Abstract

Economy constraints in Minimalist syntax are usually taken to be universal. If an economy constraint C penalizes a derivation/representation in language A, it will also do so in language B. This paper presents a type of crosslinguistic variation that casts serious doubts on this assumption, namely the distribution of resumptive relatives vis-a-vis gap relatives. It is shown that while resumption is a last resort in languages like Zurich German, i.e. occurring only when gap relatives are barred, it can be an optional strategy in languages like Hebrew/Irish, thus occurring in the same environment as gap relatives. For the first type of language this implies that gap and resumptive relatives are in the same reference set and compete and that gap relatives block resumptive relatives. Gap relatives are shown to involve movement while resumptive relatives are derived by base-generation in Zurich German. Since a different set of lexical items is involved in the two derivations the reference set must be based on identical LFs rather than identical numerations. However, once this is established it is surprising that no blocking obtains in the second group of languages. Several options to solve this problem will be evaluated. It will be shown that the variation is best modeled by means of different rankings of interacting and violable constraints. The ban against resumption will be subsumed under a general constraint that penalizes External Merge. The paper is organized as follows: section one introduces basic facts about relativization in Zurich German. Section two explains the distribution of resumptives in Zurich German relatives. Section three discusses possible analyses of resumption under the assumption that gap and resumptive relatives are based on the same numeration. Section four reviews possible economy constraints that block resumption. Section five addresses pseudo-optionality. Section six shows that cross-linguistic variation requires a different interpretation of economy constraints, and section seven concludes the paper. 1

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call