The Influence of Working Memory, Language Learning Experience and Language Background on LLAMA_F Test Performance

Abstract: This experimental study investigates how individual variation in cognitive functioning, specifically working memory, previous experience with language learning but also language background, influence language aptitude, in particular grammatical inferencing - the capacity to spot patterns in grammar - as measured by the LLAMA_F test. Firstly, the study explores individual factors hypothesised to influence grammatical inferencing performance. Secondly, it examines whether the latest version of LLAMA_F test can be considered to be language neutral comparing agglutinative and non-agglutinative languages. In a three-part online experiment, native speakers of English, Hungarian and Finnish (Ntotal = 72) completed the LEAP-Q questionnaire (language background), n-back task (working memory capacity) and the LLAMA_F test (grammatical inferencing). Working memory capacity was found to correlate positively with grammatical inferencing (r = .28, p = .018), while language learning experience did not correlate with grammatical inferencing (r = .141, p = .119). After controlling for working memory, there were significant differences in grammatical inferencing between different language groups (p = .024); with the Hungarian group outperforming both the English and the Finnish group, and no significant differences between the English and Finnish group. This indicates that language aptitude is influenced by several different factors and that LLAMA_F cannot be regarded as entirely language neutral. Theoretical and practical implications are discussed, such as considering participants’ language background when administering LLAMA_F.