Anti-transfer effects in third language acquisition

Abstract

Recent work has established that reverse transfer from L3 to L2 can occur, resulting in transfer in domains such as tense-aspect (Cheung, Matthews and Tsang 2011). A related possibility which has been less widely discussed is that L1 effects on L2 can be modulated by subsequent acquisition of an L3. Studies in the Hong Kong context suggest that this is a real possibility. L3 learners at universities who have been exposed to Cantonese and English from early on show influence of Chinese in their L2 English. For example, Hui (2010) found that speakers who learnt French as L3 were less prone to produce ill-formed subject relative clauses as in (1) in their English. (1) *The one stood outside 10B has a ponytail. Also prominent among the interlanguage features of Chinese learners are morphological errors in English, such as use of uninflected verbs referring to the past (2): (2) Last week I see a movie. In a written production task, L3 learners of German were found to produce errors of this type at approximately half the rate of bilingual controls without knowledge of German. We attribute this finding to the systematic tense inflection in German sensitizing learners to the need for morphological marking of tense. These are not merely cases of reverse positive transfer, since the effect apparently involves expunging errors resulting from L1 transfer. We therefore suggest the term anti-transfer effects for this phenomenon. We discuss how anti-transfer effects may be further investigated and their potential importance for applied linguistics.