Investigating grammatical processing in bilinguals: The case of morphological priming
João Veríssimo & Harald Clahsen (Potsdam Research Institute for Multilingualism, Germany)
Much previous work on the advantages and disadvantages of bilingualism has focussed on vocabulary or the processing of simple words. From a linguistic perspective, however, vocabulary is a rather peripheral aspect of the knowledge of language, and claims about the effects of bilingualism that are based on tasks that tap only into lexical aspects may not necessarily hold for bilingual language processing as a whole.
In this talk we will discuss methods for investigating grammatical processing in bilinguals, focusing on the use of morphological priming studies. We will present a methodological approach that relies on (i) linguistic (in our case, morphological) theory for the construction of experimental materials, (ii) a design that allows for direct (within-experiment, within-participant and within-item) comparisons of the critical conditions, and (iii) data analysis techniques that make both linear and non-linear gradient effects visible. We believe that these considerations are not only relevant for morphological priming experiments, but for studying bilingual language processing more generally.
We illustrate our approach using new data from a large-scale study with more than 90 bilingual participants from the Turkish/German community in Berlin investigating age effects in bilingual grammatical processing. The main discovery of this study was a selective age-of-acquisition (AoA) effect for grammatical processing abilities that (by using statistical modelling techniques) can be precisely delimited to a particular age range. Our participants all learnt Turkish from birth and German at different ages, spanning an AoA range from birth to about 35 years of age. We tested them on both inflectional and derivational priming in German. While derivational priming was found not to be affected by AoA and was present across the whole AoA range, inflectional priming showed a gradual AoA-related decline, but only from the ages of 5-6 onwards. Under the assumption that inflection is largely grammatical in nature (in that it simply spells out morphosyntactic features) whereas derivation is lexically-based (in that it creates new words), our results suggest that the critical period for language acquisition might be restricted to grammar, rather than applying to language as a whole. We will discuss the implications of this finding for bilingualism research, including the role of executive function and other non-linguistic factors in bilingual language processing.