| |
Abstract:
This study concerns the way in which young children access
the meanings of words that possess multiple interpretations during
on-going sentence processing. Studies with adult populations, using
sensitive on-line measures of lexical access, have demonstrated
that, independently of the constraints of prior biasing sentential
contexts, multiple meanings of an ambiguity are initially accessed
immediately upon perception of the word, followed rapidly by
selection of a single contextually relevant meaning (Swinney, 1979;
Seidenberg et al, 1979; but see also Tabossi, 1988; but, then see
Swinney, 1990). A fundamental question for language theory is
whether this apparent "modular" (automatic, reflexive, contextually
independent) access process found in the adult has a developmental
history. Is it an acquired property of comprehension or one
developed via experience? The difficulty in answering this question
is a methodological one - finding a sensitive, independent, on-line
behavioral measure of children's processing that young children can
perform. In a previous attempt to examine this question with such a
task (cross-modal picture priming; CMPP), Swinney and Prather
(1989) found evidence suggesting exhaustive access of lexical
ambiguities (again, independent of the biasing sentential context)
in children as young as approximately 5 years of age; however,
claims re modularity of lexical access were indeterminate in
younger children in that study, and since 5-year olds have
considerable experience with language, this initial work failed to
provide a convincing answer to the fundamental question. In
addition, this initial work only examine lexical access at the
point immediately following an ambiguous word, and had no measure
of possible 'delayed' access of the less frequent meanings of
ambiguities, as has been argued for in the literature.
The present study pushes the 'age' envelope in the use of
behavioral on-line measures with young children, employing the CMPP
technique with children as young as age 3:10. In addition, the
study provides for critical 'down-stream' examination of possible
delayed access to less frequent interpretations of ambiguous words.
While listening to sentences containing a lexical ambiguity for
which multiple meanings were 'known' by the child (at least
'off-line', determined via pre-tests) children ages 3:10-5:11 made
sentence-independent 'edible/nonedible' decisions to pictures
(which were related to each meaning of an ambiguous word) presented
either immediately after hearing the ambiguous word in the sentence
(*1), or 1500 msec downstream from that point (*2). Using a similar
matched sentence design (so that pictures are 'related' in one
sentence, but 'controls' in another), subjects were presented with
sentences which were strongly biased towards the primary meaning of
the lexical ambiguities such as:
As can be seen in Table 1, our data support the conclusion that
when listening to sentences containing lexical ambiguities,
children as young as three years, ten months already demonstrate
the autonomous, exhaustive automatic lexical access reflex,
immediately upon hearing the ambiguous word. Note that, as for
adults, both the more frequent and less frequent meanings are
immediately available, contrary to claims in the literature. The
downstream test point demonstrates that context has chosen the
contextually related meaning (primary meaning) shortly after the
ambiguity is heard, for these young children.
|
