The traditional term `predicate' has inherited a heterogeneous usage
in linguistic theory from its development in the history of logic and
the theory of propositions. The Aristotelian notion of proposition
divided the semantic content of a proposition into a subject term, a
predicate term, and the mode of predication, which related the
predicate to the subject. In the Fregean tradition, the desire to
find an essential property to distinguish predicate terms from
individuals led to a view of predicates as essentially semantically
`incomplete'. The incompleteness of a predicate
could come in a variety of valences, giving the standard notion of
logical predicates of varying arities. As a result, the Aristotelian
predicate became simply a 1-place predicate, the Aristotelian subject
became just the 'last argument into combination', and the binary mode
of predication was accorded no separate logical representation. Under
the view of predicates as 'propositional functions', mode of
predication was either lost as a useful concept or assimilated to
function argument application, the (partial) saturation of the
predicate's incompleteness.
We can discern this mixed notion of predication in the development of
the theory of the structural construal of predication within
generative grammar. One approach, exemplified by Williams 1980, is
Aristotelian in concentrating on a binary subject - predicate
distinction. Other approaches, exemplified by Rothstein 1983, are
more Fregean in tieing predication to $\theta$-assignment and hence to
varying arity of predicates.
The interpretation of indefinites as fundamentally one-place
predicates on individuals has made possible a number of proposals in
which the content of an argument phrase is considered "incorporated"
into the predicate by a composition operation other than normal
function application. (Cf. McNally 1992, 1995, van Geenhoven 1996)
In these analyses we see the contrast between restriction of a
predicate and true saturation of a predicate that I will suggest as
the basis for "specificity" effects.
I will then apply this view to the analysis of the two indefinite
articles in Maori as presented in Chung, Mason and Milroy 1995.
Finally, I revisit the discussion of the thetic/categorical judgement
contrast in Ladusaw 1994 and consider two alternative views of the
contrast: the one which considers it a fundamental difference in
proposition type and one which views it as a (deep) property of the
syntax-semantics interface and argue in favor of the latter.
