Connectionism as a Model of Language Learning: Parallels in Foreign Language Teaching

Abstract

new paradigm for linguistics (43: p. 871) and psychology (47: p. 73) has already arrived. This new paradigm is variously named or parallel distributed processing. As latter name suggests, new discipline comes out of computing sciences, full of terms such as connection and parallel processing elements and under leadership of two specialists in field, James McClelland and David Rumelhart. For language teachers, questions arise: What is it and what relevance does it have for classroom teachers? A direct answer to first question is perhaps difficult for, as Sampson (p. 882) points out, consumers of last paradigm for linguistics were language teachers and literary students turned linguists. They are not likely to understand new paradigm since only computer specialist is fully trained in its intricacies. However, some of terms in are not so new. In fact, connectionism derives from behaviorist era, as Lachter and Bever (p. 234) suggest: The learning models in McClelland & Rumelhart are a complex variant on traditional--or at least Hullean s-r connections formed in time. In addition, like behaviorists, The connectionist focuses on overt behavior (19: p. 89). Most important, is a term which can be used of machines that learn. It provides for computer simulation through scanning of input and adjustment of weights in until desired output is achieved. In words of Fodor and Pylyshyn (p. 6), connectionist networks can be made to learn; this is achieved by modifying on connections as a function of certain kinds of feedback. This learnability is a startling discovery, which is probably reason that Classical transformational theoreticians, if not language teachers, are considering it worthy of attention. Regarding second question posed at beginning of this article, Chomskyan linguists presently argue against viability of as a description of human psychology or language. However, to understand their arguments, some mention of early tenets of Chomsky which so immensely impacted on language teaching, in spite of Chomsky's own apparent lack of interest in this applied domain, is necessary. Furthermore, a brief update on what Chomsky sometimes now refers to as the second major conceptual shift (10 as cited in 12: p. 173), development of principles and parameters of Universal Grammar, is also in order. At same time, language practitioners are cautioned by Chomsky (11: p. 180):