A Framework for Automatic Acquisition of Croatian and Serbian Verb Aspect from Corpora

A Framework for Automatic Acquisition of Croatian and Serbian Verb Aspect from Corpora Tanja Samardˇzi´c1, Maja Miliˇcevi´c2 1Corpus Lab, University of Zurich, URPP Language and Space, Freiestrasse 16, CH-8032 Zurich, Switzerland 2Faculty of Philology, University of Belgrad, Studentski trg 3, 11000 Belgrade, Serbia tanja.samardzic@uzh.ch, m.milicevic@fil.bg.ac.rs Abstract Verb aspect is a grammatical and lexical category that encodes temporal unfolding and duration of events described by verbs. It is a potentially interesting source of information for various computational tasks, but has so far not been studied in much depth from the perspective of automatic processing. Slavic languages are particularly interesting in this respect, as they encode aspect through complex and not entirely consistent lexical derivations involving prefixation and suffixation. Focusing on Croatian and Serbian, in this paper we propose a novel framework for automatic classification of their verb types into a number of fine-grained aspectual classes based on the observable morphology of verb forms. In addition, we provide a set of around 2000 verbs classified based on our framework. This set can be used for linguistic research as well as for testing automatic classification on a larger scale. With minor adjustments the approach is also applicable to other Slavic languages. Keywords: verb aspect, event, automatic acquisition, derivational morphology, morphological segmentation, Croatian, Serbian 1. Introduction Verb aspect is a grammatical and lexical category that en- codes temporal unfolding and duration of events described by verbs. It represents a potentially interesting source of in- formation for applications that involve event analysis, such as ordering of events in text summarisation, or identifying whether a relation holds in a particular moment in relation extraction. Despite extensive and long lasting discussion in theoretical linguistics, verb aspect is considerably underex- plored as a means of event analysis in computational ap- proaches, partly due to the fact that theoretical accounts are difficult to implement in automatic analysis, and partly be- cause aspectual distinctions are not easily observable in the most widely studied languages (such as English). Unlike in most other European languages, in the Slavic language family verb aspect is observable through lexical derivations. Consider, for example, the Croatian/Serbian sentences and their translations in (1)-(2).1 The English past tense form cooked is ambiguous, potentially describ- ing a habitual activity, as in (1), or a single completed event, as in (2). The context does sometimes allow us to disam- biguate between the two interpretations (see e.g. the modi- fier often in (1)), but such clues are frequently absent (as in (2)). In Croatian and Serbian, the habitual meaning in (1) is expressed with one verb (kuhao), and the single completed event in (2) with another (prokuhao). (1) Vinston je ˇcesto kuha-o. Winston AUX often cooked. Winston often cookedI. 1Despite being two separate standard languages, both Croatian and Serbian are discussed in the paper, as their close relatedness would make separate treatment highly inefficient. All examples in the paper are well-formed sentences in both languages (with minor adjustments such as kuhati [Croatian] vs. kuvati [Serbian]). (2) Vinston je pro-kuha-o ˇcaˇsu vode. …

Content truncated to 500 words