A Relevance-theoretic Account of the Property Predication Restriction

As several linguists have pointed out, the occurrence of indefinite subject NPs (in initial or preverbal position) is more restricted than the occurrence of definite ones. There seems to be a widespread typological preference for definite preverbal subjects, and some languages (Arabic, for instance) even forbid the presence of preverbal indefinite subjects. Other languages (such as English or Spanish) do not grammaticalize that universal tendency in such a strong way, but nevertheless they do impose certain constraints on indefinite subjects as well. My aim is to show that these constraints are pragmatic in nature and thus not amenable to a purely syntactic explanation such as the one advanced in Diesing (1992), for instance. More precisely, I would like to suggest that the constraints on indefinite subjects follow from the Principle of Relevance and can ultimately be explained by means of Relevance Theory. First, I will present a sketch of the so-called Property Predication Restriction; then I will provide evidence for its pragmatic status, and a tentative explanation based on Sperber and Wilson's theory. Finally, I will try to extend the account to some additional data.