NO. 4
��REVIEWS
��number of reduplications, especially in Indo- European, whose functions we cannot even guess, but the very fact that these roots belong to the most primitive part of the vocabulary (as van Ginneken argues) would seem to sug- gest the possibility that we are here confronted with an ethno-psychological problem which the present state of our knowledge does not enable us to solve. Some types may be less mysterious than they would seem to be at first sight. If e. g. the adjectives of color, shape, and surface quality are really to be regarded as iteratives (red here and there), as Gatschet thought (Contributions to North American Ethnology II, part i, p. 276), they belong to our emphatic group; and this author's valuable information about distributive nomina actoris and acti in Klamath whose distributive meaning (" action done at different times or occasions repeatedly, habitually or gradually "; Gatschet, ibidem) suggests the idea that perhaps all reduplicated nomina actoris, agentis, and acti may originally have had this meaning.
These few remarks about some of the most common types of reduplication may suffice to show that a careful inquiry into the psycholo- gical background of the phenomenon consider- ed in its entirety may be expected to yield important results. However, such an inquiry should be founded on a somewhat complete set of data and not on a number of facts arbi- trarily selected. A very valuable foundation would e. g. be afforded by a survey of all the types reduplication of found in languages of North America, whereas a comparative treat- ment embracing such an enormous field as the one represented by Brandstetter's short paper cannot be but both incomplete and superficial.
Finally I may be allowed to remind the reader of the existence of a highly important morphological problem connected with our subject, viz. the relation between reduplica-
��tion and vocalic intermutation (" change ") in North American languages. Several years ago Uhlenbeck pointed out the probability that, wherever it presents itself, this vocalic inter- mutation has originated of reduplication attend- ed with vocalic differentiation (C. C. Uhlen- beck, Grammatical distinctions in Algonquian demonstrated especially from the Ojibway- dialect, Leyden, E. J. Brill, 1909, pp. 10-20). Though the available evidence is perhaps not yet conclusive it is not to be disputed that more recent data point in the same direction . So Boas is inclined to think that certain plural forms in the Nass river dialect which show modifications of length and accent of stem syl- lables have originated by secondary modification of reduplicated forms (Handbook Amer. Ind. Lang. I 373)- The same may be said of modi- fication of the vowel replacing distributive redu- plication in Kwakiutl (Boas, ibid., 519, 522). An interesting example of how this process may take place is to be found in Sapir's paper on noun reduplication in Comox(Canada Geologic al Survey, Memoir 63 : type IV on p. 16), in which language we also meet with nouns reduplicated to begin with and substituting for plural reduplication a change of the first stem vowel (ibid., p. 18). If it could be proved that Uhlenbeck's suggestion is true, this would be a discovery of the greatest importance, not only with regard to the North American languages under consideration, but also with a view to the problem of the qualitative " ablaut " in Indo-European, although the psychological rela- tion between the latter and the North American " change " is still obscure.
It is to be hoped that Brandstetter's descrip- tive essay is to be regarded as the precursor of a thorough inquiry in which full justice will be done to every side of the problem.
J. P. B. DE JOSSELIN DE JONG.
State Museum of Ethnography, Leiden.
��PROTAT BROTHERS, PRINTERS, MACON (FRANCE)
�� �
