Genericity and (Non)accidentalness

Abstract

This paper attempts to clarify nature of the "law-likness" or "nonaccidentalness" that generic sentences are usually claimed to express. It does so by examining the interactions of such generic sentences with a construction which seems to express "accidentalness," namely the happens to construction (as in John happens not to see well). In particular, it turns out that generics with bare plural subjects (BP generics, like Dogs have four legs), but not generics with indefinite singular subjects (IS generics, like A dog has four legs) are compatible with this construction (compare Dogs happen to have four legs vs #A dog happens to have four legs). I analyze happens to as a domain vague necessity operator, i.e. a universal quantifier over worlds, whose restriction (the domain of worlds quantified over) is systematically vague. Following Greenberg (2003Greenberg ( , 2008 I propose that a number of distributional and interpretational differences between IS and BP generics can be attributed to the fact that although both have the same basic modal quantificational semantic structure the restriction over worlds is necessarily precise in the former kind of generics but is allowed to be vague in the latter. The compatibility of BP generics with happens to is thus analyzed as a case of modal concord.

The meaning of generics

“…Generics do not convey any stable or easily specifiable information about how many members of the given kind or group have the given property (Carlson, 1977;Lawler, 1973;Leslie, 2007bLeslie, , 2008. Moreover, generics are often taken to communicate nonaccidental, quasi-essential, or principled generalizations (Carlson, 1989;Dahl, 1975;Greenberg, 2012;Pelletier & Asher, 1997;Prasada & Dillingham, 2006); however, the strength and nature of the relevant modal relation seem to vary from generic to generic (and more controversially between utterances of a single generic).…”