The Blindness Hypothesis and Individual Level Predicates

Predicates such as ‘to be tall’, ‘to be related to Chomsky’ or ‘to know Latin’, which intuitively denote permanent properties, are called individual-level predicates (henceforth: i-predicates); predicates such as ‘to be available’, ‘to talk to Chomsky’ or ‘to study Latin’, which denote properties not necessarily permanent, are called stage-level predicates (henceforth: s-predicates). An impressive list of grammatical facts have been pointed out that set the two classes of predicates apart. The overall picture is that s-predicates can do many more things than i-predicates can do: there are constructions where s-predicates are fine but i-predicates are not (e.g. temporal and locative modification, ‘there’-construction, perception sentences, etc.) and there are readings which are available with s-predicates but absent with i-predicates (e.g. the existential reading for their bare plural subjects, the episodic reading, etc.); some examples will be considered below. A theory of i-predicates should account for why i-predicates cannot do the many things that s-predicates can do. In section 2, I’ll sketch (the beginning of) a theory of i-predicates based on scalar implicatures. For instance, I’ll suggest that the sentence ‘?John is tall after dinner’ sounds odd because it triggers the implicature that John is not tall also in non-after-dinner times, which mismatches with the common knowledge that tallness is a permanent property (see Percus (1997) and Maienborn (2004) among others, for a similar proposal). The main goal of section 2 is to show that this simple idea extends also to less obvious cases, such as the restrictions on the readings of bare plural subjects of i-predicates. In order for the account to work, the following assumption is needed: that the algorithm for the computation of scalar implicatures is a purely logical algorithm, blind to common knowledge (the Blindness Hypothesis). Section 1 introduces this and some other background assumptions.