Nominal Declension in West Slavic: Variation and Abnormality

서지정보

· 발행기관 : 한국슬라브어학회
· 수록지 정보 : 슬라브어 연구 / 15권 / 1호 / 101 ~ 118페이지
· 저자명 : 최성호

초록

This paper discusses "unusal" inflection classes that are shown in Czech III declension and Polish Masculine I-declension. The Czech one illustrates inflection classes that are not lexically characterized, while the Polish masculine I-declension exhibits morphologically non-uniform affixal sets. The abnormality lies in the fact that affixal sets postulated in these two languages are either distributionally arbitrary or internally unrelated in affixal choice among cases.
In discussing the Czech example, I argued that lexically non-characterized inflection classes tend to be eliminated in favor of the most unmarked inflection class among them. The elimination of such arbitrary classes simplify the grammar, since the previously postulated morphology-based lexical features (e.g. [III-a], [III-b], etc.) is now obviated. I suggested that the innovational path takes the orders of "marked plural number" and then "unmarked singular number" and of "marked oblique cases" and then "unmarked directive cases" (see section 2.2). What this means for the synchronic grammar is that inflection classes are motivated; in other words, they are lexically characterized. The transfer of previous III-a declension nouns to current III-b (and II-b) declension nouns is abandonment of their old III-a pattern and (incomplete) acquisition of a new II-b pattern, thereby being motivated. While the III-a pattern reduces its distributional domain, it becomes motivated, too.
On the discussion of the Polish example, I suggested that the M-pattern should be subdivided into "Inanimate" and "Animate" subpatterns, since the semantico-lexical feature of Animacy is extending its selectional force beyond AS onto GS. In other words, affixal choice in GS depends on animacy opposition, just like in AS. This expansion brings about an implicational relation in affixal choice within the M-pattern: if AS shows GS syncretism, then GS selects -a. For DS, one of the two allomorphs, namely -u, is dying out, which means that nouns selecting it must be lexically so specified, thus augmenting arbitrary specification in the lexicon. This dooms to be eliminated, and that is exactly what is happening now: the productive and dominating DS affix is -owi. As for LS, its competing variants show quite different distributional statements from GS variants (cf. section 3). This entails that there is no implicational relation between GS affixal choice and other cases' affixal choice: their distributions are independent from one another. If no implicational relation is established, no affixal pattern will be established. Of course, it is possible to set up affixal sets, as in FM & CM (2000), but they are only for the sake of classification, lacking any significant generalization on affixal allomorphy.
In conclusion, I claim that the notion of "inflection class" or "affixal pattern" serves a valid analytical tool, in so far as it satisfies the two requirements: namely, "motivated distribution" and "internal implicational relation".

참고자료

· 없음