Generating and parsing clitics with GETARUN

In this paper we shall propose a principled solution to the problem of clitics from a computational point of view (Delmonte, 1992). In particular we shall deal with clitics in a specific Romance language, Italian, where they represent a highly ambiguous and ubiquitous morphological and syntactic problem. The solutions proposed will be in line with Monachesi, who however works only in a theoretical perspective (Monachesi, 1995). The syntactic framework that will be used is LFG theory: however, the details of the theoretical assumptions and the implementation can be transferred to other similar feature-based unification grammar formalisms like GPSG, HPSG, etc. As to the computational implementation, we shall be referring to a system called GETARUN, which is equipped with three main modules: a lower module for parsing where sentence strategies are implemented (Delmonte, 1990; Delmonte and Bianchi and Pianta, 1992); a middle module for semantic interpretation and discourse model construction which is cast into Situation Semantics; and a higher module where reasoning and generation takes place. We shall be referring both to the higher module and the lower one. In particular, the generation of clitics will be addressed first and then clitics parsing will be presented. We assume that from a psycholinguistic point of view, no grammar reversibility is allowed in the human language faculty, and that the language production processor is separate from the language understanding processor with which it shares a common lexicon. Language generation requires planning, a linguistic activity which is not required in the understanding process. At the same time, parsing requires setting up a number of disambiguating strategies, basically to tell arguments apart from adjuncts and reduce the effects of backtracking. The parser is a DGC depthfirst highly deterministic parser equipped with a look-ahead mechanism and a system of failures recovery based on a Well Formed Substring Table (WFST) which implements a number of sentence parsing psycholinguistic strategies (Delmonte and Dolci, 1989; Delmonte and Dolci, 1997).