A Cognitive Approach To The Mechanism Of Semantic Transfer

Semantic transfer refers to the influence of one language upon the use of another in the area of semantics. In second language (L2) pedagogy, teachers have always been complaining that learners often make lexical errors or produce semantically anomalous expressions attributable to semantic transfer from L1. As a factor pestering second language learners and teachers alike, semantic transfer is closely related to what Danesi (2000, 2003) called"SLT dilemma"—few learners could speak with the same kind of"naturalness"that is instantly detectable in native-speaker discourse after studying a language for varying periods of time. It may even be held responsible for causing the dilemma. However, semantic transfer is not given the due attention at present. Even if there is some research devoted to semantic transfer, it concentrates on revealing the manifestations of semantic transfer at the lexicosemantic level, with semantic transfer at the propositional semantic level and the mechanism that underlies semantic transfer almost totally ignored. It can be said that we are still in the dark as to how semantic transfer takes place. An investigation of the mechanism of semantic transfer (MST) therefore, is not only of high necessity but also of great theoretical and pedagogical significance. First, it may offer us the promise of overcoming or minimizing the negative effects of semantic transfer in L2 teaching and learning; second, it may enable us to gain some new insights into the intrinsic processes of SLA; third, it may conduce to developing a new view of L2 learning and teaching that remedies the deficiency of our present view of them. All of these, undoubtedly, contribute to our efforts to solve the so-called"SLT dilemma"faced by L2 pedagogy today.The present study has made an endeavor to explore how semantic transfer takes place, namely, the mechanism of semantic transfer (MST). We have first depicted a clear picture of current research on the mechanism of semantic transfer (MST) by sorting out the types of evidence existing for semantic transfer at present and by reviewing a host of insights that bear close relevancy to MST. It has been shown that there are in general five major types of evidence for semantic transfer at present. We can make recourse to"the equivalence hypothesis"(Swan 1997),"the compound conceptual system assumption"(Jarvis 1998),"the conceptualization principle"(Danesi 2003), and"thinking for speaking"(1996a) respectively to offer each of them an account. These accounts, however, are all deficient in some way. More importantly, none of these hypotheses or assumptions can offer a unified account of all the five types of evidence. This convinces us that there's no feasible conception of MST at present; it is necessary to make an attempt in this respect. Although current accounts are inadequate in general, they lend us a most significant insight: semantic transfer may be related to the'conceptualization'stage of language production.In pursuing the belief that semantic transfer may be related to the'conceptualization'stage of language production, we have first undertaken a review of the psycholinguistic and the cognitivistic conceptions of'conceptualization'. It has been shown that the psycholinguistic conception of'conceptualization', given its special purpose of modeling monolingual speech, does not fit our purpose of probing into MST in terms of it. The same is true of the psycholinguistic conception of'conceptualization', because the conception does not set the notion'conceptualization'in the framework of pscholinguistic production of language. Then we have embarked on the task of proposing a new conception of'conceptualization'that fits our purpose of discussing MST. By adopting the basic framework of Levelt's (1989) Speaking Model and by drawing heavily on insights from linguists, psycholinguists and cognitive linguists, we have accomplished a new conception of'conceptualization': Conceptualization is the mental process whereby human experiences are made sense of and transformed into verbally-encodable conceptual structures via such mental operations as domain mapping, perspective taking, schematization, categorization and so on.Based on the new conception of'conceptualization', we have made a crosslinguistic comparison of the ways speakers of different languages carry out conceptualization, namely, domain mapping, perspective taking, and schematization when expressing the same experiences. It has been shown that conceptualization is specific to the language being used. Building on the language-specificity of conceptualization, we have made the assumption that each language represents a unique mode of conceptualization; speaking in a language entails mobilizing the corresponding language's mode of conceptualization. If a speaker intends to speak in Language A but mobilizes the mode of conceptualization of Language B, he will most likely produce expressions that are structurally correct but semantically anomalous. This has led us to characterize MST as follows: Semantic transfer results from the bilingual or multilingual's mis-mobilization of the non-target language's mode of conceptualization (NTLMC). In order to explore the explanatory power of this tentative conception of MST, we have applied it to current evidence for semantic transfer. It has been shown that this tentative conception of MST can offer a reasonable account of all the semantic transfer evidence available. This shows that our conception of MST has fairly large explanatory power.With the theoretical conception of MST available, we have attempted an empirical validation of it. For this purpose, we have first made a few predictions based on our conception of MST. Then, we have cited some evidence from current literature to support these predictions. Next, we have carried out three elaborately-designed experiments to test one of the predictions. It has been shown that the results of all of the three experiments were positive, which confirmed the prediction. The conclusion we have drawn is that although the validation is preliminary, our conception of MST is valid as far as the evidence available is concerned. At the end of the present study, the relevant implications and limitations have also been dicussed.