No people take more pains to speak their language accurately than the Dakotas. Their social condition is similar to that of the Arabs, whose language has within historic observation changed more slowly than any other. The Assinniboins have been separated from the Dakotas about three centuries, perhaps a little less, possibly much more. During all this time they have been entirely separated, associating wholly with tribes speaking languages entirely different, and yet their dialect remains almost identical with the Yankton. We are then encouraged to believe that their language has not changed so rapidly as to obliterate traces of its origin.
So far as I have been able to ascertain them the most important features characteristic of the Dakotan languages generally are the following:
I. Three pronominal prefixes to verbs, i, o and wa. I, this, forms nouns of instrument. O forms nomen actionis, etc. Some Crow and Minnetaree words seem to indicate that its original form was a. Wa, meaning some or something, prefixed to transitive verbs makes them intransitive or general in their application. Wa is in Min. ma (ba, wa), in Crow, ba. Scantiness of material prevents me from more than inferring the existence of these and other prefixes in the other allied languages, from a few words apparently containing them.
II. A system of verbal prefixes used to form verbs from certain stems, regularly varied in signification, according to the prefix used. The Dakota has seven of these prefixes. The Min. has three of these almost identical in force. I should suppose that I would, with as much material, find greater similarity in the other languages, but the only one I have been able to trace at all generally is Dak yu. This merely converts the stem into a verb without changing its meaning. Dak y is nearly always represented in the allied languages so far as I have observed by r, d, l or n; so that I find it in Min. du (ru, lu, nu), Iowa, Mandan, and Crow ru, Omaha ra.
III. A reflexive pronoun tawa, Min. tama (tawa, taba), Iowa tawe, Osage tabe, forming from possessive pronouns double possessives, related to their primitives somewhat as mine to my. In some features of structure the Dakotan languages present an amazing diversity.
According to Powell (Int. to stud. Am. Lang.) a Ponka in order to say "a man killed a rabbit," would have to say "the man, he, one, animate, standing, in the nominative case, purposely, killed, by shooting an arrow, he, the one animate, sitting, in the objective case." "For the form of the verb to kill would have to be selected, and the verb changes its form by inflection, and by incorporated particles, to denote person, number and gender, as animate or inanimate, as standing, sitting or lying."
On the other hand the Dakota could not vary the form of the verb to denote any of these things except number, with reference to either subject or object. He would probably say: "Wichasta-wan mastincha-wan kte,"—"man-a, rabbit-a, kill,"—in which each word is about synonymous with its English equivalent, and case as in English denoted by position. If he wished to show that the action was done by shooting, he would probably not vary the form of the verb kill, but would use the verb kute, meaning shoot whether with arrow or bullet. Except that the Dak. order corresponds to the Icelandic the only difference in structure between the Dak. and English expression is that the Dakota word kte may mean any time, the particular time being indicated whenever desirable in all cases in Dak. as mostly in English by auxiliary verbs and adverbs. If the word man were represented by a pronoun the Dak. would be still more analytic, since its pronoun would indicate any actor, male or female, or inanimate, unless it were desirable to distinguish, in which case the distinction would be made by compounding the pronoun with a suitable auxiliary word. In this feature, often given as characteristic of American languages, is a variation the greatest possible between two languages closely related. It is also worthy of remark that the Minnetaree, which I should suppose the most analytic of the group next to the Dakota, is one of those that least resembles the Dakota in vocabulary. Some of the features often assigned as peculiarities of American languages were according to Bopp and Schleicher features of the I. E. languages in their earlier stages. Of most other features said to characterize American languages I find in Dak. but faint traces. The Dak. does have verbs nearly synonymous with go, walk, eat, drink, strike, etc. It is well supplied with purely copulative verbs. It has differentiated the various parts of speech even to the definite and indefinite article. It is sufficiently supplied with nouns denoting genera and classes. This is not a feature of recent development. A much smaller proportion of general than of special names have lost trace of origin.
The Dak does not have inclusive and exclusive plurals, etc. It does not have a multiplicity of verb forms to denote mode and tense, but when necessary does denote them with elegance and precision, by auxiliary verbs and adverbs, very much as we do in English. The Dakota is not made up chiefly of very long words. On the other hand it uses a great many little particles and connectives to express fine shades of meaning, wonderfully reminding one of the Greek. It fully agrees with other American languages in its wonderful facility for forming derivatives. The I. E. languages in their earlier stages possessed equal facility.
As a matter of fact we know scarcely anything concerning the structure of American languages aside from the Algonquin and Iroquois groups, and a very few isolated languages. They have been classified, in fact, almost entirely by examination of scanty and not very accurate vocabularies. In investigating the relations of the Dakotan to other American languages we are therefore compelled to base our conclusions chiefly on vocabulary. I once resided a year among the Chippewas, and in various ways have had much better opportunities of comparing the Dakota with the Chippewa than with any other American language. I have not been able to find a word alike in the two; and but very few words even slightly similar in sound and sense. In pronouns few languages in any part of the world are so strikingly contrasted. If I were to attempt an argument for original affinity between Dakota and Chippewa my argument would be that so great dissimilarity could not be the result of accident. Aside from the Cheyenne, an Algonkin language which has incorporated some Dakotan words, and the Pawnee group, the similarities east of the Rocky mountains are surprisingly few, though the Huron, Iroquois and Mobilian languages do not seem quite so strongly contrasted as the Algonkin. Among the Eskimo, the tribes of the Pacific Slope, Mexico, Central and South America, we occasionally find identical and not infrequently similar words. In some the resemblances seem remarkable considering the size of the vocabulary. Closer examination shows however that they are not of a kind to indicate a special relationship. They are almost exclusively confined to a few pronominal bases of very wide diffusion, and the following: 1. ata, tata. 2. papa, each meaning father; 1. ana, nana; 2. ma, mama, each meaning mother. As an example I take the base ata, tata. Dakota, ate (dialect ata); Minnetaree, ate, tata, tatish; Mandan, tata; Omaha, adi, dadi; Ponka, tade-ha; Aricaree, ate-ah; Pawnee, ate-ish.
Tuscarora ata; Cherokee e-dauda; Eskimo—Greenland ahtata, Aleutian ata, California, San Miguel tata; Mexico Aztec teta; Otomi, ta, te; Yucatan, Cakchequil tata; Central Am. Tarasca tata; Darien tauta; Eastern Peru, Mossa tata; Western Paraguay, Villela tata.