90
��INTERNATIONAL JOURNAL OF AMERICAN LINGUISTICS
��VOL. I
��more or less as in noun-compounds (i.e., i- HOUSE, for MY HOUSE, might be conceived of as a compound with merely implied genitive relation, precisely as in a form like HEN- HOUSE if interpreted as HOUSE OF HENS). There is still a further method of interpreta- tion, corresponding to the objective inter- pretation of the inactive or intransitive case given in the preceding review. This is to look upon the possessive affix as frankly objective (or dative) in character; e.g., to interpret a form like MY HOUSE as a semi- verbal HOUSE (is) TO ME. As a matter of fact, the line between such predicative forms as IT is MY HOUSE and such purely denomina- tive forms as MY HOUSE is often very difficult to draw; e.g., in Chinookan. Either of these explanations of the verbal affiliation of the possessive pronouns of so many American languages seems preferable, in my opinion, to Uhlenbeck's mystical theory of identification. The less we operate with "primitive" psy- chology, the better. Modern research is beginning to make it clear that the psychology of civilized man is primitive enough to explain the mental processes of savages.
One more point before closing. I feel that Uhlenbeck is too much inclined to look for
��functional or semantic explanations of posses- sive pronominal differentiation where purely phonetic factors are probably all that is really involved (e.g., in Washo; Salinan; Algonkin; and Takelma, aside from terms of relation- ship). A striking example of the failure to evaluate purely phonetic factors is afforded by his discussion of the Blackfoot terms isk BUCKET and its possessives (e.g., no-xk MY BUCKET). He considers the forms isk and -(o)xk as representing two etymologically unrelated stems, and connects this surprising phenomenon with such suppletive examples in Blackfoot as HORSE and MY HORSE (as also in Southern Paiute; similar cases occur fre- quently for DOG in America). It seems very much more likely to me that we are not here dealing with independent stems at all, but that an original osk was in Blackfoot regularly shifted to oxk (the back vowel and k pulling the 5 to a back position; namely, x). This explanation is practically demonstrated by comparing no-xk with Blackfoot mo-xkats-is FOOT (from Algonkin *-skat-; cf.Creemiskdt 1 LEG).
E. SAPIR.
1 Quoted from Lacombe.
�� �
