8o
��INTERNATIONAL JOURNAL OF AMERICAN LINGUISTICS
��VOL. I
��already present, and of advance statements whose validity remains to be demonstrated, I should say that the 57 distinct stocks recognized on the revised linguistic map of the Bureau may be expected to re-arrange themselves into perhaps not more than 16, or even less. Always bear in mind, however, that the great divisions recognized by Powell still have significance, only that many of them are now to be understood as major subdivis- ions of larger linguistic units. While nothing is further from my mind than to minimize the great usefulness of Powell's classification, I may be pardoned for regretting the too definitive and dogmatic form in which it was presented. This has had the effect until recently of discouraging further researches into the problem of linguistic groupings in America. It is always dangerous to erect a formidable structure on a largely negative basis, for one tends to interpret it as a positive and finished accomplishment. However, I would freely grant that the services rendered by Powell's classification have far outweighed its deterrent influence. A thoroughly revised map of linguistic stocks north of Mexico will sooner or later have to be issued; but it is as well not to be too precipitate about this, as the whole subject of the genetic classification of Indian languages is at present in a state of flux.
In reviewing the linguistic publications of the Bureau as a whole, we have a right to ask three leading questions: Is the standard of phonetic accuracy adopted in the recording of the languages adequate? Are the grammars of these languages so presented as to convey a satisfactory notion of the fundamental characteristics of their structure? and, Have various languages been treated from the com- parative standpoint, so as to suggest histori- cal perspectives transcending those obtained from the intensive study of particular lan- guages? Let us briefly consider each of these queries.
��Early in its career the Bureau outlined a phonetic alphabet, which, as compared with the best that phonetic research at the time had to offer, was quite inadequate, but which was so vast an improvement on the amateur- ish methods in vogue for recording Indian words, that its adoption must be considered a great step forward in the study of American Indian linguistics. It has undoubtedly done good work in its day, and must be taken as the basis for further improvements. However, as it was framed without any very deep knowl- edge of the actual phonetic problems pre- sented by American languages, many of which are of exceptional difficulty and com- plexity in this respect, field investigators soon found it impossible to give an even approximately adequate idea of the requisite phonetic facts without straining its resources. In this way new symbols were added from time to time by various investigators, and the accuracy of linguistic notation, limited naturally by the native abilities of the record- ers, grew apace. It is difficult to dispose of the phonetic quality of the series in a word. It is hardly fair to lay stress on the orthog- raphies of some of the earlier works; e.g., Nos. 30-34 and 44. On the other hand, I do not think one could candidly say that much even of the more recent work is as good as we should like to have it (Nos. 18, 52, and 57 probably about represent the high-water mark). The general run of the linguistic papers might be not unfairly described as "reasonably good" in phonetic respects, certainly no better.
Had a really scientific and reasonably complete phonetic alphabet been adopted earlier in the life of the Bureau, I believe the phonetic standard of some of the later linguistic work done under its auspices would have been even higher than it is. Experience shows that a field-worker tends, in his hearing 9f unfamiliar sounds, to be influenced by the standard phonetic scheme that has made itself at home in his inner ear; he will assimi-
�� �
