Punctuation as Native Language Interference

Proceedings of the 27th International Conference on Computational Linguistics, pages 3456–3466 Santa Fe, New Mexico, USA, August 20-26, 2018. 3456 Punctuation as Native Language Interference Ilia Markov INRIA Paris, France ilia.markov@inria.fr Vivi Nastase University of Heidelberg Heidelberg, Germany nastase@ cl.uni-heidelberg.de Carlo Strapparava Fondazione Bruno Kessler Trento, Italy strappa@fbk.eu Abstract In this paper, we describe experiments designed to explore and evaluate the impact of punctuation marks on the task of native language identification. Punctuation is specific to each language, and is part of the indicators that overtly represent the manner in which each language organizes and conveys information. Our experiments are organized in various set-ups: the usual multi-class classification for individual languages, also considering classification by language groups, across different proficiency levels, topics and even cross-corpus. The results support our hypothesis that punctuation marks are persistent and robust indicators of the native language of the author, which do not diminish in influence even when a high proficiency level in a non-native language is achieved. 1 Introduction Native Language Identification (NLI) – identifying the native language (L1) of a person based on his/her writing in the second language (L2) – is useful for a variety of purposes, including security, market- ing, and educational applications. The effect of native language phenomena seeping into texts pro- duced in a different language is known as language transfer (Odlin, 1989). Numerous aspects of the language have been explored for NLI – character-level language models (Ionescu et al., 2014), lexical choice (Brooke and Hirst, 2012; Lahiri and Mihalcea, 2013), grammar (Nagata and Whittaker, 2013), spelling errors (Chen et al., 2017; Koppel et al., 2005), cognates (Nicolai et al., 2013), and general etymology (Nastase and Strapparava, 2017). While punctuation has been included in some of these studies (e.g., in character-level models), its impact has not been studied. It is however an important, and often revealing, aspect of written language. For example, punctuation is a strong indicator of authorship, and has been used successfully in stylo- metric analysis for authorship attribution (Markov et al., 2017b; Grieve, 2007). More generally, from a linguistic point of view, punctuation has been disputed as following prosodic principles or as a clarifier of grammatical structure (Baron, 2001; Bruthiaux, 1993). Moore (2016) finds a common ground for these two views by observing that prosody and punctuation realize the same function – revealing/emphasizing the information structure of an utterance – in the spoken and respectively written modes of language. Since grammar and prosodic structure are language specific, indicators that reveal them would be lan- guage specific as well. As with other aspects of language, grammatical/prosodic influences from the native language may surface in the new language as particular punctuation choices. As an example, consider the following English sentence, written by a native German speaker1: I think the biggest question is , how to defin an “ enjoyed life ” . A native English speaker would not insert a comma between is and how, but it reflects correct punctu- ation usage in German: This work is licensed …

Content truncated to 500 words