| |
Abstract:
A central question in the study of language comprehension is
whether processing is carried out on-line as new input is received,
or are some aspects of processing delayed. Comprehending spoken
language, in particular, seems to present a trade-off between the
need to process input that rapidly becomes unavailable, and the
need to maintain available resources for the processing of
subsequent input. This trade-off seems to work differently in
sentence and in discourse processing. In sentence processing, most
researchers agree that syntactic information is processed
immediately, even if they disagree on the nature of this
processing. In discourse processing, however, many researchers
subscribe to the minimalist view, which maintains that all but the
most automatic and effortless processing is delayed (McKoon &
Ratcliff, 1995). Although this distinction between the processing
of syntactic structure and discourse inferences may seem clear cut,
there are several cases in which syntactic and discourse processes
are hard to separate. One such case is the interaction between
syntactic structure and the informational structure of sentences.
For example, certain syntactic constructions, such as clefts, can
affect discourse focus, which plays an important role in the
processing of anaphors.
The research reported here examined the time course of
syntactically induced focus effects in spoken language
comprehension by comparing the accessibility of referents following
two kinds of clefts: it-clefts (e.g., It was the student that
rented the car) and wh-clefts (e.g., What the student rented was
the car). The two kinds of clefts have opposite focusing effects
(the it-cleft focuses the referent the student and the wh-cleft
focuses the referent the car). This allows an assessment of focus
effects that is not confounded by word order effects (Almor, 1999).
Five experiments employed a cross-modal lexical decision task to
investigate how and when syntactic focus affects referent
accessibility. The results of these experiments indicated that
focus increased referent accessibility under one of two
conditions:
(1) When there was a subsequent definite anaphoric reference
(e.g., the lexical decision CAR was responded to faster after
subjects heard What the student rented was the car. The vehicle _
than when they heard It was the student that rented the car. The
vehicle _).
(2) When there was no anaphoric reference, but only following a
significant delay between the offset of the cleft and the
presentation of the visual target (SOA > 1000 ms).
When the delay was shorter than 1000 ms and there was no
anaphoric reference, syntactic focusing had no effect on referent
accessibility.Because changes in the accessibility of referents
seem to be the by-product of other processes, our findings likely
reflect the processing of the informational structure of sentences.
Specifically, our findings indicate that this processing may be
delayed. More generally, this study shows that in spoken language
comprehension, certain aspects of processing sentence structure can
be delayed until required by further processing (e.g., the
resolution of anaphoric reference), or until more resources become
available (e.g., when no further input is received for a long
time). We discuss the theoretical implications of these findings
under both a minimalist view of language processing, and a more
contextual view, by which the cost of various computations can vary
in different circumstances thus allowing different processing
strategies depending on the context.
Almor, A. (1999). Noun-phrase anaphora and focus: The
informational load hypothesis. Psychological Review, 106(4),
748-765.
McKoon, G., & Ratcliff, R. (1995). The minimalist hypothesis:
Directions for research. In C. A. Weaver, III (Ed.), Discourse
comprehension: Essays in honor of Walter Kintsch (pp. 97-116).
Hillsdale, NJ: Lawrence Erlbaum Associates Inc.
|
