It is also evident that things were not thought of, or talked of, out of their natural relations. There are still in most American tongues large classes of words, such as the parts of the body and terms of kinship, which cannot stand alone. They must always be accompanied by a pronoun expressing relation.
Few American tongues have any adjectives, the Cree, for instance, not a dozen in all. Prepositions are equally rare, and articles are not found. These facts testify that what are called “the grammatical categories” were wholly absent in the primitive speech of man.
So also were those adjectives which are called _numerals_. There are American tongues which have no words for any numerals whatever. The numerical concepts one, two, three, four, cannot be expressed in these languages for lack of terms with any such meaning.[355] This was a great puzzle to the missionaries when they undertook to expound to their flocks the doctrine of the Trinity. They were in worse case even than the missionary to an Oregon tribe, who, to convey the notion of _soul_ to his hearers, could find no word in their language nearer to it than one which meant “the lower gut.”
A very interesting chapter in the study of these tongues is that which reveals the evolution of specific distinctions, those inductive generalizations under which primitive man classified the objects of the universe about him. These distinctions were either grammatical or logical, that is, either formal or material. That most widely seen in America is a division of all existence into those which are considered living and those considered not living. This constitutes the second great generalization of the primitive mind, the first, as I have said, having been that into Being and Not-being. The distinctions of Living and Not-living gave rise to the _animate_ and _inanimate_ conjugations. A grammatical sex distinction, which is the prevailing one in the grammars of the Aryan tongues, does not exist in any American dialect known to me.[356]
It is true that abstract general terms are absent or rare in the most primitive tongues. On the other hand, we find in them a great many classificatory particles. These correspond only remotely to anything known in Aryan speech, and seem far more abstract than generic nouns. I will illustrate what they are by an example taken from the Hidatsa, a dialect of the Dakota.