| |
Abstract:
Neurophysiological and neuroimaging results show that
language processing is supported by various brain areas which carry
out separate working memory or processing functions. To investigate
the sources of working memory used by normals during sentence
processing, we manipulated sentence complexity (object vs. subject
relative) and phonological complexity (similar vs. dissimilar).
these factors have been claimed to burden working memory. Two
memory spans were collected which have been claimed to be
differentially sensitive to separable working memories. Analyses of
RTs and errors in a self paced grammaticality judgement task showed
no signigficant interaction of phonological and sentential
complexity. Sentence complexity interacted wity a factor high vs.
low span, based on the Daneman & Carpenter test, in the error
analysis only. Non-word span, which presumable reflects the
capacity of the phonological store, showed no correlation with
sentence complexity although there was a main effect of non-word
span on the error rates (high spanners performed worse than low
spanners). Overall, these results are more compatible with the
assumption of separable working memory resources.
|
