| |
Abstract:
We recorded event-related brain potentials (ERPs) while
participants read sentences that had either high-frequency main
verbs (talk, run), or low-frequency main verbs (brag, swim), which
were either correctly inflected (will swim), or incorrectly
inflected (*will swam). We found a large, robust main N400 effect
for verb frequency and a large, robust main P600 effect for
grammaticality, with no interaction between lexical frequency and
grammaticality. For example, ungrammatical low-frequency verbs
(e.g., *will bragged) elicited an N400 that was indistinct from the
N400 elicited by grammatical low-frequency verbs (will brag), and
they elicited a P600 that was indistinct from the P600 elicited by
high-frequency ungrammatical verbs (e.g., *will talked). Moreover,
this biphasic pattern was consistent across regular (talk, brag)
and irregular (run, swim) verb types. These results suggest that
the processing of syntactic feature values (e.g., [+past])
associated with low-frequency lexical hosts is not delayed, or
otherwise modulated, relative to the processing of syntactic
features associated with high-frequency hosts. This in turn
suggests that access to syntactic feature values during language
comprehension is not obligatorily mediated through the stored
lexical entries (e.g., lemma representations) of the word forms
that express those features superficially.
|
