EN FR
EN FR
Bilateral Contracts and Grants with Industry
Bibliography
Bilateral Contracts and Grants with Industry
Bibliography


Section: New Results

Computational morphology

Participant : Benoît Sagot.

In 2013, following previous collaborative work [92] , [105] , we have designed and developed AlexinaPARSLI in collaboration with Géraldine Walther (LLF and DDL), a formalism for encoding inflectional descriptions (lexicon and grammar) that aims at filling the gap between morphologically and typologically motivated approaches on the one hand and implemented approaches on the other hand, as will be discussed in the remainder of this section. Indeed, AlexinaPARSLI is both:

  • an implementation formalism for PARSLI, a formal model of inflectional morphology [106] that accounts for concepts underlying the canonical approach of morphological typology;

  • an extension of the Alexina lexical framework developed at Alpage for modeling lexical information and developing lexical resources. The Alexina framework now supports both morphological grammars that use the original Alexina morphological formalism as well as new grammars developed in AlexinaPARSLI.

The AlexinaPARSLI formalism and tools have been proven greatly beneficial to works both in descriptive and formal morphology, in particular in studies about Latin passivisation and Maltese verbal inflection [106] and in studies comparing the compacity of morphological descriptions [106] , [92] , [105] , as well as in NLP, for the efficient development of a large-scale and linguistically sound morphological lexicon for German (a paper describing this new lexicon is to be presented at the LREC 2014 conference).

In collaboration with Géraldine Walther and Guillaume Jacques (CRLAO, CNRS), within operation LR4.11 from strand 6 of the LabEx EFL, we have also developed two AlexinaPARSLI descriptions of (part of) the Khaling (Kiranti, Sino-Tibetan) verbal inflectional system, together with a medium-scale lexicon. Our study shows that an explicit account for the so-called direct-inverse marking, based on concepts developed within PARSLI, allows for a more compact account of this inflectional system [42] .