Beyond Concatenative Morphology: Applying OntoLex-Morph to Maltese
What it is about
OntoLex-Morph is an RDF module for morphology in lexical data, so far applied mostly to concatenative languages. The authors model the strong triliteral verbs of Maltese, using the open Ġabra lexicon (19,918 entries, about 4.5 million inflected forms) as the source. Each verb is one lexical entry with its root (k-t-b) as a linked morph, its 3SG.M perfective form as base, and paradigm rules written as a regular-expression source and replacement. Maltese digraphs (ie, gh̄) and consonant and vowel classes make these rules long and hard to read, so the paper proposes a CharacterClass construct for the module. The evaluation is qualitative: rules are shown generating the paradigms of three example verbs, and no error counts are given.
Ideas
- I1. Keep the root as its own record. Each verb points to a root id (k-t-b) and its canonical form is the base; derived nouns link to the same root. In Rust this is an integer root id with a table, as the project's rule on constants asks.
- I2. Use named character classes, not letter lists. Gutturals (א ה ח ע), liquids and weak radicals should be classes in one table, with rules written over the class names. Test it by counting rule lines for the weak-verb rules with and without classes on MACULA verb forms. The paper argues for legibility; Mova should measure the rule count.
- I3. Generate, then compare. Produce each verb's paradigm from its base form and score it against the forms attested in a corpus, by class (regular, guttural, weak). The paper's three paradigms are a starting gold split.
For Mova
I1 goes into The verb and The lexicon. I2 goes into The weak verbs, and the same classes are needed in The verb in Biblical Aramaic. I3 is a test for tools/macula compare.
Cite
@inproceedings{ionov2023beyond,
title = {{Beyond Concatenative Morphology: Applying OntoLex-Morph to Maltese}},
author = {Ionov, Maxim and Rosner, Mike},
booktitle = {Proceedings of the 4th Conference on Language, Data and Knowledge},
year = {2023},
url = {https://aclanthology.org/2023.ldk-1.40/}
}References
- J. Aquilina. 1987. Maltese-English Dictionary. Midsea Books, Valletta. find
- C. Borg and A. Gatt. 2017. Morphological analysis for the Maltese language. In Proc. 3rd Arabic NLP Workshop, pages 25–34, Valencia, Spain. Ass. Comp. Ling. find
- J. Camilleri. 2013. A Computational Grammar and Lexicon for Maltese. MSc Thesis, Chalmers University of Technology, Gothenburg, Sweden. find
- C. Chiarcos, K. Gkirtzou, F. Khan, P. Labropoulou, M. Passarotti, and M. Pellegrini. 2022. Computational morphology with OntoLex-Morph. In Proc. 8th Workshop on Linked Data in Linguistics, pages 78–86. find
- Brother FSC Henry. 1980. Grammatika Maltija. De La Salle Brother Publications, Malta. find
- B. Klimek, J. McCrae, J. Bosque-Gil, M. Ionov, J. Tauber, and C. Chiarcos. 2019. Challenges for the representation of morphology in ontology lexicons. Proceedings of eLex 2019, pages 570–591. find
- John P McCrae, J. Bosque-Gil, J. Gracia, P. Buitelaar, and P. Cimiano. 2017. The OntoLex-Lemon Model: Development and Applications. In Proceedings of eLex-2017, pages 19–21. find
- V. Ravishankar, F. Tyers, and A. Gatt. 2017. A morphological analyser for Maltese. In K. Shaalan and S. El-Beltagy, editors, Procedia Computer Science vol 117, pages 175–182. Elsevier. find
- M. Rosner and C. Borg. 2022. Report on the Maltese Language. Deliverable D1.25, European Language Equality Project. find
- B. Sagot and G. Walther. 2013. Implementing a formal model of inflectional morphology. In 3rd International W/S on Systems and Frameworks for Computational Morphology, volume 380 of Communications in Computer and Information Science, pages 115– 134, Berlin. Humboldt-Universität, Springer. *Figure 5: Conjugation of kiteb* *Figure 6: Conjugation of gh̄amel* *Figure 7: Conjugation of telaq* find
This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/aclbulk-2023-ionov-beyond-concatenative-morphology-applying-ontolex/