A Reexamination Of Bilingual Mental Lexicon And Its Implications For L2 Word Meaning Acquisition

It is widely recognized that second language (L2) vocabulary acquisition has a decisive effect on the development of second language acquisition (SLA), and thus, researches on L2 vocabulary acquisition has received more attention than ever before. However, as pointed out by some researchers (Jiang 2000; Meara 2002), studies on L2 vocabulary acquisition are largely descriptive and model-free, and those researches investigating the cognitive mechanism of its acquisition, representation and processing are scanty, with little focus on modeling bilingual mental lexicon. The preexisting bilingual mental lexicon theories and models are still primitive, with the restricted research methodology. Most of the recent researches are based on the revised hierarchical model of bilingual memory proposed by Kroll & Stewart (1994). Some assumptions in this model, like the hypothesis of independent representation of formal information in L1 and L2, and common representation of conceptual knowledge, have been confirmed and supported by other researchers. Nevertheless, the fundamental issues, like the essence of semantic representation and the relationship between the semantic representation and conceptual representation have not been made clear, either in the revised hierarchical model or in many other recent researches, in which there is still much controversy on the most fundamental issue, i.e. the overall representation pattern of L2 word knowledge. The research methodology, restricted to priming paradigm, takes the response time as the basis of bilingual mental lexicon modeling; the materials in experiments are largely common monosemous concrete nouns. Thus, it is still to be testified whether models based on the experimental conditions can apply to the general L2 learning setting, or other types of words, like abstract words, polysemous words, etc. Based on a comprehensive analysis of the present bilingual mental lexicon models and research methodology, and extended from the notion of coactivation pattern of conceptual features in distributed representation model, an extended distributed representation model of L2 words is proposed. The L2 words are assumed to be represented as the coactivation patterns of the formal representations of L2 words, their semantic representations, and the formal representations of their L1 translations. Both semantic memory and nonsemantic memory, i.e. episodic memory, are involved in L2 word representation. In this model, the notion of coactivation pattern is extended to account for the overall representation manner of L2 words, instead of only their semantic representations. Besides, the model illuminates both the facilitating and the inhibiting effects of the L1 semantic system which inevitably plays a part in L2 word acquisition, representation, as well as development.Afterwards, the model's implications for L2 word meaning acquisition are discussed. Based on the analysis of two important phenomena, i.e. the unbalanced performance of language production and language comprehension and the unavoidability of negative semantic transfer, the thesis discusses the positive and negative effects of episodic memory and the question about how to apply the episodic and semantic memory to different learning conditions.This thesis is composed of five chapters:Chapter One introduces the academic value of this research and the status quo of relevant researches, and then outlines the organization of this thesis. Chapter Two is literature review, in which the basis of bilingual research, i.e. monolingual mental lexicon researches, and main theories of the bilingual mental lexicon researches and their research methodology are summarized. Chapter Three challenges some experiment results and hypotheses based on the evaluation of the present models and research methods, for example, the local representation hypothesis of semantic knowledge, and the indirect two-stage hypothesis of the access route from formal representations of L2 words to their semantic representations. Afterwards, the dissociation of semantic representations and conceptual representations, the distributed representation pattern of semantic knowledge, and the episodic connection of formal representations of L2 words and those of the L1 translations are discussed. Chapter Four is concerned with the implications of the integration model proposed in this thesis on L2 word meaning acquisition. It is pointed out that episodic memory and semantic memory should be applied to different learning conditions to bring into full play their positive effects. Chapter Five summarizes the chief points of the thesis and offers some ideas about the directions for our further study.