Interlanguage as a projection of abstract lexical structure from L1 and TL Testing the Composite Matrix Language model of second-language acquisition

Abstract

This thesis investigates the Composite Matrix Language model of second-language acquisition (Jake, 1998) which predicts what structures may arise in the construction of interlanguage independently of the specific language pair involved. The model views second-language acquisition as an instance of language contact where the native and the target language interact at an abstract level of lexical structure preceding surface-level projections. The abstract lexical structure of lexical items is stored in the mental lexicon and includes information on three levels, namely, 1) lexical-conceptual structure, 2) predicate-argument structure, and 3) morphological-realization patterns.
The Composite Matrix Language model argues that in interlanguage construction, abstract lexical structure from the native and the target language is split up and recombined, resulting in a composite morphosyntactic frame projecting surface structure. Based on its psycholinguistic role in interlanguage construction, the model constrains the possible contribution of the native language. This thesis tests whether the predictions of the model can account for interlanguage structures in English learners of French, i.e., whether these can be explained as projections of composite lexical structure from the native and the target language.
For this purpose, spoken interlanguage data from ten British sixth-form students in their sixth year of learning French (Newcastle Corpus, Myles & Mitchell, 2013b, available via the French Learner Language Oral Corpora database, Myles & Mitchell, 2013a) are analyzed.
The results of the analysis largely confirm the predictions stated under the Composite Matrix Language model. Ambiguous and problematic structures are discussed with regard to their implications for the predictions of the model. In general, the thesis provides evidence for the universal principles of language contact assumed under the Composite Matrix Language model.