Do Sentences have Tense?

Under the classical treatment in LFG, tense is a feature encoded in morphology or in the lexicon and projected into f-structures. It is formulated, as all other features, as an attribute-value equation; the values of the tense attribute generally are the labels by which traditional grammar refers to the tenses of the specific language.

A fact often not made explicit by syntacticians is that these labels need to be interpreted in semantics. The semantics of tenses has been formulated as a theory of temporal reference since the seminal work of Reichenbach (1947). Subsequently this theory has been integrated into a theory of context change, which widens the domain of tense semantics from the sentence to discourse (Kamp 1979, Kamp and Rohrer 1983). Functional linguistics since Weinrich (1964, 1971) investigates the role tenses play in textual organization and discourse pragmatics.

The aim of this workshop is to discuss a conception in which tense and temporal reference are distributed in a new way. It is assumed that tense features belong to morphological structure and are not projected to f-structure. Instead, they are mapped onto f-structure features based on the insights of tense semantics. Since there are no tense features in f-structure, according to this approach, sentences do not have tense.

In this contribution to the workshop, I will discuss this idea from a slightly skeptical point of view.