Elamur Research⌕
☰
home / papers / 1994 / Computational Analyses of Arabic Morphology

Computational Analyses of Arabic Morphology

Paper · 1994 · George A. Kiraz

Summary

Standard two-level morphology treats a word as morphemes joined end to end. Arabic does not work that way: a stem interleaves a consonantal root such as {ktb}, a vowel melody such as {ui}, and a pattern. Kiraz shows that a multi-tape two-level formalism can describe this compactly. His formalism builds on the rule notation of Pulman-Hepple, Ruessink and Black et al., and puts each autonomous morpheme on its own lexical tape.

He writes three illustrative grammars, each following an analysis by McCarthy and Prince:

  • a CV-based grammar for verbal stems,
  • a moraic grammar for nominal stems,
  • an affixational grammar for verbs.

He also gives what he presents as the first computational treatment of the Arabic broken plural. The plural is produced by quietly deriving the singular from the lexicon and mapping it to the plural surface form.

His comparison favours moraic grammars for templatic stems and affixational grammars for a-templatic ones. The work matters because it shows that finite-state, two-level machinery can handle non-concatenative morphology without giving up its formal tractability.

Key points

  • The formalism generalises lexical expressions in rules to n-tuples, one per tape. Each tape carries its own feature structure, and the left lexical context may contain an ellipsis that skips intervening tuples.
  • The implementation (after Trost 1990) has two modules: a multi-tape two-level component for non-linear phenomena, and a shift-reduce parser with unification-based grammars for morphotactics.
  • The grammars differ in their pattern lexicons and rule granularity:
    • The CV grammar needs indexed verbal templates, which gives the largest pattern lexicon, and each rule licenses a single surface character.
    • Moraic rules license whole syllables, so fewer lexical/surface subsequences are produced, which the author ties to lower time complexity.
    • The affixational grammar keeps only one template and stores affixes in a separate lexicon.
  • Broken plural rules carry the feature [number=pl] and are obligatory. Bimoraic stems read their singular melody from tape 2 and a lexically marked plural melody from tape 4.
  • Partial vocalisation, where short vowels are omitted in writing, is not covered here; the author points to Kiraz (1994).

Interesting ideas

  • I1. Placing interleaved morphemes (root, vowel melody, pattern) on separate parallel tapes turns a non-concatenative word into several simple aligned strings, a design that could model any morphology where morphemes do not appear in sequence.
  • I2. Rules that license larger units, such as syllables instead of single characters, reduce the number of candidate segmentations and therefore the search cost; choosing the right granularity for rules is itself an efficiency decision.
  • I3. A permissive finite-state stage may overgenerate analyses on purpose (any vowel unifies), leaving a unification grammar to reject the ill-formed ones; the division of labour keeps each module simple.
  • I4. The broken plural is generated only after confirming that a valid singular exists in the lexicon, so derived forms are tied to an attested base rather than produced freely.
  • I5. The same data is analysed under three competing linguistic theories, and the resulting grammars are compared on lexicon size and rule granularity, which turns a theoretical choice into a measurable engineering one.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (tokens): English strong-verb ablaut (sing/sang/sung) could be analysed as a fixed consonant frame plus a vowel tier, so the lemmatizer can explain exactly which part signals tense.
  • I2 (tokens): morphology rules in crate en could match suffix or syllable chunks rather than letters, which would shrink the analysis lattice while keeping each step traceable.
  • I3 (tree): the tagger could pass several morphological readings forward, and parser feature constraints could prune them, with each rejection logged as an explanation.
  • I4 (back): when generating irregular plurals or derived forms, the generator could require a licensed base entry and record that derivation path, avoiding invented words.
  • I5 (learn): alternative rule sets for a component could be compared on size and number of candidates produced, not just accuracy, before one is adopted.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 2Research subjects and languages the work is about.Semitic languages · 2,908 worksMorphology · 405 works

Cite

Cite as George A. Kiraz. 1994. Computational Analyses of Arabic Morphology. arXiv:cmp-lg/9408002.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.cmp-lg/9408002 · arXiv cmp-lg/9408002

@misc{kiraz1994computational,
  title = {{Computational Analyses of Arabic Morphology}},
  author = {George A. Kiraz},
  year = {1994},
  doi = {10.48550/arxiv.cmp-lg/9408002},
  eprint = {cmp-lg/9408002},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.cmp-lg/9408002}
}

References

25 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Beesley, K. (1991). Computer analysis of arabic morphology. In Comrie, B. and Eid, M., eds., Perspectives on Arabic Linguistics III: Papers from the Third Annual Symposium on Arabic Linguistics . Benjamins, Amsterdam. ↗
  2. Bird, S. and Ellison, T. (1994). One-level phonology. Computational Linguistics , 20(1), pp. 55{90. ↗
  3. Black, A., Ritchie, G., Pulman, S., and Russel, G. (1987). Formalisms for morphographemic description. In Proceedings of the Third Conference of the European Chapter of the Association for Computational Linguistics , pp. 11{8. find
  4. Goldsmith, J. (1976). Autosegmental Phonology . PhD thesis, MIT. Published as Autosegmental and Metrical Phonology , Oxford 1990. find
  5. Kay, M. (1987). Nonconcatenative nite-state morphology. In Proceedings of the Third Conference of the European Chapter of the Association for Computational Linguistics , pp. 2{10. find
  6. Kiraz, G. (1994). Multi-tape two-level morphology: a case study in Semitic non-linear morphology. In COLING-94: Papers Presented to the 15th International Conference on Computational Linguistics . ↗
  7. Knuth, D. (1973). The Art of Computer Programming , vol. 3. Addison-Wesley. find
  8. Kornai, A. (1991). Formal Phonology . PhD thesis, Stanford University. find
  9. Koskenniemi, K. (1983). Two Level Morphology . PhD thesis, University of Helsinki. find
  10. McCarthy, J. (1981). A prosodic theory of nonconcatenative morphology. Linguistic Inquiry , 12(3), pp. 373{418. find
  11. McCarthy, J. (1993). Template form in prosodic morphology. In L. Stvan et al. , eds., Papers from the Third Annual Formal Linguistics Society of Midamerica Conference , pp. 187{218. Indiana University Linguistics Club, Bloomington. find
  12. McCarthy, J. and Prince, A. (1990a). Foot and word in prosodic morphology: The Arabic broken plural. Natural Language and Linguistic Theory , 8, pp. 209{83. ↗
  13. McCarthy, J. and Prince, A. (1990b). Prosodic morphology and templatic morphology. In Eid, M. and McCarthy, J., eds., Perspectives on Arabic Linguistics II: Papers from the Second Annual Symposium on Arabic Linguistics , pp. 1{54. Benjamins, Amsterdam. ↗
  14. Mellish, C. (1988). Implementing systemic classi cation by uni cation. Computational Linguistics , 14(1), pp. 40{51. find
  15. Narayanan, A. and Hashem, L. (1993). On abstract nite-state morphology. In Proceedings of the Sixth Conference of the European Chapter of the Association for Computational Linguistics , pp. 297{304. find
  16. Pulman, S. (1994). Expressivity of lean formalisms. In Markantonatou, S. and Sadler, L., eds., Grammatical Formalisms: Issues in Migration , Studies in Machine Translation and Natural Language Processing: 4, pp. 35{59. Commission of the European Communities. find
  17. Pulman, S. and Hepple, M. (1993). A feature-based formalism for two-level phonology: a description and implementation. Computer Speech and Language , 7, pp. 333{58. ↗
  18. Ritchie, G. (1992). Languages generated by two-level morphological rules. Computational Linguistics , 18(1), pp. 41{59. find
  19. Ritchie, G., Black, A., Russell, G., and Pullman, S. (1992). Computational Morphology: Practical Mechanisms for the English Lexicon . MIT Press, Cambridge Mass. find
  20. Ruessink, H. (1989). Two level formalisms. Technical Report 5, Utrecht Working Papers in NLP. find
  21. Sproat, R. (1992). Morphology and Computation . MIT Press, Cambridge Mass. find
  22. Trost, H. (1990). The application of two-level morphology to non-concatenative German morphology. In Karlgren, H., ed., COLING-90: Papers Presented to the 13th International Conference on Computational Linguistics , vol. 2, pp. 371{6. ↗
  23. Wehr, H. (1971). A Dictionary of Modern Written Arabic . Spoken Language Services, Ithaca. find
  24. Wiebe, B. (1992). Modelling autosegmental phonology with multi-tape nite state transducers. Master's thesis, Simon Fraser University. find
  25. Wright, W. (1988). A Grammar of the Arabic Language . Cambridge University Press, Cambridge, 3 edition. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/kiraz-1994-computational-analyses-arabic-morphology/