Whamit!

Title: (Dis)obeying the Head Final Filter: linearization or agreement?

Date and Time: *Wednesday*, October 10, 1-2pm
Location: 32-D461
Abstract:

Modifiers in many languages are subject to a well-known constraint that disallows them to be not head-final (cf. “*a proud of his son father” despite “John is proud of his son”). This constraint, which, following Williams (1982), is commonly referred to as the Head Final Filter on pre-nominal modifiers (HFF), is often believed to be universal, and in fact a version of it features as Greenberg’s (1963) Universal 21. Although there are several accounts available in the literature that attempt to derive this constraint (Abney 1987, Escribano 2004, Sheehan 2017, a.o.), none of them appears to be fully satisfactory in terms of data coverage, given that HFF does not always apply both within a language (cf. “an easy to read text” in English) and cross-linguistically, as a number of languages do not obey to it, including Bulgarian, Latin, Modern Greek, Polish, Russian, and Ukrainian. This talk will first present the results of our ongoing data collection, which seem to suggest that a reformulation of the empirical generalization is necessary, such that HFF is stated in terms of agreement richness, rather than head finality. It will then propose an alternative analysis, which covers a subset of the data and models the adjacency requirement of adnominal modifiers w.r.t. their modifiees as a result of constraints on linearization of agreement morphology. In the final part of the talk, I will discuss some possible ways to go about the remaining parts of the data, which involve zero exponence.