Second language acquisition of the dative alternation in English and Arabic : a bidirectional study

This thesis reports on an empirical investigation of native-language (L1) influence on the acquisition of second-language (L2) argument structures in which the L1 argument structures are a superset of those allowed in the L2, and vice versa. To do so, the dative alternation was adopted as a linguistic phenomenon. English allows all verbs in the Give class, Tell class and Throw class to occur in both the Prepositional Dative (PD) construction and the Double Object Dative (DOD) construction. In contrast, only some verbs in the Give class such as ‘give’ and ‘sell’ and the Tell class such as ‘tell’ and ‘show’ are allowed to appear in the DOD construction in Arabic. On the other hand, Scrambling Dative (SD) constructions are allowed in Arabic whereas they are not allowed in English. Two empirical studies investigated three questions: 1) to what extent can L2 learners realise the grammaticality of structures that are not allowed in their L1? 2) To what extent are they able to perceive the ungrammaticality of certain structures in their L2? 3) Which dative structure is acquired earlier? The first investigation was the L2 English study which explored the acquisition of the English dative alternation by native speakers of Arabic. The second investigation was the L2 Arabic study which explored the acquisition of the Arabic dative alternation by native speakers of English. The data were analysed according to four hypotheses: the Fundamental Difference Hypothesis, the Representational Deficit Hypothesis, the Subset-Superset Hypothesis, the Full Transfer and Full Access approach and the Feature Reassembly Hypothesis. The results of the L2 English study generally revealed that Arab leaners of English could not acquire what is absent in their L1 and they generally unlearn the structures that are not allowed in the L2. The results of the L2 Arabic study generally showed that English learners of Arabic could not recognise the ungrammaticality of some Arabic structures. However, they could acquire the SD structures. Overall, the bidirectional results give support to the Subset-Superset Hypothesis and the Full Transfer and Full Access approaches as L2 learners initially transfer their L1 grammar and only gradually restructure themselves and arrive at the L2 grammar, once effective positive evidence is provided.