Now the most positive evidence which philology is able to afford, in regard to the affinities of two languages, is undoubtedly that which is derived from their conformity both in vocabulary and in grammar. But it fre quently happens that one of these kinds of evidence is deficient ; and the degree of reli ance that can be placed upon the other, taken alone, must depend greatly upon the circum stances of the individual case. Thus, if there be evidence that the vocabulary of one of these languages is in a state of continual change, an entire difference of vocabularies is no obstacle to the idea of the affinity between two languages, when this is decidedly indi cated by a striking conformity in their systems of construction. On the other hand, when two languages or groups of languages differ greatly in their construction, but present a certain degree of verbal correspondence, full weight may be attached to that correspond ence, if it can be proved that it has not been the result of intercourse subsequently to the divergence of the stock, and if it can be shown to be probable that their separation took place at a period when as yet the grammatical development of both languages was in its in fancy. The first appears to be true of the American languages, which seem, as a whole, to be legitimately referable to a common stock, notwithstanding their complete verbal diversity. The second is the aspect under which it appears likely that the Indo-European or Japetic, and the Syro-Arabian or Semitic groups of languages will come to present themselves ; the results of the recent labours of Rawlinson, Layard, Botta and others, on Eastern Archaeology, tending decidedly in this direction.
Philological inquiry, then, must he looked to as the chief means of determining the question of radiation from a single centre or from multiple centres ; and although, in the present state of this department of science it would be unsafe to venture on a positive con clusion, yet the following may be considered as the principal groups under which the various languages hitherto studied may be arranged.
I. The Indo-European, sometimes termed Indo-German, frequently Japetic, and by late writers Arian, or Iranian. This group com prehends nearly all the existing languages of Europe, and those of a portion of South Western Asia.
2. The Syro-Arabian, often termed Semitic; which are spoken by a large part of the popu lation of Syria, Arabia, and Northern and Eastern Africa.
3. The Turanian, or Ugro-Tartarian; which are spoken by the (Mongolian) people of High Asia and of certain parts of Northern Europe.
4. The Seriform, or Indo-Chinese ; which are spoken by the people of South-Eastern Asia.
5. The African ; which are spoken by the people of Central and Southern Africa.
6. The Malayo-Polynesian ; which are spoken by the inhabitants of the numerous islands and island-continents of Oceania.
7. The American ; which are spoken by the inhabitants of the New World, from the Arctic Sea to Cape Horn.
Now it is not a little curious that the linguistic affinity should often be strongest, where the conformity in physical characters is slightest, and weakest when this is strongest. Thus among the Malayo-Polynesian and the American races, as already remarked, there are very striking differences in conformation, features, complexion, &c.; and 3 et the lin guistic affinity of the great mass of tribes forming each group is not now doubted by any philologist, though a doubt may still hang over some particular cases. On the other hand, the hiatus between the Turanian and the Seriform languages is very wide ; but the physical conformity is so strong between the Chinese and the typical Mongolian nations, that no ethnologist has ever thought of as signing to them a distinct origin. So, again, there would seem to be no near relationship between the American and the Turanian lan guages ; but the affinity of the two stocks appears to be established by the transition link afforded by the Esquimaux, which are Mongolian in their conformation and Ame rican in their language. The affinity of the Semitic and Japetic languages, moreover, is so deeply hidden, as to have, until recently, almost defied discovery ; and yet the people who speak them so far resemble one another in physical characters, that they have been almost invariably associated together under the general designation of the Caucasian race. The common origin of the inhabitants of the continents of Europe, Asia, and America is thus pointedly indicated by the combination of these two sources of evidence. The por tion of the Malayo-Polynesian race, that is, in nearest proximity with South-Eastern Asia, presents such a striking; resemblance in phy sical characters to the inhabitants of the neighbouring part of that great continent, that their community of origin can scarcely be doubted ; and when certain points of re semblance between some of the Oceanic and Indian dialects are taken into account, this inference receives strong confirmation. The African nations have long been regarded as the most isolated from the common centre from which all the others appear to have radiated ; but recent investigations have shown that such isolation has no real existence. For, on the one hand, there are tribes which form (like the Esquimaux) a connecting link between the Semitic and proper African fa milies, being African in their conformation, but Semitic in their language ; and, on the other, the study which has been recently be stowed on the proper African languages, especially by Dr. Latham, has shown them to have so much in common with the Semitic tongues, that, with the additional evidence derivable from community of certain usages, extending through vast areas physically iso lated from each other, it now seems impos sible to believe but that the African nations are nearly related to the Semitic, and are through them, derivable from the great Asiatic centre.