| |
Abstract:
Readers develop sentence-level representations of the
incoming string of words incrementally, as they recognize each
individual word. Temporary ambiguities, however, are frequent and
the language processing system must therefore have some way of
handling ambiguity. Look for example at sentence (1a).
1a. The actress cursed the director and the producer threw his
cigar on the floor.
Only when the word threw is encountered, it becomes clear that
the producer is not the object of cursed (as in 1b), but instead
the subject of a new, conjoined, sentence (1c).
1b. The actress cursed [the director and the producer] ...
1c. [The actress cursed the director] and [the producer threw his
cigar on the floor.]
According to Minimal Attachment and Construal theory (Frazier
& Clifton, 1996) this temporary ambiguity is dealt with by
choosing the simplest analysis in terms of syntactic structure.
Readers will thus prefer NP-coordination over S-coordination in a
sentence like (1a), and subsequently run into trouble upon
encountering the disambiguating verb threw. In a number of
experiments (self-paced reading and eye tracking) it has been shown
that this is indeed the case.
These studies, however, only looked at the processing of
sentences in isolation. In the present experiments we created a
context for target sentences like (1a) in which both actors (e.g.,
the actress as well as the producer) become the topics of a story.
See example (2).
2. Two-topic context:
When they saw the film as the director had edited it, the actress
and the producer were very angry.
Our results showed that there was no trace of processing
difficulty following such a context, while the effect was clearly
present in a control condition where neutral contexts were used. We
found this pattern of results in both off-line (sentence
completion) and on-line studies (self-paced reading and eye
tracking experiments).
These findings argue against the Minimal Attachment view on
parsing and favor context-sensitive models like, for example, the
Constraint-based model (MacDonald, Pearlmutter, & Seidenberg,
1994; Tanenhaus & Trueswell, 1995), the Concurrent Model
(Boland, 1997), and the Incremental Interactive model (Altmann
& Steedman, 1988).
References
Altmann, G. T. M., & Steedman, M. (1988). Interaction with
context during human sentence processing.
Cognition,
30, 191-238.
Boland, J. E. (1997). The relationship between syntactic and
semantic processes in sentence comprehension.
Language and Cognitive Processes,
12, 423-484.
Frazier, L., & Clifton, C., Jr. (1996).
Construal.
Cambridge, MA: MIT Press.
MacDonald, M. C., Pearlmutter, N. J., & Seidenberg, M. S.
(1994). Lexical nature of syntactic ambiguity resolution.
Psychological Review,
101, 676-703.
Tanenhaus, M. K., & Trueswell, J. C. (1995). Sentence
comprehension. In J. L. Miller & P. D. Eimas (Eds.),
Speech, Language and Communication,
pp. 217-262. San Diego, CA: Academic Press.
|
