Das Problem der "leeren Mitte" oder: Wie viele Intervalle braucht die Semantik von Phasenverben?The problem of the "emty mid" or: How many intervals does the semantics of phasal verbs need?
In this paper I argue that the default view on phasal verb semantics, conceptualising them as change-of-state entities shifting from a proposition to its negated counterpart and, consequently, operating with points of time, has to be revised by considering intervals (and not points) as the fundamental semantic reference entities; perhaps more important, the study's results make it necessary to take into account at least one more interval "in between". This interval is, I am proposing, by no means marginal, as it constitutes a relevant evaluation interval where phasal meaning proper in terms of beginning, fi nishing and continuing unfolds. Constructions with phasal verbs in general allow, as will be shown, the possibility of adverbials modifying an interval situated between the interval before and after the moment of change-of-state, which is a strong argument in favour of the proposed third interval. - On the outset of this article it is argued that mainstream approaches to phasal semantics, working with two points of time, indirectly introduce two evaluation intervals, and thereby leave open the interpretation of the mid interval. After having presented and discussed the relevant data, Dowty's influential work on verb semantics, Word meaning and Montague grammar, is discussed and it will be demonstrated that the ‘problem of the empty mid' was recognised by Dowty, but not properly integrated into his conception of interval semantics.