Portal de Periódicos da CAPES

Russian Declensional Morphemes

1953; Linguistic Society of America; Volume: 29; Issue: 3 Linguagem: Inglês

10.2307/410029

1535-0665

George L. Trager,

Lexicography and Language Studies

0. The purpose of this article is twofold: to examine the various patterns of declension found in Russian and establish the MORPHEMES in them; and, by this presentation, to exemplify a possible method of morphemic analysis. The procedure involves the identification of the MORPI-IS that occur, and their assignment as ALLOMORPHS to morphemes. The methods used are summarized and exemplified in the author's treatment of English.' Some additional explanation is in order. The forms that occur are examined in phonemically isolated words (?0.1). They are compared and then segmented mechanically at points that leave partials which recur in more than one form. Thus the forms ruka (see ?0.3 for the orthography), ridku, rdki, rukax enable us to segment into ruk-, -', -, and the partials -a, -u, -i, -ax; these can be compared with noga, ndgu, n6gi, nog'x, and are thus established as recurring; further comparison of rukax, rukam, rukamji establishes the recurring partial -a-, and the partials -x, -m, -mji; and the latter are established as recurring by comparison with n6vix, n6vim, n6vimji. After a tentative set of morphs has been established in this way, the forms are tested in comparable frames and sequences to determine which have identical functions: thus odnA + ruka, odnh + k6stj compared with odnby + ruki, odn6y + k6stji, establish the functional identity of -a in odna and in ruka, and of these instances of -a with the absence of a final vowel in k6stj, this absence being set up as a zero element, -0.1 We can then say that there is a morpheme which contains at least the allomorphs -a and -0; the morpheme can be symbolized by some arbitrary symbol, say V/-N (see ?0.3). That our segmentation is tentative can be shown by the fact that from odnby + ruki and odnby + k6stji alone we could only conclude that -oy and -i are allomorphs of a single morpheme, whereas the introduction of further evidence makes it clear (as will be shown below) that odn6y has to be analyzed as odn-0-6y-0, and that it is the final -0 that goes with -i in a morpheme V-G. In this connection it is emphasized that no zero morphemes exist, but only zero allomorphs of morphemes whose other allomorphs are not zero. Stress patterns are separate morphs, forming suprasegmental morphemes. The constant aim in this procedure is to obtain the smallest possible number of different morphemes, with one morpheme for each category throughout the analysis, and to arrive at the results wholly by inspection and commutation in similar structural frames, without the use of meaning. Since the structural