One Language to rule them all: modelling Morphological Patterns in a Large Scale Italian Lexicon with SWRL

One Language to rule them all: Modelling Morphological Patterns in a Large Scale Italian Lexicon with SWRL Fahad Khan1, Andrea Bellandi1, Francesca Frontini2, Monica Monachini1 1Istituto di Linguistica Computazionale “A. Zampolli” (ILC-CNR) Pisa, Italy 2PRAXILING UMR 5267 Univ Paul Val´ery Montpellier 3 & CNRS - Montpellier, France name.surname@ilc.cnr.it, francesca.frontini@univ.montp3.fr Abstract We present an application of Semantic Web Technologies to computational lexicography. More precisely we describe the publication of the morphological layer of the Italian Parole Simple Clips lexicon (PSC-M) as linked open data. The novelty of our work is in the use of the Semantic Web Rule Language (SWRL) to encode morphological patterns, thereby allowing the automatic derivation of the inflectional variants of the entries in the lexicon. By doing so we make these patterns available in a form that is human readable and that therefore gives a comprehensive morphological description of a large number of Italian words. Keywords: Morphology, Linked Open Data, Italian Lexicon, SWRL, SQWRL 1. Introduction The publication of lexical resources as Linked Data (LD) has recently become a important issue in computational lex- icography. However in the rush to convert all kinds of lexi- cons and dictionaries into RDF it is perhaps the case that the technical limitations of this mode of publication, as well as the new potentialities which it offers, are not as thoroughly understood as they ought to be. What, then, are some ad- vantages of using Semantic Web technologies to publish lexicons that might compensate for the loss in performance which is associated with LD with respect to other represen- tation formats? Obviously the fact that LD makes it eas- ier to link together datasets (interoperability), and to make them‘open’ and publicly available (as with Linked Open Data) is a crucial factor in its favour, but then there are also important technological reasons for publishing datasets as LD. For instance the Semantic Web gives us access to a whole ecosystem of standards, languages, and technologies which we can use to work with and to explore linked data. In this paper we look at one such language, the Seman- tic Web Rule Language (SWRL), and explore the extent to which it might potentially be able to play a useful role in the publication of lexicographic resources. We do this by detailing the conversion into RDF and publication of the morphological layer (PSC-M) of the Parole Simple Clips (PSC) Italian language lexicon. While the publication of this resource will make an important source of Italian mor- phological data freely and openly available to researchers and the wider public, the novelty of this work is in our use of SWRL to encode morphological patterns, something that allows the automatic derivation of the inflectional variants of the entries in the lexicon. By doing so we also make these patterns available in a form that is human readable and that therefore gives a comprehensive morphological de- scription of a large lexicon’s worth of Italian words. Note that we have already presented the first stage of the conver- …

Content truncated to 500 words