| Most studies of adult second language (L2) learning have focused on the transfer of first language (L1) by comparing various L1 speakers learning the same L2. No studies examined the possibility of bidirectional transfer, that is, transfer which might occur when native speakers of two different L1s learn each other's language as their L2's. The present study examined the linguistic productions of native speakers (NSs) of English learning Korean and NSs of Korean learning English for evidence of bidirectional transfer.;The study was framed within the theoretical perspective of typological markedness which predicts that less marked L1 structures are transferred to the L2 if the L2 structures are more marked than the corresponding L1 structures in the early stages of L2 development. Given the typological distinctions between Korean and English, six hypotheses were formulated to test L1 transfer with regard to four linguistic categories (i.e., word order, NP modification, pro-drop, and topic-comment structures). These were: (1) Learners' production of basic word order, SVO or SOV in the L2, would not be influenced by their L1, (2) L2 NP modification would be differentially influenced by the learner's L1 head direction, i.e., pre/simple-modification patterns would be produced in L2 target-like forms, (3) while post/complex-modification would not, (4) pro-drop would appear in ILs of Korean and English when the referent was assumed to be known, (5) pre-syntactic modes of topic-comment structure would appear in both Korean and English regardless of the learner's L1s, and (6) as NSs of a topic-prominent language. Korean NSs would not produce the subject-creating constructions of English.;Results showed that the word order hypothesis as confirmed; L1 influence did not appear in either group. The NP modification hypothesis was also confirmed; L1 influence appeared to be bidirectional. The pro-drop hypothesis was partially confirmed; English speakers of Korean employed pro-drop while Korean speakers of English did not. Thus, L1 influence seemed to be unrelated to pro-drop. The hypotheses regarding the topic-comment structures were not confirmed. Contrary to the predictions, both groups of informants produced subject-predicate structures. |
