home
events
contacts
mailing list
directions
Linguistics
Department
Stanford
University |
Stanford
Humanities Center
Mellon
Foundation
Graduate
Research Workshop Program
Stanford
Semantics and Pragmatics Workshop:
THE
CONSTRUCTION OF MEANING
Tuesday, November 11, 12noon in 460-126:
Some foundational issues for a Construction
Grammar. Mutual definition and cluster concepts.
Arnold Zwicky (Stanford)
The central problem of syntax is how the composition of an
expression is related to its
distribution. All the conceptual apparatus of syntax arises from
trying to describe how the internal syntax and the external syntax of
expressions fit together. Internally, expressions are constructed
around central words (heads, in Bloomfield's sense); the category of an
expression is the category of its head word. Externally, the
distribution of an expression is a matter of the syntactic functions it
can serve.
In this picture, the basic concepts of syntax are two kinds of
categories: classes of words (L-categories) and classes of expressions
(F-categories, the extensions of syntactic functions), which are
connected to one another via syntactic rules. Customarily, the
inventories of L-categories and F-categories are taken to be
universals, and also to be associated with prototypical meanings--for
L-categories, with prototypical denotations; for F-categories, with
prototypical semantic or pragmatic roles. So, members of the
L-category N prototypically denote objects and serve as the heads of
expressions that are, among other things, eligible to serve as members
of the F-category Subject, prototypically associated with the Agent
semantic role and the Topic pragmatic role.
A constructional approach to syntax allows for a rather different
view of L-categories and F-categories, in which they are not given
universally or associated with meaning directly, but instead stand in a
relationship of language-particular mutual definition with each other
and with constructions themselves, and receive their meanings by virtue
of the semantics and pragmatics associated with particular
constructions. Each construction has a number of "slots" associated
with it, some of which must be filled by words and some by phrases
(expressions not limited to words); for each word slot, there is then
an L-category, and for each phrase slot an F-category. (See Croft
(2001) for a version of this approach.)
The obvious objection is that the number of L-categories and
F-categories, so defined, is just enormous. For instance, what counts
as a Subject in inverted clauses in English is not quite the same as
what counts as a Subject in uninverted clauses; these are then,
literally, different F-categories.
To show how this difficulty can be resolved, I consider paradigm
classes in languages
with rich inflectional systems. If we define a paradigm class in terms
of the applicable
morphological rules, then Latin has an enormous number of noun
declensions. However, there is a very high degree of association
between the applicability of different morphological rules, with the
result that there is a small number of "rule clusters" (the major
declension classes), with most nouns falling in one of them.
So it is with L-categories and F-categories. There is a high
degree of clustering here. Yes, there are words with mixed or
intermediate properties, and words with exceptional or idiosyncratic
properties, but most words belong in one of a small number of clusters,
which are the major parts of speech for the language in question.
Please contact one of the workshop organizers
if you have suggestions for presentations or the workshop in general.
Back to the workshop homepage.
This workshop is sponsored by
the Stanford Humanities Center, and funded by a grant from the Mellon
Foundation.
|