Speaker: Martina Wiltschko (University of British Columbia)
Title: The composition of INFL: An exploration of tense, tenseless languages and tenseless constructions.
Time: Friday, September 25, 2009, 3:30pm
Place: 32-141
In this paper we argue that the functional category TENSE, sometimes viewed as the head of the clause can be decomposed. It consists of the universal functional category INFL and language specific features of temporal content: [past] and [present]. Thus, the functional category TENSE is not a primitive category of UG, but INFL is. Specifically, we argue that INFL has a universal function, namely anchoring, but that the substantive content associated with it is language specific or can be lacking alltogether.
We present three arguments for this view:
i) Arguments from language variation. We show that in some languages INFL exists without features of temporal content. Instead, in Halkomelem INFL is associated with substantive features of spatial content. In Blackfoot, INFL is associated with substantive features relating the participants of the event to those of the utterance (Ritter & Wiltschko, in press). This much establishes that the category INFL exists independent of its substantive content, at least at the initial state.
ii) Arguments from tenseless constructions. The independence of INFL from its substantive content is further supported by the existence of constructions where INFL appears without temporal content – even in a language which otherwise appears to be a tensed language: Infinitives and imperatives. This much establishes that INFL exists without substantive content even within a given language. The proposal correctly predicts that in contexts where INFL is used without substantive content the difference between English, Halkomelem and Blackfoot vanishes.
iii) Arguments from nominal licensing. Finally, we show that the licensing of nominal arguments varies with the substantive content associated with INFL. In languages with temporal features, nominal arguments are licensed via dependent marking (structural case), while in languages with spatial or participant features, nominal arguments are licensed via head-marking. This indirectly supports Pesetsky & Torrego’s 2001 idea according to which structural case reduces to tense features on D. However, since tense is not a primitive category in our analysis, we argue that case reduces instead to the substantive features that make up tense: [present = NOMINATIVE] and [past = ACCUSATIVE].
We conclude with a discussion of the implications of our proposal for our understanding of functional categories.