Second-language acquisition is subject to great variation in terms both of
ultimate attainment and of the developmental process itself. This paper exp
lores the possibility that at least part of this variation is due to a lack
stability in learners' access to UG. I report here on a "limited-scope" lo
ngitudinal study that examines this matter. The study focused on a UG princ
iple that prevents overt pronouns from having quantified NPs as antecedents
in languages (such as Japanese) that have null pronouns.
Data was collected from a group of 29 English-speaking adults learning Japa
nese as a second language at two different times. For both sessions, the L2
learners as a group exhibited a statistically significant difference in th
eir interpretation of null and overt pronominals when there was a quantifie
d antecedent. These results clearly suggest access of UG. However, when per
formance of individual L2 learners is examined, one can discern a striking
inconsistency across the two sessions. I propose that this inconsistency re
flects and instability in the relevant UG principle that results from its r
elatively infrequent use in the course of second-language acquisition.