The determinants of bilingual memory capacity
The variation in short-term memory capacity between the dominant and less-dominant languages of bilinguals has been explained in terms of a difference in the rate of subvocal rehearsal between the languages. The 11 experiments presented in this thesis examined the effect of factors other than subvocal rehearsal and word length on bilingual short-term memory processes. Experiment 1 found that the language in which bilinguals received schooling influenced memory span for Arabic numerals (e.g., 1, 2, 3, etc.), The findings of Experiments 2 and 3 showed that articulation time did not predict bilingual auditory span for digits or words, whereas Experiment 4 found that articulation time was a reliable predictor of memory span for non words. Taken together, these findings suggested the involvement of non-phonological factors, such as the strength of long-term memory representation and phonotactic knowledge in mediating the bilingual memory span effect. Experiment 5 demonstrated that cognitive demands independent of articulation influenced the processing of numerals. In addition, the finding that memory span for numerals was greater than for digit words (e.g., one. two, three, etc., Experiments 6 & 7) raised fundamental questions regarding the use of inconsistent representations of digits in previous studies of bilingual memory span. A closer examination of the numeral advantage effect (Experiments 8 & 9) suggested that this was partly mediated by a variation in the predisposition toward implicit forms of chunking between numerals and digit words. Finally, the findings of Experiments 10 and 11 suggested that when the level of bilingual fluency is low, differences in the time taken to output recall sequences is an influential factor in determining bilingual memory span. Overall, this thesis questioned the assumptions in the bilingual memory span literature that have given rise to simplistic and incomplete explanations of the factors that moderate bilingual memory span. It was concluded that a greater consideration of long-term semantic and phonotactic factors and their effect on the speed of perceptual processing, subvocal rehearsal, and output delays was necessary to provide a comprehensive account of the determinants of bilingual memory capacity.