On Practical Realisation of Autosegmental Representations in Lexical Transducers of Tonal Bantu Languages

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Abstract

A lexical transducer is a language technology resource that is typically used to predict the orthographic word forms and to model the relation between the lexical and the surface word forms of a morphologically complex language. This paper motivates the construction of tone-enhanced lexical transducers for tonal languages and gives two supporting arguments for the feasibility of finite-state compilation of autosegmental derivations. According to the Common Timeline Argument, adding a common timeline to autosegmental representations is crucial for their computational processing. According to the Compilation Argument, the compilation of autosegmental grammars requires combining code-theoretic and model-theoretic research lines.
Original languageEnglish
Title of host publicationProceedings of the Language Technologies for All (LT4All)
EditorsGilles Adda, Khalid Choukri, Irmgarda Kasinskaite-Buddeberg, Joseph Mariani, Hélène Mazo, Sakti Sakriani
Number of pages4
Place of PublicationParis
PublisherEuropean Language Resources Association (ELRA)
Publication date2020
Pages346-349
ISBN (Print)979-10-95546-33-7
Publication statusPublished - 2020
MoE publication typeA4 Article in conference proceedings
EventLT4All – International Conference on Language Technologies for All 2019: Enabling Linguistic Diversity and Multilingualism Worldwide - UNESCO Headquarters, 7, place de Fontenoy, 75007, Paris, France , Paris, France
Duration: 4 Dec 20196 Dec 2019
https://en.unesco.org/events/international-conference-language-technologies-all-lt4all-enabling-linguistic-diversity-and

Fields of Science

  • 6121 Languages
  • 113 Computer and information sciences
  • lexical transducers
  • tonal languages
  • autosegmental phonology
  • compositionality
  • origin correspondence

Cite this