|(1)||A:||Birds lay eggs.|
|B:||Mammals lay eggs, too.|
In this paper I propose a uniform account of the semantics of generics, which accounts for the quasi-existential readings as well as for the usual quasi-universal ones.
Generics are focus-sensitive: their domain is restricted by a set of alternatives, which may be provided by focus. I claim that, unlike other focus-sensitive operators, generics may, but do not have to, associate with focus. When alternatives are induced, either by focus or by other means, generics get their usual quasi-universal readings. But when no alternatives are introduced, quasi-existential readings result.
I argue that generics, unlike adverbs of quantification, do not introduce
tripartite structures directly, but are initially interpreted as cases
of direct kind predication. Only when this interpretation fails to make
sense, the generic quantifier is accommodated, and tripartite structures
result. This two-level interpretation has the effect that while adverbs
of quantification require focus to determine which elements go to the restrictor
and which to the nuclear scope, and hence must associate with focus, generics
do not, and hence allow no association with focus and quasi-existential