The two main approaches in the l iterature to the analysis of bare plural s , those of Carlson and of Kratzer/Diesing l are able to explain this observation, at least for bare plurals in subject position.2 Carlson 1 977, for example, treats I-level predicates as predicates of kinds, leading to the prediction that bare plural subj ects of I-level predicates wi l l receive only generic readings. Note that in the Carlson and KratzerlDiesing accounts, I-level predicates are identified, at least roughly, with permanent or near-permanent properties, although Carlson points out that this correspondence is not complete . A number of exceptions have been observed to the general ization that 1level predicates allow only generic readings for their bare plural subjects (see, for example, Fernald 1 994) . I t i s not very difficult to think of further exceptions . Consider the following:
[1]
Gregory Norman Carlson,et al.
Reference to kinds in English
,
1977
.
[2]
Katalin É. Kiss,et al.
On Generic and Existential Bare Plurals and the Classification of Predicates
,
1998
.
[3]
Sheila Glasbey.
Event structure in natural language discourse
,
1994
.
[4]
William A. Ladusaw.
Thetic and Categorical, Stage and Individual, Weak and Strong
,
1994
.
[5]
Robin Cooper.
Aspectual classes in situation semantics
,
1985
.
[6]
Angelika Kratzer,et al.
Stage-Level and Individual-Level Predicates
,
1995
.
[7]
J. Barwise,et al.
THE RIGHTS AND WRONGS OF NATURAL REGULARITY
,
1994
.
[8]
K. Jon Barwise,et al.
The situation in logic
,
1989,
CSLI lecture notes series.