Bilingual lexical and semantic representation of numbers

In this thesis, I have reviewed and investigated bilingual lexical and semantic number representations. In three different studies, we compared German and French lexical representations, the influence of language-dependent morpho-syntax on accessing lexical representation and the associations with lexico-semantic representations. The bilingual samples in all three studies are Luxembourgish German-French bilinguals. In the Luxembourgish education system, German is the general instruction language for the first 6 years (i.e. LM1). Then the language for math instruction switches to French for the next 4 years (i.e. LM2) and gradually becomes the general instruction. In study 1 we found an LM2 cost in lexical retrieval for the LM2 compared to the LM1, (i.e. slower number naming). In addition to the LM2 cost, French ’70s to ’90s base-20 numerals were slower to process, indicating an effect of morpho-syntax on accessing lexical representations of numbers. This independently from increasing LM2 proficiency, the results replicate on four age groups with increasing proficiency. In study 2 we investigated the morpho-syntactic effect on German-French bilinguals and language-matched monolinguals. In an auditory-visual number matching task, we manipulated the visual presentation of two-digit numbers, either mimicking LM1’s inverted unit-ten morpho-syntax or LM2’s more transparent ten-unit morpho-syntax. We found that only the LM2 was affected by the morpho-syntactic experimental manipulation Moreover, we found a bilingual lexical cost such that bilingual in the LM1 German were slower than German monolinguals. In study 3 we compared lexico-semantic associations in bilinguals with a priming distance effect paradigm. The results indicate that while both language’s equivalent primes facilitate number naming (i.e. “cinq” and “fünf” facilitate the naming of 5) the priming distance effect was only observed with LM1 primes. Since the priming distance effect arises from the association between different numerals, it is interpreted as weaker lexico-semantic associations of the LM2 compared to LM1. In a nutshell, this thesis presents empirical evidence for LM2 weaker lexical associations which are more impacted by linguistic morphos-syntax as well as weaker lexico-semantic associations compared to the LM1.