Abstract

One of the major findings in previous research on reflexive coreference in a second language is that learners of English violate the locality requirement of the reflexive form x-self more readily when it appears in the subordinate infinitival clause than when included in the subordinate tensed clause (a phenomenon called the “tensed-infinitive asymmetry”). This paper argues that the cause of this asymmetry is the difference in the factivity of the event/situation to be configured into a particular construction, rather than the difference in the syntactic type of the sentence per se. In an experiment that was designed to test the validity of this postulation, the learners tended to respond differently to the different degrees of factivity presumed in individual sentences, in line with the prediction. Thus, it seems that the learners based their interpretation on the semantics of the sentence. The theoretical as well as pedagogical implications of these findings are discussed.