88
��INTERNATIONAL JOURNAL OF AMERICAN LINGUISTICS
��VOL. I
��any affix whatever (-qso drops off: hence THY UNCLE is a simpler term than UNCLE). Further, the terms for MY FATHER and MY MOTHER are irregularly formed by adding the first person singular "objective" element -s directly to the stem, the vowel of which is lengthened (the normal affix for MY is -qsak- qas). These facts mean, for example, that while the forms for MY FATHER and THY FATHER have no suffix of physical separa- bility, and fall outside the ordinary possessive pronominal scheme, such forms as HIS FATHER, OUR FATHER, and MY UNCLE are treated, as far as the possessive pronominal affixes are concerned, like an ordinary possessed noun; in neither sets of forms is the suffix of physical inseparability in place. As far as the Nootka evidence is pertinent, it is obvious that the concepts of separability and relationship- term classification are morphologically and historically unrelated.
The pronominal distinctness of terms of relationship is not as isolated a phenomenon as Uhlenbeck implies. Wishram ' (Upper Chinook) affords us some interesting data. The possessive pronominal prefixes of terms of relationship in this language are precisely the same as for all other nouns, except for the first and second persons singular of the words for FATHER and MOTHER. In these isolated cases MY and THY are respectively expressed by -na- and -ma- instead of the normal -tc-, -k- MY and -mi- THY; the interesting point is, that -na- and -ma- are evidently closely related to the verbal pronominal prefixes n- and m-. Body-part nouns with possessives are in no way peculiarly treated in Wishram.
The combined evidence of Takelma, Yuki, Pomo, Mutsun, Nootka, and Chinookan for the occurrence of a distinctive series, some- times only preserved in very fragmentary form, of possessive pronominal affixes for terms of relationship, can hardly be set aside
1 The Paiute, Nootka, and Wishram facts are quoted from my manuscript field-notes.
��as pointing to a merely secondary reduction of the inseparable class of possessed nouns. A little reflection shows that terms of rela- tionship as modified by possessive pronouns differ from most other nouns so modified, not so much in the matter of inseparability as in the fact that in the former a personal relation is defined, while in the latter true possession or some allied concept is indicated. Thus, MY FATHER is not one who is owned by me, but rather one who stands to me in a certain relation; moreover, he may be some one else's father at the same time, so that MY FATHER has no inherently exclusive value. On the other hand, MY ARM, like MY HAT, indicates actual and exclusive possession. Hence we can readily understand both why certain non-kinship nouns that indicate rela- tionship are sometimes morphologically classed with kinship terms (e.g., FRIEND in Takelma, SWEETHEART in Nootka), and why, on the other hand, such relationship terms as do not involve an inherent or non-controllable relation frequently fall outside the true set of kinship terms (e.g., HUSBAND and WIFE are not treated like relationship terms in either Takelma or Nootka). That personal relation, not possession, is primarily expressed by the possessive pronominal affixes of relationship terms, is beautifully illustrated by the Iro- quois usage of expressing many such relations as transitive verbs; thus, one cannot say MY
GRANDFATHER Or MY GRANDSON in Iroquois,
but uses formal transitives which may be respectively translated as HE GRANDFATHERS ME or I GRANDFATHER HIM. Clearly, the morphological isolation of possessed terms of relationship finds abundant justification in psychological considerations. I would, then, in contradistinction to Uhlenbeck, allow for three fundamental types of classification of possessive pronouns in America:
1. All nouns treated alike (Yana, Southern Paiute).
2. Relationship terms contrasted with other nouns (Takelma).
�� �
