Cheikh Bamba Dione
Formal and Computational Aspects of Wolof Morphosyntax in Lexical Functional Grammar
Initial project description:
"My project aims at building grammar, lexicon, morphology for Wolof. Wolof is a Niger-Congo language spoken by ca. 10 million people in Senegal. There is currently a lack of publicly available natural language resources and applications for Wolof. Moreover, there are no tools or resources dealing with this language’s morphology or syntax, so that morpho-syntactically annotated corpora are not available. Likewise, multilingual applications linking Wolof to European languages are still out of reach.
The overall objective of my project is to study fundamental linguistic aspects of Wolof in a computational framework, with a view towards con- structing theoretical motivated computational grammars and lexicon. In addition, the current research will also be contrastive, with a view towards multilinguality. This work is embedded in CLARA (Common Language Resources and their Application) whose main objective being to create a large language research infrastructure for all of Europe.
The main methodology will be the construction of a computational grammar based on Lexical Functional Grammar (LFG) theory. An important tool supporting LFG development is the Xerox Linguistic Environment (XLE). Also, the morphology will be developed using Finite State Transducers (FST). The intended outcome will be an XLE-based LFG grammar for Wolof. Furthermore, this grammar will be used for processing real corpus data and its relevance tested on a wide range of natural Wolof examples. This parsing activity will result in a prototype parsebank for Wolof. The LFG Parsebanker tool will be used for this purpose, in cooperation with the INESS project. Thus, the overall accomplishment expected in this CLARA research is the improved accessibility of good quality language resources annotated with useful linguistic information (morpho-syntactic, semantic, pragmatic and multilingual).
The study of cross-lingual phenomena is an important objective of this research. A contrastive analysis will focus on commonalities and differences between Wolof and European languages (e.g. French, German, English and Norwegian). To this end, the parsebanking effort will be extended to a parallel parsebank aligned at phrase level. The main challenge of the proposed approach is not only to produce the first morpho-syntactical analysis of Wolof, but to perform this work in a way which is integrated and maximally compatible with comparable ongoing work at LaMoRe and in the ParGram project."
Dione has successfully defended his thesis on 3 October 2014.