Project Details
BiTPro - Bilingual Text Production: a usage-based approach to linguistic variation in translation and L2 writing
Applicant
Professorin Dr. Stella Neumann
Subject Area
Individual Linguistics, Historical Linguistics
Applied Linguistics, Computational Linguistics
Applied Linguistics, Computational Linguistics
Term
since 2024
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 542780701
How specific is translation in comparison to other tasks that involve the concurrent activation of another language while producing text? Long-held assumptions about the specificity of translation have recently been challenged by work highlighting similarities to other types of text production in language contact situations. Furthermore, different conditions of text production might have different effects on behavioral patterns in the writing process, but such differences are not consistently attested. Research comparing the different text production types to date has focused on specific factors, not least because of methodological challenges. Therefore, a comprehensive study is still missing that accounts for the relevant conditioning factors while integrating linguistic and behavioral perspectives. This project sets out to undertake just this by comprehensively investigating the differences between translated texts based on a prior text in a source language and texts produced from scratch in a foreign or second language (L2). It draws on a usage-based approach that recognizes the effect of multiple factors resulting in systematic probabilistic variation in language use. In this context, the question of directionality is of particular interest: We ask whether translations into an L2 are more similar to translations into the first language (L1), as both rely on a source text, or to texts produced from scratch in the L2 because production happens in a less dominant language. The former case would corroborate claims about the distinctive character of translation, whereas the latter would suggest that translation belongs to the same variety of language use as L2 writing. To this end, we devise an innovative multi-methods approach to data collection geared towards full comparability that combines linguistic and behavioral data. Texts in two registers are remotely elicited with a keystroke logger in the L1 German and the L2 English from professional translators. The resulting large data set is subjected to multifactorial and multivariate comparison of (i) L2>L1 translation, (ii) L1>L2 translation and (iii) L2 writing against (iv) the baseline of L1 text production. In a macroscopic linguistic analysis, patterns of variation are explored with Geometric Multivariate Analysis. Microscopic confirmatory analyses of selected linguistic features provide more detailed insight on factors driving linguistic choice in the four different tasks. Finally, the keystroke logging data is analyzed to gain insight into text production behavior linked to the linguistic patterns observed in the previous analyses. Beyond answering the above question, the results contribute novel insight into the place of translation in the context of language use in general.
DFG Programme
Research Grants
International Connection
Belgium, Brazil, Italy, Netherlands, Norway, USA