Argument structure binding and event nominal polysemy

It is well known that event nominals tend to have both an event and one or more non-event interpretations and that non-event interpretations may pick out various aspects of the event, such as the resulting object or state, the place, the instrument and so on (Apresjan 1974; Bierwisch 1990/1991, 2008; Pustejovsky 2005). However, the representation of their polysemy can still be improved, especially for that which concerns their non-event interpretations. For example, it has been assumed that Result nominalizations introduce an argument place over and above those found in the corresponding verb (see among others Asher 1993: 151). Engl. analysis for instance introduces reference to an abstract object which apparently is not originally an argument of the verb analyse. Not all Result nominals seem to behave this way, however. Some Result nominals fill an argument place of the corresponding verb: for example, building introduces reference to an entity which can be identified with the grammatical object of the verb build. Given these premises, the overall aim of this paper is to contribute to the representation of the lexical ambiguity exhibited by event nominals by looking in particular at what argument of the base verb is bound in the non-event interpretations . More specifically, I will be concerned with clarifying under what conditions the argument introduction posited for Result nominalizations takes place and when instead it does not. I will argue that this phenomenon depends to a large extent on the meaning of the underlying verb and that the base verbs of nominals which introduce reference to this extra argument are verbs of covert creation. My analysis is based on the theories of Argument structure and Event structure as developed within the Generative Lexicon model (henceforth GL) (Pustejovsky 1995), which I briefly outline in section 2. The language discussed is Italian but the proposed generalizations may easily expand on a broader perspective.