that they had once a common historical centre. Thus, even if we had no knowledge of the former existence of Latin as a political language, the resemblance of the Romanic dialects would force us to admit a political concentration of language previous to the time when this fixed and settled speech became broken up into various dialects. This resemblance between secondary dialects is a different one from that which may be observed between primary dialects, such as precede the formation of every political language. These primary dialects are earlier than the кon, just as mountain streams are earlier than rivers; the secondary dialects, on the contrary, are later, just as channels are later than lakes. Among savage tribes, where these primary dialects have never been called together into a literary system, we find, as in America, Africa, and CochinChina, that there exists so perplexing a variety of idioms, that the inhabitants of neighbouring villages are unintelligible to one another; and, in the absence of all checks on the caprices and peculiarities of individuals, old forms are changed and new forms introduced by every individual with such recklessness as to obscure for ever the traces of a primitive community of speech. In the history of the Arian family we can distinguish between several lingual centralisations. After one dialect has attracted or absorbed the floating elements of other popular dialects, and been raised to the dignity of a classical language, we see it again diverge into new branches. Latin first absorbs all the idioms of Italy, and after it has become the language of the then civilised world, it is broken up in turn into many dialects. If this political centralisation of Italy had not intervened, and if no Roman empire had brought the provinces of Italy under one common sway, the dialects of the Umbrians in the north of Italy would have developed themselves and become so different from that of the Sabines in Lower Italy as to appear to us a totally different language, differing from the Oscan at least as much as Greek from Slavonic. But these two dialects, the Oscan and Umbrian, were themselves political and literary languages, not to be compared with the unsettled idioms of savage tribes such as we find in America. If, then, we imagine a state of things where the different provinces, nay the towns and villages in the separate valleys of Italy, had each retained. its lingual independence, each continuing to use its local dialect for centuries, without any political intercourse, or common literature, political, religious, or legal, we should then find, as we do in South America, almost as many distinct languages as there are settlements. Kircher fixes the number of languages known to be spoken in South America toward the end of the seventeenth century, at five hundred; and in most cases the people who speak such idioms are said to be unintelligible to their nearest neighbours. Here, therefore, in America, we should say, that the immense diversity of dialects shows the absence of a previous political centralisation. Now, on exactly the same ground, it follows that in the Dekhan the great similarity between the different Tamulic dialects can only be the result of a former period in the history of the Tamulian speech, during which its character became fixed, grammatically and etymologically. Such a process we can only ascribe to the influence of a more comprehensive civilisation, and a more extended political and literary intercourse than is generally ascribed to the aboriginal inhabitants of India. The Tamulic dialects agree not only in roots, not only in pronouns and numerals, but in derivative words which must have been known to all before they began to diverge and grow into new dialects. Perhaps it will be possible to fix on one of these dialects as the eldest of the Tamulic sisters, and derive from it some of these words which are common to all. But even then our conclusion would be the same; for the adoption of words from one dialect into another necessitates equally the admission of a political and literary intercourse, which can only take place during a period of advancing civilisation. Another reason for supposing the Tamulic languages considerably advanced in their literary capabilities before their struggle with the Sanskrit began, may be discovered in their successful resistance against the introduction of Sanskrit elements into their grammar Although the dictionary of the Tamulic languages is as full of Sanskrit words as English is of Norman, yet the Tamulians did not give up their grammatical independence. And even the words which were adopted from Sanskrit had to submit to the genius of these dialects. With the exception of those adopted ready made, and simply transferred from Sanskrit, as Latin expressions are in English, the majority of Sanskrit terms in the Tamulic dialects has been changed to such an extent, that it is difficult sometimes to discover their foreign origin. Words simply taken from Sanskrit are, according to Ellis, called "Tatsama," i. e. equal to Sanskrit. For instance, sampadu in Telugu is the Sanskrit sampad (fortune). Appa in Telugu is the Sanskrit âpah (water). Payasu in Telugu is the Sanskrit payas (milk). Words adopted from Sanskrit with considerable phonetic changes are called "Tadbhava," i. e. produced from Sanskrit. Some of the changes which these words have undergone must be ascribed to the spoken or vulgar Sanskrit, for they depend on the same rules by which Sanskrit words are modified in the Prakrit dialects. These, the spoken or vulgar dialects of the Sanskrit, would be the most natural channels through which Sanskrit words could have reached the Tamulians. And as in French we find frequently the same Latin word under two different forms, of which the one (as for instance "rédemption") might be called a tatsama, the other (as for instance rançon") tadbhava, instances occur in the Tamulic languages where the same Sanskrit word has been adopted under two different forms. Parva in Sanskrit means a knot or joint, and, with particular reference to the moon, it means the day of the full and new moon. As these were festival days, pabba and habba in Canarese mean a "festival." But in the learned language of the Brahmans parvan came also to signify a chapter or book, and in this sense it is used in Canarese parva, section of a book. Instances where Tadbhava words in Telugu seem to have passed through Prakrit channels are the following: ፡፡ Sanskrit. brahma bamha (Vararuki, v. 47.) bomma (Brahma). bambhadu (Abhira) bapadu (a Brahman). * Cf. Weigle, Journal of the German Oriental Society, II. 265. The Prâkrit forms are given on the authority of Ellis, in his introduction to Campbell's Teloogoo grammar. Ellis must have availed himself, however, of other sources besides Vararuki. Where his forms agree with Vararuki I have added a reference to the excellent edition of this grammarian by my friend Mr. Cowell, at Oxford. Where they differ, or where they do not occur at all in Vararuki, Ellis may have followed Hemakandra, or other authorities, as he was too accurate a scholar to have formed them merely on general analogy. It will appear, even from this short list, that some phonetic changes, generally ascribed to the influence of the Telugu, can be traced back to Prakritic corruptions, but that, at the same time, the Telugu went beyond the limits of the Prakrit. Sanskrit words form so large a portion of the Tamulic dictionaries, that they are no longer considered as foreign words. Foreign words, according to Telugu grammarians, are called "Anyadesîya," i. e. of another country; and the Appakavîyam explains their origin in the following manner:-"The natives of Andhra (i. e. Telugus) having resided in various countries using Telugu terms conjointly with those of other countries, these have become Andhra terms of foreign origin." What remains, after subtracting all these extraneous ingredients, is called Desya, i. e. native words. Thus it is said, in a stanza of the Adharvana Vyâkarana with regard to Telugu:- "All the words which are in use among the several races who are aborigines of the country of Andhra' which are perfectly clear and free from all obscurity; these shine forth to the world as the pure native speech of Andhra (suddha. andhra. desyam)." There is only one more distinction made, between what are called native and vulgar words. The latter are termed grâ mya, i. e. belonging to villages, and explained by the Appakavîyam as follows:-" Such Telugu words as are commonly used by rustic folk are known as grâmyam; these lose some of their regular letters, and are not found in poetry, unless, as in abusive language, the use of them cannot be avoided." If we now look at the grammar of the Tamulic languages, we shall find at once that we have before us a system of declension and conjugation much more developed than in the Bhotiya dialects. The forms are more settled according to general grammatical categories; and although the cases, as in all Turanian languages, are formed by postpositions and are, therefore, liable to great variety, yet there exists a formal distinction between the casus rectus and obliquus. This base of the casus obliquus and the terminations of the cases, when brought in contact, are liable to phonetic changes similar to the changes of Sandhi in Sanskrit and other Arian languages, and both coalesce into one grammatical whole. This gives rise, as in some of the more advanced members of the Finnic and Tataric branches, to some real grammatical cases, which become technical, and are used in preference to mere compounds: particularly in the modern and spoken dialects, where the number of independent postpositions expressive of case is much smaller than in the ancient languages. A still greater advance toward grammatical forms is made in the conjugation. Here we find moods and tenses formed by the addition of letters and syllables which by themselves have no more meaning than any termination in Greek or Latin. The persons are expressed by pronominal terminations, and these terminations vary according to the tenses, in the same manner as in Greek and Latin. A grammar like this could only be the grammar of a civilised people. It shows signs of wear and tear, and in what it has retained as well as in what it has given up, we can discern the working of a spirit of wise economy. NINTH SECTION. Comparison of the Tamulic and Ugric Languages. Ir, therefore, we look for analogies to the Tamulic grammar in other branches of the Turanian family we should naturally take those which, like the Tamulic, have reached a certain degree of grammatical perfection. This grammatical perfection, as was stated before, consists first in the production of those formal elements which are wanting altogether in family languages, such as Chinese, and which are extremely scarce as yet in the lower Nomad languages, as in the Tungusic or in some of the Gangetic, the Lohitic, |