The present study is concerned with hole young and elderly subjects re
ad sentences which have to be immediately recalled. The task is a word
-by-word self-paced reading task followed by a verbatim recall task. T
he study aims to test the hypothesis that elderly subjects would show
a selective slowing of encoding processes (i.e., slotting more on cert
ain, parts of a sentence than on others) which would itself result fro
m a reduction of working memory resources (Stine, 1990). The study. al
so aims to test whether such a selective slowing might have an, adapta
tive value in, the sense that it could allow the elderly subjects to h
ave a recall performance similar to the young subjects' one. The data
shout a clear selective slowing of the reading process in the elderly;
in addition, this slotting seems to be partly determined bq the worki
ng memory of the subjects. The data show also that despite the slow do
wn, elderly subjects have poorer recall performance than younger subje
cts. The pattern of reading times is further interpreted as indicating
that the elderly subjects' specific problems arise from their trouble
both in identifying the syntactic phrases and in opening new semantic
structures.