The Lost Atlantis and Other Ethnographic Studies. Sir Daniel Wilson
intercourse between the Mediterranean and the western continent, have naturally turned to the Eskuara of the Basques. It is invariably recognised as the surviving representative of languages spoken by the Allophyliæ of Europe before the intrusion of Aryans. The forms of its grammar differ widely from those of any Semitic, or Indo-European tongue, placing it in the same class with Mongol, East African, and American languages. Here, therefore, is a tempting glimpse of possible affinities; and Professor Whitney, accordingly, remarks in his Life and Growth of Languages, that the Basque “forms a suitable stepping-stone from which to enter the peculiar linguistic domain of the New World, since there is no other dialect of the Old World which so much resembles in structure the American languages.” But this glimpse of possible relationship has proved, thus far, illusory. In their morphological character, certain American and Asiatic languages have a common agglutinative structure, which in the former is developed into their characteristic polysynthetic attribute. With this, the Eskuarian system of affixes corresponds. But beyond the general structure, there is no such evidence of affinity, either in the vocabularies or grammar, as direct affiliation might be expected to show. Elements common to the Anglo-American of the nineteenth century and the Sanskrit-speaking race beyond the Indus, in the era of Alexander of Macedon, are suggested at once by the grammatical structure of their languages; whereas there is nothing in the resemblance between the Basque and any of the North American languages that is not compatible with a “stepping-stone” from Asia to America by the islands of the Pacific. The most important of all the native American languages in their bearing on this interesting inquiry—those of Central America—are only now receiving adequate attention. Startling evidence may yet reward the diligence of students; but, so far as language furnishes any clue to affinity of race, no American language thus far discloses such a relationship, as, for example, enabled Dr. Pritchard to suggest that the western people of Europe, to whom the Greeks gave the collective name of Kέλται, and whose languages had been assumed by all previous ethnologists as furnishing evidence that they were precursors of the Aryan immigrants, in reality justified their classification in the same stock.
But while thus far, the evidence of language is, at best, vague and indefinite in its response to the inquiry for proofs of relationship of the races of America to those of the Old World; physiological comparisons lend no confirmation to the idea of an indigenous native race, with special affinities and adaptation to its peculiar environment, and with languages all of one class, the ramifications from a single native stem. So far as physical affinities can be relied upon, the man of America, in all his most characteristic racial diversities, is of Asiatic origin. His near approximation to the Asiatic Mongol is so manifest as to have led observers of widely different opinions in all other respects, to concur in classing both under the same great division: the Mongolian of Pickering, the American Mongolidæ of Latham, the Mongoloid of Huxley. Professor Flower, in an able discussion of the varieties of the human species, addressed to the Anthropological Institute of Great Britain in 1885, unhesitatingly classes the Eskimo as the typical North Asiatic Mongol. In other American races he notes as distinctive features the characteristic form of the nasal bones, the well-developed superciliary ridge, and retreating forehead; but the resemblance is so obvious in many other respects, that he finally includes them all among the members of the Mongolian type. If, then, the American Mongol came originally from Asia, or sprung from the common stock of which the Asiatic Mongol is the typical representative, within any such period as even earliest Phœnician history would embrace, much more definite traces of affinity are to be looked for in his language than mere correspondence in the agglutination characteristic of a very widely diffused class of speech. But we, thus far, look in vain for traces of a common genealogy such as those which, on the one hand, correlate the Semitic and Aryan families of Asia and Europe with parent stocks of times anterior to history, and on the other, with ramifications of modern centuries. We have, moreover, to deal mainly with the languages of uncivilised races. To the continent north of the Gulf of Mexico, the grand civilising art of the metallurgist remained to the last unknown; and in Mexico, it appears as a gift of recent origin, derived from Central America. The Asiatic origin of the art of Tubal-cain has, indeed, been pretty generally assumed, both for Central and Southern America; but by mere inference. In doing so, we are carried back to some mythic Quetzalcoatl: for neither the metallurgist nor his art was introduced in recent centuries. Assuming, for the sake of argument, the dispersion of a common population of Asia and America, already familiar with the working of metals, and with architecture, sculpture and other kindred arts, at a date coeval with the founding of Tyre, “the daughter of Sidon,” what help does language give us in favour of such a postulate? We have great language groups, such as the Huron-Iroquois, extending of old from the St. Lawrence to North Carolina; the Algonkin, from Hudson Bay to South Carolina; the Dakotan, from the Mississippi to the Rocky Mountains; the Athabascan, from the Eskimo frontier, within the Arctic circle, to New Mexico; and the Tinné family of languages west of the Rocky Mountains, from the Youkon and Mackenzie rivers, far south on the Pacific slope. With those, as with the more cultured languages, or rather languages of the more cultured races, of Central and Southern America, elaborate comparisons have been made with vocabularies of Asiatic languages; but the results are, at best, vague. Curious points of agreement have, indeed, been demonstrated, inviting to further research; but as yet the evidence of relationship mainly rests on correspondence in structure. The agglutinative suffixes are common to the Eskimo and many American Indian tongues. Dr. H. Rink describes the polysynthetic process in the Eskimo language as founded on radical words, to which additional or imperfect words, or affixes, are attached; and on the inflexion, which, for transitive verbs, indicates subject as well as object, likewise by addition. But, while Professor Flower unhesitatingly characterises the Eskimo as belonging to the typical North Asiatic Mongols; he, at the same time, speaks of them as almost as perfectly isolated in their Arctic home “as an island population.” Nevertheless, the same structure is common to their language and to those of the great North American families already named. All alike present, in an exaggerated form, the characteristic structure of the Ural-Altaic or Turanian group of Asiatic languages.
Race-type corresponds in the Old and New World. A comparison of languages by means of the vocabularies of the two continents, yields no such correspondence. All the more, therefore, is the American student of comparative philology stimulated to investigate the significance of the polysynthetic characteristic found to pertain to so many—though by no means to all—of the languages of this continent. The relationship which it suggests to the agglutinative languages of Asia, furnishes a subject of investigation not less interesting to American students, alike of the science of language, and of the whole comprehensive questions which anthropology embraces, than the relations of the Romance languages of Europe to the parent Latin; or of Latin itself, and all the Aryan languages, ancient and modern, not only to Sanskrit and Zend, but to the indeterminate stock which furnished the parent roots, the grammatical forms, and that whole class of words still recognisable as the common property of the whole Aryan family. Sanskrit was a dead language three thousand years ago; the English language, as such, cannot claim to have endured much more than fourteen centuries, yet both partake of the same common property of numerals and familiar terms existing under certain modifications in Sanskrit, Greek, Latin, Slavonic, Celtic, German, Anglo-Saxon, and in all the Romance languages. Thus far the American philologist has been unable to show any such genealogical relationship pervading the native languages; or to recover specific evidence of affinities to languages, and so to races of other continents. There are, indeed, linguistic families, such as some already referred to, indicating a common descent among widely dispersed tribes; but this has its chief interest in relation to another aspect of the question.
Professor Max Müller has drawn attention to the tendency of the languages of America towards an endless multiplication of distinct dialects. Those again have been grouped by the synthetic process of Hervas into eleven families: seven for the northern continent, and four for South America. But we are as yet only on the threshold of this important branch of research. In two papers contributed by M. Lucien Adam to the Congrès International des Americanistes, he gives the results of a careful examination of sixteen languages of North and South America; and arrives at the conclusion that they belong to a number of independent families as essentially distinct as they would have been “had there been primitively several human pairs.” Dr. Brinton, one of the highest authorities on any question connected with native American languages, contributed a paper to the American Antiquarian (Jan.