Learning How to Conjugate the Romanian Verb. Rules for Regular and Partially Irregular Verbs

Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics, pages 524–528, Avignon, France, April 23 - 27 2012. c⃝2012 Association for Computational Linguistics Learning How to Conjugate the Romanian Verb. Rules for Regular and Partially Irregular Verbs Liviu P. Dinu Faculty of Mathematics and Computer Science University of Bucharest ldinu@fmi.unibuc.ro Vlad Niculae Faculty of Mathematics and Computer Science University of Bucharest vlad@vene.ro Octavia-Maria S, ulea Faculty of Foreign Languages and Literatures Faculty of Mathematics and Computer Science University of Bucharest mary.octavia@gmail.com Abstract In this paper we extend our work described in (Dinu et al., 2011) by adding more con- jugational rules to the labelling system in- troduced there, in an attempt to capture the entire dataset of Romanian verbs ex- tracted from (Barbu, 2007), and we em- ploy machine learning techniques to predict a verb’s correct label (which says what con- jugational pattern it follows) when only the infinitive form is given. 1 Introduction Using only a restricted group of verbs, in (Dinu et al., 2011) we validated the hypothesis that pat- terns can be identified in the conjugation of the Romanian (partially irregular) verb and that these patterns can be learnt automatically so that, given the infinitive of a verb, its correct conjugation for the indicative present tense can be produced. In this paper, we extend our investigation to the whole dataset described in (Barbu, 2008) and at- tempt to capture, beside the general ending pat- terns during conjugation, as much of the phono- logical alternations occuring in the stem of verbs (apophony) from the dataset as we can. Traditionally, Romanian has received a Latin- inspired classification of verbs into 4 (or some- times 5) conjugational classes based on the ending of their infinitival form alone (Costanzo, 2011). However, this infinitive-based classification has proved itself inadequate due to its inability to ac- count for the behavior of partially irregular verbs (whose stems have a smaller number of allo- morphs than the completely irregular) during their conjugation. There have been, thus, numerous attempts throughout the history of Romanian Linguistics to give other conjugational classifications based on the way the verb actually conjugates. Lom- bard (1955), looking at a corpus of 667 verbs, combined the traditional 4 classes with the way in which the biggest two subgroups conjugate (one using the suffix ”ez”, the other ”esc”) and ar- rived at 6 classes. Ciompec (Ciompec et. al., 1985 in Costanzo, 2011) proposed 10 conjuga- tional classes, while Felix (1964) proposed 12, both of them looking at the inflection of the verbs and number of allomorphs of the stem. Romalo (1968, p. 5-203) produced a list of 38 verb types, which she eventually reduced to 10. For the purpose of machine translation, Moisil (1960) proposed 5 regrouped classes of verbs, with numerous subgroups, and introduced the method of letters with variable values, while Pa- pastergiou et al. (2007) have recently developed a classification from a (second) language acquisi- tion point of view, dividing the 1st and 4th …

Content truncated to 500 words