Reasoning with Polarity in Categorial Type Logic
Raffaella Bernardi · 2002
Formal linguistic literature offers several theories providing classifications of items which belong to the same semantic type but which differ in some other aspect. For example, generalized quantifiers (GQs) can be classified considering their different distribution with respect to negation [BS97]. Polarity items (PIs) differ in the sort of licensors they require [Woud94, Gia97]. Wh-phrases can be distinguished according to their sensitivity property with respect to different classes of weak-islands [SZ97]. In all these cases, the described classifications are based on semantically motivated subset relations holding among the involved items. In the talk, I will show how categorial type logic provides the right expressivity to model these phenomena. The derivability patterns among types give a precise way to (i) gain a deeper understanding of the typological classifications proposed in the literature of formal linguistics, (ii) carry out cross-linguistic comparisons, (iii) clarify the consequences predicted by the classifications opening the way to further investigations, and (iv) establish new dependency between linguistic phenomena. The talk will consist of two parts. First of all, after introducing the framework of the Lambek