Machine translation and subject enclitics in the venetan dialect

In this chapter I shall be concerned with the Venetan dialect and I will focus on work carried out under a project called “STILVEN”, which stands for Sistema di Traduzione tra la Lingua Inglese e il Veneto, financed by the Veneto Regional Council and which ultimately led to the creation of a translation application available online at http://project.cgm.unive.it/cgi-bin/stilven/moses.pl. In order to produce one such prototype, an extended number of digital resources were created, with the aim to experiment with the translation system, but also and foremost, to make such resources freely available to the linguistic community of researchers interested in the study of this dialect. All resources are now available online at the project address, on our website, http://project.cgm.unive.it/?page_id=219. They in- clude a computational lexicon, a tagged corpus, translation pairs, a parallel corpus for machine translation, frequency lists, grammatical items like minimal pairs etc. and a lot of intermediate linguistic materials that were used to build the translation system. At first I will be concerned with the experimental and application-oriented approach, by concentrating on tagging and the organization of the translator. Then I will focus on one specific linguistic phenomenon, subject enclitics, which I will describe and explain from a quantitative, application oriented and theoretical point of view. The latter will be done by referring at first to work published using a Chomskyan approach, and then I will show how the same problem can be explained within the Lexical Functional Grammar (LFG) approach.