Lecture VI. Comparative Grammar.

The genealogical classification of the Aryan languages was founded, as we saw, on a close comparison of the grammatical characteristics of each; and it is the object of such works as Bopp's “Comparative Grammar” to show that the grammatical articulation of Sanskrit, Zend, Greek, Roman, Celtic, Teutonic, and Slavonic, was produced once and for all; and that the apparent differences in the terminations of Sanskrit, Greek, and Latin, must be explained by laws of phonetic decay, peculiar to each dialect, which modified the original common Aryan type, and changed it into so many national languages. It might seem, therefore, as if the object of comparative grammar was attained as soon as the exact genealogical relationship of languages had been settled; and those who only look to the higher problems of the science of language have not hesitated to declare that “there is no painsworthy difficulty nor dispute about declension, number, case, and gender of nouns.” But although it is certainly true that comparative grammar is only a means, and that it has well nigh taught us all that it has to teach,—at least in the Aryan family of speech,—it is to be hoped that, in the science of language, it will always retain that prominent place which it has obtained through the labors of Bopp, [pg 215] Grimm, Pott, Benfey, Curtius, Kuhn, and others. Besides, comparative grammar has more to do than simply to compare. It would be easy enough to place side by side the paradigms of declension and conjugation in Sanskrit, Greek, Latin, and the other Aryan dialects, and to mark both their coincidences and their differences. But after we have done this, and after we have explained the phonetic laws which cause the primitive Aryan type to assume that national variety which we admire in Sanskrit, Greek, and Latin, new problems arise of a more interesting nature. We know that grammatical terminations, as they are now called, were originally independent words, and had their own purpose and meaning. Is it possible, after comparative grammar has established the original forms of the Aryan terminations, to trace them back to independent words, and to discover their original purpose and meaning? You will remember that this was the point from which we started. We wanted to know why the termination d in I loved should change a present into a past act. We saw that before answering this question we had to discover the most original form of this termination by tracing it from English to Gothic, and afterwards, if necessary, from Gothic to Sanskrit. We now return to our original question, namely, What is language that a mere formal change, such as that of I love into I loved, should produce so very material a difference?

Let us clearly see what we mean if we make a distinction between the radical and formal elements of a language; and by formal elements I mean not only the terminations of declension and conjugation, but all derivative elements; all, in fact, that is not radical. Our view on the origin of language must chiefly depend on [pg 216] the view which we take of these formal, as opposed to the radical, elements of speech. Those who consider that language is a conventional production, base their arguments principally on these formal elements. The inflections of words, they maintain, are the best proof that language was made by mutual agreement. They look upon them as mere letters or syllables without any meaning by themselves; and if they were asked why the mere addition of a d changes I love into I loved, or why the addition of the syllable rai gave to j'aime, I love, the power of a future, j'aimerai, they would answer, that it was so because, at a very early time in the history of the world, certain persons, or families, or clans, agreed that it should be so.

This view was opposed by another which represents language as an organic and almost a living being, and explains its formal elements as produced by a principle of growth inherent in its very nature. “Languages,”[204] it is maintained, “are formed by a process, not of crystalline accretion, but of germinal development. Every essential part of language existed as completely (although only implicitly) in the primitive germ, as the petals of a flower exist in the bud before the mingled influences of the sun and the air caused it to unfold.” This view was first propounded by Frederick Schlegel,[205] [pg 217] and it is still held by many with whom poetical phraseology takes the place of sound and severe reasoning.

The science of language adopts neither of these views. As to imagining a congress for settling the proper exponents of such relations as nominative, genitive, singular, plural, active, and passive, it stands to reason that if such abstruse problems could have been discussed in a language void of inflections, there was no inducement for agreeing on a more perfect means of communication. And as to imagining language, that is to say nouns and verbs, endowed with an inward principle of growth, all we can say is, that such a conception is really inconceivable. Language may be conceived as a production, but it cannot be conceived as a substance that could itself produce. But the science of language has nothing to do with mere theories, whether conceivable or not. It collects facts, and its only object is to account for these facts, as far as possible. Instead of looking on inflections in general either as conventional signs or natural excrescences, it takes each termination by itself, establishes its most primitive form by means of comparison, and then treats that primitive syllable as it would treat [pg 218] any other part of language,—namely, as something which was originally intended to convey a meaning. Whether we are still able to discover the original intention of every part of language is quite a different question, and it should be admitted at once that many grammatical forms, after they have been restored to their most primitive type, are still without an explanation. But with every year new discoveries are made by means of careful inductive reasoning. We become more familiar every day with the secret ways of language, and there is no reason to doubt that in the end grammatical analysis will be as successful as chemical analysis. Grammar, though sometimes very bewildering to us in its later stages, is originally a much less formidable undertaking than is commonly supposed. What is grammar after all but declension and conjugation? Originally declension could not have been anything but the composition of a noun with some other word expressive of number and case. How the number was expressed, we saw in a former lecture; and the same process led to the formation of cases.

Thus the locative is formed in various ways in Chinese:[206] one is by adding such words as ćung, the middle, or néi, inside. Thus, kûŏ-ćung, in the empire; i sûí ćung, within a year. The instrumental is formed by the preposition ẏ, which preposition is an old root, meaning to use. Thus ẏ ting, with a stick, where in Latin we should use the ablative, in Greek the dative. Now, however complicated the declensions, regular and irregular, may be in Greek and Latin, we may be certain that originally they were formed by this simple method of composition.

There was originally in all the Aryan languages a case expressive of locality, which grammarians call the locative. In Sanskrit every substantive has its locative, as well as its genitive, dative, and accusative. Thus, heart in Sanskrit is hṛid; in the heart, is hṛidi. Here, therefore, the termination of the locative is simply short i. This short i is a demonstrative root, and in all probability the same root which in Latin produced the preposition in. The Sanskrit hṛidi represents, therefore, an original compound, as it were, heart-within, which gradually became settled as one of the recognized cases of nouns ending in consonants. If we look to Chinese,[207] we find that the locative is expressed there in the same manner, but with a greater freedom in the choice of the words expressive of locality. “In the empire,” is expressed by kûŏ ćung; “within a year,” is expressed by ĭ sûí ćung. Instead of ćung, however, we might have employed other terms also, such as, for instance, néi, inside. It might be said that the formation of so primitive a case as the locative offers little difficulty, but that this process of composition fails to account for the origin of the more abstract cases, the accusative, the dative, and genitive. If we derive our notions of the cases from philosophical grammar, it is true, no doubt, that it would be difficult to convey by a simple composition the abstract relations supposed to be expressed by the terminations of the genitive, dative, and accusative. But remember that these are only general categories under which philosophers and grammarians endeavored to arrange the facts of language. The people with whom language grew up knew nothing of datives and accusatives. Everything that is abstract [pg 220] in language was originally concrete. If people wanted to say the King of Rome, they meant really the King at Rome, and they would readily have used what I have just described as the locative; whereas the more abstract idea of the genitive would never enter into their system of thought. But more than this, it can be proved that the locative has actually taken, in some cases, the place of the genitive. In Latin, for instance, the old genitive of nouns in a was as. This we find still in pater familiâs, instead of pater familiæ. The Umbrian and Oscan dialects retained the s throughout as the sign of the genitive after nouns in a. The æ of the genitive was originally ai, that is to say, the old locative in i. “King of Rome,” if rendered by Rex Romæ, meant really “King at Rome.” And here you will see how grammar, which ought to be the most logical of all sciences, is frequently the most illogical. A boy is taught at school, that if he wants to say “I am staying at Rome,” he must use the genitive to express the locative. How a logician or grammarian can so twist and turn the meaning of the genitive as to make it express rest in a place, is not for us to inquire; but, if he succeeded, his pupil would at once use the genitive of Carthage (Carthaginis) or of Athens (Athenarum) for the same purpose, and he would then have to be told that these genitives could not be used in the same manner as the genitive of nouns in a. How all this is achieved by what is called philosophical grammar, we know not; but comparative grammar at once removes all difficulty. It is only in the first declension that the locative has supplanted the genitive, whereas Carthaginis and Athenarum, being real genitives, could never be employed to express a locative. [pg 221] A special case, such as the locative, may be generalized into the more general genitive, but not vice versâ.

You see thus by one instance how what grammarians call a genitive was formed by the same process of composition which we can watch in Chinese, and which we can prove to have taken place in the original language of the Aryans. And the same applies to the dative. If a boy is told that the dative expresses a relation of one object to another, less direct than that of the accusative, he may well wonder how such a flying arch could ever have been built up with the scanty materials which language has at her disposal; but he will be still more surprised if, after having realized this grammatical abstraction, he is told that in Greek, in order to convey the very definite idea of being in a place, he has to use after certain nouns the termination of the dative. “I am staying at Salamis,” must be expressed by the dative Salamînĭ. If you ask why? Comparative grammar again can alone give an answer. The termination of the Greek dative in i, was originally the termination of the locative. The locative may well convey the meaning of the dative, but the faded features of the dative can never express the fresh distinctness of the locative. The dative Salamînĭ was first a locative. “I live at Salamis,” never conveyed the meaning, “I live to Salamis.” On the contrary, the dative, in such phrases as “I give it to the father,” was originally a locative; and after expressing at first the palpable relation of “I give it unto the father,” or “I place it on or in the father,” it gradually assumed the more general, the less local, less colored aspect which logicians and grammarians ascribe to their datives.[208]

If the explanation just given of some of the cases in Greek and Latin should seem too artificial or too forced, we have only to think of French in order to see exactly the same process repeated under our eyes. The most abstract relations of the genitive, as, for instance, “The immortality of the soul” (l'immortalité de l'âme); or of the dative, as, for instance, “I trust myself to God” (je me fie à Dieu), are expressed by prepositions, such as de and ad, which in Latin had the distinct local meanings of “down from,” and “towards.” Nay, the English of and to, which have taken the place of the German terminations s and m, are likewise prepositions of an originally local character. The only difference between our cases and those of the ancient languages consists in this,—that the determining element is now placed before the word, whereas, in the original language of the Aryans, it was placed at the end.

What applies to the cases of nouns, applies with equal truth to the terminations of verbs. It may seem difficult to discover in the personal terminations of Greek and Latin the exact pronouns which were added to a verbal base in order to express, I love, thou lovest, he loves; but it stands to reason that originally these terminations must have been the same in all languages,—namely, personal pronouns. We may be puzzled by the terminations of thou lovest and he loves, where st and s can hardly be identified with the modern thou and he; but we have only to place all the Aryan dialects together, and we shall see at once that they point back to an original set of terminations which can easily be brought to tell their own story.

Let us begin with modern formations, because we have here more daylight for watching the intricate and [pg 223] sometimes wayward movements of language; or, better still, let us begin with an imaginary case, or with what may be called the language of the future, in order to see quite clearly how, what we should call grammatical forms, may arise. Let us suppose that the slaves in America were to rise against their masters, and, after gaining some victories, were to sail back in large numbers to some part of Central Africa, beyond the reach of their white enemies or friends. Let us suppose these men availing themselves of the lessons they had learnt in their captivity, and gradually working out a civilization of their own. It is quite possible that some centuries hence, a new Livingstone might find among the descendants of the American slaves, a language, a literature, laws, and manners, bearing a striking similitude to those of his own country. What an interesting problem for any future historian and ethnologist! Yet there are problems in the past history of the world of equal interest, which have been and are still to be solved by the student of language. Now I believe that a careful examination of the language of the descendants of those escaped slaves would suffice to determine with perfect certainty their past history, even though no documents and no tradition had preserved the story of their captivity and liberation. At first, no doubt, the threads might seem hopelessly entangled. A missionary might surprise the scholars of Europe by an account of that new African language. He might describe it at first as very imperfect—as a language, for instance, so poor that the same word had to be used to express the most heterogeneous ideas. He might point out how the same sound, without any change of accent, meant true, a ceremony, a workman, and was used also [pg 224] as a verb in the sense of literary composition. All these, he might say, are expressed in that strange dialect by the sound rait (right, rite, wright, write). He might likewise observe that this dialect, as poor almost as Chinese, had hardly any grammatical inflections, and that it had no genders, except in a few words such as man-of-war, and a railway-engine, which were both conceived as feminine beings, and spoken of as she. He might then mention an even more extraordinary feature, namely, that although this language had no terminations for the masculine and feminine genders of nouns, it employed a masculine and feminine termination after the affirmative particle, according as it was addressed to a lady or a gentleman. Their affirmative particle being the same as the English, Yes, they added a final r to it if addressed to a man, and a final m if addressed to a lady: that is to say, instead of simply saying, Yes, these descendants of the escaped American slaves said Yesr to a man, and Yesm to a lady.

Absurd as this may sound, I can assure you that the descriptions which are given of the dialects of savage tribes, as explained for the first time by travellers or missionaries, are even more extraordinary. But let us consider now what the student of language would have to do, if such forms as Yeśr and Yeśm were, for the first time, brought under his notice. He would first have to trace them back historically, as far as possible to their more original types, and if he discovered their connection with Yes Sir and Yes Ma'm, he would point out how such contractions were most likely to spring up in a vulgar dialect. After having traced back the Yesr and Yesm of the free African negroes [pg 225] to the idiom of their former American masters, the etymologist would next inquire how such phrases as Yes Sir and Yes Madam, came to be used on the American continent.

Finding nothing analogous in the dialects of the aboriginal inhabitants of America, he would be led, by a mere comparison of words, to the languages of Europe, and here again, first to the language of England. Even if no historical documents had been preserved, the documents of language would show that the white masters, whose language the ancestors of the free Africans adopted during their servitude, came originally from England, and, within certain limits, it would even be possible to fix the time when the English language was first transplanted to America. That language must have passed, at least, the age of Chaucer before it migrated to the New World. For Chaucer has two affirmative particles, Yea and Yes, and he distinguishes between the two. He uses Yes only in answer to negative questions. For instance, in answer to “Does he not go?” he would say, Yes. In all other cases Chaucer uses Yea. To a question, “Does he go?” he would answer Yea. He observes the same distinction between No and Nay, the former being used after negative, the latter after all other questions. This distinction became obsolete soon after Sir Thomas More,[209] and it must have become obsolete before phrases such as Yes Sir and Yes Madam could have assumed their stereotyped character.

But there is still more historical information to be gained from these phrases. The word Yes is Anglo-Saxon, the same as the German Ja, and it therefore [pg 226] reveals the fact that the white masters of the American slaves who crossed the Atlantic after the time of Chaucer, had crossed the Channel at an earlier period after leaving the continental fatherland of the Angles and Saxons. The words Sir and Madam tell us still more. They are Norman words, and they could only have been imposed on the Anglo-Saxons of Britain by Norman conquerors. They tell us more than this. For these Normans or Northmen spoke originally a Teutonic dialect, closely allied to Anglo-Saxon, and in that dialect words such as Sir and Madam could never have sprung up. We may conclude therefore that, previous to the Norman conquest, the Teutonic Northmen must have made a sufficiently long stay in one of the Roman provinces to forget their own and adopt the language of the Roman Provincials.

We may now trace back the Norman Madam to the French Madame, and we recognize in this a corruption of the Latin Mea domina, my mistress. Domina was changed into domna, donna, and dame, and the same word Dame was also used as a masculine in the sense of lord, as a corruption of Domino, Domno and Donno. The temporal lord ruling as ecclesiastical seigneur under the bishop, was called a vidame, as the Vidame of Chartres, &c. The French interjection Dame! has no connection with a similar exclamation in English, but it simply means Lord! Dame-Dieu in old French is Lord God. A derivative of Domina, mistress, was dominicella, which became Demoiselle and Damsel. The masculine Dame for Domino, Lord, was afterwards replaced by the Latin Senior, a translation of the German elder. This word elder was a title of honor, and we have it still both in alderman, and in what is originally the same, the English [pg 227] Earl, the Norse Jarl, a corruption of the A.-S. ealdor. This title Senior, meaning originally older, was but rarely[210] applied to ladies as a title of honor. Senior was changed into Seigneur, Seigneur into Sieur, and Sieur soon dwindled down to Sir.

Thus we see how in two short phrases, such as Yesr and Yesm, long chapters of history might be read. If a general destruction of books, such as took place in China under the Emperor Thsin-chi-hoang-ti (213 b. c.), should sweep away all historical documents, language, even in its most depraved state, would preserve the secrets of the past, and would tell future generations of the home and migrations of their ancestors from the East to the West Indies.

It may seem startling at first to find the same name, the East Indies and the West Indies, at the two extremities of the Aryan migrations; but these very names are full of historical meaning. They tell us how the Teutonic race, the most vigorous and enterprising of all the members of the Aryan family, gave the name of West Indies to the country which in their world-compassing migrations they imagined to be India itself; how they discovered their mistake and then distinguished between the East Indies and West Indies; how they planted new states in the west, and regenerated the effete kingdoms in the east; how they preached Christianity, and at last practised it by abolishing slavery of body and mind among the slaves of West-Indian landholders, and the slaves of Brahmanical soulholders, till they greeted at last the very homes from which the Aryan family had started when setting [pg 228] out on their discovery of the world. All this, and even more, may be read in the vast archives of language. The very name of India has a story to tell, for India is not a native name. We have it from the Romans, the Romans from the Greeks, the Greeks from the Persians. And why from the Persians? Because it is only in Persian that an initial s is changed into h, which initial h was as usual dropped in Greek. It is only in Persian that the country of the Sindhu (sindhu is the Sanskrit name for river), or of the seven sindhus, could have been called Hindia or India instead of Sindia. Unless the followers of Zoroaster had pronounced every s like h, we should never have heard of the West Indies!

We have thus seen by an imaginary instance what we must be prepared for in the growth of language, and we shall now better understand why it must be laid down as a fundamental principle in Comparative Grammar to look upon nothing in language as merely formal, till every attempt has been made to trace the formal elements of language back to their original and substantial prototypes. We are accustomed to the idea of grammatical terminations modifying the meaning of words. But words can be modified by words only; and though in the present state of our science it would be too much to say that all grammatical terminations have been traced back to original independent words, so many of them have, even in cases where only a single letter was left, that we may well lay it down as a rule that all formal elements of language were originally substantial. Suppose English had never been written down before the time of Piers Ploughman. What should we make of such a form as [pg 229] nadistou,[211] instead of ne hadst thou? Ne rechi instead of I reck not? Al ô'm in Dorsetshire is all of them. I midden is I may not; I cooden, I could not. Yet the changes which Sanskrit had undergone before it was reduced to writing, must have been more considerable by far than what we see in these dialects.

Let us now look to modern classical languages such as French and Italian. Most of the grammatical terminations are the same as in Latin, only changed by phonetic corruption. Thus j'aime is ego amo, tu aimes, tu amas, il aime, ille amat. There was originally a final t in French il aime, and it comes out again in such phrases as aime-t-il? Thus the French imperfect corresponds to the Latin imperfect, the Parfait défini to the Latin perfect. But what about the French future? There is no similarity between amabo and j'aimerai. Here then we have a new grammatical form, sprung up, as it were, within the recollection of men; or, at least, in the broad daylight of history. Now, did the termination rai bud forth like a blossom in spring? or did some wise people meet together to invent this new termination, and pledge themselves to use it instead of the old termination bo? Certainly not. We see first of all that in all the Romance languages the terminations of the future are identical with the auxiliary verb to have.[212] In French you find—

j'ai and je chanter-ai nous avons and nous chanterons.

tu as and tu chanter-as vous avez and vous chanterez.

il a and il chanter-a ils ont and ils chanteront.

But besides this, we actually find in Spanish and [pg 230] Provençal the apparent termination of the future used as an independent word and not yet joined to the infinitive. We find in Spanish, instead of “lo hare,” I shall do it, the more primitive form hacer lo he; i.e., facere id habeo. We find in Provençal, dir vos ai instead of je vous dirai; dir vos em instead of nous vous dirons. There can be no doubt, therefore, that the Romance future was originally a compound of the auxiliary verb to have with an infinitive; and I have to say, easily took the meaning of I shall say.

Here, then, we see clearly how grammatical forms arise. A Frenchman looks upon his futures as merely grammatical forms. He has no idea, unless he is a scholar, that the terminations of his futures are identical with the auxiliary verb avoir. The Roman had no suspicion that amabo was a compound; but it can be proved to contain an auxiliary verb as clearly as the French future. The Latin future was destroyed by means of phonetic corruption. When the final letters lost their distinct pronunciation it became impossible to keep the imperfect amabam separate from the future amabo. The future was then replaced by dialectical regeneration, for the use of habeo with an infinitive is found in Latin, in such expressions as habeo dicere, I have to say, which would imperceptibly glide into I shall say.[213] In fact, wherever we look we see that, the future is expressed by means of composition. We have in English I shall and thou wilt, which mean originally I am bound and thou intendest. In German we use werden, the Gothic vairthan, which means originally to go, to turn towards. In modern Greek we find thelō, I will, in thelō dōsei, I shall give. In Roumansch we meet [pg 231] with vegnir, to come, forming the future veng a vegnir, I shall come; whereas in French je viens de dire, I come from saying, is equivalent to “I have just said.” The French je vais dire is almost a future, though originally it is vado dicere, I go to say. The Dorsetshire, “I be gwâin to goo a-pickèn stuones,” is another case in point. Nor is there any doubt that in the Latin bo of amabo we have the old auxiliary bhû, to be, and in the Greek future in σω, the old auxiliary as, to be.[214]

We now go back another step, and ask the question which we asked many times before, How can a mere d produce so momentous a change as that from I love to I loved? As we have learnt in the meantime that English goes back to Anglo-Saxon, and is closely related to continental Saxon and Gothic, we look at once to the Gothic imperfect in order to see whether it has preserved any traces of the original compound; for, after what we have seen in the previous cases, we are no doubt prepared to find here, too, grammatical terminations mere remnants of independent words.

In Gothic there is a verb nasjan, to nourish. Its preterite is as follows:—

Singular.Dual.Plural.
nas-i-danas-i-dêdunas-i-dêdum.
nas-i-dêsnas-i-dêtutsnas-i-dêduþ.
nas-i-da——nas-i-dedun.

The subjunctive of the preterite:

Singular.Dual.Plural.
nas-i-dêdjaunas-i-dêdeivanas-i-dêdeima.
nas-i-dêdeisnas-i-dêdeitsnas-i-dêdeiþ.
nas-i-dêdi——nas-i-dêdeina.

This is reduced in Anglo-Saxon to:

Singular.Plural.
ner-ë-dener-ë-don.
ner-ë-destner-ë-don.
ner-ë-dener-ë-don.

Subjunctive:

ner-ë-dener-ë-don.
ner-ë-dener-ë-don.
ner-ë-dener-ë-don.

Let us now look to the auxiliary verb to do, in Anglo-Saxon:

Singular.Plural.
didedidon.
didestdidon.
didedidon.

If we had only the Anglo-Saxon preterite nerëde and the Anglo-Saxon dide, the identity of the de in nerëde with dide would not be very apparent. But here you will perceive the advantage which Gothic has over all other Teutonic dialects for the purposes of grammatical comparison and analysis. It is in Gothic, and in Gothic in the plural only, that the full auxiliary dêdum, dêduþ, dêdun has been preserved. In the Gothic singular nasida, nasidês, nasida stand for nasideda, nasidedês, [pg 233] nasideda. The same contraction has taken place in Anglo-Saxon, not only in the singular but in the plural also. Yet, such is the similarity between Gothic and Anglo-Saxon that we cannot doubt their preterites having been formed on the same last. If there be any truth in inductive reasoning, there must have been an original Anglo-Saxon preterite,[215]

Singular.Plural.
ner-ë-didener-ë-didon.
ner-ë-didestner-ë-didon.
ner-ë-didener-ë-didon.

And as ner-ë-dide dwindled down to nerëde, so nerëde would, in modern English, become nered. The d of the preterite, therefore, which changes I love into I loved is originally the auxiliary verb to do, and I loved is the same as I love did, or I did love. In English dialects, as, for instance, in the Dorset dialect, every preterite, if it expresses a lasting or repeated action, is formed by I did,[216] and a distinction is thus established between “'e died eesterdae,” and “the vo'ke did die by scores;” though originally died is the same as die did.

It might be asked, however, very properly, how did itself, or the Anglo-Saxon dide, was formed, and how it received the meaning of a preterite. In dide the final de is not termination, but it is the root, and the first syllable di is a reduplication of the root, the fact being that all preterites of old, or, as they are called, strong verbs, were formed as in Greek and Sanskrit by means of reduplication, reduplication being one of the principal means by which roots were invested with a verbal character.[217] The root do in Anglo-Saxon is the same [pg 234] as the root thē in tithēmi in Greek, and the Sanskrit root dhâ in dadâdmi. Anglo-Saxon dide would therefore correspond to Sanskrit dadhau, I placed.

Now, in this manner, the whole, or nearly the whole, grammatical framework of the Aryan or Indo-European languages has been traced back to original independent words, and even the slightest changes which at first sight seem so mysterious, such as foot into feet, or I find into I found, have been fully accounted for. This is what is called comparative grammar, or a scientific analysis of all the formal elements of a language preceded by a comparison of all the varieties which one and the same form has assumed in the numerous dialects of the Aryan family. The most important dialects for this purpose are Sanskrit, Greek, Latin, and Gothic; but in many cases Zend, or Celtic, or Slavonic dialects come in to throw an unexpected light on forms unintelligible in any of the four principal dialects. The result of such a work as Bopp's “Comparative Grammar” of the Aryan languages may be summed up in a few words. The whole framework of grammar—the elements of derivation, declension, and conjugation—had become settled before the separation of the Aryan family. Hence the broad outlines of grammar, in Sanskrit, Greek, Latin, Gothic, and the rest, are in reality the same; and the apparent differences can be explained by phonetic corruption, which is determined by the phonetic peculiarities of each nation. On the whole, the history of all the Aryan languages is nothing but a gradual process of decay. After the grammatical terminations of all these languages have been traced back to their most primitive form, it is possible, in many instances, to determine their original meaning. This, [pg 235] however, can be done by means of induction only; and the period during which, as in the Provençal dir vos ai, the component elements of the old Aryan grammar maintained a separate existence in the language and the mind of the Aryans had closed, before Sanskrit was Sanskrit or Greek Greek. That there was such a period we can doubt as little as we can doubt the real existence of fern forests previous to the formation of our coal fields. We can do even more. Suppose we had no remnants of Latin; suppose the very existence of Rome and of Latin were unknown to us; we might still prove, on the evidence of the six Romance dialects, that there must have been a time when these dialects formed the language of a small settlement; nay, by collecting the words which all these dialects share in common, we might, to a certain extent, reconstruct the original language, and draw a sketch of the state of civilization, as reflected by these common words. The same can be done if we compare Sanskrit, Greek, Latin, Gothic, Celtic, and Slavonic. The words which have as nearly as possible the same form and meaning in all the languages must have existed before the people, who afterwards formed the prominent nationalities of the Aryan family, separated; and, if carefully interpreted, they, too, will serve as evidence as to the state of civilization attained by the Aryans before they left their common home. It can be proved, by the evidence of language, that before their separation the Aryans led the life of agricultural nomads,—a life such as Tacitus describes that of the ancient Germans. They knew the arts of ploughing, of making roads, of building ships, of weaving and sewing, of erecting houses; they had counted at least as far as one hundred. They had domesticated [pg 236] the most important animals, the cow, the horse, the sheep, the dog; they were acquainted with the most useful metals, and armed with iron hatchets, whether for peaceful or warlike purposes. They had recognized the bonds of blood and the bonds of marriage; they followed their leaders and kings, and the distinction between right and wrong was fixed by laws and customs. They were impressed with the idea of a divine Being, and they invoked it by various names. All this, as I said, can be proved by the evidence of language. For if you find that languages like Greek, Latin, Gothic, Celtic, or Slavonic, which, after their first separation, have had but little contact with Sanskrit, have the same word, for instance, for iron which exists in Sanskrit, this is proof absolute that iron was known previous to the Aryan separation. Now, iron is ais in Gothic, and ayas in Sanskrit, a word which, as it could not have been borrowed by the Indians from the Germans or by the Germans from the Indians, must have existed previous to their separation. We could not find the same name for house in Sanskrit, Greek, Latin, Slavonic, and Celtic,[218] unless houses had been known before the separation of these dialects. In this manner a history of Aryan civilization has been written from the archives of language, stretching back to times far beyond the reach of any documentary history.[219]

The very name of Arya belongs to this history, and I shall devote the rest of this lecture to tracing the origin and gradual spreading of this old word. I had intended to include, in to-day's lecture, a short account [pg 237] of comparative mythology, a branch of our science which restores the original form and meaning of decayed words by the same means by which comparative grammar recovers the original form and meaning of terminations. But my time is too limited; and, as I have been asked repeatedly why I applied the name of Aryan to that family of language which we have just examined, I feel that I am bound to give an answer.

Ârya is a Sanskrit word, and in the later Sanskrit it means noble, of a good family. It was, however, originally a national name, and we see traces of it as late as the Law-book of the Mânavas, where India is still called Ârya-âvarta, the abode of the Âryas.[220] In the old Sanskrit, in the hymns of the Veda, ârya occurs frequently as a national name and as a name of honor, comprising the worshippers of the gods of the Brahmans, as opposed to their enemies, who are called in the Veda Dasyus. Thus one of the gods, Indra, who, in some respects, answers to the Greek Zeus, is invoked in the following words (Rigveda, i. 57, 8): “Know thou the Âryas, O Indra, and they who are Dasyus; punish the lawless, and deliver them unto thy servant! Be thou the mighty helper of the worshippers, and I will praise all these thy deeds at the festivals.”

In the later dogmatic literature of the Vedic age, the name of Ârya is distinctly appropriated to the three first castes—the Brahmans, Kshatriyas, Vaiśyas—as opposed to the fourth, or the Śûdras. In the Śatapatha-Brâhmaņa it is laid down distinctly: “Âryas are only the Brahmans, the Kshatriyas, and Vaiśyas, for they are admitted to the sacrifices. They shall not speak with everybody, but only with the Brahman, the [pg 238] Kshatriya, and the Vaiśya. If they should fall into a conversation with a Śûdra, let them say to another man, ‘Tell this Śûdra so.’ This is the law.”

In the Atharva-veda (iv. 20, 4; xix. 62, 1) expressions occur such as, “seeing all things, whether Śûdra or Ârya,” where Śûdra and Ârya are meant to express the whole of mankind.

This word ârya with a long â is derived from arya with a short a, and this name arya is applied in the later Sanskrit to a Vaiśya, or a member of the third caste.[221] What is called the third class must originally have constituted the large majority of the Brahmanic society, for all who were not soldiers or priests, were Vaiśyas. We may well understand, therefore, how a name, originally applied to the cultivators of the soil and householders, should in time have become a general name for all Aryans.[222] Why the householders were called arya is a question which would carry us too far at present. I can only state that the etymological signification of Arya seems to be “one who ploughs or tills,” and that it is connected with the root of arare. The Aryans would seem to have chosen this name for themselves as opposed to the nomadic races, the Turanians, whose original name Tura implies the swiftness of the horseman.

In India, as we saw, the name of Ârya, as a national name, fell into oblivion in later times, and was preserved only in the term Âryâvarta, the abode of the Aryans. But it was more faithfully preserved [pg 239] by the Zoroastrians who migrated from India to the north-west, and whose religion has been preserved to us in the Zend-avesta, though in fragments only. Now Airya in Zend means venerable, and is at the same time the name of the people.[223] In the first chapter of the Vendidád, where Ahuramazda explains to Zarathustra the order in which he created the earth, sixteen countries are mentioned, each, when created by Ahuramazda, being pure and perfect; but each being tainted in turn by Angro mainyus or Ahriman. Now the first of these countries is called Airyanem vaêjô, Arianum semen, the Aryan seed, and its position must have been as far east as the western slopes of the Belurtag and Mustag, near the sources of the Oxus and Yaxartes, the highest elevation of Central Asia.[224] From this country, which is called their seed, the Aryans advanced towards the south and west, and in the Zend-avesta the whole extent of country occupied by the Aryans is likewise called Airyâ. A line drawn from India along the Paropamisus and Caucasus Indicus in the east, following in the north the direction between the Oxus and Yaxartes,[225] then running along the Caspian Sea, so as to include Hyrcania and Râgha, then turning south-east on the borders of Nisaea, Aria (i.e. Haria), and the countries washed by the Etymandrus and Arachotus, would indicate the general horizon of the Zoroastrian world. It would be what is called in the fourth cardé of the Yasht of Mithra, “the whole space of Aria,” vîśpem airyô-śayanem (totum Ariæ situm).[226] Opposed to the Aryan we find in [pg 240] the Zend-avesta the non-Aryan countries (anairyâo dainhâvô),[227] and traces of this name are found in the Ἀναριάκαι, a people and town on the frontiers of Hyrcania.[228] Greek geographers use the name of Ariana in a wider sense even than the Zend-avesta. All the country between the Indian Ocean in the south and the Indus in the east, the Hindu-kush and Paropamisus in the north, the Caspian gates, Karamania, and the mouth of the Persian gulf in the west, is included by Strabo (xv. 2) under the name of Ariana; and Bactria is thus called[229] by him “the ornament of the whole of Ariana.” As the Zoroastrian religion spread westward, Persia, Elymais, and Media all claimed for themselves the Aryan title. Hellanicus, who wrote before Herodotus, knows of Aria as a name of Persia.[230] Herodotus (vii. 62) attests that the Medians called themselves Arii; and even for Atropatene, the northernmost part of Media, the name of Ariania (not Aria) has been preserved by Stephanus Byzantinus. As to Elymais its name has been derived from Ailama, a supposed corruption of Airyama.[231] The Persians, Medians, Bactrians, and Sogdians all spoke, as late as the [pg 241] time of Strabo,[232] nearly the same language, and we may well understand, therefore, that they should have claimed for themselves one common name, in opposition to the hostile tribes of Turan.

That Aryan was used as a title of honor in the Persian empire is clearly shown by the cuneiform inscriptions of Darius. He calls himself Ariya and Ariya-chitra, an Aryan and of Aryan descent; and Ahuramazda, or, as he is called by Darius, Auramazda, is rendered in the Turanian translation of the inscription of Behistun, “the god of the Aryans.” Many historical names of the Persians contain the same element. The great-grandfather of Darius is called in the inscriptions Ariyârâmna, the Greek Ariaramnēs (Herod, vii. 90). Ariobarzanēs (i.e. Euergetēs), Ariomanes (i.e. Eumenēs), Ariomardos, all show the same origin.[233]

About the same time as these inscriptions, Eudemos, a pupil of Aristotle, as quoted by Damascius, speaks of “the Magi and the whole Aryan race,”[234] evidently using Aryan in the same sense in which the Zend-avesta spoke of “the whole country of Aria.”

And when, after years of foreign invasion and occupation, Persia rose again under the sceptre of the Sassanians to be a national kingdom, we find the new national kings the worshippers of Masdanes, calling [pg 242] themselves, in the inscriptions deciphered by De Sacy,[235] “Kings of the Aryan and un-Aryan races;” in Pehlevi, Irân va Anirân; in Greek, Ἀριάνων καὶ Ἀναριάνων.

The modern name of Irán for Persia still keeps up the memory of this ancient title.

In the name of Armenia the same element of Arya has been supposed to exist.[236] The name of Armenia, however, does not occur in Zend, and the name Armina, which is used for Armenia in the cuneiform inscriptions, is of doubtful etymology.[237] In the language of Armenia, ari is used in the widest sense for Aryan or Iranian; it means also brave, and is applied more especially to the Medians.[238] The word arya, therefore, though not contained in the name of Armenia, can be proved to have existed in the Armenian language as a national and honorable name.

West of Armenia, on the borders of the Caspian Sea, we find the ancient name of Albania. The Armenians call the Albanians Aghovan, and as gh in Armenian stands for r or l, it has been conjectured by Boré, that in Aghovan also the name of Aria is contained. This seems doubtful. But in the valleys of the Caucasus we meet with an Aryan race speaking an [pg 243] Aryan language, the Os of Ossethi, and they call themselves Iron.[239]

Along the Caspian, and in the country washed by the Oxus and Yaxartes, Aryan and non-Aryan tribes were mingled together for centuries. Though the relation between Aryans and Turanians is hostile, and though there were continual wars between them, as we learn from the great Persian epic, the Shahnámeh, it does not follow that all the nomad races who infested the settlements of the Aryans, were of Tatar blood and speech. Turvaśa and his descendants, who represent the Turanians, are described in the later epic poems of India as cursed and deprived of their inheritance in India. But in the Vedas Turvaśa is represented as worshipping Aryan gods. Even in the Shahnámeh, Persian heroes go over to the Turanians and lead them against Iran, very much as Coriolanus led the Samnites against Rome. We may thus understand why so many Turanian or Scythian names, mentioned by Greek writers, should show evident traces of Aryan origin. Aspa was the Persian name for horse, and in the Scythian names Aspabota, Aspakara, and Asparatha,[240] we can hardly fail to recognize the same element. Even the name of the Aspasian mountains, placed by Ptolemy in Scythia, indicates a similar origin. Nor is the word Arya unknown beyond the Oxus. There is a people called Ariacœ,[241] another called Antariani.[242] A [pg 244] king of the Scythians, at the time of Darius, was called Ariantes. A cotemporary of Xerxes is known by the name of Aripithes (i.e. Sanskrit, aryapati; Zend, airyapaiti); and Spargapithes seems to have some connection with the Sanskrit svargapati, lord of heaven.

We have thus traced the name of Ârya from India to the west, from Âryâvarta to Ariana, Persia, Media, more doubtfully to Armenia and Albania, to the Iron in the Caucasus, and to some of the nomad tribes in Transoxiana. As we approach Europe the traces of this name grow fainter, yet they are not altogether lost.

Two roads were open to the Aryans of Asia in their westward migrations. One through Chorasan[243] to the north, through what is now called Russia, and thence to the shores of the Black Sea and Thrace. Another from Armenia, across the Caucasus or across the Black Sea to Northern Greece, and along the Danube to Germany. Now on the former road the Aryans left a trace of their migration in the old name of Thrace which was Aria;[244] on the latter we meet in the eastern part of Germany, near the Vistula, with a German tribe called Arii. And as in Persia we found many proper names in which Arya formed an important ingredient, so we find again in German history names such as Ariovistus.[245]

Though we look in vain for any traces of this old national name among the Greeks and Romans, late researches have rendered it at least plausible that it has [pg 245] been preserved in the extreme west of the Aryan migrations, in the very name of Ireland. The common etymology of Erin is that it means “island of the west,” iar-innis, or land of the west, iar-in. But this is clearly wrong.[246] The old name is Ériu in the nominative, more recently Éire. It is only in the oblique cases that the final n appears, as in regio, regionis. Erin therefore has been explained as a derivative of Er or Eri, said to be the ancient name of the Irish Celts as preserved in the Anglo-Saxon name of their country, Íraland.[247] It is maintained by O'Reilly, though denied by others, that er is used in Irish in the sense of noble, like the Sanskrit ârya.[248]

Some of the evidence here collected in tracing the ancient name of the Aryan family, may seem doubtful, and I have pointed out myself some links of the chain uniting the earliest name of India with the modern name of Ireland, as weaker than the rest. But the principal links are safe. Names of countries, peoples, rivers, and mountains, have an extraordinary vitality, and they will remain while cities, kingdoms, and nations pass away. Rome has the same name to-day, and will probably have it forever, which was given to it by the earliest Latin and Sabine settlers, and wherever we find the name of Rome, whether in Wallachia, which by the inhabitants is called Rumania, or in the dialects of the Grisons, the Romansch, or in the title of the Romance languages, we know that some threads would lead us back to the Rome of Romulus and Remus, the stronghold of the earliest warriors of Latium. The ruined city near the mouth of the Upper Zab, now [pg 247] usually known by the name of Nimrud, is called Athur by the Arabic geographers, and in Athur we recognize the old name of Assyria, which Dio Cassius writes Atyria, remarking that the barbarians changed the Sigma into Tau. Assyria is called Athurâ, in the inscriptions of Darius.[249] We hear of battles fought on the Sutledge, and we hardly think that the battle field of the Sikhs was nearly the same where Alexander fought the kings of the Penjáb. But the name of the Sutledge is the name of the same river as the Hesudrus of Alexander, the Śatadru of the Indians, and among the oldest hymns of the Veda, about 1500 b. c., we find a war-song referring to a battle fought on the two banks of the same river.

No doubt there is danger in trusting to mere similarity of names. Grimm may be right that the Arii of Tacitus were originally Harii, and that their name is not connected with Ârya. But the evidence on either side being merely conjectural, this must remain an open question. In most cases, however, a strict observation of the phonetic laws peculiar to each language will remove all uncertainty. Grimm, in his “History of the German Language” (p. 228), imagined that Hariva, the name of Herat in the cuneiform inscriptions, is connected with Arii, the name which, as we saw, Herodotus gives to the Medes. This cannot be, for the initial aspiration in Hariva points to a word which in Sanskrit begins with s, and not with a vowel, like ârya. The following remarks will make this clearer.

Herat is called Herat and Heri,[250] and the river on [pg 248] which it stands is called Heri-rud. This river Heri is called by Ptolemy Ἀρείας,[251] by other writers Arius; and Aria is the name given to the country between Parthia (Parthuwa) in the west, Margiana (Marghush) in the north, Bactria (Bakhtrish) and Arachosia (Harauwatish) in the east, and Drangiana (Zaraka) in the south. This, however, though without the initial h, is not Ariana, as described by Strabo, but an independent country, forming part of it. It is supposed to be the same as the Haraiva (Hariva) of the cuneiform inscriptions, though this is doubtful. But it is mentioned in the Zend-avesta, under the name of Harôyu,[252] as the sixth country created by Ormuzd. We can trace this name with the initial h even beyond the time of Zoroaster. The Zoroastrians were a colony from northern India. They had been together for a time with the people whose sacred songs have been preserved to us in the Veda. A schism took place, and the Zoroastrians migrated westward to Arachosia and Persia. In their migrations they did what the Greeks did when they founded new colonies, what the Americans did in founding new cities. They gave to the new cities and to the rivers along which they settled, the names of cities and rivers familiar to [pg 249] them, and reminding them of the localities which they had left. Now, as a Persian h points to a Sanskrit s, Harôyu would be in Sanskrit Saroyu. One of the sacred rivers of India, a river mentioned in the Veda, and famous in the epic poems as the river of Ayodhyâ, one of the earliest capitals of India, the modern Oude, has the name of Sarayu, the modern Sardju.[253]

As Comparative Philology has thus traced the ancient name of Ârya from India to Europe, as the original title assumed by the Aryans before they left their common home, it is but natural that it should have been chosen as the technical term for the family of languages which was formerly designated as Indo-Germanic, Indo-European, Caucasian, or Japhetic.