Predicting Native Language from Gaze

Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 541–551 Vancouver, Canada, July 30 - August 4, 2017. c⃝2017 Association for Computational Linguistics https://doi.org/10.18653/v1/P17-1050 Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 541–551 Vancouver, Canada, July 30 - August 4, 2017. c⃝2017 Association for Computational Linguistics https://doi.org/10.18653/v1/P17-1050 Predicting Native Language from Gaze Yevgeni Berzak MIT CSAIL berzak@mit.edu Chie Nakamura MIT Linguistics chienak@mit.edu Suzanne Flynn MIT Linguistics sflynn@mit.edu Boris Katz MIT CSAIL boris@mit.edu Abstract A fundamental question in language learn- ing concerns the role of a speaker’s first language in second language acqui- sition. We present a novel methodol- ogy for studying this question: analysis of eye-movement patterns in second lan- guage reading of free-form text. Using this methodology, we demonstrate for the first time that the native language of English learners can be predicted from their gaze fixations when reading English. We pro- vide analysis of classifier uncertainty and learned features, which indicates that dif- ferences in English reading are likely to be rooted in linguistic divergences across native languages. The presented frame- work complements production studies and offers new ground for advancing research on multilingualism.1 1 Introduction The influence of a speaker’s native language on learning and performance in a foreign language, also known as cross-linguistic transfer, has been studied for several decades in linguistics and psy- chology (Odlin, 1989; Martohardjono and Flynn, 1995; Jarvis and Pavlenko, 2008; Berkes and Flynn, 2012; Alonso, 2015). The growing avail- ably of learner corpora has also sparked interest in cross-linguistic influence phenomena in NLP, where studies have explored the task of Native Language Identification (NLI) (Tetreault et al., 2013), as well as analysis of textual features in relation to the author’s native language (Jarvis and Crossley, 2012; Swanson and Charniak, 2013; Malmasi and Dras, 2014). Despite these advances, 1The experimental data collected in this study will be made publicly available. the extent and nature of first language influence in second language processing remains far from be- ing established. Crucially, most prior work on this topic focused on production, while little is cur- rently known about cross-linguistic influence in language comprehension. In this work, we present a novel framework for studying cross-linguistic influence in language comprehension using eyetracking for reading and free-form native English text. We collect and ana- lyze English newswire reading data from 182 par- ticipants, including 145 English as Second Lan- guage (ESL) learners from four different native language backgrounds: Chinese, Japanese, Por- tuguese and Spanish, as well as 37 native English speakers. Each participant reads 156 English sen- tences, half of which are shared across all partic- ipants, and the remaining half are individual to each participant. All the sentences are manually annotated with part-of-speech (POS) tags and syn- tactic dependency trees. We then introduce the task of Native Language Identification from Reading (NLIR), which re- quires predicting a subject’s native language from gaze while reading text in a second language. Fo- cusing on ESL participants and using a log-linear …

Content truncated to 500 words