Can Alternations Be Learned? A Machine Learning Approach To Romanian Verb Conjugation

Proceedings of Recent Advances in Natural Language Processing, pages 539–544, Hissar, Bulgaria, 12-14 September 2011. Can Alternations Be Learned? A Machine Learning Approach To Romanian Verb Conjugation Liviu P. Dinu Faculty of Mathematics and Computer Science University of Bucharest ldinu@ funinf.cs.unibuc.ro Emil Ionescu Faculty of Letters University of Bucharest emilionescu@ 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 University of Bucharest mary.octavia@ gmail.com Abstract In this paper we look at the conjugation of the Romanian verb, in particular, at its ir- regularities, from a machine learning point of view. Our attempt is to predict the pres- ence or absence of any alternation in the stem (apophony), using n-gram represen- tations of the infinitive. We combine for- mal labelling mechanisms with learning methods in order to build a general con- jugational model. 1 Introduction The problem that we approached in this paper deals with phonological alternations in the stem of the Romanian irregular verbs during their conju- gation. What we attempted to investigate, using machine learning techniques, was whether there is reason to believe that a pattern can be identi- fied in the conjugation of the Romanian verb and whether that pattern could be learnt through auto- matic means such that, given the infinitive form of a verb, its correct conjugation could be produced. Like other Romance languages, Romanian has tra- ditionally received a Latin-inspired classification of verbs into 4 conjugational classes (or some- times 5, where the 4th conjugation is divided be- tween verbs with the infinitive ending in i and ˆı, respectively) based on the ending of their infini- tival form alone (Costanzo, 2011). However, this infinitive-based classification has often been found inadequate due to the many conjugational patterns that have been found in each class and to its inabil- ity to account for the behavior of partially irregular verbs (those whose stem has a smaller number of allomorphs than the completely irregular such as “a fi”) during their conjugation. There have been, thus, numerous attempts throughout the history of Romanian Linguistics to give other conjugational classifications. Lombard (1955) combined the traditional 4 infinitive-based conjugational classes with the in- formation related to the variation in the suffix re- ceived by 1st and 4th conjugational class verbs in the indicative first person singular form and came up with 6 classes. Other classifications based on the way Romanian verbs conjugate include (Ciompec et. al., 1985 in Costanzo, 2011), who proposed 10 conjugations, and Felix (1964) who came up with 12 conjugations by looking at the in- flection of the verbs and the number of allomorphs of the stem. Moisil (1960) proposed 5 regrouped classes of verbs and introduced the method of let- ters with variable values, in his computer scien- tific effort toward a “mechanical grammar”. Pa- pastergiou et al. (2007) have developed a classifi- cation from a (second) language acquisition point of view, dividing the 1st and 4th traditional classes into 3 and respectively 5 subclasses, each …

Content truncated to 500 words