Tobias Scheer (Nice)

On the difference between the lexicon and computation (regarding Slavic yers)

The talk discusses three concerns regarding Gouskova & Becker's (2016) analysis of Russian yers that is based on two sublexicons (yer-bearing morphemes and morphemes without yers), to which lexically specific constraints make reference. First, the basic generalization about Slavic yers that is expressed by the established analysis (Lower) is missed, and it is shown that the major objection against this analysis disappears when the existence of final empty nuclei is recognized. Second, generalizations about the lexicon (i.e. where exactly yers occur) and properties of the computational system (i.e. the mechanism that decides which yers appear on the surface) are two different things. Their confusion in Gouskova & Becker undermines arguments that are meant to speak to yer vocalization when in fact they concern lexical distributions. Third, generalizations about the sonority profile of yer deletion-created clusters are central for Gouskova & Becker. They are irrelevant for actual human learners (children or adults, as opposed to machine learners that are used in Gouskova & Becker), though, and the experimental evidence adduced by the authors precisely shows that speakers happily lexicalize and compute sequences (such as yerCC) which are absent from the lexicon. The gaps at hand are thus accidental, rather than systematic.

Gouskova, Maria & Michael Becker 2016. Source-oriented generalizations as grammar inference in Russian vowel deletion. Linguistic Inquiry 47: 391-425.