subject, object and verb in one word, and if for any reason the object is not included, the scheme of the sentence is still maintained in the verb, and the object is placed outside, as in apposition, without case ending, and under a form different from its original and simple one.

This will readily be understood from the following examples from the Mexican language.

The sentence ni-naca-qua, is one word and means “I, flesh, eat.” If it is desired to express the object independently, the expression becomes ni-c-qua-in-nacatl, “I it eat, the flesh.” The termination tl does not belong to the root of the noun, but is added to show that it is in an external, and, as it were, unnatural position. Both the direct and remote object can thus be incorporated, and if they are not, but separately appended, the scheme of the sentence is still preserved; as ni-te-tla-maca, literally, “I, something, to somebody, give.” How closely these accessories are incorporated is illustrated by the fact that the tense augments are not added to the stem, but to the whole word; o-ni-c-te-maca-e, “I have given it to somebody;” when the o is the prefix of the perfect.

In these languages, every element in the sentence, which is not incorporated in the verb, has, in fact, no syntax at all. The verbal exhausts all the formal portion of the language. The relations of the other words are intimated by their position. Thus ni-tlagotlaz-nequia, I wished to love, is literally “I, I shall love, I wished.” Tlagotlaz, is the first person singular of the future, ni-nequia, I wished, which is divided, and the future form inserted. The same expression may stand thus: ni-c-nequia-tlagotlaz, where the c is an intercalated relative pronoun, and the literal rendering is, “I it wished, I shall love.”

In the Lule language the construction with an infinitive is simply that the two verbs follow each other in the same person, as caic tucuec, “I am accustomed to eat,” literally, “I am acustomed, I eat.”

None of these devices fullfils all the uses of the infinitive, and hence they are all inferior to it.

In languages which lack formal elements, the deficiency must be supplied by the mind. Words are merely placed in juxtaposition, and their relationship guessed at. Thus, when a language constructs its cases merely by prefixing prepositions to the unaltered noun, there is no grammatical form; in the Mbaya language e-tiboa is translated “through me,” but it is really “I, through;” l’emani, is rendered “he wishes,” but it is strictly “he, wish.”

In such languages the same collocation of words often corresponds to quite different meanings, as the precise relation of the thoughts is not defined by any formal elements. This is well illustrated in the Tupi tongue. The word uba is “father;” with the pronoun of the third person prefixed it is tuba, literally “he, father.” This may mean either “his father,” or “he is a father,” or “he has a father,” just as the sense of the rest of the sentence requires.

Certainly a language which thus leaves confounded together ideas so distinct as these, is inferior to one which discriminates them; and this is why the formal elements of a tongue are so important to intellectual growth. The Tupis may be an energetic and skillful people, but with their language they can never take a position as masters in the realm of ideas.

The absence of the passive in most, if not all, American tongues is supplied by similar inadequate collocations of words. In Huasteca, for example, nana tanin tahjal, is translated “I am treated by him;” actually it is, “I, me, treats he.” This is not a passive, but simply the idea of the Ego connected with the idea of another acting upon it.

This is vastly below the level of inflected speech; for it cannot be too strenuously maintained that the grammatical relations of spoken language are the more perfect and favorable to intellectual growth, the more closely they correspond to the logical relations of thought.

Sometimes what appears as inflection turns out on examination to be merely adjunction. Thus in the Mbaya tongue there are such verbal forms as daladi, thou wilt throw, nilabuite, he has spun, when the d is the sign of the future, and the n of the perfect. These look like inflections; but in fact d, is simply a relic of quide, hereafter, later, and n stands in the same relation to quine, which means “and also.”

To become true formal elements, all such adjuncts must have completely lost their independent signification; because if they retain it, their material content requires qualification and relation just as any other stem word.

A few American languages may have reached this stage. In the Mexican there are the terminals ya or a in the imperfect, the augment o in the preterit, and others in the future. In the Tamanaca the present ends in a, the preterit in e, the future in c. “There is nothing in either of these tongues to show that these tense signs have independent meaning, and therefore there is no reason why they should not be classed with those of the Greek and Sanscrit as true inflectional elements.”24-*

§ 13. Psychological Origin of Incorporation.

This Incorporative plan, which may be considered as distinctive of the American stock of languages, is explained in its psychological origin by Humboldt, as the result of an exaltation of the imaginative over the intellectual elements of mind. By this method, the linguistic faculty strives to present to the understanding the whole thought in the most compact form possible, thus to facilitate its comprehension; and this it does, because a thought presented in one word is more vivid and stimulating to the imagination, more individual and picturesque, than when narrated in a number of words.25-*

But the mistake must not be made of supposing that Incorporation is a creative act of the language-sense, or that its products, the compounds that it builds, are real words. Humboldt was careful to impress this distinction, and calls such incorporated compounds examples of collocation (Zusammensetzung), not of synthesis (Zusammenfassung). On this ground, he doubted, and with justice, the assertion of Duponceau, that the long words of the Lenape (Delaware) dialect are formed by an arbitrary selection of the phonetic parts of a number of words, without reference to the radical syllables.25-† He insisted, as is really the case, that in all instances the significant syllable or syllables are retained.

§ 14. Effect of Incorporation on Compound Sentences.

As has been seen, the theory of Incorporation is to express the whole proposition, as nearly as possible, in one word; and what part of it cannot be thus expressed, is left without any