The syllabicity constraint in Mandarin and Wenzhounese

An OT-LFG account

Abstract

In Mandarin and Wenzhounese, a syllabicity constraint penalises syntactically well-formed VPs if the VP consists of a disyllabic verb and a monosyllabic object. There is quantitative evidence that the syl- labicity constraint is stronger in Mandarin and weaker in Wenzhounese. Grounded on these empirical data, this paper makes three theoretical points. First, this apparently syntax-sensitive constraint can be formalised in a purely phonological way, so LFG’s modular architecture can be maintained. Second, the stronger/weaker effect of the syllabicity constraint requires a gradient view of grammaticality, which can be captured by combining LFG with Stochastic Optimality Theory. Third, Optimality-Theoretic constraints might violate modularity by simultaneously referring to syntactic and phonological information. I will show that faithfulness constraints of this format do not violate modularity, while markedness constraints of this format does.