senex, Lat. 
 
 1^ — nava, (Am. Cosh.) or «1<3 : navah. 
 Greek, veog ; Latin, novus, 
 Germ, neu, new ; Russian, novaii. 
 Celtic, newydh, Welsh ; nuadh, Erse. 
 
 The following are chiefly adjectives common to the 
 Celtic and the Greek languages. 
 
 alius 
 
 claudus 
 
 alh, (W.) 
 
 eile, (E.) 
 
 aWos 
 
 coch 
 
 
 KOKKLVOS 
 
 cloff 
 
 
 \(»>kbs 
 
 medhws 
 
 misgeach 
 
 ixidvcos 
 
 melus, sweet 
 
 milis 
 
 juetA.to-o-0) 
 
 melyn, yellow 
 
 
 ixrikivos 
 
 tlawd 
 
 
 rd\as 
 
 caled 
 
 
 XaA.€7ros 
 
 car 
 
 chara 
 
 Xa/)tety 
 
 carus 
 
THE CELTIC NATIONS. 
 
 89 
 
 trist 
 
 tuirseach 
 
 Tpvaaos 
 
 tristis 
 
 bvr 
 
 gear 
 
 KVpTOS 
 
 brevis 
 
 either 
 
 
 kripoi 
 
 caeteri 
 
 ambylus 
 
 
 ap,(BX.vs 
 
 
 dilys 
 
 
 bTJkos 
 
 j 0€pfJ.bs 
 
 
 twyni 
 
 
 ^ difiepos 
 
 
 iachaus 
 
 
 Irj'Cos 
 
 
 Parag. 2. pronouns. 
 
 The personal pronouns must be given in full 
 when we proceed to the inflections of verbs, which 
 are conjugated by means of them. It is only requi- 
 site at present to anticipate the remark, of the truth 
 of which the reader will be afterwards convinced, 
 that the personal pronouns in the Celtic dialects 
 consist of the very same elements, and these but 
 slightly modified, which pervade all, or nearly all, 
 the other languages referred to the Indo-European 
 class. 
 
 The possessive pronouns are in the Celtic formed, 
 as in other languages, by a modification in the end- 
 ing of the corresponding personal pronouns. 
 
 Thus in Welsh. 
 
 Personal pronouns. Possessives. 
 
 1 Sing. mi, I, becomes mau 
 
 2 — ti, thou, tau 
 
 3 — masc. ev, in Erse se ei 
 
 fem. hi, in Erse si ei 
 
 1 Plur. 
 
 ni, we 
 
 2 — 
 
 chwi, 
 
 3 — 
 
 hwy or 
 hwynt 
 
 ein 
 eich 
 
 eu 
 
 The interrogative pronouns serve to exemplify 
 
90 
 
 EASTERN ORIGIN OF 
 
 the remarks made on the interchange of consonants 
 in chapter 1. section 2. 
 
 Interrogative Pronoun. 
 
 In Greek, tls tL 
 
 N. B. The existence of the interrogative particles nws, not, 
 &c. renders it probable that there was an older Greek interro- 
 gative pronoun corresponding, as nis, ni 
 
 In Latin, 
 
 quis quae 
 
 quid 
 
 
 qui 
 
 
 
 In Erse, 
 
 kia 
 
 
 kidh 
 kad 
 
 In Sanskrit, 
 
 kah 
 
 ka 
 
 kim 
 
 In Welsh, 
 
 pwy 
 
 pa. 
 
 Parag. 
 
 3. 
 
 PARTICLES. 
 
 ni — na (Welsh) 
 
 
 V7] 
 
 na Sansk. 
 
 yna 
 
 
 tva 
 
 
 wng, yng, near 
 
 
 kyyvs 
 
 
 agaws, or agos, prep. 
 
 
 eyyvs 
 
 
 cyd, cyda, pron. ciida 
 
 
 Kara 
 
 
 am, round 
 
 
 afjLfpl 
 
 um in German. 
 
 heb, without 
 
 
 airo 
 
 ab, abs 
 
 oc, out of 
 
 
 
 ex 
 
 trwy 
 
 
 
 through, durch 
 
 yn 
 
 
 
 
 neu, particle of affirmation 
 
 vat 
 
 
 cyn, with, cum, con, Lat. 
 
 avv, Gr. 
 
 ^rjr, sum, Sansk 
 
 either 
 
 
 arep 
 
 
 di (insep. part.) di, dis, 
 
 Lat. 
 
 b\ 
 
 
 dyre, veni 
 
 
 btvpo 
 
 
 evo 
 
 
 a/Jia 
 
 
 etto 
 
 
 hi 
 
 yet 
 
 mo, negative) 
 
 J 
 
 \ixa 
 
 
 after ni ) 
 
 
 \0V fJLCL 
 
 
 blaen 
 
 
 7Tkr]V 
 
 
THE CELTIC NATIONS. 91 
 
 CHAPTER IV. 
 
 Proofs of a common origin derived from the grammatical struc- 
 ture of the Celtic and other Indo-European languages. 
 
 Section i. Review of the preceding facts and inferences. 
 Introductory remarks on the personal inflections of verbs. 
 
 1 HE instances which have been pointed out in the 
 last chapter, to which I believe that it would be 
 easy to make great additions, are sufficient to prove 
 that there is an extensive affinity in the component 
 vocabularies of the Celtic dialects and those of the 
 other languages with which they have been com- 
 pared. The examples of analogy already adduced 
 are by far too numerous and too regular, or in ac- 
 cordance with certain general observations, to be the 
 result of mere chance or accidental coincidence. It 
 must likewise be remarked that they are found in 
 that class of words which are not commonly derived 
 from one language into another. I allude particu- 
 larly to such terms as denote the most familiar ob- 
 jects and relations, for which no tribe of people is 
 without expressive terms. When such relations as 
 those of father, mother, brother, and sister are ex- 
 pressed by really cognate words, an affinity between 
 the several languages in which these analogies are 
 found is strongly indicated. The same remark may 
 be made in respect to the names of visible bodies 
 and the elements of nature, such as sun, moon, air, 
 sky, water, earth. Lastly, the inference is confirmed 
 by finding many of the verbal roots of most frequent 
 occurrence, as the verb substantive, and those which 
 express generation, birth, living, dying, knowing, 
 
92 EASTERN ORIGIN OF 
 
 seeing, hearing, and the like, to be common to all 
 these languages. 
 
 It may be remarked, that in the Celtic language, 
 as well as in the Persian, and in some German dia- 
 lects, the Sanskrit and Greek words are represented 
 by terms in a shortened and broken form, which 
 have lost the regularity and beauty of their termi- 
 nations. Yet there are several instances in which 
 the Celtic words resemble more nearly their Sanskrit 
 analogues than those belonging to other European 
 languages, as the terms tad and brawd for tatah and 
 bhrata, meaning father and brother. In many ex- 
 amples the Sclavonic dialects and the Persian lan- 
 guage display the transition from the form of words 
 peculiar to the Sanskrit to that of the northern Eu- 
 ropean idioms. The root shru or sru, meaning 
 to hear, becomes in Russian slu ; but in Greek 
 and in Celtic kXv and clyw, or clu. Ashwah, a 
 horse, becomes asb in Persian, and in Erse each. 
 Shukarah, a hog, is in Persian khuk, and in Welsh 
 hwch. In most cases we discover something to con- 
 firm the laws of deviation laid down in the preced- 
 ing chapters, according to which it appears that 
 words derived by the western from the eastern 
 languages are changed in a peculiar way. The most 
 general of these alterations is the substituting of 
 guttural for sibilant letters, which by the Celtic dia- 
 lects is made almost uniformly, and very frequently 
 by the Greek and the Teutonic. 
 
 There is a still more striking resemblance in the 
 grammatical forms of these languages, which I shall 
 now endeavour to point out. 
 
 Professor Murray has attempted to illustrate the 
 grammatical structure of the European languages 
 
THE CELTIC DIALECTS. 98 
 
 from a quarter to which few persons would have 
 been inclined to look with any hope of success for 
 the means of its elucidation, I mean the Teuto- 
 nic idioms, and even some of the modern dialects 
 of the German language. It would really appear 
 that in these idioms some words, affording traces of 
 ancient forms and derivations, have still survived, 
 which can no longer be recognised in the classical 
 languages of India, of Greece, and of Italy. In the 
 following pages it will more evidently appear, if I 
 am not mistaken, that from the Celtic dialects a part 
 of the grammatical inflections, and that a very im- 
 portant part, common to the Sanskrit, the iEolic 
 Greek, the Latin, and the Teutonic languages, is 
 capable of an elucidation which i$? has never 
 yet received. This can only be accounted for by 
 the remark that the Celtic people have been more 
 tenacious of the peculiarities of their language, as 
 they have been in many respects of their customs 
 and manners, than the other nations of Europe. 
 
 The mode of conjugating verbs appears to be es- 
 sentially the same in all these languages. It consists 
 partly in certain variations indicating time and 
 mood, and partly in the addition of particular end- 
 ings, by which the differences of number and per- 
 son are denoted. The former class of variations 
 will be considered in the sequel. At present I shall 
 investigate the nature and origin of the personal 
 terminations, or of those increments or suffixes which 
 the verbal roots receive for the purpose of distin- 
 guishing the person and number. It will appear 
 that these are all pronominal suffixes, or abbreviated 
 or otherwise modified pronouns. This has been 
 conjectured and shewn to be probable by many phi- 
 
94 EASTERN ORIGIN OF 
 
 lological writers, but the proof has always been de- 
 fective in several particulars, because this subject 
 has not been surveyed in a sufficiently comprehen- 
 sive manner, and with attention to all the evidence 
 which can be brought to bear upon it, and especially 
 to that portion which is derivable from a compari- 
 son of the Celtic dialects. 
 
 In proceeding to this investigation, I shall in the 
 first place shew by examples what are the charac- 
 teristic endings of the different persons of the verb 
 in several languages. 
 
 SECTION II. 
 
 Personal endings of the Sanskrit verbs. 
 One system of personal terminations belongs to 
 all Sanskrit verbs, and the differences of conjugation 
 which are distinguished by grammarians consist in 
 the changes which the verbal roots undergo. The 
 following is an example displaying the terminations 
 of the present tense as they are subjoined to the 
 verbal root tud, to strike, in Latin tundo. 
 
 
 1. Person. 
 
 2. Person. 
 
 3. Person. 
 
 Singular, 
 
 Tudami 
 
 Tudasi 
 
 Tudati 
 
 Dual, 
 
 Tudavas 
 
 Tudat'has 
 
 Tudatas 
 
 Plural, 
 
 Tudamas 
 
 Tudafha 
 
 Tudanti. 
 
 This verb belongs to those classes of roots which 
 insert a vowel a between the theme and the per- 
 sonal endings. Others subjoin these endings imme- 
 diately. The personal endings alone are as follows: 
 1. Person. 2. Person, 3. Person. 
 
 bing. 
 
 -mi 
 
 -si 
 
 -ti 
 
 Dual, 
 
 -vas 
 
 -t'has 
 
 -tas 
 
 Plur. 
 
 -mas 
 
 -fha 
 
 -anti. 
 
THE CELTIC NATIONS. 
 
 95 
 
 The same terminations belong to the future tenses 
 as to the present; but those tenses which have the 
 augment prefixed to the verb have the personal end- 
 ings, as in Greek, in a more contracted form. The 
 following is the first preterite of the verb tudami, 
 corresponding closely to the Greek imperfect a . 
 
 Praeteritum augmentatum 1. 
 
 1. Person. 2. Person. 3. Person. 
 
 Sing. Atudam Atudas Atudat 
 
 Dual, Atudava Atudatam Atudatam 
 
 Plur. Atudama Atudata Atudan. 
 
 The abbreviated personal endings in Sanskrit 
 verbs are as follows : 
 
 Sing. 
 
 -am 
 
 -s 
 
 -t 
 
 Dual, 
 
 -va 
 
 -tarn 
 
 -tam 
 
 Plur. 
 
 -ma 
 
 -ta 
 
 -an b 
 
 There is another form of the indicative tenses in 
 the parasmaipadum, or active voice, which it may be 
 right here to exhibit. It is that of the reduplicated 
 preterite, formed by rules nearly the same as those 
 of the preterperfect in Greek verbs. The prseteritum 
 reduplicatum of the verb tud or tudami is as fol- 
 lows : 
 
 
 1. Person. 
 
 2. Person. 
 
 3. Person. 
 
 Sing. 
 
 Tut6da c 
 
 Tutodit'ha 
 
 Tutoda 
 
 Dual, 
 
 Tutudiva 
 
 Tutudafhus 
 
 Tutudatus 
 
 Plur. 
 
 Tutudima 
 
 Tutuda 
 
 Tutudus. 
 
 N. B. It may be observed that the vowel of the root u is 
 changed into o in this instance by the form termed guna, of the 
 
 a I represent the augment 3f by an a in following sir W. 
 Jones's orthography ; but it might perhaps as correctly be re- 
 presented by the Greek e. 
 
 b Bopp, Gram. Crit. Sansk. p. 144. c In Latin tutudi. 
 
96 EASTERN ORIGIN OF 
 
 influence of which we trace the result in the Greek reduplicate 
 preterite of the old form, commonly termed the prcterperfect of 
 the middle voice. We shall observe likewise the influence of 
 guna to be very extensive in the inflections of verbs in the dif- 
 ferent European languages. 
 
 SECTION III. 
 
 Terminations characteristic of the persons of the Greek verb. 
 
 Of the two principal forms of inflection by which 
 Greek verbs are conjugated, one, viz. that of verbs 
 in [Mi, corresponds nearly with the Sanskrit. There 
 are strong reasons for believing this to be an ancient 
 and perhaps the original method of conjugating 
 verbs used in the Greek language a , independently 
 of the circumstance that it so nearly resembles the 
 forms of the Sanskrit. This conjugation comprises 
 the verb substantive and a great many old and very 
 anomalous and defective verbs, and those of very 
 common and familiar occurrence b . The conjugations 
 of verbs in « are so much more regular, that they 
 bear the appearance of a designed and systematic 
 scheme introduced for the sake of simplifying the 
 inflections of the language. The Doric form of the 
 verbs in //.< will probably serve to exemplify the per- 
 sonal endings as they existed in the earliest state of 
 the Greek language of which we can obtain any 
 knowledge. The following is the Doric form of the 
 verb laTYj^i in the present tense : 
 
 a Matthiae indeed seems inclined to believe that there was a 
 still older form of Greek verbs than those now extant, and that 
 the termination was in &>. 
 
 b As (pTjfii, ei/xi, irjfii, ^[xai, &C. 
 
THE CELTIC NATIONS. 
 
 97 
 
 Person. 
 
 8. Person 
 
 tOTCl? 
 
 toTari 
 
 lararov 
 
 tararov 
 
 la-rare 
 
 icrravrt. 
 
 1. Person. 
 
 Sing. tora/ut 
 
 Dual, 
 
 Plur. torajuter 
 
 The irregular and defective verbs often display 
 the primitive verbal forms of a language better than 
 those which are regular, and of the former the verb 
 substantive, which is anomalous in many languages, 
 may be considered as the most ancient. The follow- 
 ing is the present tense of the verb substantive in its 
 oldest forms. 
 
 
 
 Singular. Old form. 
 
 1. 
 
 etjott, 
 
 Doric form e/x/xt, originally (?) eo-fu 
 
 2. 
 
 5> 
 
 ets; 
 
 in Homer, Pindar, Theocritus, ecro-i 
 
 3. 
 
 kcrri 
 
 hrrl 
 
 1. 
 2. 
 
 
 Dual. 
 
 icTov 
 
 iarov 
 
 3 
 
 £(TTOV 
 
 ecrrbi' 
 Plural. 
 
 1. 
 
 eo-juer, 
 
 Doric dfA€$, originally (?) eoyxc? 
 
 2, 
 
 core 
 
 eOT6 
 
 3. 
 
 eurt 
 
 Doric €vrl 
 
 It seems from this statement, that the following 
 are the personal endings of the verb substantive in 
 the present tense, subjoined immediately to the ver- 
 bal root. 
 
 Dual. Plural. 
 
 Singular. 
 
 1. 
 
 -fXL 
 
 2. 
 
 -(TL 
 
 3. 
 
 -TL 
 
 -re 
 
 These will be seen on comparison to be nearly 
 identical with the endings of Sanskrit verbs. A like 
 analogy may be traced by comparing the abbreviated 
 
 H 
 
98 EASTERN ORIGIN OF 
 
 form of the augmented preterite with the Greek pre- 
 terimperfect and the preter reduplicate with the Greek 
 preterperfect ; but as it is not my principal aim to 
 trace the analogies of the Sanskrit and Greek lan- 
 guages, I shall say nothing further on this subject. 
 
 SECTION IV. 
 
 Personal endings in Latin verbs. 
 
 It is probable that the first person of the present 
 tense in the indicative mood of Latin verbs termi- 
 nated originally in m instead of o. This results 
 from a comparison of the endings of the other per- 
 sons in the various tenses of the indicative mood, 
 and from the analogy afforded by the first person 
 in the subjunctive mood. The supposition is con- 
 firmed by the actual existence of old and anomalous 
 forms, sum or esum, and inquam, where the ter- 
 mination in m is still extant. 
 
 The personal endings in Latin verbs, exclusive of 
 the imperative mood, may be comprised in the fol- 
 lowing forms. 
 
 Singular. 
 
 1. 
 
 -am 
 
 -em 
 
 -im 
 
 % 
 
 -as 
 
 -es 
 
 -is 
 
 s. 
 
 -at 
 
 -et 
 Plural. 
 
 -it. 
 
 1. 
 
 -amus 
 
 -emus 
 
 -imus 
 
 2. 
 
 -atis 
 
 -etis 
 
 -itis 
 
 3. 
 
 -ant 
 
 -ent 
 
 -unt, int 
 
THE CELTIC NATIONS. 99 
 
 SECTION V. 
 
 Terminations which distinguish the persons of verbs in the Teu- 
 tonic dialects. 
 
 A new and very important light has been thrown 
 on the structure and history of the Teutonic lan- 
 guages by the researches of Professor Jacob Grimm. 
 I shall endeavour to abstract in a short compass 
 some of the results of his inquiries which relate to 
 the inflections of verbs in the oldest of these lan- 
 guages, as being closely connected with the subject 
 now under consideration. 
 
 The verbs are divided in all the Teutonic dialects 
 into two classes, chiefly distinguished from each 
 other by the manner in which they form the past 
 tense and participle. These different modes of in- 
 flection are termed by Dr. Grimm respectively the 
 strong and weak conjugations. The former is sup- 
 posed by that writer to be more ancient than the 
 other, and to be in fact the genuine and primitive 
 method by which the German nations distinguished 
 the times and modes of action and of passion in the 
 use of verbs. In this first method a great propor- 
 tion of the original and peculiar roots of the Teu- 
 tonic dialects were conjugated; but its use has given 
 way in a great degree to a different scheme of in- 
 flection, which of late has become prevalent, as being 
 more in harmony with the genius of modern lan- 
 guage. The latter is supposed to be more recent in 
 its origin, and it comprises, besides many primitive 
 German roots, all foreign words which have been 
 adopted into the vocabulary of the Teutonic nations. 
 The English reader will have an idea of the strongly 
 and weakly inflected conjugations by observing that 
 all those verbs belong to the former which make 
 
 H 2 
 
 itofC. 
 
100 EASTERN ORIGIN OF 
 
 the past tense and participle by changing the vowel 
 of the monosyllabic root, as speak, spake, spoken ; 
 while the inflection of praise, praised, praising, ex- 
 emplifies the weak conjugation. In the Moeso-Gothic, 
 which preserves the oldest forms of the Teutonic 
 languages, there is, in addition to the change of 
 vowel which characterises the past tense, a redupli- 
 cation of a part of the root a . There are twelve 
 forms belonging to the strongly inflected verbs, and 
 three or four of the other class. As the character- 
 istic parts of the verbs of each conjugation Dr. 
 Grimm has given the indicative mood, present tense, 
 first person singular, the first person singular and 
 plural of the past tense, and the participle and infi- 
 tive mood. 
 
 As the subject of the present chapter is the cha- 
 racteristic endings of persons and numbers, I should 
 not have touched upon any thing which relates to 
 the formation of tenses and moods, until I come to 
 the proper place for that inquiry, had it not been 
 for the circumstance that the personal endings them- 
 selves are different in the several modes of conjuga- 
 tion. As I wish to include the endings belonging 
 to both systems, I found it necessary to explain, in 
 the first place, the principle by which they are dis- 
 tinguished from each other. I shall now extract a 
 table of the terminations belonging to each form as 
 laid down by Dr. Grimm, beginning with the Moeso- 
 Gothic verbs. 
 
 1. Strongly inflected conjugation of Moeso-Gothic 
 verbs. 
 
 The following verbs will afford a specimen of this 
 
 a This was observed by Hickes. See his Moeso-Gothic Gram- 
 mar. Thesaur. Ling. Sept. torn. i. 
 
THE CELTIC NATIONS. 
 
 101 
 
 inflection, and display in an interesting manner some 
 of the oldest forms belonging to our own language 
 or that of our Teutonic ancestors. 
 
 1. Slepa, I sleep; saizlep, I slept; saizlepum, we 
 
 slept ; participle, slepans. 
 
 2. Laia, I laugh; lailo, laughed; lailoum; laians. 
 
 3. Svara, I swear ; svor, I swore ; svorum ; s va- 
 
 rans. 
 
 In this instance, and in the six last of the strongly 
 inflected conjugations, the verb merely changes the 
 radical vowel, and has no reduplication. 
 
 Paradigm of the personal endings of verbs of the 
 strongly inflected conjugations. 
 
 1. 
 
 Person. 2. 
 
 Person. 
 
 3. Person 
 
 Indie. Pres. Sing. 
 
 -a 
 
 -is 
 
 -ith 
 
 Dual 
 Plur. 
 
 -6s 
 -am 
 
 -ats 
 -ith 
 
 
 -and 
 
 Pret. Sing. 
 
 
 -t 
 
 
 Dual 
 
 -u (?) 
 
 -uts 
 
 
 
 Plur. 
 
 -urn 
 
 -uth 
 
 -un 
 
 Subjunct. Pres. Sing. 
 
 -au 
 
 -ais 
 
 -ai 
 
 Dual 
 
 -aiva (?) 
 
 -aits 
 
 
 
 Plur. 
 
 -aima 
 
 -aith 
 
 -aina 
 
 Pret. Sing. 
 
 -jau 
 
 -eis 
 
 -ei 
 
 Dual 
 
 -eiva 
 
 -eits 
 
 
 
 Plur. 
 
 -eima 
 
 -eith 
 
 -eina 
 
 Imperative Sing. 
 
 
 
 
 
 
 Dual 
 
 
 
 -ats 
 
 
 
 Plur. 
 
 -am 
 
 -ith 
 
 
 
 Infinitive -an ; Particip. pres. -ands ; Particip. pret. -ans. 
 
 N. B. The mark . . . indicates that no additional ending is 
 
 subjoined to the verb, and the mark that the form for which 
 
 it stands is wanting. 
 
 H 3 
 
102 
 
 EASTERN ORIGIN OF 
 
 The following paradigm illustrates the weakly in- 
 flected conjugation, of which the verb sokjan may 
 serve as an example, sokjan or sokyan is suchen, to 
 seek ; sokja, I seek ; sokida, I sought ; sokjands in 
 the participle. 
 
 T J' 
 
 •x 
 
 1. 
 
 Person. 
 
 2. Person. 
 
 3. Person 
 
 Indicat. 
 mood 
 
 I Pres. T. 
 
 Sing. 
 Dual 
 
 (vowel) 
 -6s 
 
 -s 
 -ts 
 
 -th 
 
 
 
 Plur. 
 
 -m 
 
 -th 
 
 -nd 
 
 
 Pret. T. 
 
 Sing. 
 
 -da 
 
 -des 
 
 -da 
 
 
 
 Dual 
 
 
 
 -deduts 
 
 
 n 1 • 
 
 "V 
 
 Plur. 
 
 -d&dum 
 
 -deduth 
 
 -dedun 
 
 Subjunct 
 mood 
 
 ' V Pres. T. 
 
 Sing. 
 Dual 
 
 (vowel) 
 
 -s 
 -s 
 
 (vowel) 
 
 
 
 Plur. 
 
 -ma 
 
 -th 
 
 -na 
 
 
 Past T. 
 
 Sing. 
 
 -dedj£n 
 
 -dedeis 
 
 -dedi 
 
 Imperat. 
 mood 
 
 } 
 
 Dual 
 Sing. 
 Dual 
 
 
 
 -dedeits 
 (vowel) 
 -ts 
 
 -dedeina 
 
 TnTinitivf 
 
 i 
 
 Plur. 
 
 -m 
 
 -th 
 
 
 
 JL lllllll 11 \ v_ 
 
 mood 
 
 
 -n 
 
 
 
 
 Participl 
 ParticipJ 
 
 e present, 
 
 le past, 
 
 -nds 
 -ths 
 
 
 Next to the Moeso-Gothic Dr. Grimm has ranked 
 in the affiliation of Teutonic languages the Old 
 High German, the characteristics of which approach 
 most nearly to those of the Gothic forms. Under 
 this denomination of Alt-hoch-deutsch or Old High 
 German, it must be observed that the remains of 
 several dialects are comprised, which were nearly al- 
 lied, but yet probably differed from each other in pe- 
 culiarities now scarcely to be ascertained. Among 
 these were the idioms of the Franks, Bavarians, and 
 
THE CELTIC NATIONS. 103 
 
 Allemanni, and perhaps of other tribes between the 
 seventh and eleventh centuries, of which specimens 
 are preserved in the remains of Keros and Notker, 
 and in the extant works of Ottfried. From these, 
 and from some other relics of the period above men- 
 tioned, this ancient form of the High German lan- 
 guage has been made up and restored by the accu- 
 rate researches of Dr. Grimm, 
 
 Forms of the verb in the Old High German. 
 Forms of the strongly inflected conjugation. 
 
 "2? JP-T.Sing 1 ' 
 
 Plur. 
 
 Past T. Sing. 
 
 ^ , . . Plur. 
 
 mood j _ n s 
 
 \ Plur. 
 
 Past T. Sing. 
 
 Plur. 
 
 Imperative Sing. 
 
 Plur. 
 
 Infinitive -an. Part. pres. -anter. Part, preter. -aner. 
 
 As an example of this conjugation we may take 
 the following: 
 
 Slafu, I sleep. 
 Sliaf, I slept. 
 Slkfumes, we slept. 
 Slafaner, having slept. 
 
 H 4 
 
 erson. 2. 
 
 Person. 
 
 3. Person 
 
 -u 
 
 -is 
 
 -it 
 
 -ames 
 
 -at 
 
 -ant 
 
 
 -i 
 
 ... 
 
 -uraes 
 
 -ut 
 
 -un 
 
 -e 
 
 -&s 
 
 -e 
 
 -ernes 
 
 -et 
 
 -en 
 
 -i 
 
 -is 
 
 -i 
 
 -imes 
 
 -It 
 
 -in 
 
 — — — 
 
 _3t 
 
 
 
104 
 
 EASTERN ORIGIN OF 
 
 Paradigm of the weakly inflected verb of the Old 
 High German. 
 
 1st Person. 2d Person. 3d Person. 
 
 ^mood' }Pres.T.Sing. -u(-m) 
 
 Plur. -mes 
 
 Pret. T. Sing. -ta 
 
 Q , . . Plur. -tumes 
 
 mood* j Pres * T " Sin ^- ( vowel ) 
 
 Plur. -mes 
 Pret. T. Sing. -ti 
 
 Plur. -times 
 
 Imperative Sing. 
 
 Plur. 
 
 Infinitive -n. Part. pres. -nter, -ter. 
 
 -s 
 
 -t 
 
 -tos 
 
 -tut 
 
 -s 
 
 -t 
 
 -tis 
 
 -tit 
 
 (vowel) 
 
 -t 
 
 -t 
 
 -nt 
 
 -ta 
 
 -tun 
 
 (vowel) 
 
 -n 
 
 -ti 
 
 -tin 
 
 Dr. Grimm has added an analysis of the gramma- 
 tical forms in the other dialects belonging to the 
 Teutonic family of languages, viz. the Old Saxon, 
 the Anglo-Saxon, the Old Frisian, the Old Norse or 
 northern dialect of the Voluspa and the Edda, the 
 Middle High German, the Middle Netherlandish, 
 the modern High German, the modern Netherland 
 dialect, the modern English, the Swedish, and the 
 Danish. The comparison of these varying forms of 
 one original speech is extremely interesting to the 
 philologer, and indispensable to those who wish to 
 be thoroughly and fundamentally acquainted with the 
 relations of our own mother tongue, but it would 
 be foreign to my present design to pursue this sub- 
 ject further. I shall here add merely an outline of 
 the personal endings of the Gothic and Old High 
 German verbs in comparison with each other, con- 
 fining myself to the present tense. 
 
THE CELTIC NATIONS. 105 
 
 Personal endings of the Moeso-Gothic and Old 
 High German verbs in the present tense. 
 1st Pers. Sing, a vowel, (often -a or -u,) or -6111, or -em. 
 2d Pers. Sing, -is, -es, -6s. 
 3d Pers. Sing, -ith, -it, -et, -6t. 
 1st Pers. Plur. -m, -am, -ames, -ernes, -omes, &c. 
 2d Pers. Plur. -ith, -it, -et, -ot. 
 3d Pers. Plur. -nd, -and, -ant, -ent, -ont. 
 
 Examples of these terminations which so much resemble the 
 classical languages, and which are now lost to so great a degree 
 in the Germanic dialects, occur in the following verses of a 
 translation of that magnificent hymn of the ancient church, the 
 Te Deum, which I copy from Hickes's Thesaurus, 
 i . Thih Cot lopemes, Te Deum laudamus, 
 
 Thih Trutinan gehemes, Te Dominum confitemur, 
 
 Thih euuigan Fater, Te eeternum Patrem, 
 
 Eokiuuelih erda uuirdit. Omnis terra veneratur. 
 
 2 . Thir alle engila, thir himila, Tibi omnes angeli, tibi cceli, 
 Inti alio kiuualtido, Et universal potestates, 
 Thir Cherubim inti Seraphim Tibi Cherubim et Seraphim 
 Unbilibanlieheru stimmo fo- Incessabili voce proclamant, 
 
 raharent, 
 
 3. Uuiher, uuiher, uuiher, Sanctus, sanctus, sanctus, 
 Truhtin Cot herro, Dominus Deus Sabaoth. 
 Folliu sint himila inti erda Pleni sunt cceli et terra 
 Thera meginchrefti tiurda Maj estate glorise tuse. 
 
 thinera. 
 
 SECTION VI. 
 
 Personal endings of verbs in the Sclavonian dialects and in the 
 Persian language. 
 
 As the Sclavonian dialects constitute one import- 
 ant branch of the European languages, they must 
 not be entirely passed over in a treatise, the object 
 of which is to point out and illustrate the relations 
 of these idioms to each other. I shall, however, 
 
106 EASTERN ORIGIN OF 
 
 confine myself to one dialect belonging to this divi- 
 sion, and on this I shall touch but briefly. The 
 following examples will afford my readers a speci- 
 men of the inflection of verbs in the Russian lan- 
 guage, so far as the personal endings are con- 
 cerned ; and they will be sufficient to shew, that these 
 terminations belong to the generally prevailing sys- 
 tem which we have traced in other languages. 
 
 The Russian verbs are complicated in other re- 
 spects, but their personal terminations present very 
 little variety. In several tenses these endings are 
 entirely wanting, and the personal pronouns alone 
 distinguish the modifications of meaning; but the 
 present tense has a perfect inflection. The following 
 is the present tense of the verb stoyu, I stand a . 
 Singular. Plural. 
 
 1. ya stoyu mi stoim 
 
 2. ti stoish vi stoite 
 
 3. on' stoit oni stoyat. 
 
 The following paradigm of the terminations of 
 Russian verbs in the two forms which differ most 
 widely from each other is given by Professor Vater 
 in his excellent Russian Grammar. 
 
 First Form. Fifth Form. 
 
 Singular. Singular. 
 
 1. -yu 1. -u 
 
 2. -esh % -ish 
 
 3. -et 3. -it. 
 Plural. Plural. 
 
 1. -em 1. -im 
 
 2. -ete 2. -ite 
 
 3. -yut. 3. -yat b . 
 
 a Elemens de la Langue Russe par M. Charpentier. Petersb. 
 1768. p. 148. 
 
 b Dr. Johann Severin Vater's Praktische Grammatik der Rus- 
 sischen Sprache, p. 88. 
 

 THE CELTIC NATIONS. 107 
 
 The Persian verbs display the same general ana- 
 logy ; their terminations are even more nearly allied 
 to those of the Teutonic verbs than the Sclavonian. 
 Of this the reader will judge from the present tense 
 of the verb substantive, which is regarded as a model 
 for the variations of the persons in all tenses. 
 Sing. 1. -am 2. -iy 3. -est 
 
 Plur. 1. -Im 2. -id 3. -end. 
 
 The following is the preterite of the verb buden, 
 and may serve as an example of past tenses in ge- 
 neral. 
 
 Sing. 1. -budem 2. -budi 3. -bud 
 
 Plur. 1. -budim 2. -budid 3. -budend. 
 
 SECTION VII. 
 
 Terminations characterising the persons and numbers of verbs 
 in the Celtic languages. 
 
 I now proceed to the personal endings of verbs in 
 the Celtic language ; and as they appear to have been 
 preserved in a more complete state in the Welsh 
 than in any other dialect of this language, I shall 
 take the Welsh verbs at present as my principal 
 subject. In the sequel, the formations peculiar to 
 the Erse will be examined, and compared with those 
 belonging to other idioms. 
 
 It has been observed, that the Teutonic verbs have 
 only one form for the future and the present tense. 
 The same remark applies to the Welsh ; for the Welsh 
 language, except in the instance of the verb substan- 
 tive, which has two distinct forms, one for the pre- 
 sent and the other for the future tense, has only one 
 modification of the verb, which is used to represent 
 both. In the German dialects the single form above 
 
108 EASTERN ORIGIN OF 
 
 referred to is properly a present tense ; but the Welsh 
 grammarians consider that their language has only 
 a future, and say that the future is put for the pre- 
 sent. It is however used as such in cases where 
 no license of expression or trope can have place, as 
 in the Creed : " Credav yn Nuw Dad," Credo in 
 Deum Patrem. 
 
 The Welsh verbs present a considerable variety 
 in their terminations, as the following examples will 
 shew. 
 
 First Form, Future Tense, of the verb caru, to love. 
 Singular. Plural. 
 
 1 carav carwn 
 
 2 ceri, i. e. keri cerwch 
 
 3 car carant 
 
 It must be noticed that the third person of the 
 future tense is the root of the verb. The endings 
 of the other persons are pronominal suffixes, as we 
 shall clearly perceive in the sequel. The termina- 
 tion of the first person in av is equivalent to amk, 
 or the v to a soft m. In the present tense of verbs 
 in the Erse dialect the corresponding termination is 
 always aim or im. 
 
 Second Form, Preterperfect Tense a . 
 Singular. Plural. 
 
 1 cerais carasom 
 
 2 ceraist carasoch 
 
 3 carodh carasant 
 
 a It may be worth while to add the same tense of the verb 
 substantive, as it displays somewhat more strikingly the affinity 
 of the Celtic to other European inflections. 
 
 i bum fui buom fuimus. 
 
 2 buost fuisti buoch fuistis. 
 
 3 bu fuit buont fuerunt. 
 
THE CELTIC NATIONS. 109 
 
 Third Form, Preterpluperfect Tense. 
 Singular. Plural. 
 
 1 carwn carem 
 
 2 carit carech 
 
 3 carai carent 
 
 The preterpluperfect caraswn changes its endings 
 exactly as the foregoing. 
 
 There are some other varieties, an account of 
 which will be given in a following chapter. It is 
 only my object at present to deduce a general prin- 
 ciple. 
 
 The following are the terminations. I begin with 
 the plural, as presenting more regularity. 
 Plural Terminations. 
 1st Form. 2d Form. 3d Form. 4th Form. 
 
 1 -wn -om -em -ym 
 
 2 -wch -och -ech -ych 
 
 3 -ant -ant -ent -ynt 
 
 The fourth column contains a modification used 
 by the poets. 
 
 The terminations proper to the singular number 
 are as follows: 
 
 1st Form. 2d Form. 3d Form. 
 
 1 -av -ais -wn 
 
 2 -i -aist -it 
 
 3 no addition 
 to the root. 
 
 It will be observed at once that there is sufficient 
 resemblance between these inflections and those of 
 other Indo-European languages to connect them in- 
 dubitably with that class. This is particularly ma- 
 nifest in the plural endings. There are some appa- 
 rent anomalies, but these will be explained in the 
 sequel, and will be found illustrative of the general 
 result to be deduced. 
 b Note, dh, commonly written dd, is pronounced as th in other. 
 
 1 \ -odhb _ ai 
 
 i 
 
110 EASTERN ORIGIN OF 
 
 CHAPTER V. 
 
 Of the personal pronouns in the Indo-European languages, and 
 of the derivation of the personal terminations of verbs. 
 
 Section I. Personal pronoun of the first person. 
 
 .HAVING examined in the preceding chapter the 
 systems of terminations which characterise the per- 
 sons of verbs, I now proceed to compare with them 
 the personal pronouns still extant in the same lan- 
 guages, and to shew that the endings of verbs which 
 distinguish the persons and numbers are supplied by 
 abbreviated forms of those pronouns subjoined to 
 the verbal roots. In what degrees the pronouns be- 
 longing to each language have contributed to the 
 formation of these endings or suffixes will appear in 
 the course of the following investigation. 
 
 Personal pronoun of the first person in the Sanskrit, Greek, 
 
 Latin, Russian, Moeso-Gothic, and Old High German 
 
 languages. 
 
 Singular. 
 Nominative. 
 nS^Hv — aham; eyav, eycb, 'icaya, wvya; ego; ya, Russ.; 
 ik, Goth, ih, O. H. Germ. 
 
 Genitive. 
 *{?{ — mama and 3T — mai ; peSev, epeo, pov ; mei ; 
 menya, Russ.; meina, Goth.; min, O. H. Germ. 
 
 Dative. 
 
 H \^\ *£ — mahyam and H" — mai; efuv, e/ao/, fxoi; mihi ; 
 mne, Russ.; mis, Goth.; mir, O. H. Germ. 
 
THE CELTIC NATIONS. Ill 
 
 Accusative. 
 *n*^ — main and *TT — m &; e/xe, //i, e/x/v; me; menya, 
 Russ.; mik, Goth.; mih, O. H.Germ. 
 Ablative. 
 E{7^ — mat ; car e/xot) ; a, me, &c. 
 
 Instrumental. 
 *\*\ | — maya; me, Lat.; mnoiu and mnoi, Russ. 
 
 Locative. 
 HK| — mayi ; in me. 
 
 Propositi ve. 
 nine, Russ. 
 
 Dual. 
 Nominative. 
 s$\ m \\ — avam; a/x^e, vm 9 vu; vit, Goth.; wiz (?) 
 O. H. Germ. 
 
 Genitive. 
 
 *\ \*\A l*V — avayos and *T — nou ; vtotv, vw ; ugkara, 
 Goth.; unchar, O. H.Germ. 
 Dative. 
 
 3n^T^r^ — avabhyam, *T — nou; vmv, v£v; ugkis, 
 Goth.; unch, O. H. Germ. 
 Accusative. 
 3fT^7^r — avam and «T — nou ; a^/xe, vai', vS> ; ugkis, 
 Goth.; unch, O. H. Germ. 
 
 Ablative and Instrumental. 
 ^ N I ^T^— avabhyam. 
 
 ^^ Locative. 
 
 s*i!qeJTF_avayos. 
 
 Plural. 
 
 Nominative. 
 
 ^^1^ — vayam ; a/xe$, i. e. vames, wees, 07/xe/V ; nos ; 
 
 mi, Russ.; veis, Goth.; wir, O. H. Germ. 
 
112 EASTERN ORIGIN OF 
 
 Genitive. 
 3TT^TT^?T — asmakam ; apuv, yj^uv ; nas, Russ.; un- 
 sara, Goth.; unsar, O. H. Germ.; our. 
 
 Dative. 
 3fT5T^5pT — asmabhyam, •Tff — nas; apiv, fan; no- 
 bis; nam, Russ.; unsis, Goth.; uns, O. H. Germ. 
 
 Accusative. 
 3f^(7 ? [ — asman, «1*^ — nas ; a/x//.a$-, fakaq ; nos ; 
 nas, Russ.; unsis, (uns,) Goth.; unsih, Old H. 
 Germ. 
 
 Ablative. 
 
 s5H*| \<\ — asmat ; a<j> fawv ; a nobis. 
 
 Instrumental. 
 
 ^(t*i|[ *TH — asmabhis ; nobis ; nami, Russ. 
 
 Locative. 
 
 3fWTW — asmasu ; in nobis. 
 
 Prsepositive Case, 
 nas, Russ. 
 
 Note. An attentive examination will enable the reader to 
 ascertain, that, notwithstanding the great variety of these 
 pronouns and their inflections, a few common elements are 
 the foundation of them all. A satisfactory analysis of the 
 Sanskrit pronoun has been given us by Professor Bopp, who 
 has dissected the elements which enter into its declension 
 with his usual ingenuity and discernment. His object is 
 the Sanskrit pronoun, but his remarks may tend to eluci- 
 date the corresponding forms in all the cognate languages. 
 He observes that the Sanskrit aham, ego, which is quite un- 
 connected with its oblique cases, consists of two elements, 
 viz. 3f^T ? ah and 3f51" ? am ; the latter is a mere termina- 
 tion, occurring as such in other pronouns : ah is the root. 
 Compare it with ih, ik, ek, ego, h being considered as a gut- 
 tural consonant. The oblique cases in the Sanskrit pronoun 
 
THE CELTIC NATIONS. 113 
 
 are derived from two similar roots 5^ — ma and 3-f — mai, 
 which, however, have no existence as distinct words in the 
 Sanskrit language. We may observe that from a root al- 
 lied to the last, the oblique eases in the European languages 
 are formed. This root is not to be found as an independent 
 word, or as a nominative case in any of those idioms of 
 which the pronouns have been hitherto compared. We 
 shall discover it in the Celtic. 
 
 The plural nominative is c|" — vai, prefixed to the above- 
 mentioned termination am. The plural oblique cases come 
 from an etymon common to all these languages, but not ex- 
 isting in any of them as a distinct word. From it we de- 
 rive the Russian nas, and v&'i and nos in Greek and Latin. 
 We shall find this etymon to be the Celtic nominative plural. 
 Asman and a/x/xe (originally aoyxe, as also fyxj^e was w/utc ?) 
 contain an epen thesis of sma. 
 
 M. Bopp supposes the endings of the cases to have been 
 formed by involved prepositions; as asmabhis, nobis, from 
 the Sanskrit preposition abhi, added to the elements of the 
 word. The same termination is to be traced in the Latin 
 nobis, and perhaps in the Greek fjfuv, which may have been 
 originally fjfutylv or ayniufriv. On this subject M. Bopp re- 
 fers to a dissertation of his own on the origin of cases, in 
 the Abhandlungen der Historisch-Philologischen Klasse der 
 K.akad. der Wissenschaften, (viz. at Berlin,) ann. 1826. 
 
 SECTION II. 
 
 Pronoun of the second person. 
 
 Singular. 
 
 Nominative. 
 
 ^TT — tuarn or twam ; tv, <tv, rvyoc; tu; tii, Russ.; 
 
 thu, M. Goth.; du, O. H. Germ. 
 
 Genitive. 
 
 rl^ — tava or n — tai ; Teo, t€v$, o-eo, &c. ; tui ; tebya, 
 
 Russ.; theina, Goth.; din, O. H. Germ. 
 
 i 
 
114 EASTERN ORIGIN OF 
 
 Dative. 
 
 *P^*1 — tubhyam and n — tai ; t\v, Teh, rot; tibi ; 
 tebe, Russ.; thus, Goth.; dir, O. H. Germ. 
 
 Accusative. 
 
 tTT*£ — tuam and f^\ | — twa ; tv, ere; te ; tebya, 
 Russ.; thuk, Goth.; dih, O. H. Germ. 
 
 Ablative. 
 
 ^r\ — tuat ; a te.' 
 
 Instrumental. 
 
 f^H I — twaya ; te ; toboyu, toboi, Russ. 
 
 Locative. 
 m U( — tuayi. 
 
 Praepositive. 
 tebe, Russ. 
 
 Dual. 
 
 Nominative. 
 2J^"TET — yuvam; v^e, a<j>£i, o-<£a>; yut, (?) Goth.; yiz. 
 iz, (?) O. H. Germ. 
 
 Genitive. 
 ^qej |H — yuvayos, TT*T — vam ; igqvara, Goth. ; 
 inchar, O. H. Germ. 
 
 Dative. 
 2J^TP"2J3T — yuvabhyam, ^ (H^ — vam ; Z^fxe k.t.X. 
 igqvis, Goth.; inch, O. H. Germ. 
 
 Accusative. 
 Sanskrit and Greek the same as the nom.; Gothic 
 and Old High German the same as the dative. 
 
 Ablative and Instrumental. 
 2jq |±3JET — yuvabhyam . 
 
 Locative. 
 ^q^JM^ — yuvayos. 
 
THE CELTIC NATIONS 115 
 
 Plural. 
 
 Nominative. 
 
 ^^IH, — yiiyani ; ^ €€ ^» fy* € $j *•*". A.; vos ; vii, Russ.; 
 yus, Goth.; ir, O. H. Germ. 
 
 Genitive. 
 ^p^n^JT — ynshmakam and ^H — vas ; tyxeajv, S^»v ; 
 vostrum, vestrnm ; vas, Russ.; i'zvara, Goth.; 
 iwar, O. H. Germ. 
 
 Dative. 
 
 ^ ^H^H, — ynshmabhyam, ^~H — vas ; l^h ; vobis ; 
 
 vam, Russ.; i'zvis, Goth.; iu, O. H. Germ. 
 
 Accusative. 
 ^H \*\ — yushman, ^T — vas ; vpeas, l[x<x$ ; vos ; 
 vas, Russ.; izvis, Goth.; iwih, O. H. Germ. 
 
 Ablative. 
 S^JTr^ — yushmat ; a vobis. 
 
 Instrumental. 
 ^^ IW[ — yushmabhis ; vobis ; vami, Russ. 
 
 Locative. 
 ^H l*J — yushmasu. 
 
 Praepositive. 
 vas, Russ. 
 
 Note. The pronouns of the second person are susceptible 
 of an analysis similar to that of the preceding, as may be 
 seen by the reader of Bopp's critical observations on these 
 pronouns in his Sanskrit Grammar. The cases of the sin- 
 gular number are formed from the elements R" — tu, (as no- 
 minative twam or tu-am,) and ^~ — tua, or ^cf — tuai. The 
 dual cases are formed from dJc| , yuva, and agree in ter- 
 minations with those of the pronoun of the first person. 
 The plural cases are formed from 7T — yu, and from c| ^, 
 vas, or vos. 
 
11G EASTERN ORIGIN OF 
 
 SECTION III. 
 
 Pronouns of the third person. 
 The pronouns of the third person are still more 
 varied and numerous in their inflections than the 
 preceding. The personal pronouns of the third 
 person, which are properly so termed, and chiefly in 
 use as such, appear to have little or no relation either 
 to the corresponding personal pronoun in Sanskrit, 
 or to the personal endings of verbs. But there are 
 some other words in these languages, which, though 
 chiefly used as demonstrative pronouns or definite 
 articles, appear to have been originally personal 
 pronouns. For example, the definite article in 
 Greek was used, as Matthiae has observed, for olrog, 
 and was in fact a pronoun. It bears also in its 
 forms a near analogy to the Sanskrit personal pro- 
 noun. The Gothic demonstrative pronoun or article 
 sa, so, thata is closely allied to both of these, and 
 all the three were apparently the same word very 
 slightly modified. The Latin pronoun approaching 
 most nearly to these is iste. I shall collate the 
 forms of all of them, that the reader may perceive 
 their affinities. It must be observed, that the chief 
 reason for selecting these rather than any other pro- 
 nouns of the third person is the circumstance, that 
 the verbal endings of the third person which have 
 been traced in the preceding chapter are perhaps 
 formed by suffixes, or abbreviations of them, and 
 are quite unconnected with those personal pronouns, 
 which in the actual state of the respective languages 
 are more regularly used as such. 
 
 Matthiae has conjectured that the primitive form 
 of what is called the definite article in Greek was 
 ro$, ry, to : but the analogy of permutations indicates 
 
THE CELTIC NATIONS. 
 
 117 
 
 the aspirate to have taken place rather of a sibilant 
 than of a dental, and it is probable that ao$, aa, to, 
 was the form which preceded the present one. Yet 
 the sigma is peculiar to the masculine and feminine 
 nominative, and the real etymon of the pronoun 
 must have been in Greek similar to the root which 
 exists in Sanskrit and the other cognate languages. 
 rfrT — tat, is the nominal root, as given by gramma- 
 rians, but the real etymon, as Professor Bopp has 
 observed, must have been rT — ta, and rT — ta, to, te, 
 and tha seem to have been the roots in the Sanskrit, 
 Greek, Latin, and Teutonic languages. Iste is pro- 
 bably a compound word formed of is the personal 
 pronoun and an old demonstrative, strikingly ana- 
 logous to the Sanskrit personal pronoun. 
 
 
 
 Singular. 
 
 
 
 Nominative. 
 
 
 Masculine. 
 
 Feminine. 
 
 Sanskrit - 
 
 r^-sas, 
 
 \ or 
 I ST:— sab. 
 
 [• *TT-sa 
 
 Greek 
 
 6 or crb 
 
 tj or aa 
 
 Gothic 
 
 sa 
 
 so 
 
 0. H. G. 
 
 der 
 
 diu 
 
 Latin 
 
 is-te 
 
 is-ta 
 
 
 riT^T— tas 
 
 Genitive. 
 
 Sanskrit 
 
 ya rlHIItf- 
 
 Greek 
 
 TO 10, TOV 
 
 TCLS, TTJ$ 
 
 Gothic 
 
 this 
 
 thizos 
 
 0. H. G. 
 
 des 
 
 dera 
 
 Latin 
 
 istius 
 
 istius 
 I 3 
 
 Neuter. 
 
 r\r[— tat 
 
 t6 
 
 thata 
 daz 
 is-tud 
 
 ■tasyas r1^| — tasya 
 roto, tov 
 this 
 des 
 istius 
 
118 
 
 
 EASTERN ORIGIN OF 
 
 
 
 
 
 Dative. 
 
 
 
 *^ 
 
 >moi 
 
 rT^f — tasyoi 
 
 ■*k 
 
 Sanskrit 
 
 rUH—tas 
 
 r\ tH — tasmoi 
 
 Greek 
 
 ro) (rwt) 
 
 
 rq, rr\, (rat) 
 
 ro) (tgh) 
 
 Gothic 
 
 thamma 
 
 
 thizai 
 
 thamma 
 
 O. H. G. 
 
 demu 
 
 
 deru 
 
 demu 
 
 Latin 
 
 isti 
 rT^T— tam 
 
 [ 
 
 istae (istai) 
 Accusative. 
 
 isti 
 
 Sanskrit 
 
 ri|^— tarn 
 
 rfr^— tat 
 
 Greek 
 
 TOV 
 
 
 rrjv, rav 
 
 V 
 
 TO 
 
 Gothic 
 
 thana 
 
 
 tho 
 
 thata 
 
 O. H. G. 
 
 den 
 
 
 dia 
 
 daz 
 
 Latin 
 
 istum 
 
 • 
 
 istam 
 Ablative. 
 
 istud 
 
 Sanskrit r\**\ \t\ — tasmat rK<£| |^ — tasyas r\&\ \T\ — tasmat 
 
 Instrumental. 
 
 Sanskrit rvT — taina rT^TT — tava <i*1 — taina. 
 
 Latin isto ista isto 
 
 Locative. 
 Sanskrit rnW^ — tasmin r1^i)|^ — tasyam rTf^T'T — tasmin 
 
 Dual. 
 
 Nominative and Accusative. 
 
 Sanskrit nT — tou H — ta i H — tai 
 
 Greek ro> ra ro> 
 
 Genitive and Dative in Greek. 
 Instrumental, Dative, and Accusative in Sanskrit. 
 
 Sanskrit rTT^TR^— tabhyam 
 
 Greek rolv tqxv roiv 
 
THE CELTIC NATIONS. 
 
 119 
 
 
 
 Plural. 
 
 
 
 
 Nominative. 
 
 
 Sanskrit 
 
 H— tai 
 
 r1|^— tas 
 
 rlM—tani 
 
 Greek 
 
 oi (crot) 
 
 at (cat) 
 
 TO. 
 
 Gothic 
 
 tMi 
 
 thos 
 
 tho 
 
 0. H. G 
 
 die 
 
 dio 
 
 diu 
 
 Latin 
 
 isti 
 
 istse 
 
 ista 
 
 
 
 Genitive. 
 ^7^7*^— tasam 
 
 ■s 
 
 Sanskrit 
 
 rj b| j *4 — taisham 
 
 r| ^ | *\ — taisham 
 
 Greek 
 
 TU>V 
 
 T<S V 
 
 t£>v 
 
 Gothic 
 
 thize 
 
 thizo 
 
 thize 
 
 0. H. G. 
 
 dero 
 
 dero 
 
 dero 
 
 Latin 
 
 istorum (istosum) 
 
 istarum (istasum) 
 
 istorum (istosum) 
 
 
 ^"^Jlf — taibhyus 
 
 Dative. 
 
 
 Sanskrit 
 
 rTP^IlT-- tabn y us H^^— taibhyu 
 
 Greek 
 
 rot? 
 
 rats 
 
 rots 
 
 Gothic 
 
 thaim 
 
 thaim 
 
 thaim 
 
 0. H. G. 
 
 dem 
 
 dem 
 
 dem 
 
 Latin 
 
 istis (istobus) 
 
 istis (istabus) 
 Accusative. 
 
 istis (istobus) 
 
 Sanskrit 
 
 r\\*\ — tan 
 
 rTT^— tas 
 
 rt M — tani 
 
 Greek 
 
 TOVS 
 
 TCLS 
 
 TCL 
 
 Gothic 
 
 thans 
 
 thos 
 
 tho 
 
 0. H. G. 
 
 die 
 
 dio 
 
 diu 
 
 Latin 
 
 istos 
 
 istas 
 
 ista 
 
 
 Ablative in 
 
 Sanskrit same as Dative. 
 
 
 
 Instrumental. 
 
 
 Sanskrit 
 
 r\ f^— tois 
 
 rnfH^T— tabhis 
 
 OF— tois 
 
 Greek 
 
 rots 
 
 rats 
 
 rots 
 
 Latin 
 
 istis 
 
 istis (istabus) 
 Locative. 
 
 istis 
 
 Sanskrit 
 
 rl N — taishu 
 
 r| |H — tasu 
 
 n^T — taishu. 
 
120 EASTERN ORIGIN OF 
 
 SECTION IV. 
 
 General observations on the preceding facts. 
 
 It will be apparent on a survey of the foregoing 
 table, that the pronouns of the third person bear as 
 near a relation to each other, in the several lan- 
 guages compared, as do those of the first and se- 
 cond. These relations, as well as the analogies dis- 
 coverable in the former tables, are chiefly evident in 
 the oblique cases, though by comparing the nomina- 
 tives merely they might be recognised. 
 
 But in the multiplicity of terminations which the 
 declension of these pronouns displays, it will be in 
 vain to look for the pronominal suffixes of the sys- 
 tem of verbs. The variety of endings precludes the 
 hope of any certain discoveries in this respect. And 
 if we confine our examination to the nominative 
 cases of the pronouns, which alone can be taken into 
 the account with strict propriety, we find only one 
 which contains exactly the ending connected with 
 the personal verb. In all the languages compared 
 in the preceding tables, the termination of the first 
 person plural is in amus, ames, ame, or am. This 
 in the older forms of the Greek language is the pro- 
 noun of the corresponding person. If in other in- 
 stances such a correspondence were discoverable, the 
 problem which refers to the actual origin of the verbal 
 inflections would be solved. But this unfortunately 
 is not the case; and hence many philological writers 
 and grammarians still deem it uncertain on what 
 principle these varieties in the endings of the verb 
 were really formed ; and those who consider them 
 as dependent upon pronominal suffixes have been 
 rather inclined to lay down this position as a 
 probable one, than as established by decisive proofs. 
 
THE CELTIC NATION-. 121 
 
 In this state of the question it is fortunate that 
 there is one idiom in which the personal pronouns, 
 as well as the verbal suffixes, have been preserved in 
 a form apparently much less altered from their ori- 
 ginal one, than in any of the more celebrated and 
 classical dialects, in which philologists have in ge- 
 neral sought the means of elucidating the structure 
 of language. I allude to the Celtic dialects, and 
 particularly to that still spoken by the Welsh peo- 
 ple, but which is found in a much more perfect 
 state in the productions of British writers coeval 
 with, or even of greater antiquity than the oldest 
 compositions of the Anglo-Saxons. The preserva- 
 tion of the pronouns in the Welsh language during 
 so long a period of time has perhaps resulted from 
 the circumstance, that in that idiom they are un- 
 declinable words, whereas in most of the European 
 dialects they are susceptible, as we have seen, of co- 
 pious inflection and variety of endings. The ter- 
 minations of words in general are but little capable 
 of change in the Celtic idioms, as indeed are those 
 idioms themselves, of which the people appear ever 
 to have been remarkably tenacious. It would per- 
 haps not be going too far to say, that no language 
 in Europe has undergone so little change in an 
 equal space of time as the Welsh sustained during 
 the centuries which intervened between Aneurin 
 and Lhywarch and the period when the sacred 
 scriptures were translated into it. To whatever cir- 
 cumstances the fact is to be attributed, it seems to 
 be certain, as I hope to make it sufficiently apparent, 
 that the Celtic idioms preserve, in a more perfect 
 state than any other languages of Europe or Asia, 
 
122 EASTERN ORIGIN OF 
 
 the original pronouns of which abbreviated forms 
 enter as suffixes into the inflections of verbs through 
 the numbers and persons. 
 
 That the resources afforded by the Celtic dialects 
 have not yet been applied to the elucidation of 
 grammatical forms in the European languages in 
 general, has arisen, as I apprehend, from the fact 
 that inquiries of this description have been pursued 
 chiefly by German scholars, who, owing to local cir- 
 cumstances, have been little acquainted with these 
 provincial idioms of the British isles. It will be my 
 endeavour, in the course of the following investiga- 
 tion, to supply the deficiency ; but before I enter 
 upon this part of my task, I shall beg leave to set 
 before my readers some passages from Professor 
 Grimm's Analysis of the Teutonic Languages, in 
 order to shew how far the inquiry respecting the 
 origin of verbal inflections has already advanced, 
 and what remains to be done, or to be attempted. 
 
 The characteristic terminations of person and 
 number in the Teutonic verbs, which, as we have 
 seen, have such endings closely analogous to those 
 of other European languages, are thus deduced by 
 Professor Grimm. 
 
 The ending of the first person singular seems, as 
 he observes, to have been originally M. This, how- 
 ever, is in many instances defective, and has been 
 more lately softened into N. The second person 
 singular is characterized by a final s ; the third 
 person by th. 
 
 The first person plural added originally to the 
 final M of the singular number an s, (with a vowel 
 interposed,) which however was gradually dropped. 
 
THE CELTIC NATIONS. 123 
 
 The second person plural had TH, like the third 
 person singular, adding perhaps an s, (with an in- 
 terposed vowel,) which was afterwards rejected. 
 
 Lastly, the third person plural had ND, of which 
 the D again is in many instances defective. 
 
 The dual seems originally to have had vs in the 
 first person, and ts in the second, and in the third 
 to have been wanting. 
 
 The imperative mood in the strongly inflected 
 conjugation entirely rejects any personal inflection, 
 and it makes the second person dual and plural, as 
 well as the first person plural, perhaps also the first 
 person dual, like the indicative, while it always 
 wants the first and third person singular. The cha- 
 racteristic of the infinitive mood is the consonant N, 
 which however is wanting in several dialects. It 
 may be remarked, that the indicative mood has the 
 personal characteristics in a more complete state 
 than the conjunctive mood, and the present tense 
 than the preterite tense. Moreover, the first and 
 third persons of the preterite invariably want the 
 characteristic consonant, and the D subjoined to the 
 N of the third person plural in the present tense is 
 wanting in the preterite, the modification already 
 induced in the verb itself, in the formation of the 
 preterite tense, being sufficient for distinguishing the 
 sense in conversation a . 
 
 In another part of his work the same writer 
 makes the following observations, with a view to 
 elucidate, as far as the languages within his scope 
 would afford opportunity, the origin of these in- 
 flections. 
 
 He says, " The personal characteristics in the 
 a Grimm's Deutsche Grammatik. th. i. p. 835 — 6. 
 
124 EASTERN ORIGIN OF 
 
 conjugation of verbs allow of a satisfactory compa- 
 rison with the personal pronouns, the relations of 
 which are blended in the idea of the verb. Some- 
 thing is really explained by this comparison. Some 
 parts of the personal pronoun destitute of gender 
 offer themselves in a striking manner : what is less 
 obvious we must endeavour to restore from the cor- 
 rupt state partly of the pronoun, and partly of the 
 verbal inflection, the variations of which have been 
 for an indefinite time increasing each in its own 
 way, without regard to the original connection be- 
 tween them. Sometimes the forms of the pronoun b 
 may be conjectured from the verb, and sometimes 
 those of the verb from the pronoun ; the third per- 
 son is for obvious reasons the most obscure, of which 
 the pronoun destitute of gender has undergone the 
 greatest change, has become defective in some cases, 
 and in some instances has been entirely lost, while 
 the pronoun of the third person having gender 
 shews no relation to the verbal inflection. The cha- 
 racteristic terminations of the third person, singular 
 and plural, viz. D and nd, appear to me quite in- 
 explicable by means of the German pronouns. The 
 M of the first person singular is more tractable. 
 From pentames in the first person plural I infer a 
 more ancient meis, instead of the Old High Ger- 
 man veis, and trace from meis, mis, wis, win. 
 The termination of the second person in th is 
 clearly related to the pronoun thu, and affords 
 room for conjecturing an older, thjus, (thyiis,) in- 
 stead of jus, for the second person plural. Lastly, 
 the characteristic endings of the dual v and ts have 
 
 b The difficult anomalies of which are observed in page 813 
 of Grimm's Grammatik. 
 
THE CELTIC NATIONS. 185 
 
 a relation to the dual pronominal forms vit and jut, 
 (originally juts). The examination of foreign lan- 
 guages anciently connected will help to support 
 these conjectures . 
 
 I shall examine whether this subject will admit 
 of further elucidation from the extant forms of the 
 Celtic verbs and pronouns. 
 
 SECTION V. 
 
 Of the Celtic pronouns. 
 
 The Celtic dialects, having no declension of the 
 pronouns, properly so termed, supplies the deficiency 
 in a manner similar to that adopted in the Hebrew 
 and other cognate languages. They have two series 
 of personal pronouns, the distinct or entire pro- 
 nouns, which are chiefly used as nominative cases, 
 or as accusatives after verbs, and a class of abbre- 
 viated pronouns used in regimen particularly after 
 prepositions, and answering the purpose of the ob- 
 lique cases of pronouns in other languages. I shall 
 first give a table of the entire pronouns, as they exist 
 in both of the principal branches of the Celtic lan- 
 guage. 
 
 Paragraph 1. 
 Entire personal pronoun in the Erse. 
 
 The entire personal pronouns in the Erse are as 
 follows : 
 
 First Person. 
 
 Me, I or me. 
 
 Sinn, we ; inn, secondary form, the initial s being 
 changed for H and at length omitted b . 
 
 c Grimm, p. 1052. 
 
 b In chap. i. sect. 1. the reader will find an explanation of 
 what is meant by the secondary forms of initial consonants. 
 
126 EASTERN ORIGIN OF 
 
 Second Person. 
 Tu, thou ; thu, i. e. t'hu, secondary form 
 Sibh, you; ibh, secondary form. 
 
 Third Person. 
 Singular Masculine. 
 Se, he ; e, secondary form. 
 
 Singular Feminine. 
 Si, she; i, secondary form. 
 
 Plural Common. 
 Siad, they ; iad, secondary form. 
 
 Paragraph 2. 
 Entire pronouns personal in Welsh. 
 First Person. . 
 Mi, I ; reduplicate form, myvi, eywye ; secondary 
 form of initial, vi and i. 
 
 Ni, we ; reduplicate form, nyni. 
 
 Second Person. 
 Ti, thou ; reduplicate form, tydi ; secondary form 
 of initial, di and thi. 
 
 Chwi, you ; reduplicate form, chwychi. 
 
 Third Person, Masculine Singular. 
 Ev, eve, evo, ve, vo, e, o. 
 All these various words occur for he and him. 
 
 Note. The Welsh translators of the holy Scriptures con- 
 sidered eve as a nominative case preceding the verb, and 
 they used ev for the accusative; but in this they are said by 
 the most learned of the Welsh grammarians to be in oppo- 
 sition as well to the common usage of the Welsh language 
 as to the authority of the old poets b . 
 
 b Antiquse Linguae Britannicse Rudimenta, auctore Joh. Da- 
 vies. Editio altera, Oxonii, 1809. p. 84. 
 
THE CELTIC NATIONS. 127 
 
 With more probability, eve and evo have been considered 
 by grammarians as reduplicate forms, the simple pronoun 
 being ev 9 or rather e. Analogy leads us to suppose that the 
 original state of this pronoun was in Welsh as it is in Erse, 
 se, and, the initial being softened, he, which was afterwards 
 written c. 
 
 Feminine Singular. 
 
 Hi, she ; reduplicated, hihi. 
 
 Note. The same rule of analogy above referred to proves 
 that hi was derived from a primitive form si, whence hi, as 
 in Erse. 
 
 Plural Common. 
 
 Hwy and hwynt ; reduplicate form, hwynt hwy. 
 
 Note. There is reason to suspect that hwy and hwynt 
 were in like manner originally swy and swynt, though this 
 ancient form is no longer extant even in Erse. But of this 
 there is no proof, but that which is afforded by analogy. 
 
 Paragraph 3. 
 
 Pronouns in regimen or Pronominal Suffixes. 
 
 Such is the entire and proper form of the personal 
 pronouns in the Celtic dialects, and they probably 
 represent a very old or the primitive state of these 
 parts of speech in the Indo-European languages. 
 It may indeed in many instances be observed, that 
 the Celtic pronouns are the nominatives from which 
 the oblique cases in those languages may be regu- 
 larly formed, whereas these cases, in several ex- 
 amples that might be adduced, have little or no affi- 
 nity to the vocables which now stand to them in the 
 relation of nominatives. The real nominatives ap- 
 pear to have been lost, and other words substituted 
 in their places, but in the Celtic, which has no de- 
 clension of pronouns, the original forms, perhaps in 
 consequence of this very circumstance, have been 
 preserved. 
 
128 EASTERN ORIGIN OF 
 
 But besides the series of forms above given, the 
 pronouns in the Celtic language are also found in a 
 state considerably modified by composition or con- 
 struction with other words. The preceding are all 
 separate and complete words by themselves ; those 
 to which I now allude are abbreviated, or modified 
 and affected in orthography by the words which are 
 immediately prefixed. I shall shew this by ex- 
 amples, and take, in the first place, the pronouns as 
 governed by and blended with some of the preposi- 
 tions c . 
 
 The following are the forms in which the per- 
 sonal pronouns appear when following the preposi- 
 tion at, to. 
 
 mi or vi ^ 
 
 
 ( av 
 
 
 as attav, to me. 
 
 ti or di 
 
 
 at 
 
 
 attat, to thee. 
 
 evo 
 
 > is changed into - 
 
 I aw and 
 ° 
 
 attaw ) 
 
 [■to him. 
 atto J 
 
 hi J 
 
 
 ti 
 
 
 atti, to her. 
 
 ni ^ 
 
 
 Tom 
 
 
 as attorn, to us. 
 
 chwi > 
 
 becomes 
 
 i och 
 
 
 attoch, to you. 
 
 hwynt J 
 
 
 (ynt 
 
 
 attynt, to them. 
 
 The preposition tan, under, changes them in a si- 
 milar manner, as 
 
 1 tanav tanom. 
 
 2 tanat tanoch. 
 
 3 tano and tani tanynt. 
 Rhwng, between changes them as follows : 
 
 1 rhyng-ov rhyng-om. 
 
 % rhyng-ot rhyng-och. 
 
 {rhyng-dho rhyng-dh-ynt or 
 
 rhyng-dhi rhyng-th-ynt. 
 
 c In what remains to be said on the subject of the pronouns, 
 I shall, to avoid perplexity, confine myself to the Welsh dialect 
 of the Celtic, premising that in the Erse dialect very nearly the 
 same facts are to be observed. 
 
THE CELTIC NATIONS 129 
 
 Yn, in, changes them thus : 
 
 vnov, in me. ynom, in us. 
 
 ynot, in thee. ynoch, in you. 
 
 yntho, or ) . . 
 yndho, J m J>thynt,or 
 
 ynthi, or ) • , } ,, . 
 
 di ' f m r " ly ndn y nt ? in tnem - 
 
 Trwy, through, alters them thus : 
 
 trwyov, trwyom. 
 
 trwyot, trwyoch. 
 
 trwydho 
 trw 
 
 ydho ) n 
 
 ydhi, 1 trwydhynt. 
 
 Wrth, by, thus : 
 
 wrthyv, wrthym. 
 
 wrthyt, wrthych. 
 
 wrtho, 
 
 wrth 
 
 . ' r wrthynt. 
 
 The preceding are all very analogous, but another 
 form occurs in the combination of the pronouns with 
 the preposition i, to, of which it is important to take 
 notice. 
 
 1. im" or ym", to me. in' or yn", to us. 
 
 2. it 1 or yt"*, to thee. iwch, to you. 
 
 3. idho, to him. idhynt, to them, 
 idhi, to her. 
 
 Nor are these mutations of the personal pronouns 
 confined to the instance of their combinations with 
 prepositions. They are thus compounded with the 
 possessive pronoun or adjective eidho, own. 
 
 eidhov, my own. eidhom, our own. 
 
 eidhot, thy own. eidhoch, your own. 
 
 eidho, his own. eidhont 
 
 eidhi, her own. eidhynt 
 
 J- their own. 
 
 The Welsh grammarians deduce analytically the 
 
 K 
 
130 EASTERN ORIGIN OF 
 
 following series of forms under which the personal 
 pronouns occur when thus modified by the preceding 
 words. 
 
 ( av, ov, yv, or m\ 
 at, ot, yt, or t\ 
 aw, o, or dho. 
 i or dhi. 
 
 ti 
 
 evo 
 
 hi Y becomes 
 
 ni 
 
 chwi 
 
 hwynt 
 
 om, ym, or n\ 
 och, ych, or ch. 
 . ynt, sometimes dhynt. 
 
 The reader can hardly fail to be struck with the 
 very obvious relation which discovers itself between 
 this series of pronouns and the personal endings of 
 the Welsh verbs, of which the different forms were 
 given in the preceding chapter. The comparison of 
 the two tables will at once prove that the termina- 
 tions of the verbs are in fact a series of pronominal 
 suffixes, and the problem which regards the origin 
 of these personal inflections may be considered as 
 solved, in so far as it regards the Welsh and the 
 other dialects of the Celtic language. There is in- 
 deed in Welsh a considerable variety in the per- 
 sonal terminations of the verbs, and this may be 
 supposed with probability to have been a conse- 
 quence of the poverty of the Celtic language in re- 
 spejct to the conjugations in temporal and modal in- 
 flections, or in those changes by which the differences 
 of mood and tense are indicated. In these modifi- 
 cations the Celtic has fewer resources than many 
 other languages ; and it was probably found neces- 
 sary to supply the deficiency by a considerable va- 
 riety in the personal endings, which in some mea- 
 sure help to characterise the tenses. There is not, 
 however, in these a greater diversity than among 
 
 - I 
 
THE CELTIC NATIONS. 131 
 
 the abbreviated pronouns, and nearly all the verbal 
 terminations are to be found in the preceding table. 
 Tli is I shall now shew by a comparison of the verbal 
 endings with the pronouns. 
 
 Paragraph 4. 
 
 Comparison of the personal endings of verbs with 
 the contracted forms of the pronouns. 
 It may be remembered that in a former section 
 the personal endings of the verbs in the Welsh lan- 
 guage were said to be reducible for the most part to 
 four, or rather three principal forms. These are as 
 follows : 
 
 First form. Second form. Third form. 
 
 Sing. Sing. Sing. 
 
 1. -av -ais -wn 
 
 2. -i -aist -it 
 
 3. root simply -odh -ai. 
 
 Plur. Plur. Plur. 
 
 1. -wn -om -em or ym 
 
 2. -wch -och ~ech or ych 
 
 3. -ant -ont -ent or ynt. 
 
 If the reader will only compare this table with 
 that of abbreviated pronouns contained in the end of 
 the last paragraph he will perceive at once their re- 
 lation. 
 
 The plural terminations are precisely the pro- 
 nouns. The first set presents the greatest variety, 
 but even these are traced among the pronouns ; in' or 
 yn' and iwch, being the forms which the pronouns ni 
 and chwi assume after the proposition i. The first 
 of these, in' or yn', seems a more natural change of 
 ni, than the more usual om or ym, which is so re- 
 
 K 2 
 
132 EASTERN ORIGIN OF 
 
 mote from ni as to give rise to a suspicion that the 
 Welsh language had once a pronoun resembling the 
 «/xe or ape$ of the Greeks, and that this has been lost, 
 notwithstanding the permanent character of the Cel- 
 tic dialects. 
 
 The personal endings in the singular number are 
 more various, but they are still analogous to the ab- 
 breviated pronouns. In the first form for example, 
 which is that of the future or present tense, the first 
 person has the ending in av, which the pronoun mi 
 or vi generally assumes in regimen, as above shewn. 
 The ais of the second form is not pronominal, but an 
 inflection characteristic of the tense, the syllable ais 
 or as being introduced in the past tenses of the Cel- 
 tic verb, nearly as the od or ed in the Teutonic con- 
 jugations ; it is brought in before the pronominal 
 termination, as in the plurals carasom, carasoch, 
 carasant. The second person, in two out of the 
 three forms, has the abbreviated pronoun as a suffix, 
 either in it or t. In the first form, the ending i 9 
 though it does not appear among the abbreviated 
 pronouns, is the termination of the separate pronoun 
 of the second person, and this is therefore probably a 
 suffix. The third person is differently constituted. 
 In the first form of the verb, as in carav, ceri, car, 
 from the word caru to love, or in bydhav, bydhi, 
 bydh, from the verb substantive, the third person is 
 merely the verbal root used, as in the Semitic lan- 
 guages, without any suffix. In the third form ai 
 was perhaps ae, and derived from eve, or its mo- 
 dification as used in regimen e. The ending in odJi 
 seems anomalous in the Welsh language, though it 
 nearly resembles the termination of the third person 
 
THE CELTIC NATIONS. 133 
 
 in other idioms, as the Teutonic aith, or ot*. The 
 Welsh suffix, if formed regularly from the pronoun, 
 would be in o or aw ; and this actually occurs in the 
 future tense of the subjunctive mood, which does not 
 fall under either of the forms above stated, but has 
 the singular number thus : 
 
 1. bydhwyv, 2. bydhych, 3. bydho. 
 
 Paragraph 5. 
 General result in respect to the Celtic verbs. 
 
 On a review of this analysis it appears clearly that 
 the Welsh verbal terminations are in general merely 
 abbreviated or modified pronouns, affixed to the 
 verbal roots ; and this conclusion does not rest merely 
 upon a probable conjecture, on which the gramma- 
 rians of other Indo-European languages have been 
 obliged to found it, but on the more substantial fact, 
 that the very terminations in question are actually 
 to be identified with the pronouns as they are used 
 on other occasions in an abbreviated form. 
 
 SECTION VI. 
 
 Conclusions respecting the personal inflections of verbs in the 
 other Indo-European languages. 
 
 As it has been, I trust, satisfactorily proved that 
 the inflection of verbs in the Welsh language con- 
 sists in the addition of pronominal suffixes to the 
 verbal roots, and as in a former section sufficient evi- 
 dence appeared, of the affinity and original sameness 
 of the verbal inflections in all these languages, we are 
 entitled to infer without hesitation, that in the other 
 
 a Is it the abbreviated form of the pronoun (dho) reversed ? 
 K 3 
 
134 EASTERN ORIGIN OF 
 
 languages which helong to this stock, the verbs are 
 inflected on the same principle, and that although 
 in many instances they are no longer extant, pro- 
 nouns formerly existed in all these idioms more or 
 less analogous to the Welsh pronouns. 
 
 It will be worth while to go a little more mi- 
 nutely into this consideration. 
 
 1. The pronoun of the third person plural in 
 Welsh is hwynt in the entire form, and ynt in the 
 contracted one, which as a verbal suffix is ynt, ent, 
 ont, ant. In the other languages the terminations 
 of the verb are as follows : 
 
 In Greek, ovti, em, av, ovto, k. t. A. 
 In Sanskrit, anti, unt, an. 
 In Latin, ant, ent, unt, anto, ento, &c. 
 In Teutonic, and, aina, ont, ant, on, &c. 
 
 These languages have no personal pronoun now 
 extant similar to hwynt or ynt ; but, from the consi- 
 derations above adverted to, it is probable that such 
 a pronoun existed in them. 
 
 2. The Welsh separate pronoun of the first person 
 plural is ni, which is to be recognised in other lan- 
 guages in the oblique, if not in the nominative cases b . 
 The contracted form of this pronoun in n' enters 
 into some of the Welsh tenses as a suffix, but most 
 of them have the other Welsh pronoun of this per- 
 son, om or ym. This, as we have observed, can 
 scarcely be derived from ni, but rather comes from 
 some separate pronoun originally common to all 
 
 a Viz. in Sanskrit, *1 1 — nau. 
 
 in Greek, vai 
 
 in Latin, nos 
 
 in Russian, nas. 
 
THE CELTIC NATIONS. 135 
 
 these languages, which must have been analogous to 
 the Greek ape or a/xe^, or perhaps a plural formed 
 from the nominative singular mi. However this may 
 have been, the termination, am, em, ym is really a 
 contracted pronoun in the Welsh language, and 
 must have existed as such in the cognate idioms. 
 The following endings may therefore be regarded as 
 pronominal suffixes : 
 
 In Greek, apeg, opes. 
 
 In Sanskrit, amah, or amas, am. 
 
 In Latin, amus, emus, imus, umus. 
 
 In Teutonic, ames, omes, aima, am. 
 
 3. The separate pronoun of the second person 
 plural in Welsh is chwi, and the abbreviated one 
 och or yen, which, as we have seen, is also the suffix 
 in the endings of verbs for this person. All the 
 other Indo-European languages have a dental con- 
 sonant in the place of the Welsh guttural or palatine 
 letter, as in the 
 
 Teutonic dialects, aith, ith, uth, ot, et. 
 
 In Sanskrit, at'ha, t'ha, or ta. 
 
 In Greek, are, ere, re, 
 
 In Latin, atis, ate, etis, ete, itis, ite. 
 
 What the separate pronoun was in these languages 
 from which the termination of the verb is contracted, 
 we can scarcely hope to discover ; but the fact being- 
 proved that the Celtic verb is here formed by means 
 of a pronominal suffix, we may infer from analogy 
 that the same construction holds in the other lan- 
 guages. 
 
 4. We have seen that the separate pronoun of the 
 first person singular in Welsh is mi or vi, and the 
 
 K 4 
 
136 EASTERN ORIGIN OF 
 
 constructive pronoun av, ov, yv, or m\ The verbal 
 suffix is av or yv: in the Erse dialect it is am, aim, or 
 im. In most other languages m is the characteristic 
 consonant of this person, with or without a subse- 
 quent vowel, as, 
 
 In Greek, fxi, as eifjj, ti8yj[jli. 
 
 In Sanskrit, mi, or m, as bhavami, abhavishyam. 
 
 In Latin, m, as inquam, sum, amabam. 
 
 Although the pronouns extant in these languages 
 do not come so near to the above terminations 
 as the Welsh mi, vi, and m', still they may account 
 for it tolerably well. 
 
 In Greek and in Latin, the eyw or ego probably 
 gave rise to the ending of verbs in o, which is per- 
 haps a later form than the termination in mi. 
 
 In other instances the first person singular has no 
 addition to the simple verb, or to the common cha- 
 racteristic of the tense. The verb was used in this 
 state either with the separate pronoun or without 
 any. The other persons are marked by characteris- 
 tic additions, and it was sufficient for the first to be 
 without any suffix. But while the Teutonic dialects 
 have the first person in this state, the Celtic dialects, 
 like the Semitic languages, have the third person 
 most frequently in the simple state, or in that which 
 is nearest to the verbal root. 
 
 b This is contrary to the opinion of some eminent gramma- 
 rians, who regard the form in /xi as more recent than that in o. 
 Before this opinion can be allowed to be probable, some answer 
 must be given to the question, How it can have happened that 
 the newer forms in the Greek language should resemble those 
 of the Sanskrit so much more than the older ones, as they would 
 do on the hypothesis? 
 
THE CELTIC NATIONS. 137 
 
 5. The second and third persons singular end as 
 
 follows. 
 
 
 Second person. 
 
 Third person. 
 
 Greek, 
 
 at or ?, 
 
 ■n, to, or €i. 
 
 Sanskrit, 
 
 si or s, 
 
 ti, or t. 
 
 Latin, 
 
 s. 
 
 t. 
 
 Teutonic, 
 
 ais, is, es. 
 
 eith, eth, t. 
 
 Here there seems to be, at the first view of the 
 subject, an interchange of pronouns; for the pronoun 
 of the second person, in its usual state, was in all 
 these languages nearly the same as in Welsh. It 
 was tv, tu, twam, (i. e. radically tu,) thu, in Greek, 
 Latin, Sanskrit, and Gothic respectively : and the 
 pronoun of the third person is sah, and sa in San- 
 skrit and Gothic. But we may observe in the first 
 place that the original form of the third person was 
 in all these languages to, tali, te, or at least a t with a 
 vowel adjoined. This is indicated by the analogy 
 of the neuter gender and the oblique cases. The 
 Greek was originally, 
 
 Tog, Ta, to. 
 
 > te, ta, tud, 
 
 The Latin with 
 is prefixed ( ? ) 
 
 r™ o. i tos or ) 
 
 The Sanskrit, , h ta, tot, or tud. 
 
 The Gothic, tha, tho, tha, or thata. 
 
 A modification of these pronouns, according to the 
 rule adopted in the other persons, would produce the 
 endings of verbs in the third person singular exactly 
 as they are above laid down. 
 
 In those instances in which the third person of the 
 verb has an ending in a vowel, we may account for 
 the peculiarity by supposing, either that the suffix 
 
138 EASTERN ORIGIN OF 
 
 has been omitted, as it was above shewn to be 
 in some languages, or that a contracted pronoun, 
 akin to the e of the Welsh or the e of the Greek 
 language, has been used. 
 
 The personal pronouns of the second and third 
 persons are so nearly alike, that it was found neces- 
 sary to distinguish the verb connected with each by 
 some discriminating mark ; and this was easily done 
 by taking a form of one personal pronoun, which 
 was perhaps originally only a dialectic difference, 
 but in which the sibilant consonant is substituted 
 for a dental one. It is well known that $ and * are 
 easily interchanged, as when the Greek a a- is trans- 
 muted into tt, and in the present instance tv has 
 been actually changed for crv. There being two 
 forms of the pronoun, a sibilant and a dental one, 
 the former was preferred for the characteristic of 
 the second person in those instances in which the 
 dental had been appropriated to the third person. 
 
 We shall endeavour in the sequel to make some 
 of these remarks more certain and explicit, by an 
 examination of particular tenses and a comparison 
 of the different forms of verbs in the several lan- 
 guages of the same stock. 
 
THE CELTIC NATIONS. 1559 
 
 CHAPTER VI. 
 
 Of the Inflection of Verbs through Tenses and Moods. 
 Section I. General view of the subject. 
 
 X HE observations comprised in the two last chap- 
 ters relate merely to the personal endings of verbs, 
 or to those inflections which serve to distinguish 
 their different persons and numbers. The modify- 
 ing principles, on which depends the discrimination 
 of moods and tenses, yet remain to be analysed and 
 compared. These are two distinct subjects of in- 
 quiry. I have been induced to enter into the former 
 at some length for two reasons. The principal of 
 these is, the convincing proof which the inflections 
 already surveyed appear to furnish of a deeply- 
 rooted affinity between the Celtic dialects and the 
 other languages of Europe and Asia which have 
 been compared with them. Another motive has 
 been the hope of throwing some light on the gram- 
 matical principles governing the inflection of verbs in 
 all these idioms. How far this attempt has been 
 successful my readers must judge. I ought now to 
 proceed to the more arduous task of examining the 
 structure of verbs through their different moods and 
 tenses, and of tracing the relations which the latter 
 bear to each other in different languages. But this 
 endeavour is in the outset obstructed by great, and, 
 I fear, as yet hardly surmountable difficulties. The 
 structure of the Teutonic languages, and the analo- 
 gies of these to the Greek, Latin, and Sanskrit, have 
 already occupied the attention of several accurate 
 
140 EASTERN ORIGIN OF 
 
 and ingenious writers, who have been mentioned in 
 the preceding pages ; but the Celtic dialects may be 
 said to furnish a new and almost unexplored field. 
 The attempts which have been made to elucidate 
 their etymology have been, with very few exceptions, 
 remarkably unsuccessful ; and it will perhaps be long 
 before any person, possessed of the requisite opportu- 
 nities for performing this task in the best manner, 
 may be found ready to undertake it. In this de- 
 ficiency of materials, and in the want of any correct 
 arrangement of such as exist, I am aware that I 
 enter upon the remainder of my inquiry under very 
 unfavourable circumstances. I am quite unable to 
 proceed in the investigation of the Celtic language 
 and its affinities with that accuracy and lucid ana- 
 lysis, which Professors Bopp and Grimm have ap- 
 plied to the idioms which they have examined and 
 compared. Perhaps, indeed, the subject itself does 
 not admit of such illustration. I expect, however, 
 to furnish proofs which shall be deemed a sufficient 
 groundwork for the inferences to be founded upon 
 them. 
 
 I shall enter upon this part of my subject, as in 
 the former instance, by examining the particular 
 features in other idioms, which I mean afterwards 
 to compare with those of the Celtic dialects. 
 
 SECTION II. 
 
 Modifications of Verbs common to the Sanskrit and Greek. 
 
 The most striking and extensive marks of rela- 
 tionship are to be traced between the Sanskrit con- 
 
THE CELTIC NATIONS. 141 
 
 jugations of verbs and the Latin and Greek, but 
 particularly between the Sanskrit verbs and the 
 Greek verbs in pi. The Sanskrit verbs may in- 
 deed be said to be governed by the same laws of 
 conjugation as the latter. But, in making this re- 
 mark, we must distinguish three series or different 
 sets of Greek verbs in /x<, and allow the two former 
 to constitute in some respects an exception to this 
 analogy, though in another point of view they will 
 be found to confirm it. The first are those verbs 
 which, besides the characteristic endings of this con- 
 jugation, have also a reduplication of the first syl- 
 lable, or an addition which is a substitute for one, as 
 the verbs tiSyj^i, &/&»jk/, and Janifu. There are verbs 
 in Sanskrit which have a similar reduplication : it 
 is not, however, a general character, but the mark 
 of a particular conjugation, which ranks as the third 
 in the series of ten classes. Verbs of the third con- 
 jugation reduplicate the first consonant, or make 
 some equivalent prefix. Thus from the verbal 
 root da, to give, comes the verb dadami, dadasi, da- 
 dati, answering to &/&«/*/, $/&»$-, &/&a>o-/ or &/W/, which 
 prefixes this reduplication through the four first 
 tenses formed like the three first in Greek from the 
 present tense, but has no remarkable peculiarity in 
 the other parts of the verb. Many other Sanskrit 
 verbs, which do not belong to the third conju- 
 gation, undergo a modification not unlike that of 
 the root of lo-Typi. Thus from the root ^7 — 
 sht'ha, to stand, comes the present tense tisht'hami, 
 tisht'hasi, tisht'hati. A second class of Greek verbs 
 in pi insert the syllable w between the verbal root 
 and the personal endings, as the verb ^evy-w-fja. 
 There is likewise a particular class of Sanskrit verbs 
 
142 
 
 EASTERN ORIGIN OF 
 
 analogous to these, and having the same distinction, 
 which is considered as the character of the fifth 
 conjugation : but these are, as in Greek, a compa- 
 ratively small number of verbs. Those Greek verbs 
 in pi, however, which bear the nearest resemblance 
 to the generality of Sanskrit verbs, are such as make 
 no prefix to the initial of the root, nor any insertion, 
 but merely add the personal termination. Among 
 these we reckon </>^/x/, dico ; ic^p, scio ; elfxi, sum ; 
 and elpi, vado. Some of the same class are obsolete 
 in the present tense, and only used in the aorist, as 
 fiyfjii, proficiscor, and yypypi senesco; and this last may 
 be compared to the Sanskrit verb jarami, of the 
 same meaning. 
 
 Doric forms. 
 
 yvjpaTi. 
 
 
 Sanskrit. 
 
 Greek. 
 
 Si. 
 
 jarami 
 
 ytpyfAi 
 
 
 jarasi 
 
 7W 
 
 
 jarati 
 
 yyjpvja-i 
 
 Du 
 
 . jaravas 
 
 
 
 jarat'has 
 
 yvjpxTov 
 
 
 jaratas 
 
 y/jpocTOV 
 
 PL 
 
 jaramas 
 
 yYjpafxev 
 
 
 jarat'ha 
 
 yvjpaTt 
 
 
 jaranti 
 
 yv\pa<Ji 
 
 yripapeg (?) 
 
 yypaVTi. 
 
 Of the Prceterita Augmentata, or Pretet imperfect 
 Tense and Aorists. 
 
 There are two preterite tenses in Sanskrit verbs, 
 which are deserving of particular notice, as they are 
 formed in a manner very similar to that of two 
 tenses of the Greek verb. One of them is analogous 
 to the imperfect, and the other to the aorists; and 
 
THE CELTIC NATIONS. 143 
 
 there is no reason that forbids their being distin- 
 guished by these terms. 
 
 1. The imperfect is formed from the present tense 
 by prefixing an augment, and abbreviating the per- 
 sonal endings. The augment is the first short vowel 
 3f — a, which corresponding with the short vowels 
 of the Greeks might be represented indifferently by 
 a or e. Thus are formed from 
 
 tudami "| f atudam 
 
 tudasi Y <J atudas 
 
 tudati J [atudat. 
 
 2. The aorist has three, or rather, according to 
 professor Bopp's division, seven forms. Of these it 
 is observed, that the four first agree more or less 
 closely with the Greek first aorist, the fifth and sixth 
 with the second aorist, and the seventh, which, be- 
 sides an augment, admits a reduplication of the first 
 syllable, with the preterpluperfect. Thus in the 
 four first some make this tense by inserting s, or the 
 syllable is, or sa, or sas, between the root and the 
 personal endings, and by prefixing the augment to 
 the root, the vowel of which undergoes a change by 
 the forms guna and vriddhis. The root kship, pre- 
 sent tense kshipami, makes the aorist akshoipsam. 
 This is one of the examples given by Bopp, and the 
 analogy is more striking, if the words are written as 
 the Greeks would have written them, thus ; 
 
 Root. Present tense. Aorist. 
 
 The two succeeding forms of the aorist differ from 
 the imperfect tense very nearly in the same manner 
 in which the second aorist in Greek differs from the 
 Greek imperfect. 
 
144 EASTERN ORIGIN OF 
 
 We shall find the insertion of s, sa, or as, to be a 
 method used in other instances for the formation of 
 tenses with a past signification. 
 
 SECTION III. 
 
 Forms common to Greek, Latin, and Sanskrit. 
 The preceding modifications of the verb are pecu- 
 liar to the Sanskrit and the Greek languages. The 
 Latin has an imperfect resembling the Greek in 
 meaning, though formed by a totally different mode 
 of inflection, but it has nothing analogous to the 
 aorist. All these three languages agree in the cir- 
 cumstance, that they frequently modify the present 
 tense by an insertion of particular consonants be- 
 tween the verbal root and the personal endings, 
 which consonants are dropped in the further conju- 
 gation of the verb. Such are the ktk 9 or isc, in such 
 verbs as avaA/Wy and frigesco, the av, in alMvo^ai, 
 afxapTavw, and the numerous verbs resembling them, 
 which are analogous to the verbs of the ninth con- 
 jugation in Sanskrit : the latter insert na between 
 the root and the termination in the present tense a . 
 These insertions are retained only in those tenses 
 derived immediately from the present, as the imper- 
 fect is in Latin and Greek : they are dropped in the 
 preterperfect and other forms of the verb. 
 
 The Teutonic language wants all these and many 
 other variations : it has no tense formed by a mo- 
 dification of the present. " The capability of flection 
 " in the German verbs seems," says professor Grimm, 
 " to have been greatly impaired. Of the passive 
 
 a As krinami, krinasi, &c. from the root kri, pret. chi-kri-ya. 
 
THE CELTIC NATIONS. 145 
 
 *' voice the last remains disappear with the Moeso- 
 " Gothic : the middle voice is every where wanting, 
 u with the exception of a reflected form in the old 
 " northern dialect, which is in some degree analogous 
 " to a middle voice. Four moods exist ; the infini- 
 " tive, imperative, indicative, conjunctive, but there 
 " is no optative. What is most to be regretted is the 
 u loss of many tenses : only a present tense and a pre- 
 " terite remain to us : the future, and all the different 
 " degrees of the past signification, can no longer be 
 " expressed by a change in the form of the verb it- 
 " self b ." 
 
 SECTION IV. 
 
 Formation of the preterperfect in Greek, Latin, Sanskrit, and 
 Teutonic verbs. 
 The preterperfect seems originally to have been 
 formed on the same principle in the Greek, Latin, 
 Sanskrit and Teutonic languages. Many changes 
 in the forms of verbs have been produced by the 
 addition of auxiliaries, or of particles inserted in or 
 added to the root, but the preterperfect, in that me- 
 thod of conjugation which appears to have been the 
 primitive one, is an inflection properly so termed. 
 A partial repetition of the verbal root itself seems to 
 have been originally adopted to denote a past signifi- 
 cation, implying the act to have been done and com- 
 pleted. 
 
 Paragraph 1. 
 
 Preterite in Teutonic Verbs. 
 
 The preterite of the strongly inflected conjugation, 
 
 says Dr. Grimm, " must be considered as a chief 
 
 " beauty of our language, as a character intimately 
 
 b Grimm's Deutsche Grammatik. Theil. I. 
 
 L 
 
146 EASTERN ORIGIN OF 
 
 " connected with its antiquity and its whole consti- 
 " tution. Independently of the inflection in the 
 " endings, of which the nature has already been 
 " pointed out, it affects the roots themselves, and 
 " that in a double manner ; either the beginning 
 " of the root is repeated before itself, or the vowel 
 " of the root, whether initial or medial, is modi- 
 " fied. The Gothic language yet retains both me- 
 " thods, it reduplicates and modifies : sometimes 
 " it applies both methods at once. Reduplication 
 " never affects the terminating consonants of the 
 " root. In the other Teutonic dialects reduplication 
 " disappears, if we except slight traces, and instead 
 " of it an unorganized diphthong has been formed, 
 " the doubling of the consonant being no more 
 " thought of. 
 
 " The reduplicating conjugation leaves the vowel 
 " sound of the root unaltered, and only puts the 
 " doubled syllable before the singular and plural of 
 " the preterite both indicative and conjunctive, but 
 " not before the participle. The modifying conju- 
 " gation never leaves the vowel of the present tense 
 " unaltered in the preterite. On this principle are 
 " formed six reduplicating conjugations, and six of 
 " the latter description." I must refer the reader 
 for further particulars to Dr. Grimm's work, and 
 shall here give merely an example of each of these 
 twelve conjugations, all belonging to the strong or 
 primitive Gothic inflection. Under each verb are in- 
 serted the present indicative, the preterite first per- 
 son singular and first person plural, and the par- 
 ticiple. 
 1. Conjugation. Salta — salio, I leap. 
 
 Salta ; pret. saisalt, saisaltum ; saltans. 
 
THE CELTIC NATIONS. 147 
 
 & Maita ; abstido, I cut. 
 
 Maita; pret. mairaait, maimaitum ; maitans. 
 
 3. Hlaupa ; curro, I run. 
 
 Hlaupa ; pret. hlaihlaup, hlaihlaupum ; hlaupans. 
 
 4. Slepa, dormio, I sleep. 
 
 Slepa ; pret. saislep, saislepum ; slepans. 
 .5. Laia, irrideo, I laugh. 
 
 Laia ; pret. lailo, lailoum ; laians. 
 
 6. Greta, ploro, I weep or grete, Scottish. 
 Greta ; pret. gaigrot, gaigrotum ; gretans. 
 
 Sixth Conjugation without reduplication. 
 
 7. Svara, juro, I swear. 
 
 Svara; pret. svor, svorum ; svarans. 
 
 8. Skeina, luceo, I shine. 
 
 Skeina; pret. skain, skinum ; skinans. 
 
 9. Liuga, mentior, I lie. 
 
 Liuga ; pret. laug, lugum ; lugans. 
 
 10. Giba, do, I give. 
 
 Giba ; pret. gab, gebum ; gibans. 
 
 11. Stila, furor, I steal. 
 
 Stila ; pret. stal, stelum ; stalans. 
 
 12. Hilpa, adjuvo, I help. 
 
 Hilpa ; pret. halp, hulpum ; hulpans. 
 
 I have taken these examples in full, because the 
 analogies which they display to the forms of our 
 tongue are interesting to English readers. The 
 Moeso-Gothic is far more perfect in its inflections 
 than any other language of the same stock. In the 
 later dialects they gradually fade away, but still the 
 remains are reducible to the same system. 
 
 The weaker form of verbal inflections is, according 
 to Grimm, modern in comparison of the stronger 
 form. For the grounds of this opinion I must refer 
 
 L 2 
 
148 EASTERN ORIGIN OF 
 
 the reader to his work a . The preterite is formed 
 by the insertion of a syllable of which d is the con- 
 sonant, and this is regarded by Grimm as an abbre- 
 viated auxiliary, derived from the verb to do, which 
 in the Old High German is tuon. Thus sokida, I 
 sought, is I "seek — did." The inflections of this verb 
 tuon are very distinct in the Old High German, and 
 appear to give some probability to Dr. Grimm's con- 
 jecture as to the origin of the preterperfect tense, 
 in the form which has become most frequently used 
 in the modern German and English languages b . 
 
 Paragraph 2. 
 
 Preterperfect tense in Greek and Sanskrit 
 
 The formation of the preterperfect tense in Greek 
 and in Sanskrit is on principles so similar, that it 
 requires more care to sum up the points in which 
 they differ than those in which they agree. In both, 
 the root, which frequently consists of one syllable, is 
 preserved nearly in its entire state, with a final 
 short vowel added to it, and a short syllable gene- 
 rally prefixed, which is termed the reduplication. In 
 Greek the vowel of the reduplication is always e, but 
 in Sanskrit the vowel is the short one corresponding 
 with that of the verbal root : thus from the root 
 rJ4, — TV7r > which forms a verb of nearly the same 
 meaning in Greek and in Sanskrit, the preterite 
 is in one language rervyra, and in the other tii- 
 topa. The reduplicated consonant in Greek is a te- 
 nuis, and in Sanskrit the tenuis or media, when 
 the root begins with an aspirate, and when double 
 consonants are the initials, the first is alone redupli- 
 
 a Deutsche Grammatik. th. i. p. 1040. 
 
 b See p. 885, 1039, 1042, of Grimm's Deutsche Grammatik. 
 
THE CELTIC NATIONS. 149 
 
 cated. In this last case the two languages just 
 mentioned have consulted euphony more than either 
 the Latin or Teutonic, both of which repeat the double 
 consonants. The principal vowel of the root under- 
 goes a change in the preterperfect, which in Sanskrit 
 is termed guna and vriddhis. This is analogous to 
 the corresponding change in the Greek old preter- 
 perfect, in such forms as fAepova and ol^a, and to the 
 changes above alluded to in the Teutonic and Latin 
 preterites. 
 
 The following words will serve to exemplify this 
 form of Sanskrit verbs : 
 
 Root. Third person Present. Preterite. 
 
 bhri bharati babhara. 
 
 tri tarati tatara. 
 
 tup tupati tiitopa. 
 
 Paragraph 3. 
 Of the preterperfect in Latin verbs. 
 Many Latin verbs form the preterperfect by re- 
 duplication, and there is reason to believe that the 
 number was originally much more considerable. 
 We may regard it as probable, that this was the 
 oldest form of the preterperfect tense in the Latin 
 language, as well as in the cognate idioms. 
 
 In some respects the Latin reduplicated preterite 
 agrees with the Teutonic, in others with the San- 
 skrit, and with both more nearly than with the 
 Greek. 
 
 1. Aspirate and double consonants are redupli- 
 cated in Latin as in Moeso-Gothic verbs ; this is 
 avoided in Greek and Sanskrit. In the Moeso-Gothic 
 we have from the verb sJcaidan, scheiden, to sepa- 
 rate, the following : 
 
 Skaida ; skaiskaid, skaiskaidum ; skaidans. 
 L 3 
 
150 EASTERN ORIGIN OF 
 
 In Latin we have in like manner : 
 sciscidi from scindo. 
 spondeo spospondi not sospondi. 
 
 fallo fefelli not Trefelli. 
 
 2. The vowel of the reduplicated syllable is in 
 Greek always e, in Moeso-Gothic ai. In Latin as 
 well as in Sanskrit verbs the vowel of the verbal 
 root is reduplicated ; as in Sanskrit, the verbal root 
 mad, makes mamada, 
 
 lish 
 
 lilaisha, 
 
 tup 
 
 tutopa, 
 
 in Latin, 
 
 
 pedo 
 
 pepedi, 
 
 mordeo 
 
 momordi, 
 
 tundo 
 
 tutudi, 
 
 curro 
 
 cucurri c , 
 
 In Latin however the reduplicated syllable fol- 
 lows the quantity of the verbal root ; in Sanskrit it 
 is always short, whatever may be the quantity of the 
 root. 
 
 The following are some of the examples of redu- 
 plication yet remaining in Latin verbs. 
 
 memim, 1. 
 
 e. memem. 
 
 spospondi. 
 
 momordi. 
 
 
 pepedi. 
 
 peperi. 
 pupugi. 
 
 
 poposci. 
 sciscidi. 
 
 tetigi. 
 
 totondi. 
 
 cucurri. 
 
 
 pepigi, i. e. pepegi 
 
 pepuli. 
 
 cecidi. 
 
 cecini, i. e. 
 
 ceceni. 
 
 cecidi. 
 
 didici. 
 
 
 tutudi. 
 
 fefelli. 
 
 
 
 c The apparent exceptions from this rule seem to admit of ex- 
 planation. See Grimm, th. i. p. 1055. 
 
 d Pango and Tvr)yvv^i being variations of Trrjyco. 
 
THE CELTIC NATIONS. 151 
 
 Dr. Grimm has remarked that verbs which change 
 a short vowel in the root or present tense into a long 
 e in the preterperfect had originally a reduplication. 
 Pango, or rather pago, makes pepigi, but compingo 
 makes compegi. This proves the analogy of the two 
 forms ; and on the model of pago, pepegi, contracted 
 to pegi, we have 
 
 
 capio, 
 
 cepi. 
 
 
 ago, 
 
 egi. 
 
 frango, i 
 
 i. e. frago, 
 
 fregi 
 
 
 facio, 
 
 feci. 
 
 
 jacio, 
 
 jeci. 
 
 
 lego, 
 
 legi. 
 
 
 emo, 
 
 emi. 
 
 
 venio, 
 
 veni. 
 
 
 edo, 
 
 edi. 
 
 
 sedeo, 
 
 sedi. 
 
 
 fugio, 
 
 fugi. 
 
 It is observed, in confirmation of this remark, that 
 these verbs have in many instances a reduplication, 
 or, what is allied to it, an internal inflection, in the 
 cognate languages ; as cepi resembles hqf in Gothic; 
 fugi, 7r€(f)vya; legi, \e\oya; and venio, veni, the Moeso- 
 Gothic verb which is analogous to come and came. 
 
 The custom of reduplication in forming the pre- 
 terite fell into disuse ; supplementary methods were 
 found to answer the same purpose, the principal of 
 which were the following. 
 
 1. The insertion of the letter s before the final i, 
 as in 
 
 repo, 
 
 reps; 
 
 lego, 
 
 lexi. 
 
 
 L 4 
 
152 EASTERN ORIGIN OF 
 
 2. The insertion of the syllable av, Iv, mutable 
 into u, as 
 
 am-av-i. 
 
 aud-iv-i. 
 
 doc-u-i. 
 The former of these methods is analogous to an 
 inflection, of which we trace an extensive use, in the 
 cognate languages ; the latter is quite peculiar to 
 the Latin. The av has been thought to be allied to 
 the bo and bam of the future and imperfect preterite. 
 
 SECTION V. 
 
 Of the remaining forms of the Verb — Potential, Optative, and 
 Conjunctive moods — Future tenses — Middle and Passive 
 voices. 
 
 Most of the remaining forms of the verb appear 
 to be simple inflections properly so termed, and not, 
 as some have suspected, compound words. From this 
 remark we must make an exception, as far as regards 
 the pronominal suffixes, on which so much has al- 
 ready been said ; for these are, as the reader is well 
 aware, abbreviated words brought into composition 
 with the verbs. With this exception, the moods and 
 tenses of verbs which are now to be considered may 
 be looked upon as formed in all probability by simple 
 inflection. There are indeed some of these forms 
 which have been thought by late writers to have 
 derived their peculiar shades of meaning, in relation 
 to time and mode, from the insertion or addition of 
 significant particles, or other words of a similar use a . 
 
 a I allude to Professor Bopp's opinion and to some other si- 
 milar conjectures. According to Bopp's, the future tenses are 
 compounds of a verbal root, or of an attributive vocable and 
 
THE CELTIC NATIONS. 153 
 
 But the instances in which this can be supposed 
 with any degree of probability are, in comparison 
 with others, very few, and the inference with respect 
 to them is but doubtful at best. And in by far the 
 greater number of examples composition of words 
 seems to be out of the question ; and it is evident 
 that a mere inflection has been employed, the ori- 
 ginal or simple verb having been by design some- 
 what modified in pronunciation, or by the addition 
 or insertion of a consonant or vowel, so as to impart 
 to it a sense in some mode or circumstance different 
 from the primitive one. 
 
 Thus it has often been remarked, that it is a cha- 
 racter common to the conjunctive, potential, and op- 
 tative forms of the verb to change proper vowels, 
 and especially short vowels, into diphthongs. On 
 this fact a somewhat whimsical theory has been 
 founded by the learned and fanciful Professor Mur- 
 ray. He~says, " The subjunctive of all Greek, Latin, 
 " and Teutonic verbs arose from laying an emphasis 
 " expressive of the conditional state of the mind on 
 " the last syllable of the verb immediately before 
 " the personal pronoun. This emphasis not only 
 " drew the accent to the syllable, but also extended 
 " it by the insertion of e or o short, the consequence 
 
 certain forms of the verb substantive. This conjecture has been 
 supported with great ingenuity, and has even in its favour a 
 great number of coincidences. But I think it is by no means 
 established. Those, however, who are not acquainted with Pro- 
 fessor Bopp's able attempts to analyse the conjugations of verbs 
 will find their trouble amply repaid in reading his " Conjugation- 
 " system der Sanskritsprache," and likewise his observations 
 subsequently published in the Annals of Oriental Litt. and still 
 more fully in his Grammatica Critica Linguae Sanskritae. 
 
154 EASTERN ORIGIN OF 
 
 " of protracted pronunciation." " The voice was 
 " kept up, and this inserted vowel gradually slid 
 " into union with that which supported the pronoun, 
 " and formed with it a long sound, expressive of sus- 
 " pense and incomplete indication." " The optative 
 " of all tenses had a similar origin. In wishing, we 
 " dwell on the word, and give it an unusual empha- 
 " sis, the sign of strong, lingering, ardent desire." 
 " In grief the emphasis is long, and uttered with a 
 " wailing, melancholy tone. The connection between 
 " grief 'and desire is close and obvious : 
 
 ws dirore KpetovT ' AfxapvyK^a Ochttov 'E7retot — 
 
 " The effects of this state of mind on the medium 
 " of thought, are that the vowels are protracted, 
 " while the consonants rather sink and vanish." 
 
 Whatever may be thought of this explanation, the 
 fact to which it relates is undoubtedly observed in 
 the conjugations of Latin, Greek, Sanskrit, and Teu- 
 tonic verbs. In all these, the substitution of diph- 
 thongs, and longer or more numerous vowel sounds, 
 is characteristic of the optative, potential, and con- 
 junctive moods. 
 
 Paragraph 2. 
 Of the Future Tenses. 
 
 The formation of future tenses deserves a parti- 
 cular notice. 
 
 Proper future tenses formed by inflection are en- 
 tirely wanting in the Teutonic languages. In Latin, 
 Greek, and Sanskrit they are yet extant ; and in all 
 these, analogies are to be traced in their formation. 
 
 The Sanskrit has two distinct future tenses, which 
 are formed as follows. 
 
THE CELTIC NATIONS. 155 
 
 The first future, or perfect future, is formed by- 
 adding td to the verbal root, or rather by inserting 
 it between the root and the pronominal suffix. This 
 syllable is often preceded by an additionally inserted 
 I, and in some of the persons it is tas rather than 
 td. Thus from the root 2pcJ — yach, or rather the 
 verb yachami, aheu, we have yachitasmi, ah^o-a, or, 
 as I here divide the words, 
 
 Present, yach-a-mi. 
 Future, yach-i-tas-mi. 
 
 The second future instead of the syllable ta or 
 tas, inserts sya, as yachisyami, or 
 
 yach-i-shya-mi. 
 
 The ^ — s is converted into %\ — sh b by the San- 
 skrit rules of euphony. 
 
 Future Tenses in Greek. 
 
 The first future in Greek is formed in a manner 
 very similar to that of the second future in San- 
 skrit. 
 
 The termination of the first future seems origin- 
 ally, as it is observed by Matthise, to have been the 
 same throughout, e<ru from co. Thus we find oAeVco 
 from oAci>, apeo-a from apw. The middle voice okeo-ofxat 
 would indicate an active oAeVo/x*, which perhaps once 
 existed, and which would be the regular form if the 
 suffix fu had been preserved through the different 
 tenses in Greek as it has been in Sanskrit. 
 
 b The Sanskrit sibilant SJ i s neither s nor sh. It is said to be 
 pronounced by passing the voice, with the tip of the tongue ap- 
 plied to the fore part of the palate, and is represented in Wil- 
 kins's Grammar by s. Sh is perhaps the mode of expressing it 
 most nearly in our orthography. 
 
15(3 EASTERN ORIGIN OF 
 
 The first future consists therefore in Greek in the 
 insertion of e$ before the pronominal suffixes, in 
 analogy with the Sanskrit sya or ishya. 
 
 It may be observed, that the terminations of this 
 form of the verb, both in Greek and in Sanskrit, are 
 identical with the future tense of the verb substan- 
 tive, in Greek eaofjuu and in Sanskrit syami. This 
 is the principal foundation for the hypothesis of Pro- 
 fessor Bopp, who considers many modifications of 
 attributive verbs to be derived from a composition 
 of a verbal root with the tenses of the verb substan- 
 tive. If other tenses corresponded with the termi- 
 nations of the verb substantive so closely as the fu- 
 ture, there would be sufficient evidence to support 
 this opinion. At present, we can only regard it as 
 an ingenious conjecture. The Celtic language, how- 
 ever, presents a feature which gives it a degree of 
 additional probability : to this we shall have occa- 
 sion hereafter to advert. 
 
 2. The second future in Greek is a slight inflec- 
 tion of the present, as key®, 7ri6S>, from \eyco 9 vetOu. The 
 present tense is often used for the future by the 
 poets ; and this form seems to have been originally a 
 mere change in the accent or emphasis of the pre- 
 sent, designed thus to mark a variety in the sense. 
 Some grammarians c have indeed maintained that 
 the supposed second future is merely a first fu- 
 ture in a different form, in which case there would 
 be no second future in Greek. This would be con- 
 trary to the analogy of the cognate languages. 
 However, it must be allowed that there is not in 
 Greek, as there is in Latin and in Sanskrit, a second 
 
 c See Dawes, Miscellanea Critica, p. 372. 
 
THE CELTIC NATIONS. 157 
 
 future distinguished from the first by a difference of 
 sense. 
 
 Latin Futures. 
 
 The future tenses in the Latin language are 
 formed in a great variety of ways. 
 
 1. The most simple form is a slight modification 
 of the present, regain, reges, reget being substituted 
 for rego, (which, according to the analogy already 
 pointed out, was perhaps originally regim,) regis, 
 regit. This recalls those languages in which the 
 present tense is used for a future, and the British 
 future creclav is nearly like it. It is still more 
 closely allied to the conjunctive present regam, re- 
 gas, regat. A slight difference in pronunciation was 
 adopted, to mark these varieties in the meaning or 
 in the relations of the verb to time and mode. This 
 is an instance of simple inflection. Here is no place 
 for the hypothesis of compound verbs, or of particles 
 introduced and interpolated. 
 
 2. Another mode of giving to verbs a future sig- 
 nification adopted by the Latin grammarians was 
 that of inserting a syllable, a method analogous to that 
 practised in Greek and in Sanskrit conjugations ; 
 but instead of the ea or <r of the Greek futures, and 
 the sya or ishya of the Sanskrit, the Latins inserted 
 er between the verb and the pronominal suffix, the 
 verb retaining the characteristic of the preterper- 
 fect tense ; as 
 
 rex-er-o, rex-er-imus. 
 
 audiv-er-o, audiv-er-imus. 
 
 The substitution of r instead of s appears in this 
 
 instance, as in a great many inflections in the Latin 
 
 language, to be an innovation, originating either 
 
 from some preference on account of euphony, or from 
 
158 EASTERN ORIGIN OF 
 
 accidental variation of utterance. The old Latin 
 form was in all probability similar to the Greek, es 
 being the syllable interpolated. Rexeso and rexesi- 
 mus gave a harsh and difficult sound, and were soft- 
 ened to rexero and rexerimus d . 
 
 3. A third species of future tense in Latin verbs 
 arises from the insertion of b or rather of ab, eb, or 
 ib. Even this has been referred to the composition 
 of an attributive root with the verb substantive. 
 Recourse has been had here to the verb fuo, analo- 
 gous to bhu or be, as in the former instance to the 
 verb 3fH — as, esse, or elvai. The terminations of 
 the Latin future amabo, bis, bit, are supposed by 
 Professor Bopp to have been derived from a Latin 
 future of the verb substantive analogous to the An- 
 glo-Saxon beo, bys, byth. 
 
 It is remarkable, with respect to both the Latin 
 futures, that a slight change merely of the termination 
 gives rise to a form of the verb, which has a preter- 
 ite signification, as amavero (perhaps originally 
 written amaverim, but subsequently distinguished 
 from that conjunctive form) makes amaveram, and 
 amabo (amabim?) makes amabam. This circum- 
 stance tends to render it most probable that there is 
 nothing in the inserted ba or bo, which by itself 
 gives either a past or future sense, and that it is 
 merely through conventional use that any precise 
 distinction of time is obtained from this modification 
 of the verb. In Sanskrit and in Greek, as well as in 
 Latin, the future tense is converted into a preterite 
 by a change of the termination ; this is connected in 
 
 d On the substitution of r for s, I must refer the reader to 
 sect. 6. of chapter ii. in which this subject is considered. 
 
THE CELTIC NATIONS. 159 
 
 the two last mentioned languages with the addition 
 of the augment 
 
 The facility indeed with which these preterite 
 forms of the verb, either with a slight modification, 
 or without any change, are adopted with a future 
 meaning, seems remarkable, when we first con- 
 template it ; but the frequent occurrence of a parallel 
 fact in different languages is calculated to lessen our 
 surprise. What can be more strange and apparently 
 anomalous than the changes connected in Hebrew 
 with the use of the Vau Conversivum ? In Hebrew 
 the future imperative, optative, and potential forms 
 are all identified, or nearly so, with each other, or 
 one modification of the verb answers to all. 
 
 In like manner we find corresponding forms of 
 the verb, which may be considered as the same iden- 
 tical inflections of the verbal root, (if we make allow- 
 ance for the varieties of orthography and termina- 
 tion proper to the cognate languages, or for the 
 stated interchange of letters between them,) denot- 
 ing in one language a future, and in another a po- 
 tential, or even a preterite tense, with something of 
 the potential or conjunctive signification attached to 
 it. Thus the Sanskrit future, the Greek aorist op- 
 tative, and the Latin preterpluperfect have nearly 
 the same endings. From the Sanskrit root yach we 
 have in the future, 
 
 yach -ishyami. 
 
 From airea, aor. 1. opt. an- -YjaaifJLi. 
 
 From quaero in Latin, quaes-issem. 
 
 In the plural, yach -ishyamus. 
 
 air -YjvaifAtv, or fxeg. 
 quaes-issemus. 
 
 Here ishyami, ecra/p, and issem, in which the e 
 
160 EASTERN ORIGIN OF 
 
 makes emus and etis long, and was nearly equi- 
 valent to a diphthong, afford striking analogies, and 
 are in fact the same form with slight varieties of 
 orthography, but somewhat greater variety in sig- 
 nification. 
 
 Paragraph 3. 
 Middle and Passive Voices. 
 
 The Sanskrit has in its verbs three voices nearly 
 corresponding with the Greek. 
 
 The voices in Sanskrit corresponding with the 
 Greek middle and passive form themselves by a si- 
 milar change of termination with that of the verbs 
 in pi : as %!$»*§, or rather &*&yn, is changed into &/&GTa/, 
 so bhavati is converted into bhavatai. I cannot but 
 believe that the original form of the verb in San- 
 skrit was 
 
 bha vam ai, ^ ( bha vami . 
 
 bhavasai, >- from ■< bhavasi. 
 bhavatai, J l bhavati. 
 
 but the first person is, according to the established 
 inflections of the Sanskrit language, bhavai, instead 
 of bhavamai. 
 
 I shall not pursue further at present the inflec- 
 tions of verbs in the different voices. The reader 
 will find enough to answer my chief design in a 
 succeeding chapter, in which examples of the re- 
 gular verbs are inserted. 
 
 The termination most characteristic of passive 
 tenses in Latin, viz. r, must here be mentioned, as 
 it serves as a point of comparison between the Celtic 
 and the Latin verbs. 
 
 Another point in which we shall find a relation 
 
THE CELTIC NATIONS. 161 
 
 between the Latin and the Celtic verbs, as likewise 
 between the Greek and Celtic, is the defective 
 state of the inflection of persons in the passive 
 tenses. But we shall again take notice of this cir- 
 cumstance in its proper place. 
 
 M 
 
162 EASTERN ORIGIN OF 
 
 CHAPTER VII. 
 
 Illustration of the principles developed in the preceding chap- 
 ter. Conjugation of the verb substantive and of attributive 
 verbs, both in the other Indo-European languages and in the 
 Celtic dialects. 
 
 Section I. General Remarks. Analysis of the Verb Substan- 
 tive in several languages. 
 
 J. HE preceding remarks will perhaps be deemed 
 sufficient to explain the general principles of verbal 
 inflection in the languages to which they refer ; but 
 before I can proceed to my ultimate object, which is 
 to compare the Celtic verbs with those of the idioms 
 supposed to be cognate with the Celtic language, it 
 is requisite to illustrate the principles now developed 
 by some particular examples. I shall with this view 
 lay before my readers a brief analysis of the verb 
 substantive in Sanskrit, pointing out in the first 
 place the agreements of the Sanskrit with the other 
 languages generally allowed to be allied to it. I 
 shall afterwards endeavour to illustrate in a similar 
 manner the Celtic inflections, and to shew that they 
 manifestly partake in the same general analogies. 
 
 I have already observed, in the list of verbal roots 
 contained in a preceding section, that there are in 
 Sanskrit two verbs substantive, of which cognates 
 are found in various idioms. They are the verb 
 3TtT5T — asmi, from the root as, corresponding with 
 esse or sum in Latin and elp) in Greek, and H^TPT — 
 bhavami from the root bhu, allied to the old Latin 
 verb fuo, and in the sense of oriri, nasci, which 
 also belongs to this Sanskrit root, to the verb <f>vu 
 or (pvvai in Greek. There is, I believe, no language 
 in which both of these verbs are extant in a com- 
 a Rosen, Radices Sanskritse, p. 52, 53. 
 
THE CELTIC NATIONS. 163 
 
 plete state, but they are least defective in Sanskrit, 
 which has lost many tenses of the verb asmi, though 
 it has preserved the whole of bhavami. The Per- 
 sian has two corresponding verbs, buden, and am re- 
 sembling asmi : they are both defective, and each 
 contributes some tenses towards the conjugation of 
 the verb substantive, which is thus made up of their 
 fragments. The Sclavonian verb substantive is 
 formed from similar elements ; the present tense and 
 those dependent on it are nearly allied to asmi, and 
 the past tenses are derived from but' or bud' the 
 cognate of buden or bhu. The Teutonic languages 
 display the same formation: ist or is belongs to the 
 former; beon, be, been, to the latter element. In 
 Latin fuo and esse are combined in a similar man- 
 ner. The Celtic language, as I shall shew in the 
 following section, has one of these verbs in a more 
 perfect state than any other language except the 
 Sanskrit. The verb bu or bydh, corresponding with 
 bhu or buden, is nearly complete, if not entirely so; 
 but there are only fragments, as in other languages, 
 which resemble the cognates of asmi. 
 
 Paragraph 1. 
 Verb asmi and its cognates. 
 I shall now compare the principal parts of the 
 verb asmi, and subjoin some corresponding forms in 
 the cognate languages. 
 
 Present tense. 
 1. In Sanskrit. 
 
 First person. Second person. Third person. 
 
 Sing. asmi asi asti 
 
 Plur. smah or smus st'ha santi 
 
 N. These plural forms were originally (?) 
 asmus ast'ha asanti. 
 
164 EASTERN ORIGIN OF 
 
 2. 
 
 In Greek 
 
 according to the old forms 
 
 Sing. 
 
 €fJ.[JLl 
 
 eaai 
 
 5 V 
 
 €(TTl 
 
 Plur. 
 
 elfins 
 
 €(TT€ 
 
 S. In Latin. 
 
 €VTl 
 
 Sing. 
 
 esum 
 
 es 
 
 est 
 
 Plur. 
 
 sumus 
 
 estis 
 4. In Persian. 
 
 sunt 
 
 Sing. 
 
 am 
 
 iy 
 
 est 
 
 Plur. 
 
 Im 
 
 id 
 
 end 
 
 5. In Sclavonian b . 
 Sing. yesm' yesi vest' 
 
 Plur. yesmi yeste sut'forjesut' 
 
 6. In Lettish or Lithuanian . 
 Sing. esmi essi esti 
 
 Plur. esme este esti 
 
 7. In Moeso-Gothic. 
 Sing. i'm is ist 
 
 Plur. siyum siyuth sind 
 
 which according to Dr. Grimm was originally in 
 the plural, 
 
 isum isuth isind 
 
 or, isam isith isand. 
 
 It is at once evident, that all these are slight mo- 
 difications of the same element conjugated by means 
 of the same suffixes. The variation between the 
 different languages is not exceeding such as exists 
 between proximate dialects of the same speech. 
 
 The preterimperfect tense is not to be traced 
 with so much regularity. It is in Sanskrit, 
 Sing. asam asis asit 
 
 Plur. asma asta asan. 
 
 b Grimm, I. p. 1064. Vater. p. 98. c Grimm, ibid. 
 
THE CELTIC NATIONS. 1G5 
 
 In Latin, esam was probably the old form of eram, 
 for B, as we have seen, was often changed into r, and 
 esam would regularly form esem in the subjunctive, 
 which is actually found, 
 
 Sing. esam esas esat 
 
 Plur. esamus esatis esant. 
 
 The second form of the verb in the arrangement 
 adopted by Sanskrit grammarians is the potential. 
 The potential form of the verb asmi bears a strong 
 analogy to the old Latin potential siem, and, as M. 
 Bopp has also shewn, to the Moeso-Gothic potential. 
 
 Singular. 
 
 Sanskrit, 
 
 syam 
 
 syas 
 
 syat 
 
 Latin, 
 
 siem 
 
 sies 
 
 siet 
 
 Gothic, 
 
 d siyau 
 
 siyais 
 Plural. 
 
 siyai. 
 
 Sanskrit, 
 
 syam a 
 
 syata 
 
 syus 
 
 Latin, 
 
 siemus 
 
 sietis 
 
 sient 
 
 Gothic, 
 
 siyaima 
 
 siyaith 
 
 siyai] 
 
 It may be observed that all these words have lost 
 the initial vowel a or e, and that if it were restored 
 the preceding forms would bear a near analogy to 
 ecraifxi, which, though not extant, would be a regu- 
 lar derivative from the etymon of co-opai. 
 
 The Sanskrit verb asmi has no future, but M. 
 Bopp conjectures with great probability, that syami, 
 the adjunct by which a future tense is formed in 
 attributive verbs, is in fact only the obsolete future 
 
 d This form is considered by Hickes (Thesaur. Ling. Sept. 
 torn. I.) as a future tense, but Dr. Grimm has shewn that the 
 Teutonic dialects have no simple future, properly so termed. 
 The potential is, however, used for a future by Ulphilas. See 
 Bopp, in Annals of Oriental Litt. p. 49. 
 
 M 3 
 
166 EASTERN ORIGIN OF 
 
 of the verb asmi. A fact strongly favouring this 
 hypothesis is, that a tense of this verb exists in San- 
 skrit, and is recognised as such, which is only used 
 in forming the preterperfect tense of certain verbs. 
 Asa, asit'ha, asa, is termed the third preterite or 
 aorist of asmi e . It is joined with karayam from 
 the verb karomi, facio, creo, and forms karayamasa, 
 fecit, creavit. 
 
 There is only one other tense of the verb asmi, 
 which is the imperative. 
 
 Sing. asani aidhi astu 
 
 Plur. asama sta santu. 
 
 Compare astu, with cforw, esto ; sta with eore, este, 
 and santu, with sunto. The second person, aidhi, 
 bears, as we shall see, a strong analogy to some of 
 the modifications of the verb substantive in the 
 Celtic. 
 
 Paragraph 2. 
 
 Verb bhavami and its cognates. 
 I shall now give the principal parts of the San- 
 skrit verb bhavami, which is entire, though its cog- 
 nates in most other idioms are only extant in frag- 
 ments. 
 
 1. Present tense. 
 Sing. bhavami bhavasi bhavati 
 
 Plur. bhavamus bhavath'a bhavanti. 
 This tense exemplifies the personal endings in their 
 complete state. 
 
 2. Potential. 
 Sing. bhavaiyam bhavais bhavait 
 
 Plur. bhavaima bhavaita bhavaiyus. 
 This form corresponds with the Greek in cu/xt, ais, at, 
 aifJL€V, cure, cutv. 
 
 e Wilkins's Sanskrit Grammar, p. 187. 
 
THE CELTIC NATIONS. 107 
 
 3. Imperative. 
 Sing. bhavani bhava bhavatii 
 
 Plur. bhavama bhavata bhavantu. 
 
 4. First Preterite or Imperfect. 
 Sing. abliavam abhavas abhavat 
 
 Plur. abhavam abhavata abhavan. 
 
 This tense has the abbreviated form of personal end- 
 ings. 
 
 N. The preceding forms, considered as derived from 
 the present tense, display that modification in the 
 root (as bhava for bhu) which marks the different 
 conjugations. The remainder, which may be com- 
 pared with the tenses of iot^/xi formed from oraco, 
 have the root in its original state, and with respect 
 to these tenses, there is only one conjugation in San- 
 skrit verbs f . 
 
 5. Second Preterite or Aorist. 
 Sing. abhuvam abhus abhut. 
 
 Personal endings in the abbreviated form. 
 This corresponds with £<pvv, t(f)vs, tcpv. 
 
 f The ten conjugations or forms of Sanskrit verbs differ from 
 each other only in those tenses which are formed from the pre- 
 sent. The first modifies the root and interposes a between it 
 and the suffixes, as bhav-a-mi from bhu. The second subjoins 
 the suffixes immediately, as asmi from as. The third reduplicates 
 in the present, as dadami (St'So/xt) from da. The fourth inserts 
 ya between the root and suffixes, as damyami (domo) from dam 
 (tame, domitus). The fifth inserts nu after the root, and corre- 
 sponds exactly with the conjugation of ^vy-vv-fxi. The sixth 
 modifies the termination of the root somewhat differently from 
 the first, as mriyatai (moritur) from the root mri. The seventh 
 inserts n in the root, as in Latin we find jungo from the root 
 jug-um. The eighth resembles the fifth, but adds only u to 
 roots ending in n. The ninth adds na to the root, and is analo- 
 gous to the form of (pdava, eXeei'i/co. The tenth inserts i in the 
 root, and inflects like the first. These analogies have been pointed 
 out fully by Dr. Murray (Hist, of European Languages). 
 
 M 4 
 
168 EASTERN ORIGIN OF 
 
 6. Reduplicate Preterite. 
 Sing. babhuva babhuvit'ha babhuva 
 
 Plur. babhuvima babhuva babhuvus. 
 
 This corresponds very closely with the Greek itityva, 
 
 7T€(f)VaS, irtyvt, 7T€(j)Va}X€V, 7T€(f)VaT€, TT€(f)VaO-L. 
 
 Babhuvima in its termination is allied to fuvimus from 
 fuot\ 
 
 7. First Future. 
 
 Sing. bhavatasmi. 
 The personal endings nearly as in the present tense. 
 
 8. Precative. 
 
 Sing. bhuyasam. 
 The personal endings abbreviated. 
 
 9. Second Future. 
 
 Sing. bhavishyami. 
 
 The personal endings in full. In Latin fuissem nearly 
 approaches to the above form, or perhaps more 
 closely to the following : 
 
 10. Conditional. 
 
 Sing. Abhavishyam. 
 Personal endings abbreviated. 
 
 Infinitive Mood. 
 Bhavitum. 
 
 The infinitive mood in Sanskrit bears an analogy to the 
 Latin supine. 
 
 Participles. 
 Adverbial Participle, bhiitva. 
 Passive, bhuta. 
 
 Pret. Reduplicate, babhuvas. 
 
 b As by Ennius, " Nos sumu' Romani, qui fuvimus ante Ru- 
 dini." 
 
THE CELTIC NATIONS. 169 
 
 Present, bhavat. 
 
 Future, bhavishyat. 
 
 The participles are given without their termina- 
 tions, which are added in declining them, and re- 
 semble those of adjectives. 
 
 The preceding forms are given by grammarians 
 as those of the active voice; but this verb is inflected 
 through two other voices, analogous to the middle 
 and passive. Of these I shall only extract the pre- 
 sent, the reduplicate preterite, and the participles. 
 
 Middle Voice or Atmanaipadum. 
 Sing. bhavai bhavasai bhavatai 
 
 Dual, bhavavahai bhavait'hai bhavaitai 
 
 Plur. bhavamahai bhavadhvai bhavantai. 
 In this we have only to supply the first personal ending 
 Sj~ — mai, which the analogy of the other persons 
 seems clearly to suggest, and the whole form will 
 correspond nearly with the Greek. 
 
 Reduplicate Preterite. 
 
 Sing, babhuvai babhuvishai babhuvai 
 
 Dual, babhuvivahai babhuvat'hai babhuvatai 
 
 Plur. babhuvimahai babhuvidvai babhuvirai 
 
 Passive Present Tense. 
 Sing. bhuyai bhuyasai. 
 
 Personal endings as in the middle voice. 
 
 Reduplicate Preterite. 
 Same as in the middle Voice. 
 Participles present in the middle voice : 
 Bhavamanah bhavamana bhavamanam. 
 Reduplicate Preterite : Babhuvanah, a, am. 
 Future : bhavishyamanah. 
 Passive Present : bhuyamanah, a, am. 
 The terminations would be represented correctly 
 
170 EASTERN ORIGIN OF 
 
 thus; fxdvo$, pava, [xavovi and it is needless to remind 
 the reader of the near correspondence of these forms 
 with the Greek. 
 
 SECTION II. 
 
 Analysis of the Celtic Verb Substantive. 
 
 The verb substantive in Welsh grammars appears 
 to a learner as if made up of the fragments of two 
 or three defective roots, like the verbs substantive 
 of other European languages. But in reality there 
 is in the Welsh a verbal root, which is cognate with 
 the Sanskrit bhu and the Persian buden, and which 
 is like the former, perfect, or very nearly so, having 
 as many extant forms as the Welsh verbs generally 
 possess. This verb is in the infinitive mood bod, 
 and bod may perhaps be regarded as the root, 
 although Dr. Davies gives that term to the third 
 person singular of the preterite, which is bu, fuit. 
 The third person of the future is, however, often 
 the root of Welsh verbs, and this in the verb sub- 
 stantive is B ydh, erit. Bydh, if not the root, is the 
 basis on which most of the modifications of this verb 
 are formed. 
 
 Regular verbs have in Welsh, besides the infini- 
 tive and imperative moods, five distinct tenses or 
 forms : these are two futures, one of which is indi- 
 cative and the other conditional or subjunctive, a 
 preterimperfect, preterperfect, and preterpluperfect 
 tense. All these forms are extant in the verb bod. 
 They are as follows. 
 
THE CELTIC NATIONS. 171 
 
 Paragraph I. 
 1. Future Indicative. 
 Sing. bydhav bydhi bydh. 
 
 Plur. bydhwn bydhwch bydhant. 
 
 N. It must be observed that the Welsh y in the penultima is 
 a short u. The ending av in Welsh stands for am in Erse, and 
 the t' is equivalent to mh, or is a secondary form of m. 
 
 N. It may be worth while, before we proceed further, to 
 compare with this the future tense of the verb sub- 
 stantive in the Russian, as a specimen of resemblance 
 in one of the eastern branches of the European lan- 
 guages. 
 
 Sing. budu budet' budut' 
 
 Plur. budem' budet e budut'. 
 
 Compare also the potential form of the verb 
 buden, to be, in Persian : 
 
 Sing. budemi budi budi 
 
 Plur. budimi budidi budendi. 
 
 2. Future Potential, Conditional or Subjunctive. 
 
 Sing. bydhwyv bydhych bydho 
 
 Plur. bydhom bydhoch bydhout. 
 
 This form is varied as follows : 
 
 Sing. bythwyv bythych bytho, &c. 
 
 and contractedly thus : 
 
 Sing. bwyv bych bo 
 
 Plur. bom boch bout. 
 
 N. Compare with the preceding the indefinite or sub- 
 junctive form in the Persian, which is also termed a 
 future. 
 
 Sing. buvem buvi buved 
 
 Plur. buvim buvid buvend. 
 
172 EASTERN ORIGIN OF 
 
 3. Preterimperfect. 
 Sing. bydhwn bydhit bydhai 
 
 Plur. bydhem bydhech bydheut. 
 
 This likewise is contracted by dropping the dh, as 
 bawn for bydhwn. 
 
 N. This form is considered by Dr. Davies as appro- 
 priated to the subjunctive mood, oedhwn which is 
 derived from another root being used in the indica- 
 tive. In regular verbs, in general, this tense belongs 
 rather to the subjunctive than the indicative. 
 N. Compare with the preceding the preterite of the 
 verb substantive in Persian. 
 
 Sing. budem budi bud 
 
 Plur. budim budit bud end. 
 
 4. Preterperfect tense. 
 Sing. 1. bum, poetice buum and buym. Lat. fui 
 (olim fuim?) 
 
 2. buost fuisti. 
 
 3. bu fuit. 
 Plur. 1. buom fuimus. 
 
 2. buoch fuistis. 
 
 3. buont and buant fuerunt. 
 
 N. The relation of these inflections to the Latin is 
 
 obvious. 
 In Greek and in Sanskrit the forms most allied to this 
 preterite are the aorists, as, 
 
 Sing. abhuvam abhus abhut 
 
 €(pvv €(pvg e<pv, 
 
 Plur. abhuma abhuta abhuvan, 
 
 ecf>vfA€v €<f>vT€ Itpvaav. 
 
 5. Preterpluperfect Tense. 
 The preterpluperfect tense bears also a near re- 
 semblance to the corresponding form in Latin, and 
 this is still more striking if we restore the s in the 
 
THE CELTIC NATIONS. 
 
 173 
 
 place of r, where we have reason to believe that it 
 originally stood. 
 
 Latin. Welsh. Latin. Welsh. 
 
 fuesam bhnaswn fuesamus bhuesym 
 
 fnesas bhnasit fuesatis bhuesych 
 
 fnesat blmasai fuesant bhuesynt. 
 
 Imperative Mood. 
 
 Sing. 2. bydh, be thou. 3. bydhed, boed, bid, 
 Plur. bydhwn. bydhwch. bydhant. 
 
 Infinitive Mood, 
 bod. 
 Persian, buden ; Russian, buit'. 
 
 The preceding are all the forms properly belong- 
 ing to the verb bod in the active voice, and, as be- 
 fore observed, they are as many as belong to any 
 regular verb in Welsh. 
 
 Note, Before we proceed further, it will be worth while to 
 compare the present tense of the verb substantive in the 
 Erse dialect of the Celtic, with a corresponding form in the 
 Sclavonic language. The Erse has a present tense properly 
 so called, although it is wanting in Welsh. 
 
 The infinitive mood and root of this verb in Erse 
 is beith, to be. The following is the negative form 
 of the present tense. 
 
 1. In the Erse, properly so termed, or Irish Celtic. 
 
 Sing. Plural. 
 
 1. ni fhuilhim ni fhuilmid 
 
 2. ni fhuilhir ni fhuilthidh 
 
 3. ni fhuilh ni fhuilidh b 
 
 b Gaelic (i. e. Irish) Grammar by E. O'C , printed by J. 
 
 Barlow, Dublin, 1808. 
 
174 EASTERN ORIGIN OF 
 
 2. In the Gaelic of Scotland. 
 Sing. Plural. 
 
 1. ni bheil mi ni bheil sinn 
 
 2. ni bheil thu ni bheil sibh 
 
 3. ni bheil e ni bheil iad c 
 
 Conditional form of the verb buit', to be, in Russian. 
 
 Sing. Plural. 
 
 1. ya bui buile mu bui buili 
 
 2. til bui buile vu bui buili 
 
 3. one bui buile oni bui buili d 
 
 Passive Voice. 
 The Celtic grammarians, like the Sanskrit, dis- 
 tinguish passive forms of the verb substantive. The 
 passive voice, however, in the Celtic, has only the 
 third person singular throughout the moods and 
 tenses. 
 
 1. Future Indicative, 
 bydhir. 
 
 2. Future Subjunctive. 
 
 bydher, byther, and contractedly, baer. 
 
 Note. With byther or bydhir the Latin futurus is evi- 
 dently cognate. 
 
 3. Preterimperfect. 
 bydhid, contracted beid. 
 
 Note. Compare the Sanskrit, bhuyatai. 
 
 c Grammar prefixed to the Gaelic Dictionary published by 
 order of the Highland Society. 
 
 d Etemens de la Langue Russe, Petersbourg. 1768. p. 133. 
 Praktische Grammatik der Russischen sprache, von D. Johann 
 Severin Vater. Leipzig. 18 14. Tab. 7. 
 
THE CELTIC NATIONS. 175 
 
 4. Preterperfect. 
 buwyd. 
 
 Note. Compare the Sanskrit bhutwa. 
 
 5. Preterperfect. 
 buasid and buesid. 
 
 Note. Compare in Latin fuisset or rather fuesit, the old 
 form of fuerit. 
 
 Paragraph 2. 
 
 Of defective verbs used as verbs substantive in the 
 Celtic dialects. 
 
 Besides the verb bod, which we have compared 
 with its cognates, there are other defective verbs in 
 the Celtic dialects used as parts of the verb substan- 
 tive. In the Welsh it has been remarked that re- 
 gular verbs want a present tense properly so termed. 
 In fact, the Welsh grammarians give the denomina- 
 tion of a future to a particular form of the verb, 
 which is used with both a future and present signi- 
 fication ; and it is perhaps somewhat doubtful to 
 which tense it properly belongs. That it is capable 
 of expressing a present signification, without any 
 metaphor or reference to the future, is fully evident 
 from the instances adduced by the Welsh gramma- 
 rian Dr. Davies, who observes, that in the Creed, 
 the expression " Credo in Deum Patrem" is ren- 
 dered by " Credav yn Nuw Dad," and that in con- 
 versation " Mi a welav" means " I see," and " Beth 
 " medhi di," " what sayest thou ?" 
 
 The following forms are considered as belonging 
 to the present tense. 
 
 1. Sydh, and by apocope sy; est, is. This is used 
 indefinitely in all numbers and persons. 
 
176 EASTERN ORIGIN OF 
 
 >• only the third person extant. 
 
 2. Mae, est 
 Maent, sunt, 
 
 3. Oes, est. This has no inflections, and is used 
 
 only as a third person singular. 
 
 4. Yw, pronounced yu ; est. This is also used in 
 
 the same way as sydh, and it is likewise in- 
 flected, as follows : 
 
 Singular. 
 
 1. wyv, perhaps originally ywyv. 
 
 2. wyt or wyd. 
 
 3. yw. 
 
 Plural. 
 
 1. ym 2. ych 3. ynt. 
 
 There is another form of yw, with a sort of re- 
 duplication, thus : 
 
 Sing. 1. yd wyv 2. yd wyt 3. ydyw. 
 Plur. 1. ydym 2. ydych 3. ydynt. 
 
 There is a poetical form yttwyv, yttwyt, &c. 
 
 Passive Form. 
 
 Welsh grammarians distinguish likewise a passive 
 form of this verb. It is as follows : 
 
 In the poets, and in the dialects of South Wales 
 and Powys, ys. 
 
 The poets have also ydis ; it is commonly ydys. 
 
 Note. It must be observed that the Welsh ys is pro- 
 nounced like us in English, and exactly as the root ^tt — 
 as, in Sanskrit. 
 
 The same root slightly modified, viz. is or isi, is 
 extant in the Erse and Scottish dialects of the 
 Celtic e ; as 
 
 e Lhuyd's English-Irish Dictionary. 
 
THE CELTIC NATIONS. 177 
 
 Sing, is mi, or is misi, I am, 
 
 is tu, tliou art. 
 
 is e, he is. 
 
 Plur. 1. is simi. 2. is sibh. 3.isiad f . 
 Preterimperfect tense, in Welsh. 
 
 1. Active or variable form. 
 Sing. 1. oedhwn. 2. oedhit or 3. oedli. 
 
 oedhyt. 
 
 Plnr. 1. oedhym or 2. oedhych or 3. oedhynt or 
 oedhem. oedhech. oedhent. 
 
 2. Passive or invariable form, 
 oedhid. 
 
 Cognates. 
 The form ys, which is perhaps the real etymon, 
 is precisely the root in Sanskrit and in the Euro- 
 pean languages, which, adding the pronominal suf- 
 fix always wanting in Welsh in the third person 
 singular, as well as in the passive form, make of the 
 same word, eo-n, est, ist. 
 
 Note. It is remarkable that the verb substantive has 
 forms appropriated to the present tense, while all other 
 Welsh verbs are destitute of them. This circumstance may 
 be accounted for. There being two distinct verbs substan- 
 tive, and each having that form which is used, as we have 
 seen, in the generality of verbs, sometimes with a future 
 and sometimes with a present signification, practice, founded 
 on convenience, at length appropriated the use of onj of 
 them to the future, and the other to the present tense. 
 Hence bydh came to express " it zvill be" and sydh, ys, and 
 oes, " it is" while, in respect to attributive verbs, which 
 have a single form, the ambiguity still remains. 
 
 f Shaw's Analysis of the Gaelic Language. 
 N 
 
178 EASTERN ORIGIN OF 
 
 SECTION III. 
 
 Inflection of a regular verb in Welsh through its moods and 
 
 tenses. 
 
 The root or origin of a verb in Welsh is, as the 
 learned Dr. Davies remarked, for the most part a 
 noun, as dysc, doctrina ; dyscais, docui ; car, amicus; 
 carav, amo vel amabo. This substantive, adds the 
 same writer, is generally identical with the third 
 person singular of the future indicative, (as in He- 
 brew the third of the preterite is the root,) or with 
 the second of the imperative, which forms are for 
 the most part the same. 
 
 In some verbs, however, the third person of the 
 preterite is the root, as aeth, daeth. 
 
 Indicative Mood, Present Tense. 
 There is, according to the Welsh grammarians, 
 no present tense in attributive verbs, and this tense 
 is supplied by a circumlocution, as 
 
 wyv yn caru, literally, el[u kv t» <f>i\e7v, I am loving. 
 
 Note. All the other tenses may be formed by a similar 
 circumlocution. 
 
 Preterimperfect Tense. 
 Sing. 1. carwn 2. carit 3. carai 
 
 Plur. 1. carem 2. carech 3. carent. 
 
 A tense seldom used in the indicative. 
 
 Preterperfect Tense. 
 Sing. 1. cerais, i.e.kerais 2. ceraist 3. carodh 
 Plur. 1. carasom 2. carasoch 3. carasant 8 . 
 
 a See Dr. Davies's Grammar, entitled, Antiquse Linguae Britan- 
 nicee Rudimenta, from which, and from the grammar prefixed to 
 Richards's Dictionary, the following as well as the preceding 
 conjugations of Welsh verbs are extracted. 
 
THE CELTIC NATIONS. 
 
 179 
 
 The principle on which this and the following 
 tense are formed seems to be the insertion of the 
 syllable as or ais between the root and the personal 
 endings, and a change of the radical vowel in a mode 
 analogous to the Sanskrit form guna. Both these 
 changes have been traced in the inflections of verbs 
 in the Sanskrit and European languages. And the 
 modification of the vowel by guna is, in Sanskrit, as 
 in Welsh, confined to particular persons in the 
 tense. 
 
 Preterpluperfect Tense. 
 Sing. 1. caraswn 2. carasit 3. carasai 
 Plur. 1. carasem 2. carasech 3. carasent. 
 
 Future Tense. 
 Sing. 1. carav 2, ceri 3. car 
 
 Plur. 1. carwn 2. cerwch 3. car ant. 
 
 Imperative Mood. 
 
 Sing. 1. 2. car 3. cared 
 
 Plur. 2. carwn 2. cerwch 3. car ant. 
 
 Potential, Optative, and Subjunctive Mood. 
 
 Present Tense wanting. 
 
 Preterimperfect Tense. 
 
 Sing. 1. carwn 2. cerit 3. carai 
 
 Plur. 1. carem 2, carech 3. carent. 
 
 Poetic Form. 
 Plur. 1. cerym 2. cerych 3. cerynt. 
 
 Preterperfect and Preterpluperfect. 
 
 Sing. 
 
 1. caraswn 
 
 2. carasit 
 
 3. carasai 
 
 Plur. 
 
 1. carasem 
 
 2. carasech 
 
 3. carasent 
 
 
 Or, 
 
 Poetic Form. 
 
 
 Plur. 
 
 1. caresym 
 
 2. caresych 
 
 N 2 
 
 3. caresynt 
 
180 EASTERN ORIGIN OF 
 
 Future Tense. 
 Sing. 1. carwyv 2. cerych 3. caro 
 
 Plur. 1. carom 2. carech 3. caront. 
 
 Infinitive Mood, 
 caru, amare. 
 
 This one form, taking various prefixes, as yn caru, 
 in amando, serves the purpose of Infinitive, Gerunds, 
 and Supines. 
 
 Passive Voice. 
 The Welsh language has a proper passive voice b ; 
 but the tenses have no variety of endings to distin- 
 guish the persons. 
 
 Indicative Mood, Present Tense. 
 Formed by a circumlocution, the infinitive used 
 as a gerund, being constructed with the passive form 
 of the verb substantive ; as 
 
 f vy ngharu, amor, 
 yr ydys yn-< dy garu, amaris. 
 lei garu, amatur. 
 
 Note. Literally elfu kv ra> fxov c/nA.eiz>, &c. 
 
 The plural is similarly formed ; but the future is 
 likewise used as a present tense. 
 
 b In this as well as in the number of tenses in the active 
 voice, the Celtic language is richer than the Teutonic. In the 
 latter, according to Dr. Grimm, the Moeso-Gothic is the only 
 dialect that preserves any remains of a passive voice ; and in 
 that only the present tense indicative and subjunctive is ex- 
 tant. The following words are examples : galeikoda, it is liken- 
 ed ; galeikozau, thou mayest be compared ; haitanda, we are 
 called -, halyaindau, they may be hidden. Grimm's D. Gram. p. 
 855. There are likewise some indications of a middle voice in 
 the Gothic version. 
 
THE CELTIC NATIONS. 181 
 
 Preterimperfect Tense. 
 Cerid vi, ti, ev, ni, chwi, hwynt. 
 
 Preterperfect Tense. 
 Carwyd vi, ti, &c. Amatus, sum, es, est, &e. 
 
 Preterpluperfect Tense. 
 Carasid, or caresid, vi, ti, &c. 
 
 Future Tense. 
 Sing. Cerer vi, ti, ev. 
 Plur. Cerir ni, chwi, hwynt. 
 
 Imperative Mood, 
 Sing, and Plur. Carer, vi, di, ev, ni, chwi, hwynt. 
 
 Potential Mood, Present Tense. 
 Sing, and Plur. Cerir vi, di, &c. 
 
 Preterimperfect Tense. 
 Sing, and Plur. Cerid vi, di, &c. 
 
 Preterpluperfect Tense. 
 Sing, and Plur. Caresid vi, di, &c. 
 
 Future Tense. 
 Sing, and Plur. Carer vi, di, &c. 
 
 Participles. 
 ri dhyn, amans homini. 
 Caredig < gan dhyn, amatus ab hornine. 
 
 tdyn, amatus vel dilectus hominis. 
 
 Caradwy, amandus. 
 
 Note. This form is nearly analogous to the Sanskrit ad- 
 verbial participle bhu-twa. 
 
 Such are the inflections of passive verbs in the 
 Welsh language. They contain but a few instances 
 
 N 3 
 
182 EASTERN ORIGIN OF 
 
 of interpolated syllables, and those have been already- 
 remarked. The greater part of these inflections con- 
 sists, excluding the personal endings or affixes, of 
 slight variations in the final syllables, and chiefly in 
 the vowels, very analogous to the changes which 
 distinguish the moods and tenses of the passive voice 
 of Latin verbs, particularly in the third conjuga- 
 tion. 
 
 SECTION IV. 
 
 Conjugation of a regular Verb in Erse. 
 
 Present Tense, Indicative Mood. 
 Verb, Cesaim or kesaim, I torment. 
 
 Note. The root of the verb is said to be the first person 
 of the present tense, the last syllable being cut off. 
 
 Sing. 1. Cesaim. 2. cesair. 3. cesaidh. 
 
 Plur. 1. Cesamaid or) 
 
 2. Cesthai. 3. cesaid. 
 cesam 
 
 Preterite. 
 
 Sing. 1. Do chesas. 2. chesas. 3. ches. 
 
 Plur. 1. Do chesamar^ , fchesadar 
 
 I 2. chesa- 1 
 
 or V- „ S.< or 
 
 do chessam ) tchessad. 
 
 Note. It may be perceived that the form of the present 
 cesaim nearly corresponds with that which the Welsh gram- 
 marians term a future tense terminating in av, and that the 
 preterite in as agrees with the Welsh preterite in ais. The 
 Erse language has adopted a peculiar form for a future 
 tense, made by inserting a syllable fa between the root and 
 the personal endings. This insertion, however, is not used 
 in all verbs. 
 
rn 
 
 THE CELTIC NATIONS. 183 
 
 Future Tense. 
 
 Sing. 1. Cesfad. 2. cesfair. 3. cesfaidh. 
 
 Plur. 1. Cesfamaid or ) ■ 
 
 c h 2. cesiaidhe. 3. cesfaid. 
 
 cesfam. ) 
 
 Imperative. 
 
 Sing. 1. 2. ces. 3. cesadh. 
 
 Plur. 1. Cesam. 2. cesaidhe 3. cesaid or 
 
 cesadis. 
 Potential Mood. 
 
 Preterimperfect Tense. 
 Sing. 1. Do chesfainn. 2. chesfa. 3. chesfadn. 
 Plur. 1. Do chesfamair. 2. chesfaidhe. 3. chesfaidis. 
 
 Infinitive Mood. 
 Cesadh or do chesadh, to torment. 
 
 Note. The different tenses have another form without the 
 personal endings, in which case the personal pronouns are 
 immediately subjoined. 
 
 Passive Voice. 
 
 Indicative mood, Present Tense. 
 Cestar me, thu, e, inn, ibh, iad. 
 
 Note. As in Welsh, only one form in the passive for all 
 the pronouns. 
 
 Preterite. 
 Do chesadh me, thu, &c. 
 
 Future. 
 Cesfaidher me, thu, &c. or cesfar me 
 
 Imperative. 
 Cestar me, thu, &c. 
 
 Potential, Preterimperfect, 
 Do chesfaidhe me, thu, &c. 
 
 N 4 
 
184 EASTERN ORIGIN OF 
 
 Infinitive. 
 Do bheit cesta, to be tormented. 
 
 Participle. 
 Cesta, tormented. 
 
 For the varieties and irregularities of verbs in 
 the Erse, as well as of the Welsh, I must refer to 
 the grammars of those languages. 
 
 SECTION V. 
 
 Concluding observations on the Celtic verbs, with general re- 
 marks on the grammatical peculiarities of the Celtic lan- 
 guages. 
 
 The observations made in the two last chapters 
 allow us to conclude that the inflection of verbs in 
 the Celtic dialects, excluding for the present the con- 
 sideration of suffixes, or the systems of personal end- 
 ings, which were previously compared, is founded 
 on principles similar to those which prevail in the 
 Sanskrit and in several European languages. The 
 Celtic verbs do not display any traces of the preter- 
 ite by reduplication, which is so remarkable a fea- 
 ture in the eastern branches of the Indo-European 
 stock of languages, and which is also found in the 
 Latin and Teutonic ; but they change the middle 
 vowels in a mode analogous to that which these 
 four languages possess, under the form termed guna 
 by Sanskrit grammarians, and they interpolate simi- 
 lar consonants or syllables for the purpose of distin- 
 guishing moods and tenses, the varying terminations, 
 particularly in the passive voice, being closely ana- 
 logous to those of the other old European idioms, 
 and especially to the Latin. 
 
THE CELTIC NATIONS. 185 
 
 When we connect the consideration of these ana- 
 logies with the results formerly obtained on compar- 
 ing the systems of personal endings or suffixes, it 
 will perhaps not be going too far to say, that the 
 whole structure of inflections in the Celtic dialects is 
 founded on principles similar to those which are the 
 groundwork of verbal conjugations in the other lan- 
 guages compared with them. 
 
 The principal affections which words undergo in 
 the construction of sentences in the Celtic languages, 
 may be referred to two heads ; first, interchanges 
 between cognate letters on a principle which we 
 have compared with that of sandhi ; and secondly, 
 the inflections of verbs. In these consists a great 
 part likewise of the peculiarity of the Sanskrit lan- 
 guage a . In both respects there is a remarkable 
 congruity between the Celtic and the Sanskrit. 
 There is a third series of variations in words com- 
 mon to the Sanskrit and several European idioms, 
 in which the Celtic dialects are more defective than 
 some other ancient languages of Europe and of the 
 East, I mean the declensions of nouns. 
 
 Welsh nouns make their plural number nearly on 
 the same principle as several of the European lan- 
 guages. They add terminations in i, au, ion, &c. 
 and they vary the interior vowels of words. 
 
 Welsh nouns have no cases properly so called, but 
 the want of them is supplied by prepositions which 
 have not coalesced with the words governed by them, 
 as they appear in other languages to have done in 
 such a manner as to give origin to cases b . 
 
 a The different forms of samasa and sandhi occupy a consi- 
 derable space in the Sanskrit grammars of Yadaraja and Vopa- 
 daiva. 
 
 b Such at least, according to the opinion defended by Bopp, 
 
186 EASTERN ORIGIN OF 
 
 In the Erse dialect nouns have a very peculiar 
 mode of declension. The following may serve as an 
 example : 
 
 An bard, a poet. 
 
 Sing. 
 
 Nom. 
 
 an bard, 
 
 PJur. 
 
 Norn. 
 
 na baird, 
 
 
 Gen. 
 
 an bhaird, 
 
 
 Gen. 
 
 na mbhard, 
 
 
 Dat. 
 
 cTn mbard, 
 
 
 Dat. 
 
 o na bardaibh 
 
 
 Ace. 
 
 an bard, 
 
 
 Ace. 
 
 na barda, 
 
 
 Voc. 
 
 a bhaird. 
 
 
 Voc. 
 
 a bharda. 
 
 It is worth while to notice particularly the dative 
 plural, which generally terminates in aibh, though 
 this perhaps admits of a variety, for it is given by 
 Lluyd in uibh. The terminations in uibh or aibh 
 are plainly related to the old Latin dative, in obus 
 and abus, which was probably the genuine and ori- 
 ginal form of this case in Latin. The Sanskrit 
 datives plural end in abhyus or abhyah, or at least in 
 bhyus after a vowel, as TJST^HT — raj abhyus ; 
 Latin, regibus ; Erse, righaibh or rioghaibh. 
 
 SECTION VI. 
 
 General Inference. 
 
 I have thus laid before my readers the most ob- 
 vious and striking analogies between the Celtic dia- 
 
 is the origin of Sanskrit cases, and therefore also of Greek and 
 Latin, which so nearly resemble them. Professor J. Grimm, 
 however, who has examined, with a view to this question, the 
 cases of the Mceso-Gothic and other Teutonic dialects, seems in- 
 clined to a different opinion, as far as those languages are con- 
 cerned. He concludes his inquiry into this subject (Bedeutung 
 der casusflexion) with the remark — Die Casuszeichen bleiben 
 mir ein geheimnisvolles element das ich lieber jedem worte 
 zuerkennen will, als es von einem auf alle ubrigen leiten. Th. I. 
 P. 835. 
 
THE CELTIC NATIONS. 187 
 
 lects, and the languages which are more generally al- 
 lowed to be of cognate origin with the Sanskrit, Greek, 
 and Latin. On the facts submitted to them, they will 
 form their own conclusion. Probably few persons 
 will hesitate in adopting the opinion, that the marks 
 of connexion are too decided and extensive to be re- 
 ferred to accident or casual intercourse, that they 
 are too deeply interwoven with the intimate structure 
 of the languages compared, to be explained on any 
 other principle than that which has been admitted 
 by so many writers in respect to the other great 
 families of languages belonging to the ancient popu- 
 lation of Europe, and that the Celtic people them- 
 selves are therefore of eastern origin, a kindred tribe 
 with the nations who settled on the banks of the 
 Indus, and on the shores of the Mediterranean and 
 of the Baltic. It is probable that several tribes 
 emigrated from their original seat in different stages 
 of advancement in respect to civilization and lan- 
 guage, and we accordingly find their idioms in very 
 different degrees of refinement; but an accurate ex- 
 amination and analysis of the intimate structure 
 and component materials of these languages, is still 
 capable of affording ample proofs of a common 
 origin. 
 
 My present inquiry has been professedly confined 
 to language ; and I must refer to my former work for 
 the confirmations which the inferences now deduced 
 may obtain from other sources. 
 
NOTE ON THE SEMITIC LANGUAGES. 
 
 AT the conclusion of a work designed to illustrate the 
 mutual affinity of those idioms which are termed collec- 
 tively Indo-European, it will not be improper to add a few 
 remarks on the relation between the latter and two other 
 families of languages, which have co-existed with them 
 from the earliest periods of history. 
 
 One of these is the class of idioms termed by German 
 philological writers Semitic languages. This designation 
 was, I believe, first suggested by Eichhorn, who has re- 
 marked that the three principal branches into which the 
 idioms belonging to this class divide themselves, viz. the 
 Hebrew or the dialect of Palestine and Phoenice, the Arabic, 
 and the Aramaean or northern Semitic spread over Syria 
 and Mesopotamia, are as nearly related to each other as the 
 Ionic, iEolic, and Doric dialects of Greek a . The term 
 Semitic has been thought by some to be objectionable, on 
 the ground that several of the nations who spoke the 
 languages so denominated, in common with the descend- 
 ants of Shem, were of Hamite origin, as the Phoenicians or 
 Canaanites. It has, however, got into general use, and 
 must therefore be retained. Schlozer, the learned editor of 
 Nestor's annals, has proposed on similar grounds to name 
 the Indo-European dialects Japetic languages, most of 
 the nations by whom they are spoken having descended, as 
 it is generally believed, from Japhet. We might perhaps, 
 with less hesitation, apply the term Hamite to the third 
 family of languages, to which I have alluded. I refer prin- 
 cipally to the dialects of the old Egyptian speech, the 
 Coptic, Sahidic, and Bashmuric, including conjecturally, 
 
 a Einleitung in das Alte Testament, von Joh. G. Eichhorn. B. I. p. 49. 
 Dritt. Ausg. 
 
 b A. L. Schlozer, von den Ohaldaeern, Repertorinm fur biblische und mor- 
 genlamdische literatur. th. 11. 
 
190 NOTE ON THE 
 
 until the mutual relations of these languages shall have 
 been more fully investigated, several idioms spoken by 
 races of Africa, in whose history marks are to be found of 
 connection with the ancient subjects of the Pharaohs. 
 One of these is the dialect of the Nouba, Bar&bra or Ber- 
 berins of the Upper Nile, a race who strikingly resemble 
 the ancient Egyptians in their physical characters, as we 
 know by comparing the present Berberins with the paint- 
 ings and mummies preserved in the Egyptian catacombs c . 
 They are probably the offspring of the ancient Ethiopians 
 of Meroe, who in a later age were the subjects of queen 
 Candace. Another race, much more extensively spread in 
 Africa, are the descendants of the Libyans. The Showiah, 
 spoken by the Kabyles among the Tunisian mountains, 
 and the Amazigh of the Berbers and the Shilha of mount 
 Atlas, are dialects of their language, which has been traced 
 from the Oasis of Siwah to the Atlantic ocean, and which 
 seems also to have been the idiom of the Guanches in the 
 Canary islands, whose curiously desiccated mummies bear, 
 as Blumenbach has shewn, much resemblance to those 
 of the Egyptians, and indicate a very ancient connection 
 among the tribes of northern Africa in arts and customs d . 
 
 It seems to be the prevalent opinion among philological 
 writers of the present time, that the three classes of lan- 
 guages above referred to, namely, the Indo-European, the 
 Semitic, and the Egyptian dialects, are entirely uncon- 
 nected with each other, and betray no traces even of the 
 most remote affinity. A late writer, whom I have before 
 cited, seems to hold this opinion in a very decided manner 
 in reference to the Semitic and Indo-European idioms. 
 
 " It has been asserted," says Col. Vans Kennedy, " that 
 " the Greek contains more Arabic words than is generally 
 " supposed ; but until these words are produced, and their 
 " identity established, I must doubt the correctness of this 
 " assertion ; for I have never been able to discover any such 
 " identical terms. " He concludes, that " the portion of 
 
 c Researches into the Physical History of Mankind, vol. i. 
 d Blumenbach's Decades Craniorum. 
 
SEMITIC LANGUAGES. 191 
 
 u Asia which comprises Arabia and Syria was peopled, or 
 " at least inhabited from time immemorial, by a distinct 
 " race of men, who spoke a language peculiar to them- 
 " selves ; and that this language, with its cognate dialects, 
 " has been at all times confined to these countries, and 
 " that it never has extended its influence beyond their 
 " limits, except to a small part of Africa." 
 
 It must be allowed, that the Semitic dialects constitute a 
 very distinct department of languages, which can by no 
 means be associated or brought into the same class with the 
 Indo-European idioms \ yet it is by far too much to affirm 
 that there are no traces of connection between the two 
 classes. In the preceding remarks upon the Indo-European 
 languages, some futures have been pointed out which dis- 
 play a remarkable analogy to the well-known characters of 
 the Hebrew and its cognate dialects ; I shall only instance 
 the system of pronominal suffixes. This is one point in 
 which the Celtic, at the same time that it appears to be 
 the least artificial and grammatically cultivated of the 
 Indo-European languages, forms an intermediate link be- 
 tween them and the Semitic, or perhaps indicates a state 
 of transition from the characters of one of these classes of 
 languages to those of the other. 
 
 In my work on the Physical History of Mankind, I ven- 
 tured to remark, that a very considerable number of the 
 vocables belonging to the Semitic dialects may be recog- 
 nised in some of the Indo-European languages. It would 
 be foreign to the object of the present work to enter at 
 large into a proof of this opinion ; but I shall here adduce a 
 few instances of undoubtedly cognate words, which will 
 be sufficient to render it probable that a much larger 
 number may be discovered by an extensive and accurate 
 research. 
 
 Among the first ten numerals there are a few terms 
 which appear to be cognate. 
 
 Semitic Dialects. Indo-European Languages. 
 
 i. echad, Heb. aika, Sansk. 
 
 yik or eek, Pers. 
 
li): 
 
 NOTi; OX THtt 
 
 Semitic Dialects. Indo-European Languages 
 
 3. Ordinal in Chald. 3. Ordinal in Sanskrit. 
 
 ^■TPTTI — tlithay, (Dan. ii. 39.) tritaya. 
 
 6. shesh, Heb. shash, Sansk. 
 
 7. shevang, Chald. seven, sibun, &c. 
 
 The following are some verbal roots and nouns which 
 are evidently of the same origin. Among them are verbs 
 which nearly resemble the two verbs substantive already 
 traced in the Indo-European language. 
 
 r^2L — bith from J"fi3, — buth, beith, Erse. 
 
 Chald. to tarry, dwell, (Dan. vi. 
 18.) often used in the Targum 
 for YO' I n A ra °i c this word is 
 C>Ij — bat, or *^\aj — beit, to 
 tarry, be situated e . 
 
 But the verb in Hebrew 
 which closely corresponds with 
 the Indo-European verb sub- 
 stantive, and in fact identical 
 with it, is tiJ* — yesh, he is ; in 
 
 Arab, (j*^! — is f . 
 
 khol, (whole, all) 
 
 hor, horim, hori, (mountain) 
 laish, (lion) 
 leom, (people) 
 luach, (a stone table) 
 
 loang, (throat, swallow) 
 tor, tori, (Chald.) 
 keran, (Chald.) 
 giivra, (Chald.) 
 
 bydh, bod, W. 
 buden, Pers. 
 bhu, Sansk. 
 be, beon, Teut. 
 
 It can hardly be doubted 
 that ffi is a real cognate of the 
 Indo-European verb. See p. 
 83. above. 
 is, Erse. 
 
 ys, Welsh, 
 as, Sanskrit. 
 &c. &c. 
 oXos. 
 h61h, Welsh. 
 
 opos, opoi. 
 
 Xecov. 
 
 Xeo)?. 
 
 lhech, (a flat stone.) 
 
 \10OS. 
 
 lung, lingua. 
 ravpos, ravpoi. 
 cornu, Kepas. 
 gwr, vir. 
 
 e Buxtorf. Lex. Heb. p. 69. Michaelis Supplem. in Lex. Heb. voce J^^lS." 
 f Gesenius's Lexicon, Cambridge edition. See Genes, xxviii. 16. Deuteron. 
 xxix. 17, &c. p. 316. 
 
SEMITIC LANGUAGES. 
 
 193 
 
 } 
 
 Semitic Dialects. 
 
 terez, in Chald. WW— wrgha 
 C% being, however, often mu- 
 table into d, which would make 
 aerda.) 
 ^y — gnabi (clouds) 
 
 *^V-» pronounced Nep in Ori- 
 
 gen's Hexapla, Hos. ii. I. (a 
 
 youth.) 
 
 mV2 — naaerah, puella. 
 
 ganaz 
 
 chetoneth 
 
 sepel (Jud. v. 25.) (a cup) 
 
 yayin 
 
 Here we find *> standing for 
 the digamma or vau. An in- 
 sertion of the vau will convert 
 many Hebrew into Indo-Euro- 
 pean words, as 
 
 2. yadang (know,) in Pih. yid- 1 
 dang. j 
 
 3. halak 
 
 4. rong (evil) 
 
 5. chiva (an animal) chavah, \ 
 life. J 
 
 6. ragang 
 
 kum, (arise, come) 
 
 laat, (to hide, secret) 
 
 arar, aru, ar, (curse) 
 
 ad 
 
 lakak, also lakhak and likhak 
 
 TIM— ud 
 
 thiggenu (Gen. iii. 5.) 
 
 tardemah ex radam 
 
 moth 
 
 moth and 1 
 
 meth J 
 
 olem, (age) 
 
 Indo-European Languages. 
 erda. 
 erth. 
 
 dhara, daiar. 
 terra. 
 
 nabhah, Sansk. 
 nubes. 
 
 narah, Sansk. 
 avi]p. 
 
 narl, Sansk. 
 
 ydvos. 
 
 simpulum . 
 
 vidan, eldciv. 
 
 o?Sa, vaida. 
 
 walk. 
 
 wrong. 
 
 vivo, viva, 
 
 jiva, Sansk. 
 
 pr)ywpi, frango, i. e. frago. 
 
 komm, come. 
 
 lateo, Xrj6e. 
 
 apa, apdopai. 
 
 ad, at. 
 
 liha, Sansk. \eLxa>, lick. 
 
 udus, ud, Sansk. vbeop, &c. 
 
 diyydvere, (Gen. Hi. 5.) 
 
 traum, dream, 
 motus. 
 
 meath JErse, to die. 
 meatham -> 
 
 olim, Lat. 
 
 o 
 
194 NOTE ON THE SEMITIC LANGUAGES., L A r 
 
 Semitic Dialects. Indo-European Lan^naires. 
 
 charats, (cleave, wound slightly, ") x a P"°' (r(B - 
 
 Gesenius) -* scratch. 
 
 H^^— laghah, (to babble) 
 
 t^v — laghaz, (speak barba- 
 rously) 
 
 ^7 — laghag, (laugh and speak ? 
 
 unintelligibly) 
 
 In all these we recognise 
 
 one element. 
 
 The same element in XaK<?a>, 
 laugh, liicheln, loquor ? 
 
 PRONOUNS. 
 
 atta, pron. (thou) tu. 
 
 ta, suffix. ta, t'ha, suffix in Sanskrit. 
 
 hi, (she) hi, si. 
 
 hu, (he) evo. 
 
 anu, suffix nu. ni, nos, nau. 
 
 No sufficient comparison of the Egyptian and other 
 Northern African dialects with each other and with the 
 Semitic languages has been made to allow of any ge- 
 neral statement as to their relations. I may however ob- 
 serve, that those who have denied that any affinity can here 
 be traced appear rather to have presumed the fact than to 
 have proved it. The affinity of some striking words 
 among the personal pronouns in the Egyptian and Hebrew 
 languages is such as to excite a strong suspicion that more 
 extensive resemblances exist, though it does not appear 
 probable that the idioms of Northern Africa are even so 
 nearly related to the Semitic, as the latter are to the Indo- 
 European languages. 
 
 ERRATA. 
 
 P. 10. discovered read dissevered 
 P. 7^. display read displays. 
 P. 93. are read is. 
 they have read it has. 
 
LIBRARY OF CONGRESS 
 
 II llll II II III 
 
 003 097 242 9