Predicational 'BE'

Susan Rothstein
Bar-Ilan University
rothss@ashur.cc.biu.ac.il


I present a unified account of the meaning of predicational 'be' which explains its behaviour both in simple predicational sentences such as (1) and so-called 'agentive' contexts such as (2):

(1)

Mary is very clever.

(2)

(a) Mary is being very clever. (b) She made Mary be very clever.

The account rests on identifying a fundamental aspectual dichotomy between verbs and adjectives: verbs have a davidsonian eventuality argument and denote sets of atomic, countable eventuality elements. Adjectives have their denotation in the mass domain, and denote sets of non-atomic states. I argue that BE, like all verbs, introduces a davidsonian eventuality argument, but that, being a copula, it doesn't ascribe any property to that eventuality or determine any participants via theta-marking. Instead it has the role of mapping from the mass to the count domain: it takes as an argument a set of states, denoted by AP, and yields a set of eventualities which instantiate the set of states, in other words, eventualities in which the states hold. I represent the meaning of BE as in (3), where S is a variable over sets of states:

(3)

$S$e.Inst(e,S) ("$" = lambda)

Eventualities, including those denotated by complex verbs of the form BE + AP, have a number of properties which true states do not have, indicating the 'count' status of the former; in particular, they can be temporally located and counted.

I show that the semantic distinction between the denotations of AP and [be+ AP] allows us to explain the agentivity effects in (2) as a pragmatic effect resulting from the interaction of the small clause and progressive constructions respectively with the semantic properties of BE.

---

Last updated July 20, 1997 by rblight@mail.utexas.edu Return to main program