Speaker: Mitya Privoznov (MIT)
Title: Russian stress in inflectional paradigms
Date/Time: Monday, May 15, 5:00–6:30pm
Location: 32-D461
Abstract:
All analyses of Russian stress agree that it is contrastive. This is easily shown by the existance of minimal pairs, like zámok ‘castle’ vs. zamók ‘padlock’. Hence, at least some information about stress has to come from the lexical entries for individual morphemes. The question is: what kind of information? Most analyses, starting with Jakobson (1963), Zaliznyak (1967) and Halle (1973), assume that there is an underlyingly specified feature [+/-stress], cf. Zaliznyak (1985), Melvold (1990) and Alderete (1999). These theories distinguish between three types of morphemes in Russian: stressed, unstressed and “right-stressed”. The latter type is underlyingly accented, but, in Halle (1973)’s terms, it invokes a specific rule that shifts stress from the morpheme itself to the next syllable to the right (cf. Zaliznyak (1985)’s rightward marking). In this talk I am going to argue that instead of introducing the right-shifting stress rule we could assume that the stress feature, apart from having [+str] and [-str] values, can also be unspecified. This will give us the desired three-way distinction: morphemes can be stressed, unstressed and unspecified for stress. I am going to show that in a combination with StressLeft constraint (cf. Melvold (1990)’s BAP principle) and some independently needed auxiliary assumptions about morpheme dominance, this would derive the desired result.