CHAPTER XXV.

ETHNOGRAPHIC HYPOTHESES: MIGRATIONS.

THE ethnology of the New World is unquestionably simpler than that of Europe or Asia, in its freedom from complicated elements which retard our study of the latter alike in their ancient and modern aspects. Nevertheless, this may be more apparent than real. Our knowledge of history prevents our under-estimating Pelasgian or Etruscan, Basque, Magyar, or Celtic elements of diversity. Ignorance may be the cause of our overlooking or under-estimating diversities among American languages as great as the German and Euskara, or the Sanscrit and the Chinese. America, indeed, appears to have its monosyllabic Otomi and Mazahui, with their analogies to the Chinese, and their seemingly radical contrast to that polysynthetic structure which appears to be as predominant throughout the New World as Indo-European affinities are characteristic of the languages of Europe. But we scarcely know yet how justly to estimate the amount of difference; for Mr. Schoolcraft affirms, as a conclusion to which his intimate familiarity with the Algonquin dialects had led him, that they betray evidence of having been built up from monosyllabic roots. If this be indeed demonstrable in any other than the vague sense in which it may be stated of every tongue, the same conclusion will apply to other American languages. Nearly all the Chippewa root

words, he observes, are of one or two syllables; and Gallatin has shown that the same may be affirmed to a great extent of the Mexican, if the pronominal adjuncts and the constantly recurring terminations are detached from the radix. But the polysyllabic characteristics of the Algonquin exceed even those of the Esquimaux. Holophrasms are common in all its dialects, compounded of a number of articulations, each of which is one of the syllables of a distinct word; and the whole undergoes grammatical changes as a verbal unit. This, therefore, is a condition widely diverse from that of the monosyllabic languages, even where, as in the Otomi, many compounded words occur in the vocabulary. But after making every allowance for unknown nations and tongues, and misinterpreted or unappreciated elements of difference among the varieties of man in the New World, the range of variation appears to extend over a smaller scale than that of Europe or Asia, or even of Africa; while he is everywhere found there under much less diversified modifications of civilized or savage life than on the old historic continents. The original centres of population may have been manifold; for the evidence of the lengthened period of man's presence in America furnishes abundant time for such operations of climatic influences, direct or indirect intercourse, or even positive intermixture, to break down strongly-marked elements of ethnic diversity. Nevertheless, after carefully weighing the various kinds of evidence which have been glanced at in previous chapters, they all seem to resolve themselves into three great centres of propagation, of which the oldest and most influential belongs to the southern and not to the northern continent. The routes originally pursued in such immigrations may have been various, and it is far from impossible that both southern and northern immigrants entered the con

tinent by the same access. Such, however, is not the conclusion to which the previous investigations appear to me to point. If we adopt the most favoured theory, that the New World has been entirely peopled from Asia, through Behring Straits, then the Patagonian should be among the oldest, and the Esquimaux the most recent of its immigrant occupants. But that which seems theoretically the easiest is by no means necessarily the most probable course of migration; and many slight indications combine to suggest the hypothesis of a peopling of South America from Asia, through the islands of the Pacific.

The tendency of philological inquiry, as directed to the peculiar grammatical structure and extreme glossarial diversities of the American languages, was at first to isolate them entirely, and to exaggerate their special phenomena into widely prevalent linguistic features, common to the New World and utterly unknown elsewhere. In this the philologist only pursued the same course as the physiologist, the attention of each being naturally attracted chiefly by what was dissimilar to all that had been observed elsewhere. But as physiological investigations have extended, their disclosures prove less conclusive in the support they yield to the favourite theory of an essential isolation and ethnic diversity for the American man. Increasing knowledge of his languages tends rather to diminish the proofs of that radical difference from all other forms of human speech which was at first too hastily assumed. The synthetic element of structure, though very remarkable in the extent of its development, has many analogies in ancient languages, and is embraced in the grammatical process of all inflectional tongues. But beyond this, important elements of relationship appear to be traceable between languages of America and those of the Polynesian family.

Gallatin early drew attention to certain analogies in the structure of Polynesian and American languages as deserving of further investigation; and pointed out the peculiar mode of expressing the tense, mood, and voice of the verb, by affixed particles, and the value given to place over time, as indicated in the predominant locative verbal form. The peculiar substitution of affixed particles for inflections, especially in expressing the direction of the action in relation to the speaker, is common to the Polynesian and the Oregon languages, and also has analogies in the Cherokee.1 Subsequent observations, though very partially prosecuted, have tended to confirm this idea, especially in relation to the languages of South America, as shown in their mode of expressing the tense of the verb; in the formation of causative, reciprocal, potential, and locative verbs by affixes; and the general system of compounded word structure. The incorporation of the particle with the verbal root appears to embody the germ of the more comprehensive American holophrasms. But here again, while seeming to recover links between Polynesia and South America, we come on the track of affinities no less clearly Asiatic. Striking analogies have been recognised between the languages of the Deccan and those of the Polynesian group, in which the determinate significance of the formative particles on the verbal root equally admits of comparison with peculiarities of the American languages. On this subject the Rev. Richard Garnett remarks that most of the languages of the American continent respecting which definite information has been acquired, bear a general analogy alike to the Polynesian family and the languages of the Deccan, in their methods of distinguishing the various modifications of time; and he adds: "We may venture to assert in general terms that a South American 1 American Ethnological Transactions, vol. ii. p. cliv.

"1

verb is constructed precisely on the same principle as those in the Tamul and other languages of Southern India; consisting, like them, of a verbal root, a second element defining the time of the action, and a third denoting the subject or person." Such indications of philological relation of the islands of the Polynesian archipelago and the American continent to Southern Asia, acquire an additional interest when taken in connexion with remarkable traces of megalithic sculpture and of ancient stone structures in the Pacific, long ago noted by Captain Beechey on some of the islands nearest to the coasts of Chili and Peru, and more recently observed on Bonabe and other islands lying off the Asiatic shores. Some of those have already been referred to in their general bearings on oceanic migration, and on the probability of an era of insular civilisation, during which maritime enterprise may have been carried out on a scale unknown to the most adventurous of modern Malay navigators.

The affinities recognisable between Polynesian and American arts manifestly belong to a remote past; and the character of such philological relations as have been indicated fully accord with this. The direct relationship of existing Polynesian languages is not Mongol but Malay; and this is for the most part so well defined as to indicate migrations from the Asiatic continent to the islands of the Pacific at periods comparatively recent; whereas the diversity of those of America, and their essentially native vocabularies, prove that the latter have been in process of development from a remote period free from all contact with tongues which, as we see, were still modelling themselves according to the same plan of thought in the clustering islands of the Pacific. But the American languages present a widely

1 Proceedings of the Philological Society, vol. i. p. 271.