1. The Anchoring Condition

In this paper we assume without further discussion that in main declarative clauses of all natural languages events must be anchored to the utterance or some other salient reference point – we refer to this as the Anchoring Condition (cf. Enç 1987: 642). In particular, we follow Stowell 1995 and Zagona 1990, among others in assuming that anchoring proceeds temporally via the syntactic category tense and its temporal arguments. That is, the event time (EvT) is anchored to the utterance time (UttT) or some other reference time (RefT) (Enç 1987). In this paper we focus on root sentences where events are anchored to the utterance. When we talk about such events, we relate them to the utterance by asserting when, relative to the utterance, the event happened. We further follow Demirdache and Uribe-Etxebarria 2000 in assuming that tense is a temporal predicate of (non-)coincidence (in the sense of Hale 1986). According to this view, a sentence in the present tense (1)a can be represented as in (2)a, where present tense corresponds to a predicate of coincidence asserting that the event time coincides with the utterance time (i.e. it happens now).1 In contrast, a sentence in the past tense (2)a can be represented as in (2)b, where the past tense corresponds to a predicate of non-coincidence, asserting that the event time does not coincide with the utterance time (i.e., it happens not now but then, which is interpreted as past 2). (1) a. Konrad is playing the guitar. b. Mika danced