There is an orthodoxy in formal semantics, which is to treat bare plural generics (sentences of the form, ‘Ks are F’) as covertly quantificational, in the style of Lewis’s analysis of adverbs of quantification (1975). On the covert quantification account, generics have a tripartite structure, along with some kind of ‘quasi-universal’ quantificational force. The only alternative which has drawn serious attention is the kind-predication account, on which generics have the simple form of the predication of the name of a kind. In this paper, I sketch the beginnings of an alternative regimentation of bare plurals, which takes their surface form at face value and analyses them in terms of plural predication, by analogy with plural definite descriptions.
Like generics, plural definite descriptions exhibit a tolerance for exceptions. This property is known as non-maximality (Brisson, 2003). Appending ‘all’ to a definite plural has a “maximising effect" (Dowty, 1987): ‘the townspeople are asleep’ admits exceptions, whereas ‘all the townspeople are asleep’ does not. ‘All’ has the same effect on generics: ‘dogs have four legs’ admits exceptions, whereas ‘all dogs have four legs’ does not. I suggest that these are cases of a single pragmatic phenomenon; the members of each pair of sentences have the same truth conditions, but ‘all’ reduces how much “pragmatic slack" they allow (Lasersohn, 1999). Through the analogy with definite plurals, we can account for three puzzling features of bare plurals: (i) tolerance of exceptions; (ii) context-sensitivity and (iii) interaction with negation.
See the DPhil Seminar website for details.
DPhil Seminar Convenor: Mariona Miyata - Sturm
