| |
Abstract:
According to two-stage theories (e.g., Frazier, 1979; 1987;
Rayner et al., 1983) processing difficulty arises when an initial
analysis, made on the basis of a restricted range of information,
has to be reanalysed. In contrast, constraint-based theories claim
that processing difficulty arises from a process of competition
when two or more analyses are about equally activated by multiple
constraints (MacDonald, 1994; McRae et al., 1998; Spivey-Knowlton
& Sedivy, 1995; Tabor et al., 1997). In recent papers (Traxler
et al., 1998; van Gompel et al., 1998), we proposed the
unrestricted race model, in which multiple constraints can be used
to make an initial analysis, but in which processing difficulty
arises due to reanalysis rather than competition.
Experiment 1 tested sentences such as (1a-c). As indicated by
plausibility pre-tests, (1a) is globally ambiguous, whereas (1b)
and (1c) are semantically disambiguated as low and high attachment
respectively.
(1a) The advisor of the mayor that had
been driven to the meeting had a lot of problems.
(1b) The village of the mayor that had
been driven to the meeting had a lot of problems.
(1c) The mayor of the village that had
been driven to the meeting had a lot of problems.
Reading times on the region
to the meeting
were longer in (1b-c) than (1a). These results are consistent with
our other findings (Traxler et al., 1998; van Gompel et al., 1998).
They provide support for the unrestricted race model, according to
which readers initially adopt each analysis half the time, and thus
have to reanalyse half the time in both (1b) and (1c). The results
are inconsistent with constraint-based theories that predict
competition in (1a), but not in (1b-c). The garden path theory
predicts that low attachment is always initially adopted (due to
late closure) and therefore, reanalysis should only have occurred
in (1c).
In Experiment 1 the processor may have dropped one analysis in
the ambiguous region
(that had been),
before the disambiguation
(driven)
was reached. Hence, there may have been initial competition in the
ambiguous region, but when the disambiguation was reached, only one
analysis was available. To test this possibility, we removed the
ambiguous region in Experiment 2. The results were similar to
Experiment 1: reading times on
to the meeting
were longer in (1b-c) than (1a). Once more, this supports the
unrestricted race model, but poses problems for constraint-based
theories.
References
Frazier, L. (1979).
On Comprehending Sentences: Syntactic Parsing Strategies.
Bloomington, IN Indiana University Linguistics Club.
Frazier, L. (1987). Sentence Processing: A Tutorial Review. In M.
Coltheart (Ed.),
The Psychology of Reading, Attention and Performance XII,
p. 559-586. Hillsdale, NJ: Erlbaum.
MacDonald, M. C. (1994). Probabilistic constraints and syntactic
ambiguity resolution.
Language and Cognitive Processes,
9, 157-201.
McRae, K., Spivey-Knowlton, M. J., & Tanenhaus, M. K. (1998).
Modeling the influence of thematic fit (and other constraints) in
on-line sentence comprehension.
Journal of Memory and Language,
38, 283- 312.
Rayner, K., Carlson, M., & Frazier, L. (1983). The interaction
of syntax and semantics during sentence processing: Eye movements
in the analysis of semantically biased sentences.
Journal of Verbal Learning and Verbal Behavior,
22, 358-374.
Spivey-Knowlton, M. J., & Sedivy, J. C. (1995). Resolving
attachment ambiguities with multiple constraints.
Cognition,
55, 227-267.
Tabor, W., Juliano, C., & Tanenhaus, M. K. (1997). Parsing in
a dynamical system: An attractor-based account of the interaction
of lexical and structural constraints in sentence processing.
Language and Cognitive Processes,
12, 211-271.
|
