changes, and the transparency of the method of composition, make it clear that we are performing here an analytical process, that does not need to have as its counterpart the synthesis of independent elements. The same question may also be raised in regard to phonetic modifications of the stem, which may be secondary, and due to the influence of changing accents in composition or to vanished component elements, while they may also be primary phenomena.
This problem is in a way identical with the whole question of the relation between word and sentence. Here also American languages may furnish us with much important material that emphasizes the view that the unit of human speech as we know it is the sentence, not the word.
The problems treated in a linguistic journal must include also the literary forms of native production. Indian oratory has long been famous, but the number of recorded speeches from which we can judge their oratorical devices is exceedingly small. There is no doubt whatever that definite stylistic forms exist that are utilized to impress the hearer; but we do not know what they are. As yet, nobody has attempted a careful analysis of the style of narrative art as practised by the various tribes. The crudeness of most records presents a serious obstacle for this study, which, however, should be taken up seriously. We can study the general structure of the narrative, the style of composition, of motives, their character and sequence; but the formal stylistic devices for obtaining effects are not so easily determined.
Notwithstanding the unsatisfactory character of the available material, we do find cases in which we may at least obtain a glimpse of the intent of the narrator. In many cases metaphorical expressions occur that indicate a vigorous imagination. Not much material of this character is available, but what little we have demonstrates that the type of metaphor used in different parts of the continent shows characteristic differences. It would be interesting to know in how far these expressions have become purely formal without actual meaning, and in how far they reflect an active imagination.
Evidence is not missing which shows that the sentence is built up with a view of stressing certain ideas or words by means of position, repetition, or other devices for securing emphasis. There are curious differences in the tendency to fill the discourse with brief allusions to current ideas difficult to understand for any one who is not versed in the whole culture of the people, and the enjoyment of diffuse, detailed description. Collectors of texts are fully aware that in the art of narrative there are artists and bunglers in every primitive tribe, as well as among ourselves. At present there is hardly any material available that will allow us to characterize the tribal characteristics of the art of narrative.
The most promising material for the study of certain aspects of artistic expression are the formal elements that appear with great frequency in the tales of all tribes. Most of these are stereotyped to such an extent, that little individual variation is found. Even in poorly recorded tales, written down in translation only, and obtained with the help of inadequate interpreters, the sameness of stereotyped formulas may sometimes be recognized. Conversation in animal tales and in other types of narrative, prayers and incantations, are probably the most important material of this character.
Attention should also be paid to the existing forms of literature. The narrative is of universal occurrence, but other forms show a much more irregular distribution. The psychological basis of the trivial American anecdote is not easily understood. The connotation of meaningless syllables that occur in songs, the frequent use of distorted words in poetry, and the fondness for a secret language, including obsolete, symbolic, or arbitrary
