| |
Abstract:
The mental representation of lexical concepts is of
fundamental importance for theories of language representation
and processing. That is because lexical concepts are said to be
the building blocks of language comprehension, thus the elements
upon which we construct mental representations of utterances.
During the last three decades, there have been many proposals for
the representation of lexical concepts and virtually all of them
posit that lexical concepts are decomposed in deeper levels of
syntactic representation (Levin and Rappaport Hovav, 1994).
In this paper we challenge the lexical-decompositional view in
light of experiments that suggest that causative lexical concepts
do not decompose during language comprehension. The experiments
relied on two contrasts between four types of constructions.
First, we compared verbs that take infinitival complements (e.g.
expect) with verbs that take direct objects (persuade); second,
we compared Causatives (e.g., kill) with "simple" Transitives
(e.g., kiss). The assumption was that there is a "shift" between
superficial and deep syntactic relations of subject-verb-object
triads in the Expect class (NP raising) but not in the Persuade
class; the same shift, by hypothesis, would occur in the
Causative class (if lexical causatives decompose), but not in the
Transitive class (Fodor et al., 1980).
The experiments employed three types of priming techniques.
The first set employed the Masked Priming of a Word Probe
technique (MPWP; de Almeida, 1998), in which priming effects are
taken to reveal the structural distance (relatedness) between
verb and noun object in the underlying representation of the
sentences. In the Expect vs. Persuade comparison, RT's to probes
were faster in the Expect class than in the Persuade class. In
the Causative vs. Transitive comparison, however, there was no
significant difference between RT's to probes in the two types of
constructions. In a second series of experiments, we employed a
"classical" masked priming task (Forster et al., 1987) in which
only masked primes (verbs) and targets (nouns) were presented for
lexical decision. Again, there was a significant difference
between classes of verbs and their argument-taking properties in
the Expect vs. Persuade case, but not in the Causative vs.
Transitive case. With a third technique, a cross-modal version of
MPWP (with sentences presented aurally), results pointed in the
same direction as in the other experiments.
In conclusion, we defend an atomic theory for the
representation of lexical concepts (Fodor, in press). The
implications for cognitive science and for psycholinguistics are
quite straightforward: If lexical concepts do not decompose, they
are elementary mental representations -- thus, the very elements
upon which we build representations of linguistic
expressions.
de Almeida, R. G. (1998). The Representation of Lexical
Concepts: A Psycholinguistic Inquiry. PhD Dissertation, Rutgers
University.
Fodor, J. A. (in press).
Concepts: Where Cognitive Science Went Wrong.
Oxford: Oxford University Press.
Fodor, J. A., Garrett, M. F., Walker, E. C. T., and Parkes, C.
H. (1980). "Against definitions."
Cognition,
8, 263--367.
Forster, K. I., Davis, C., Schoknecht, C., and Carter, R.
(1987). "Masked priming with graphemically related forms:
Repetition or partial activation?"
The Quarterly Journal of Experimental Psychology,
39A, 211--251.
Levin, B., and Rappaport Hovav, M. (1994) "A preliminary
analysis of causative verbs in English."
Lingua,
92, 35--67.
|
