Constructional Transfer in L2 Learning: The Role of Entrenchment

Abstract

The interlanguage of L2 learners is shaped by various factors, perhaps most noticeably transfer from the L1 to the L2 and overgeneralization in the L2. Transfer has traditionally been treated as an all-or-nothing phenomenon by focusing on L2 structures that either have or do not have an equivalent in the L1. Only recently, researchers working in usage-based construction grammar have taken into account frequencies and association strengths of L2 structures. These studies suggest that entrenchment plays a similar role in L2 learning as it does in L1 acquisition. This paper extends this research by systematically taking into account the potential role of the entrenchment of L1 structures and its interaction with L2 entrenchment. We report a translation study and two acceptability rating studies that investigate these factors using the case of the ditransitive construction in the interlanguage of German learners of English as a Foreign Language. Ditransitive constructions exist in both languages but the German construction has a broader meaning (it can encode almost any type of three-participant event) than the English one (which is restricted to events of caused reception and has a large number of idiosyncratic restrictions on individual verbs). This semantic difference leads to differences in the collostructional preferences of the constructions in the two languages (i.e., the set of verbs that they are associated with positively or negatively), allowing us to investigate the extent to which the transfer of L1 usage patterns to the L2 depends on the entrenchment of verb-construction combinations in the L1, and the extent to which transfer can be inhibited by the positive or negative entrenchment of verb-construction associations in the L2. Our results suggest an influence of both L1 entrenchment and L2 entrenchment (positive and negative) on linguistic production and linguistic representation respectively, such that learners transfer L1 associations between verbs and constructions to their L2, especially if these L1 associations are strong, and that they seem to be sensitive to negative L2 entrenchment, especially if the corresponding L1 associations are only weakly entrenched. However, differences between the translation study and the judgment studies suggest that learners may not acquire the specific semantic restrictions of the English ditransitive at all, relying on verb-specific associations to particular argument structures in the L2 instead and resorting to their L1 knowledge when such verb-specific associations are not available to them.