Aspectual Shifts With and Without Type Conflict

The referents of temporally determined sentences are usually talked of as situations (Vendler (1976)). The standard language internal testing procedure for telling apart telic accomplishments and achievements (??) from atelic states and activities (??) uses grammaticality effects in connection with temporal adverbial modification. (1) arrive *for an hour / in an hour play the sonata *for an hour / in an hour (2) be quiet for an hour / *in an hour swim for an hour / *in an hour However, after closer inspection, the marked phrases, far from being ungrammatical, quite naturally display several kinds of readings. Depending on the type of the situation, the focus of the modifier can be switched to the pre-phase, preparatory phase, or result state of the situation, or there appear special effects such as iteration. This phenomenon is known as aspectual shift. Aspectual shifts usually are theoretically handled on the analogy with typecoercion in programming languages (Moens/Steedman (1988)). However, none of the approaches available so far can offer a full theoretical explanation and compositional derivation of the shifted readings (e.g. Pulman (1997), de Swart (1998)). Moreover, there are cases where aspectual shifts take place without there being any conflict in situational type. Take, for instance, the pre-phase-reading of the left-peripheral adverbial construction in (??). Those cases are not treated