von Studnitz, Roswitha E.; (2001) An investigation of control processes in bilinguals, using visual word recognition tasks. Doctoral thesis (Ph.D), UCL (University College London).

Text An_investigation_of_control_pr.pdf Download (15MB)

Abstract

How is competition between bilinguals' languages resolved? This thesis contrasts two main accounts - an internal (Bilingual Inhibitory Control [BIA] model, Dijkstra & van Heuven, 1998) and an external account (Inhibitoiy Control [IC] model, Green, 1998). The internal account proposes varying the activation of languages within the lexico-semantic system; the external account proposes task schemata acting differentially to the signals emerging from this system. On-line visual word recognition tasks (with German/English bilinguals) measured RTs and errors. Experiments la/1b: switching languages incurred a greater cost when language was task relevant than when it was not ('language-specific' vs. 'language-general' lexical decision task), consistent with both external and internal accounts. Experiment 2 sought to distinguish between the two accounts by refining Experiment 1. In a language-specific lexical decision task language-switch costs were reduced when this task followed the language-general task, but only for 'pure-HF' stimulus lists, not for 'mixed' (HF / LF) lists. Such patterns are consistent with an external, but not an internal account, and provide insight into lists' effects on decision criteria. Experiment 3: animacy decisions (for German and English words) produced a language-switch cost only in the first half of the experiment, and only for response- repetition trials, not for response-switch trials. This is consistent with an external, but not an internal account, and shows the influence of language signals on decisions even when task-irrelevant. Experiments 4a/4b: English-specific lexical decisions incurred a cost on interlingual homographs (vs. matched controls). The degree of interference was greater if pure German words (to be rejected) were included than when none were included. This enhanced interference diminished with time (Experiment 4a). Interference was less when indviduals were informed about the presence of IHs than when not informed (Experiment 4b). IH interference and 'carry-over' effects dissociated, consistent with an external but not an internal account. Overall, the data are consistent with a non-selective access account. They further indicate that language activation levels can be raised exogenously, but not reduced endogenously. A mechanism reacting differentially to signals from the lexico-semantic system can account for the observed patterns. This mechanism appears to be influenced by task demands, experience, item constraints and available information. The external account is parsimonious, proposing the same mechanism for resolving competition between languages, as for control of action. In the text the 3rd person singular 'he' is used from time to time, which should be taken to mean both 'he' and 'she'. The studies undertaken here were given ethical permission by the Joint UCL/UCLH Committees on the Ethics of Human Research: Committee Alpha.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item