| |
Abstract:
We investigated the role of short-term memory for a word's
phonological code in recovering from errors in meaning selection
during sentence comprehension. We induced errors by presenting
sentences in which biased ambiguous words were preceded by neutral
context and later disambiguated toward their less frequent meaning.
Under these circumstances, readers initially select the more
frequent (wrong ) meaning (Duffy, Morris, & Rayner, 1988) and
must then recover from this misinterpretation when they encounter
the disambiguating information. We hypothesized that these recovery
processes retrieve a short-term memory for the ambiguous word's
pronunciation (its phonological code) to re-access the word and
select the intended meaning. If this is the case, recovery
processes will be more difficult for heterophones because the
reader initially selects not only the wrong meaning but also the
wrong pronunciation for the word. As a result, the short-term
phonological code available to recovery processes is incorrect.
Thus we included both heterophones (two meanings and two
pronunciations) and homophones (two meanings and one pronunciation)
among our ambiguous target words. We also varied whether or not the
target word was in focus. Birch and Garnsey (1995) suggested that
focusing a word enhances memory for its phonological properties. If
this is the case, then placing the ambiguous word in focus could
affect the error recovery process.
Participants read sentences containing three types of target
words: biased homographic heterophones, biased homographic
homophones, and unambiguous control words. Words were arranged in
triples matched on length and frequency. In sentence contexts,
target words were preceded and immediately followed by neutral
context; disambiguating information instantiating the less frequent
meaning of the ambiguous targets appeared later in the sentence.
Additionally, target words were either in focus (using the
it-
cleft construction) or not. Readers read the sentences while their
eye movements were monitored. Sample stimuli are presented below;
target word and disambiguating region are underlined.
Initial results indicate that reading times are inflated in the
heterophone conditions compared with the other conditions in both
first pass reading times for the disambiguating region and second
pass measures. This suggests that the phonological code is involved
in error recovery. When words are in focus, readers spend more time
on first pass in the disambiguating region in all three conditions.
Second pass times elsewhere in the sentence were shorter when
targets were in focus and ambiguous, suggesting that focus serves
to reduce regressions and re-reading in error recovery.
References
Birch, S. L., & Garnsey, S. M. (1995). The effect of focus
on memory for words in sentences.
Journal of Memory and Language,
34, 232-267.
Duffy, S. A., Morris, R. K., & Rayner, K. (1988). Lexical
ambiguity and fixation times in reading.
Journal of Memory and Language,
27, 429-446.
|
