Page:IJAL vol 1.djvu/97

 NO. I

��REVIEWS

��8 9

��3. Possessed nouns classified into insepa- rable (comprising chiefly body-parts and terms of relationship) and separable (Chima- riko).

Sometimes types 2 and 3 intercross, when we get the triple classification of languages like Sioux and Haida.

Uhlenbeck's desire to look upon insepara- bility as the most fundamental concept involved in the so-called possessive relation is evidently largely determined by reasons of a speculatively psychological order. He notes with justice that the possessive pro- nouns of the inseparable category are gener- ally simpler than those of the separable cate- gory; that the latter are, indeed, frequently derivatives from the former. From this he argues that originally only inseparable nouns (body-part nouns and terms of relationship) had possessive affixes at all. Further, aside from certain exceptions (Miwok, Mutsun, Chumash), he finds that where, as is generally the case, the possessive pronouns are related to the pronominal affixes of the verb, they agree in form, not with the subjective or energetic, but, on the whole, with the objec- tive or casus inertia. The evidence for this important and well-known fact is drawn from Tlingit, Haida, Tsimshian, Chinook, Chim- ariko, Maidu, Yuki, Pomo, Muskhogean, and Siouan, to which we might add Shoshonean and Nootka.

Uhlenbeck's psychological interpretation of this fact, as well as of the greater primitive- ness of the possessive pronominal affixes of inseparable nouns, is given at the close of the paper: "Where there is identity of the posses- sive elements with inert personal elements, there can hardly be any talk of real 'posses- sion,' seeing that, where real 'possession' is involved, we should rather expect similarity of possessive with energetic elements, as opposed to a distinct series of inert personal pronouns or personal affixes. If, now, we recollect the excellent remarks of Lucien Levy-Bruhl on 'possession' in Melanesia, and bear in mind

��that, for example, in Dakota a noun with inseparably-possessive affixes has entirely, or nearly so, the form of a conjugated adjective, or, aside from the, in Dakota, differently placed pronominal element, of a verbalized independent noun, we shall not go wrong in recognizing in the so-called possessively in- flected noun an identifying expression. A [Dakota] form [meaning 'my heart'] thus does not signify 'my heart' in the manner of our civilized languages, but indicates the identity of myself with the one heart with which I, and no other, stand in the closest relation. Similarly the inclusive [Dakota form meaning 'child of us two'] is not so much 'child of us two' as indeed 'the child that we both are,' 'the phase of us two which is the child.' But it is impossible to transcribe into modern words the thoughts and feelings of 'primi- tives,' even though we are perhaps able to think and feel ourselves into them."

This psychological interpretation strikes me as extreme, the more so as I see no conclusive reason for assuming that possessive pro- nominal affixes were originally not employed with separable nouns. If we interpret Uhlenbeck's casus inertia, as suggested in the preceding review, as a neutral form of no intrinsic case significance, then the identifica- tion of a functional possessive with a specifi- cally intransitive or inactive case is arbitrary. As a matter of fact, in quite a number of American languages we find that the posses- sive affixes, while generally closely related to a series of pronominal affixes in the verb, are composed of a distinctively possessive element of non-personal significance and a pronominal element proper. This is the case, for instance, in Nootka and most of the Takelma possessive affixes. In such cases the possessive affix must naturally be periphrastically inter- preted: MY as OF ME, BELONGING TO ME. Where the sign of general possessive relation is lacking, the pronominal affix can be con- ceived of as standing in an implicit position- determined genitive relation to the noun,

�� �