Modeling the Impact of Syntactic Distance and Surprisal on Cross-Slavic Text Comprehension

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 7368–7376 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 7368 Modeling the Impact of Syntactic Distance and Surprisal on Cross-Slavic Text Comprehension Irina Stenger, Philip Georgis, Tania Avgustinova, Bernd Möbius, Dietrich Klakow Saarland University, Collaborative Research Center (SFB) 1102: Information Density and Linguistic Encoding, Campus A 2.2, 66123 Saarbrücken, Germany Project C4: INCOMSLAV – Mutual Intelligibility and Surprisal in Slavic Intercomprehension ira.stenger@mx.uni-saarland.de, {pgeorgis, avgustinova, moebius}@lst.uni-saarland.de, dietrich.klakow@lsv.uni-saarland.de Abstract We focus on the syntactic variation and measure syntactic distances between nine Slavic languages (Belarusian, Bulgarian, Croatian, Czech, Polish, Slovak, Slovene, Russian, and Ukrainian) using symmetric measures of insertion, deletion and movement of syntactic units in the parallel sentences of the fable “The North Wind and the Sun”. Additionally, we investigate phonetic and orthographic asymmetries between selected languages by means of the information theoretical notion of surprisal. Syntactic distance and surprisal are, thus, considered as potential predictors of mutual intelligibility between related languages. In spoken and written cloze test experiments for Slavic native speakers, the presented predictors will be validated as to whether variations in syntax lead to a slower or impeded intercomprehension of Slavic texts. Keywords: syntactic distance, surprisal, asymmetric intelligibility, Slavic intercomprehension 1. Introduction 1.1 Background Intercomprehension (Doyé, 2005), semi-communication (Haugen, 1966), lingua receptiva (Rehbein et al., 2012), mutual intelligibility (Gooskens and van Heuven, 2021) or receptive multilingualism (Braunmüller and Zeevaert, 2001) refers to situations where people take advantage of similarities between their native language (L1) and a closely related non-native language (L2) to comprehend speech or text in this L2. Certain degrees of mutual intelligibility are typically due to existing and perceived similarities at different levels of linguistic structure. Investigations of (dis)similarities largely consider lexical, phonetic/phonological, and orthographic distances to predict the ability to comprehend (closely) related languages (Gooskens and van Heuven, 2021). Jágrová et al. (2017) have found that despite small lexical distances measured in terms of the proportion of non-cognates, Czech and Polish (both West Slavic, using Latin script) are orthographically more distant from each other than Bulgarian and Russian (respectively South and East Slavic, both using Cyrillic script). Stenger (2019: 245) shows the Levenshtein distance (Levenshtein, 1966) of cognates to be a reliable predictor of orthographic intelligibility of Slavic languages written in Cyrillic script to Russian native speakers. The role of syntactic (dis)similarities for Slavic intercomprehension is investigated to a considerably lesser extent. Golubović (2016: 105) uses the trigram approach (Nerbonne and Wiersma, 2006; Lauttamus, Nerbonne, and Wiersma, 2007) to illustrate the role of syntactic distance as a predictor of mutual intelligibility in written and spoken cloze tests. Stenger and Avgustinova (2021) adapt measures of word movement, insertion, and deletion from (Heeringa et al., 2017) to investigate the impact of syntactic predictors on contextualized cognate recognition in Slavic. The number of insertions or deletions to obtain the closest wording in subjects’ language reveals a negative effect on intelligibility only as a tendency. The movement measure, however, did not influence …

Content truncated to 500 words