The Comparative Fallacy: An Analysis and Discussion of Selected Research

Ever since the introduction of the “independent grammars assumption,” whereby a child is said to have its own grammar and not just an imitation of adult language, first language (L1) acquisition researchers such as Martin Braine (1963) and others have constructed grammars for children’s languages rather than treat them as defective adult grammars. This is based on the view that language learners create grammars of their own, rather than master the target grammar imperfectly. Thus, at any moment, a child’s grammar is an independent system. In second language acquisition (SLA) research this notion has led to several slightly variant conclusions. On the one hand, second language (L2) learners should be treated as already having a language of their own, not just as poor beginning learners of the target language, and should be encouraged to create a system from experience and internal resources rather than have the target language thrust down their throats by teachers (Johnson & Johnson, 1998). In reality, however, as Cook (1999) asserts, “SLA research has often fallen into the comparative fallacy (BleyVroman, 1983) of relating the L2 learner to the native speaker” (p. 189). In other words, the learner’s language is typically explained by reference to the target language system, rather than treated as an independent set of rules and performance characteristics. The aim of this paper is to examine two empirical studies in SLA from the perspective of the comparative fallacy: the studies of Dulay and Burt (1974) and Mendizabal (2001), both of which address L2 learners’ morpheme acquisition order. The first study investigated whether L1 transfer plays a significant role in determining this order. The second study examined whether L1 and L2 acquisition orders coincide. Both studies will be reviewed and analyzed with particular attention paid to the issues of measurement and multicompetent language use in connection with the comparative fallacy. In this regard, the implications of acknowledging the distinctiveness of multicompetent language use for both empirical research and L2 teaching practice will be discussed.