|
Abstract:
Can age differences in the comprehension of grammatically
complex sentences be explained by age differences in working
memory? Old adults were compared with young adults who read
sentences with and without additional working memory load
(remembering a 3 digit number). Additionally, old adults were
compared with young adults who read sentences with reduced
presentation time.
We presented sentences comprising a main clause (MC) and a
center-embedded relative clause (RC) in a RSVP format and asked
questions about who did something or to whom something was done.
The sentences varied in five orthogonal dimensions: (a) subject-
vs. object-initial MC, (b) subject- vs. object-initial RC, (c)
early vs. late RC position, (d) nouns with 1 vs. 2 thematic roles,
and (e) MC vs. RC questions. Since the sentences varied in
difficulty, one would expect age by complexity interactions in all
critical conditions according to span
x
difficulty interactions reported for example by Just and Carpenter
(1992).
Although there was a substantial age-related main effect (16%
difference in overall accuracy), only one interaction was
significant: The age difference was more pronounced for
object-initial than for subject-initial MCs.
Age simulation with additional memory load:
Young adults read the center-embedded relative clauses while
remembering a 3 digit number. This manipulation reduced their
overall accuracy to the old adults' level, but old adults were
significantly worse both in object-initial MCs (as in Experiment 1)
and in object-initial RCs. The results show that working memory
load for young adults can be used to equate overall accuracy but
does not lead to a processing pattern identical to that observed
for old adults.
Age simulation with manipulation of presentation time:
In this experiment we used the same sentences but reduced
presentation time for young adults from 750 ms to 312 ms per word.
Again young adults' overall accuracy was reduced to that of the old
adults. The interaction of age by MC-initiality mentioned above
disappeared and no new interactions with age were induced. Thus,
age simulation with reduced presentation time was successful.
The data suggest that processing speed rather than working
memory capacity is responsible for old adults' difficulties with
center-embedded relative clauses.
References
Just, M. A., & Carpenter, P.A. (1992). A capacity theory of
comprehension: Individual differences in working memory.
Psychological Review,
99, 122-149.
Kliegl, R., Fanselow, G., Schlesewsky, M., & Oberauer, K. (in
prep.). Age simulation of syntactic complexity effects with
manipulations of presentation rate.
|
