2ND LANGUAGE PROCESSING OF SUBORDINATE CLAUSES AND ANAPHORA - 1ST LANGUAGE AND UNIVERSAL INFLUENCES - A REVIEW OF FLYNN RESEARCH

Flynn has reported that the second‐language processing of subordinate clauses and of anaphora is affected by the basic word order (head direction) of the learner's native language. This phenomenon is claimed to be a prediction of Universal Grammar and of Flynn's formulation of a parameter‐setting model of second‐language acquisition. This paper explores both the theoretical status of the prediction and the factual status of the phenomenon. The principal conclusions are that the prediction cannot derive from any current version of the theory of Universal Grammar, and that there is no evidence in Flynn's research for a processing difference based upon L1 grammatical contrasts. The clearest results are subject to more plausible interpretations, involving general considerations of online sentence processing that are prerequisite to other psycholinguistic explanations. The L1‐based differences reported by Flynn are evidently artifacts of experimental method and inappropriate analysis, especially of the incorrect use of analysis of covariance in an attempt to correct for a mismatch between experimental groups. Flynn's version of a head direction parameter linked to anaphora interpretation is thus without foundation in linguistic theory and, furthermore, is founded upon empirical observations of doubtful applicability and replicability. However, Flynn's research is still of considerable importance, for three reasons. First, the study of pronoun processing ease in language acquisition is of intrinsic psycholinguistic interest. Whereas a fair amount of information about anaphora in child language acquisition exists, there is little in adult language acquisition. Second, Flynn's research is of methodological interest, especially with regard to the techniques used to deal with the crucial problem of comparing groups with different L1s and at different proficiency levels. Third, Flynn's claimed findings are relevant to the general issue of the effect of L1/L2 differences and of universal processing strategies in second‐language acquisition. Thus, despite our general reservations as to the theoretical validity of this research, we will attempt to review the major claims and findings, adhering as much as possible to Flynn's own formulations. © 1990 Language Learning Research Club, University of Michigan