civilized nations; and that, as between hunters and hunters there are rarely any objects of exchange, numerals cannot in America have been borrowed by one tribe from another. The pronouns of the first and second person belong also in the Indian languages to the class of primitive words. No definitive opinion can, for want of sufficient materials, be formed with respect to prepositions.

An apprehension of being deceived by false etymologies, or accidental coincidences, has perhaps led into a contrary error. The only case where any language has been placed as belonging to a certain family, without conclusive proof, is that of the Minetares. But there are several, and particularly the Choctaw and Muskhogee, which have been set down as forming distinct families, that will probably be found, on further investigation, to belong to the same. Some of the vocabularies are not sufficiently copious; in many instances, affinities will be discovered through the medium of kindred dialects; and, in order to have a full view of the subject, we should have not only a small collection of primitive words, but dictionaries including derivatives. Thus, for instance, the numeral, one, in Choctaw is, achufa, and, in Muskhogee, humma, between which there is not the slightest affinity. Yet it is revealed by the Choctaw word for once, which is himmunna. It is therefore highly probable, that the number of distinct families of Indian languages will be found to be less than has been here stated; though, at the same time, that of subordinate dialects is undoubtedly greater.

The diversity which does actually exist proves only, that the separation of some of the Indian nations took place in very early times; and the difficulty of accounting for it is not greater here than on the other continent. We find there, in one quarter, the Sanscrit and the Chinese in juxtaposition, and, in another, the Basque surrounded by languages of Latin origin. The same cause, which produced that effect, may, under different circumstances, have given rise to ten, instead of two totally distinct languages. In point of fact, the number does not appear to be greater in North America than in Africa, in the northeastern parts of Asia, or in the Oceanic region. The varieties of languages and of dialects must be more numerous amongst uncivilized tribes, principally those in the hunter state, necessarily subdivided into small communities, than in populous nations united under one government. Public speaking

in their councils is the only standard of language of our Indians. None can become fixed and stable, until that character has been imparted to it by the art of writing and the influence of powerful writers. We have proofs of the multitude, at least, of dialects, which will spring out of an oral language, in those of Germany and of Italy, and in the Patois of France. These are indeed but varieties of the French, with a greater or less residue of Latin or of the other more ancient language of Gaul;* but they still differ (much more in words than in grammatical forms), and are perpetuated, notwithstanding the long-continued influence of a common government and of a common written language. To those obvious causes of a tendency to produce changes, we may add, that inflected languages seem to be more liable to alterations, than those which, like that spoken in China, consist principally of monosyllables.

Although, for a proper study of the character of a language, a dictionary could afford but little aid, if it did not include derivatives and compound words, even our meagre vocabularies, if thoroughly investigated, might offer interesting results. Thus for instance, a single glance at the table of numerals shows, that all the Indian nations have resorted to a decimal numeration. But an examination of several of the languages will afford proof, that they must at first have counted by fives, instead of tens. Thus, in the Choctaw, the numerals seven and eight, untuklo and untuchina, are evidently derived from tuklo two and tuchina three, meaning respectively, five and two, five and three. The same will be found in various other languages, and particularly in those belonging to the AlgonkinLenape. A further investigation will also show, that, although the Knistinaux, Chippeways, Algonkins, and Abenakis use for the unity the word peyac, or paizhik, instead of nequit or ngut, as the other nations of the same stock, they must originally have had also the last word; since their numeral six is, in all of them, derived from it. It is probable, that those two

* In seventy-five French Patois, of which specimens have been lately published, no greater grammatical variations are to be found than the union of the pronoun with the verb, such as Soui for Je suis. In one, in the Ardennes, the pronoun coalesces with the noun, as in our Indian languages; M'per, and S'per, for Mon père and Son père. Mr. Heckewelder's apostrophe has, probably for the same purpose, been used by the French writer.

words were used, as a and one are, in the English language; and Mr. Schoolcraft corroborates that which with me could only be a conjecture.

It will also be found, that, in the Knistinaux and the Chippeway, the initial m is often prefixed to the noun, instead of the pronominal characteristics n, k, w, when such nouns are taken in an absolute or abstract sense, as, miskcewon, nose, miskotick, forehead, meeton, mouth, meepit, teeth, &c.; which seems to corroborate the existence of a definite article mo, discovered by Mr. Du Ponceau in Eliot's translation of the Bible.

Another feature, which may be discovered by the vocabularies, consists in the different names, by which all the Indian nations distinguish the various degrees and modifications of relationship, such as the elder brother, and the elder sister, as distinguished from the younger ones; paternal, or maternal uncle, &c. But what is remarkable, as a feature common to all, is, that women use different words from men for those purposes; and that the difference of language, between men and women, seems, in all the Indian languages, to be confined to that species of words, or others of an analogous nature, and to the use of interjections.

It is perhaps less, however, in dictionaries, than by an investigation of grammatical forms and structure, that we must study the philosophy of language and the various ways, in which man has applied his faculties to that object. We may discover in their Relations, that the Jesuits had analyzed the two principal languages spoken in Canada. The venerable Eliot had in his Grammar, published in 1666, exhibited the most prominent features of the Massachusetts dialect. And we have long been in possession of good grammars of several of the languages of Mexico and South America by the Catholic missionaries. But it was not, till after the publication of the more popular works of Egede * and of Crantz, that public attention was attracted by the peculiar character of the Karalit or Eskimau language. And the first inference was, that the Eskimaux must have been a colony from Europe, or from some other civilized country, and a distinct race from the other American Indians. In the year 1819, Mr. Du Ponceau, after having elicited with much labor, from Mr. Heckewelder, the principal features of the Delaware, and compared it with the

* Alluding to his account of Groenland, rather than to his Grammar.

Eskimau, with the languages of South America, and with the scanty specimens within his reach of those of our own Indians, submitted to the further investigation of the learned the three following propositions, to wit:

1. That the American languages in general are rich in words and in grammatical forms, and that, in their complicated construction, the greatest order, method, and regularity prevail.

2. That these complicated forms, which he calls polysynthetic, appear to exist in all those languages from Greenland to Cape Horn.

3. That these forms appear to differ essentially from those of the ancient and modern languages of the old hemisphere. The last proposition does not fall within the scope of this essay, and is far beyond my very limited knowledge of languages. All the information, connected with the first proposition, which could be obtained, has been collected, and will be found in a condensed form in the annexed grammatical notices and specimens of conjugations. But the inquiry has, with a single exception, been confined to the languages of our own Indians; and the result, so far as it goes, fully confirms the first two propositions of Mr. Du Ponceau; although I think, that there is less of method and regularity in the Delaware and other dialects of the Algonkin-Lenape, than in some of the other Indian languages.

Yet the materials are very incomplete; although we may perceive the general features, we cannot yet deduce with sufficient precision the rules of grammar or of the composition of words; and there is some difficulty in discriminating between. the specific characters which distinguish certain languages, and the general features which belong to all. But we are at least justified in asserting, that such a general character does exist, that it applies to all those American languages which have been sufficiently investigated, and that it seems to prove, beyond a doubt, that common origin, which could not be discovered in vocabularies so entirely different from each other. It is not however intended to assert, that all the American languages, without exception, possess that general character. It would indeed appear more astonishing, to find them all belonging to one and the same family, than to discover some, like the Chinese in Asia, and the Basque in Europe, of a structure altogether differing from the general mass.

The fundamental characteristic of the Indian languages of

America appears to be a universal tendency to express in the same word, not only all that modifies or relates to the same object, or action, but both the action and the object; thus concentrating in a single expression a complex idea, or several ideas among which there is a natural connexion. All the other features of the language seem to be subordinate to that general principle. The object in view has been attained by various means of the same tendency and often blended together: a multitude of inflections properly so called; a still greater number of compound words, sometimes formed by the coalescence of primitive words not materially altered, more generally by the union of many such words in a remarkably abbreviated form; and numerous particles, either significative, or the original meaning of which has been lost, prefixed, added as terminations, or inserted in the body of the word.

The modern languages of Europe generally, and none more than the English, have substituted, for the inflections of the ancient languages, auxiliary verbs and separable prepositions; and the inflections or compounded words, in the classical languages, bear no proportion in point of number to the multiplied forms and combinations exhibited by those of the Indians.

Notwithstanding this great apparent complexness, all these various forms, either of inflected or compounded words, must necessarily have their foundation in analogy, modified by euphony: but they render a competent acquirement of the language extremely difficult to a foreigner; and even after this object has been attained, more by routine than in any other way, it must be no easy task for the student, to analyze the words, to reduce them to their proper elements, to class them in conformity with the genius of the language, and to convey to others his knowledge with method and sufficient perspicuity.

This remains to be done for almost every Indian language; and we can, in the mean while, only try to give some imperfect notions of the most general features which appear to have been ascertained.

Number and Gender.

There is a great variety in the Indian languages with respect to Genders and Number.

Like all others, they have various distinct words, expressive of the differences of sex in the human species, in reference