may be said to consist in flexibility and elasticity. And if I were to designate in the same way the principle of most of the leading non-Sanscritic features in the etymological department, I should call it analytical distinctness: flexibility, elasticity, analytical distinctness-and are not these the qualities which most nearly represent the character of the whole Celtic nation?-But the idea I have touched upon in the phrase analytical distinctness requires some farther explanation, for which I must solicit the kind attention of my hearers. When we compare our modern European languages, the English and French for instance, with the ancient, especially the Latin and Greek, we are struck by one marked difference in their grammatical characters, namely, the different manner in which they express relative or incidental notions or ideas. By the term relative or incidental we designate and distinguish from the other great class of notions, which we call substantive, all those notions or ideas which, at the same time that they exclusively represent phanomena of a certain general and categorical meaning, moreover represent each of them, not with reference to itself, but only to two or several other phænomena which of course always belong to the class of substantive notions. For instance, in the sentence, the horse is struck by a spear,-equus tangitur telo-the three substantive notions of which, as of its substantial elements, the proposition is composed, are expressed by the words horse, struck, spear, whereas the four particles the, is, by, a, express the relative or incidental notions of the sentence, which evidently does not receive from them the addition of any new independent element, but merely the connexion and determination of the three above-mentioned. And the equivalent Latin sentence which I have mentioned will at once have directed the attention of my hearers to the nature of the difference which we have stated to exist between ancient and modern languages, in expressing relative or incidental notions. The notions in the above sentence belonging to that class are in English rendered by four separate and auxiliary words placed beside the principal, whereas in Latin they are rendered through the inflexion, as it is generally called, of the latter. But what is inflexion? It is a system of etymological combinations, by which any one of those elementary parts of imitative articulation which (by a metaphorical term referring to the analogy existing between the development of plants and words) are usually called roots, and more especially any one of those roots which express substantive ideas, and which for this reason we may call substantive roots, becomes, in connected speech, regularly allied with one or several of another class of roots which differ from the former, both in form and meaning, the one being generally slighter than that of substantive roots, and consisting not, as most of those, in a double, but in a simple articulation, the other (the meaning) being always that of an incidental or relative idea. The place occupied by the incidental root may be either before or after the substantive root: in the former case it is called prefix, and in the latter, which is by far the more general, suffix. And having thus defined the term inflexion-which in a more appropriate sense refers particularly to the mode of interchange which takes place between several incidental roots as becoming alternately attached to one substantive root-we may say that the great difference alluded to, between ancient and modern languages, consists in the former expressing incidental notions by auxiliary words, and the latter by auxiliary roots: for instance, in the example above given, the notions expressed in English by the words the, is, by, a, are expressed in Latin respectively by the three suffixes (one of them double) us, it-ur, o. The comparative advantages and disadvantages of these two methods may be easily understood. The one, uniting the incidental with the substantive notion under the same emission and intonation of voice, and blending both, for the mind as well as the ear and the eye, into one organized whole, composed, as it were, both of an etymological and a phonic arsis and thesis, is more fit to exercise the synthetic and artistic capabilities of the human intellect, of which moreover it favours the development, by perfecting what may be called the objective beauty of language, inasmuch as, through the varying union of a series of suffixes with one unchanging root, it endues the process of inflexion with the appearance of vital activity. The other method, which gives distinct breath and accent to each incidental notion, and so both to the corporeal and intellectual eye is constantly renewing that difficult process of the understanding, through which the primitive root, which always involved a full sentence, has decomposed itself into its logical elements, is better calculated for the exercise of the analytical and discriminating powers of the intellect, and as it prevents the meaning even of the slightest imitative sound from being obscured, serves to quicken the consciousness of each minute member of the sentence, and thus to augment the subjective force of the language. And now which of these two methods is the more ancient? In the Teutonic languages it is certain that the analytical tendency which now predominates in their etymological department is not the primitive one, inasmuch as it is not found in their most ancient dialect, the Gothic, which has nearly all the synthetic habits of the Sanscrit and the Latin: and hence, in every language in which the analytical method of declension and conjugation has been observed, it has been suspected by modern philology to be the effect of decomposition. But the case is different with the Celtic, which by its entire structure, as well as by its history, lays claim to a much higher antiquity than the Teutonic, and reaches back to an epoch in the history of human speech anterior, as we may infer from philosophical considerations, to that of the synthetic principle represented by the Sanscrit, and during which the analytical principle must have prevailed. This conclusion is fully borne out, and confirmed as a fact, by one of the greatest discoveries of modern philology, that of the Old Egyptian. This language, at the same time that it shows in a considerable portion of its grammatical features-especially the formation of roots, the choice and specification of their meaning, and the system of conjugation—a decided primitive affinity to the Sanscrit, in another manifests an almost total absence of the observances of etymological synthesis, so systematically carried out by the younger language, thus proving that the decomposition which has taken place in the Teutonic languages, with reference to the Sanscrit, is, with reference to the more ancient mothertongue, only a kind of return to their original state. And I have no doubt that this return has been effected not more by their instinctive tendency to recover the lost perception of the meaning of most of the incidental roots than by the influence of the Celtic, which in all its non-Sanscritic features most strikingly corresponds with the Old Egyptian. This correspondence refers first, to a considerable number of specified roots and words, which, as far as I am aware, belong exclusively to those two languages; e. g. Eg. rā, sun. aah, moon. val, eye. Ir. lã, day. eagh, moon. W. syw, bright, clean. sew-yd, stars. syw-ed, astronomy. gwel-ed, to see. mas, to suckle, young, child. Ir. meas, child; W. Ir. meas, child; W. moes, suckling, nursing, education (coll. Lat. mos). Secondly, to several incidental roots of great import in the etymological department: e. g. 3 pers. masc. Eg. ef, o; W. ev, o. 3 pers. suffix. Eg. f (ai-f, he goes); W. f (ai-ff). 2 pers. masc. singular and plural Eg. k (ai-k, thou goest); W. ch (ae-ch, you were going); el-och, thou didst go. Indefinite auxiliary verb. Eg. ar, au (ar ai-f); W. yr, a (yr ai-ff). Thirdly, to the system of combining, in the form of suffixes, the personal pronouns with the prepositions, a usage similar to that which prevails in the Hebrew, where personal pronouns are suffixed to substantive nouns, but which is more remarkable in a linguistical point of view, inasmuch as it implies the consciousness of the primitive meaning of prepositions, which was always that of substantive nouns; e. g. Eg. (a)r-of, ar-o, toward him. W. ar-n-o (the n is genitive pre position), upon him. Fourthly and principally, it refers to the expressing incidental notions by roots, in the character of separate and independent words, which are used in Sanscrit to express the same notions, but as suffixes and prefixes, and in a much more limited signification. Thus, in the conjugation of the verb, the three persons, which the Sanscrit regularly expresses by the personal pronouns combined, under the form of suffixes, with the verbal root, are expressed in Celtic sometimes in the same way, but in other cases by the same pronominal roots under the form of separate auxiliary words, which may be placed indifferently either before or after the verbal root: a flexibility of expression to which the Egyptian supplies a parallel, the use of the pronoun, as suffix, belonging to the sacred, and as prefix to the demotic (popular) dialect of this language; e. g. Eg. sacred dialect, ai-f, ai-k, ai-a (it, is, eo). demotic dialect, ef-ai, ek-ai, ei-ai. W. can-a-vi (Godod. 612) or can-a-v (canam). cen-i-t(i) (canes). can-o (cecinerit). can-er vi, ti, evo (canor, caneris, canitur). canu yr wyv (canere sum) alternating with can-wyv (cano). I. can-aim (cano); can-t-ar me (canor). Thus the Welsh indefinite auxiliary verb a, to go, to be, which, even as the corresponding Egyptian au (Coptic o), is placed before substantive verbs (verba concreta) to mark the indefinite mood, appears in Sanscrit and Greek as the well-known augment: e. g. Welsh, a ddysg-odd, he did teach; Sansc. a-diks,-a-ta; é-dídače*. *The author of this paper was the first, as far as he knows, to indicate this origin of the augment, in an article on two ancient Italian inscriptions (inserted in the Münchner Gelehrte Anzeigen, April, 1843) and afterwards in the Wiener Jahrbücher. And thus also, to mention an instance which I have already alluded to, the word n, which, alternating with m (and undoubtedly identical with the word m, ma, place), serves in Egyptian as a preposition to denote all cases, though particularly the genitive, serves in Celtic (where it is generally contained, as we have seen, in the transmutation of the initial) to denote exclusively the genitive, and more particularly the genitive plural, to which in Sanscrit, Greek and Latin, it has been regularly limited: e. g. Egypt. nef n anach, breath of life; sont n ataf, avenger of his father; fitw m ah-aya, four (of) oxen.- Welsh, Caer-n-arvon, town of Arvon; ar-n-av (vide supra), saith niwrnod (instead of "saith n diwrnod," septem dierum).-Irish, iar n-dilinn, after the deluge (originally, in the back, west, of the deluge), na-ndia (pronounced na-n-ia), of the days.-Sanse. diu-n-am.-Lat. die-r-um (coll. Old High Ger. kep-on-o; A.-Sax. giv-en-a, of the gifts). Now, weighing all these affinities of the Celtic with the Egyptian on the one side, and the Sanscrit on the other, I believe we may be justified in saying that it occupies a place in history between both, and marks an intermediate stage in the development of human language, and more especially of the Japhetic, between the analytical fluidity of its genial infancy and that beautiful synthetic consistence, so to speak, of its vigorous maturity, as we find it represented in the Sanscrit. The intermediate position which we have assigned to the Celtic, with respect to the different epochs of the Japhetic languages, it still holds, as regards the relation of this family with the Semitic and Finish, both of which participate in many of its non-Sanscritic features. It appears to be also by this internal relationship, much more than by external contact, that we must explain the resemblance of many Celtic elements with those of two languages, both of which seem to belong to a Celto-Finish branch, I mean the Basque and Etruscan*. And if at present, once more passing our eye successively over the fields of ethnology, history of language and philosophy of grammar, we take one full view of all the light which falls on them from the study of the Celtic, we may perhaps find it excusable, that the Celtomanians were so much dazzled by it as to fancy that in that language they had discovered the mother-tongue of mankind as well as the key of all worldly and divine science; though all that they have achieved in carrying out their fancies has certainly been only to obscure, and for a long time to discredit, that very study to which they attached such supreme importance. As a fourth argument in favour of that study, and perhaps the most weighty of all, I might add, the access which it opens to the study of modern Celtic literature, a literature as interesting in an artistical as an historical point of view, and which, by a long series of poetical and historic works, some as ancient perhaps as the fifth century of our era, and all abounding in the genuine features of native art and inspiration, exhibits the Celtic origin of two of the principal elements of modern European literature, rhyme and allegorical romance. But the very copiousness of this subject forbids me to enter upon it on this occasion. I shall conclude this discourse with the fervent hope, in which I sincerely trust my hearers participate, that the two dialects of this highly important and beautiful language, still extant in the island of Pryd, may never cease to subsist, but may be maintained, both by their own vital energy and by the enlightened care of the government, as constituting one of the most precious, as well as most ancient gems in the Imperial Crown. * As regards the advantage which may be derived from the Celtic for the elucidation of the Etruscan, vide the article last quoted in the Münchner Gelehrte Anzeigen. On the Relation of the Bengali to the Arian and Aboriginal Languages of India. By Dr. MAX MÜLLER. THE interest which the Bengali language presents to oriental scholars, and which induces them to devote their time to the study of this Indian dialect, may be viewed under three different beads, as practical, literary, and linguistical. On the first point, it is hardly necessary to enter into any details. The English people, who have been called to rule the destinies of more than a hundred millions of souls in the East, one-tenth of whom make use of the Bengali as their vernacular dialect, have well understood the duties of those who have been appointed to govern this great oriental empire. Great exertions have been made to give sufficient training to those who are destined to execute the various duties connected with the internal government of India ; and it has not been thought enough that they should receive such an education as would entitle them to employments in their own country, but it has been felt that it was peculiarly incumbent upon them to study the languages of the people over whom they were to be placed, not as the sons of a foreign and conquering nation, to raise taxes, to punish disobedience, and to suppress every trace of national feeling, but as men devoted to the higher object of inspiring confidence, of winning affection, and of promoting for the benefit of the native population the benign influence of European civilization. With this view of the mission which the English people have been desirous of fulfilling in India, it could not be considered enough for an officer to understand just so much or so little of Persian and Hindustani, as to decipher representations and complaints, or to convey official decrees to a subject people. For though these two languages may have some claim to be regarded as the official languages of India, particularly among the higher classes of the natives, yet they are, like the French in Europe, unknown to the great mass of the population, and of little use therefore for the ordinary purposes of daily life. Although, then, a prejudice may have prevailed for some time against the study of the vernacular dialects spoken in the large and densely peopled districts of India, it was soon acknowledged, that for local communication and for an immediate and effective intercourse with the people, a knowledge of provincial languages like the Bengali, Marathi, Telugu, Tamil, Carnatika and Cingalese, was of no less importance and necessity than that of the more fashionable Persian and Hindustani. But, as Professor Wilson, the distinguished president of our Section, whose name is as much cherished by the natives of India as it is esteemed by the learned men of Europe, well remarks, it is not enough to understand the language of a people; the people themselves must be understood with all their popular prejudices, their daily observances, their occupations, their amusements, their domestic and social relations, their local legends, their national traditions, their mythological fables, their metaphysical abstractions, and their religious worship. The best means of acquiring such a knowledge is generally to be found in the literature of the people. It is however necessary to confess that upon this point, namely, the literary interest of the language, the Bengali is poor, and inferior in this respect to most of the other vernacular languages. There existed, indeed, scarcely anything worthy to be called literature in Bengali before the settlement of the missionaries in Bengal, and it is due to their unwearied exertions that the Bengali has become in any sense a literary language, and has arrived at a certain degree of grammatical regularity. Nor need we be surprised at this, when we |