Using LatInfLexi for an Entropy-Based Assessment of Predictability in Latin Inflection

摘要: This paper presents LatInfLexi, a large inflected lexicon of Latin providing information on all the wordforms 3,348 verbs and 1,038 nouns. After description structure resource some data its size, procedure followed to obtain from database Lemlat 3.0 morphological analyzer is detailed, as well choices made regarding overabundant defective cells. The way in which LatInfLexi can be exploited order perform quantitative assessment predictability verb inflection then illustrated: results obtained by computing conditional entropy guessing content paradigm cell assuming knowledge one wordform or multiple are presented turn, highlighting descriptive theoretical relevance analysis. Lastly, envisages advantages an inclusion into LiLa base, both for base itself.