| |
Abstract:
Miyake et al. (1994) suggest that activation of relevant
information is the mechanism underlying working memory (WM) in
language processing whereas Gernsbacher and Faust (1991) claim it
is the inhibition of irrelevant information. This issue was
investigated by measuring ERPs during the reading of sentences
containing an ambiguous word (i.e. 'Ton'). High and low span
subjects read sentences containing an ambiguous word at the second
position, a nominal pre-disambiguation at the 5th position and a
final disambiguation using a verb at the 6th position. Four types
of experimental sentences were used: 1. Der Ton wurde vom Saenger
gesungen. (The tone was by the singer sung.) dominant
predisambiguation, dominant disambiguation 2. Der Ton wurde vom
Toepfer gebrannt. (The clay was by the potter glazed.) subordinated
predisambiguation, subordinated disambiguation 3. Der Ton wurde vom
Saenger gebrannt. (The clay was by the singer glazed.) dominant
predisambiguation, subordinated disambiguation 4. Der Ton wurde vom
Toepfer gesungen. (The tone was by the potter sung.) subordinated
predisambiguation, dominant disambiguation. There was a larger N400
for the subordinated predisambigiation noun in high span readers
indicating that they inhibited the secondary meaning to a larger
extent than low span readers. The ERP-effect on the final
disambiguation showed that high span readers suppressed the
irrelevant meaning very quickly whereas low span readers did not.
The results therefore seem to support the inhibition model.
|
