| |
Abstract:
Studies in sentence comprehension (e.g. Pearlmutter, Garnsey,
& Bock 1999; Nicol, Forster, & Veres, 1997) have shown that
subject-verb agreement is an early and integral component of
language comprehension. Our studies address the issue of whether
the mechanisms of language comprehension manipulate number
representations independently of the individual lexical items on
which number information is encoded.
Our experiments tested agreement comprehension using a
cross-modal lexical decision task: Participants listened to
preambles like "Whenever there's a jailbreak, the trooper with the
bloodhound(s)". At the offset of each preamble, a letter string was
presented visually for lexical decision. In the critical trials,
the letter string was a verb that was morphosyntactically
compatible with the subject NP in the preamble (e.g. "captures").
After the lexical decision was made, the participants listened to
the remainder of the sentence ("the fugitive"). On about half of
the trials, participants were required to answer follow-up
questions about whether the sentence made sense.
It has been shown that decisions to verbal material are
sensitive to the verb's grammatical compatibility with the global
syntactic environment (Tyler & Marslen-Wilson, 1977). Our first
experiment was designed to test whether this particular cross-modal
technique would also be sensitive to the morphosyntactic properties
of local nouns when the verb and head noun agree. If so, we should
observe a processing cost in the conditions where the head noun and
the local noun were mismatched in number. This is in fact what we
observed. These results establish that this experimental paradigm
is sensitive to the effects of locally distracting material on
agreement calculations for grammatical sentences.
Using the same experimental methodology, Experiment 2 was
designed to test the hypothesis that the interference that is
observed when a head and local noun differ in number arises because
of a purely syntactic agreement process that on occasion mistakenly
copies the number features of a local noun onto the subject NP. In
this account, the semantic properties of the local noun are
entirely irrelevant to the agreement process. In contrast, if the
calculation of agreement is _lexically_driven_ in the sense that
non-syntactic properties of the local nouns can influence the
outcome of the calculation, then potentially distracting local
nouns should be more likely to interfere with the agreement process
when their semantic properties make them plausible subjects. The
preambles in Experiment 2 were therefore modified so as to contain
local nouns whose semantic properties made them plausible vs.
implausible subjects of the verbs (e.g. "the girl with the
rabbit(s)/basket(s) hops"). Response times showed an overall effect
of head noun/local noun mismatch as well as an effect for the
semantic fit between the local noun and the main verb. These
results provided positive evidence that the calculation of
agreement can be affected by the semantic content of lexical
sources, and not merely by their morphosyntactic specification for
number.
Tyler, L.K., & Marslen-Wilson, W.D. (1977). The on-line
effects of semantic context on syntactic processing. Journal of
Verbal Learning and Verbal Behavior, 16, 683-692.
Nicol, J.L., Forster, K.I., & Veres, C. (1997). Subject-verb
agreement processes in English. Journal of Memory and Language, 36,
569-587.
Pearlmutter, N.J., Garnsey, S.M., & Bock, K. (1999). Agreement
processes in sentence comprehension. Journal of Memory and
Language, 41, 427-456.
|
