Crosslingual Language Varieties: A Multifaceted Investigation

Most people in the world today use more than one language in the course of their daily lives, and it is estimated that most children today grow up with exposure to two or more languages. Multilingualism has been recognised as an educational and social goal, especially by the European Union. In addition to individual multilingualism, culture, technology and information are increasingly characterised by a globalisation of resources, mediated by translation of materials originating in a wide variety of languages. This innovative and interdisciplinary proposal brings together researchers and methods from corpus linguistics, psycholinguistics and computer science to achieve a better understanding of how individuals use the various languages at their disposal. The findings of the proposed research
could then be used to inform language education and translation studies.

The proposed research program will simultaneously investigate several varieties of the language produced by bilinguals using the same set of multi-disciplinary methods in order to establish common characteristics as well as meaningful differences among these varieties. We will use qualitative and quantitative computational methods on language corpora and complement our findings with psycholinguistic experiments.

We will uniformly investigate various manifestations of crosslingual language varieties, including both translations and the language of non-native speakers at different levels of expertise and proficiency. These varieties will be contrasted with the language of native speakers producing original (i.e., not translated) utterances. In particular, we will address the following research questions: (i) What characteristics are common to the various crosslingual language varieties and distinguish them from native language? (ii) What properties distinguish crosslingual language varieties from each
other? (iii) Are such differences language-pair specific, or are they "universal"? Answers to these questions will forward our understanding of the cognitive and computational processes supporting crosslingual language varieties, and will illuminate the specific circumstances that lead to similarities and differences between individuals and language settings.

The proposed research is novel in two important ways: (i) we will address several crosslingual language varieties, including translations and advanced non-native language, under a unifying umbrella, whereas existing approaches typically focus only on a single variety; and (ii) we will utilise three different and complementary methodologies (statistical analysis of large corpora with a modelling component, fine-grained linguistic analysis of deeply annotated
smaller corpora, and psycholinguistic experiments) to investigate these language varieties. This broad approach, which cuts across established disciplinary boundaries, holds the promise of going beyond previous findings, and having a strong theoretical and practical impact.