BILINGUAL LEXICAL RETRIEVAL.

Title

Identifier

AAI8112738

identifier

8112738

Creator

ALTENBERG, EVELYN P.

Contributor

Prof. Helen S. Cairns

Date

1981

Language

English

Publisher

City University of New York.

Subject

Language, Linguistics

Abstract

Two questions about bilingual language processing were investigated: (1) Does the bilingual have one unified language processing system, used for processing in both languages, or two processing systems, one for each language?, and (2) If the bilingual has two language processing systems, is one "shut off" while the other is "on," or are both systems always available and/or operating? These two questions were examined with respect to one aspect of bilingual processing: bilingual lexical retrieval.;It was demonstrated that a monolingual lexical retrieval system must have at least a scanner, containing a knowledge of phonotactic constraints, and a lexicon, containing linguistic tags (words) and their semantic representations. A search model of monolingual lexical retrieval was assumed.;Experiments 1 and 2 investigated the question of whether the bilingual has one or two sets of phonotactic constraints. German-English bilinguals rated the acceptability of visually presented nonwords as new words of German or new words of English. The nonwords were either legal in both languages (e.g., FLESS), illegal in both languages (e.g, TLON), legal only in English (e.g., TWIF), or legal only in German (e.g., ZWOK). It was found that bilinguals rated the last two groups differently, depending on whether the items were rated as potential English or potential German words. It was also found that the bilinguals' ratings in English were the same as the ratings of a monolingual English control group. Thus, the bilingual has a knowledge of two sets of phototactic constraints, one for each language, and the constraints for English are the same as those of English monolinguals.;Experiments 3 and 4 investigated the question of whether the bilingual's two lexical retrieval systems (i.e., two sets of phonotactic constraints and two lexicons) are always available during processing. The experiments were lexical decision tasks, in which the nonwords were the same four groups used in the rating task. The lexical decision task is sensitive to phonotactic constraints, since monolingual reaction time is faster for illegal nonwords than for legal nonwords.;The major finding was that bilinguals had a different overall pattern of reaction times to the items than did the monolingual control group. This would not be expected if the bilingual simply "shuts off" the retrieval system of one language while processing in the other. Thus, both lexical retrieval systems must be operating during bilingual lexical retrieval.;Another finding was that there was no difference in overall reaction time between monolinguals and bilinguals. One can thus reject any model of bilingual lexical retrieval which assumes that bilinguals always add an extra step during retrieval, which monolinguals do not. Further, the bilingual must organize the words from both languages into two separate lexicons, rather than one large lexicon, since one large lexicon would incorrectly predict longer bilingual reaction times, due to longer search times.;A number of possible models of how bilingual lexical retrieval might work, if both systems are operating during processing, were examined. It was concluded that a parallel exhaustive model best predicts the results. According to the parallel models, the bilingual has two lexical retrieval systems and both systems are always simultaneously operating during lexical retrieval. Further, according to the parallel exhaustive model, the bilingual does not make a decision in the lexical retrieval task, or during language processing, until input from both lexical retrieval systems has been received.

Type

dissertation

Source

PQT Legacy CUNY.xlsx

degree

Ph.D.

Program

Linguistics

Item