NO. 4
��REVIEWS
��307
��it is lacking. It seems highly probable, moreo- ver, that a number of other /- prefixes (verbal and local) that Mason discusses in the progress of his sketch are etymologically identical with the nominal t- (e. g. conditional t-, la-, p. 44). It is most plausibly interpreted as a kind of nominal article of originally demonstrative force (cf. Hokan demonstrative stem * ta ; this fuller form seems to be found in Salinan en- clitic -ta NOW). It offers a striking and probably significant analogy to Washo d-, similarly pre- fixed to both primary and derivative nouns. The possessive pronominal prefixes of Salinan offer important analogies to the corresponding elements of other Hokan languages, notably Chimariko and Washo ; the lack of a distinct pronominal prefix for the first person singular is paralleled, it would seem, in Yuman.
In discussing the pronominal system of Sali- nan, Mason points out the presence of six more or less distinct series of elements : the indepen- dent personal pronouns ; the "proclitic" series, which might better have been frankly recogni- zed as constituting true prefixes (they occur only as verb subjects and are closely connected with the stem, whose initial vowels they sometimes displace) ; the objective elements, suffixed to the verb ; the locative series (e. g. NEAR ME, TO HIM); the possessive prefixes ; and the enclitic subjects. The last of these, however, are merely a secondarily abbreviated set derived from the independent pronouns. Of the others, the objective series stands out, for the most part, as distinctive, the others show consider- able interrelationship. The locative series, in particular, is evidently closely related, not, as Mason remarks, to the independent series, but to the " proclitics " and possessives. It is com- pounded of the pronominal element proper and a preceding k-, ke-, evidently an old locative or objective particle (cf. Yana objective and locative particle gi") ; hence, e. g., -k'e ME (loca- tive) and -ho HIM (locative) are to be analyzed
��as k(e)-'e TO-ME and ke-o TO-HIM (such a form as Mason's tewa'kok'e NEAR ME is most easily interpreted as f-e*r/b-'eTHE-pROXiMITY TO-ME). The close parallelism between the first person singular and plural forms in Salinan is charac- teristic of other Hokan languages ; the contrast of the e (/) or zero of the singular with the a of the plural is strikingly reminiscent of Chi- mariko.
In the section on " temporal proclitics " (pp.. 3 4, 35) there is betrayed a certain incom- pleteness or haltingness of analysis which is in evidence also elsewhere in the book. Phoneti- cally, this comes out in the author's treatment of the pronominal prefix or initial vowel of .the stem, which is often mistakenly, I imagine, drawn to the proclitic. To say that " the prefix ma- probably differs only phonetically from me- [WHEN]" (p. 25) is misleading. Such examples as me-yam WHEN I SEE and ma-yaL WHEN WE GO suggest strongly the analysis m-e-yatn and m-a-yaL with the regular " pro- clitic " pronouns e- i and a- WE. Morphologi- cally, Mason does not seem to realize the pro- bable denominating, in part demonstrative, origin of his temporal proclitics. They are only secondarily subordinating elements. Such a form as be'-ya WHEN i WENT (better b-c'ya or contracted be-eyd) is, without doubt, an indi- cative -fya i WENT subordinated by the demons- trative stern pe, pa " the, that " ; THAT I-WENT, whence WHEN i WENT, is a method of subor- dination that seems to be paralleled by like constructions in Yana and is strongly reminis- cent of Siouan .
The use of the perplexing verbal prefixes p- and k- (pp. 38, 39) suggests a fundamental generic classification of verbs. Mason himself doubtfully describes the p- verbs as transitives, the k- verbs as intransitives (e. g. k-enai TO HURT ONESELF, p-enai TO WOUND). This is the most obvious explanation but there are many difficulties in the way of its acceptance. That
�� �
