| |
Abstract:
Category ambiguities (e.g., bark) have different meanings
that are associated to a different form class. The selection of the
contextually appropriate meaning of these category ambiguities has
been proposed to be dependent upon both the lexical frequency of
the alternative meanings of the ambiguities and upon contextual
constraint. To directly assess the relative influence of contextual
constraint and meaning frequency on category ambiguity resolution
we asked 12 subjects to listen to naturally produced sentences in
four context conditions, followed by the same target word. Sentence
contexts were semantically neutral but syntactically constraining.
(Note: Target words are in capitals: Concordant/Concordant Control:
The blind man ran his hands across the bark/silk TREE;
Discordant/Discordant Control: Without apparent reason they started
to bark/shiver TREE.) The selectional status of the ambiguous words
was directly inferred from the amplitude of ERPs to the sentence
final words. Results indicated that both meanings of ambiguous
words were activated; the ERPs to ambiguous words were more
negative than to the non-ambiguous controls. In addition, ERPs to
ambiguous words were more negative when the local syntactic
structure of the sentence biased toward the less frequent meaning
of the ambiguity than when it biased toward the more frequent
meaning. These results indicate that syntactic structure alone
cannot guide meaning selection, and that lexical frequency
influences relative activation of category ambiguities.
|
