Elamur Research⌕
☰
home / papers / 2018 / Automatic Annotation of Locative and Directional Expressions in Arabic

Automatic Annotation of Locative and Directional Expressions in Arabic

Paper · 2018 · Rita Hijazi, Amani Sabra and Moustafa Al-Hajj

Summary

The paper tackles automatic recognition of spatial language in Arabic, focusing on expressions of location and direction. This area has received little attention for Arabic, and it is hard because many prepositions also carry temporal or abstract meanings. The authors first carry out a descriptive linguistic analysis of Arabic spatial prepositions, using the classic split between topological relations (internal localisation) and projective relations (external localisation). They organise the findings into a semantic map of the spatial domain. They then encode the map and the analysis as local grammars in NooJ, a platform for finite-state transducers, so that matching text segments are tagged automatically. The data is a 2015 Arabic novel by Antoine Douaihy, chosen because narrative prose is dense with spatial prepositions, adverbs and deictics. The authors describe the recognition results as very encouraging. They also argue that syntactic analysis is indispensable for this kind of annotation. The work matters as an early, explicit, rule-based resource for spatial semantics in Arabic.

Key points

  • The method is rule-based: NooJ local grammars and finite-state transducers, guided by a hand-built semantic map of spatiality.
  • The corpus is the novel "La naufragée du lac Murray" (Antoine Douaihy, 2015). Of the corpus, 75% was used to build the lexical lists and patterns, and the remaining 25% was held out for evaluation.
  • The relation taxonomy has two branches:
    • Topological relations, split into support and inclusion.
    • Projective relations, split into distance and orientational. Orientational relations place the target on a lateral, vertical or frontal axis.
  • The analysis draws several fine distinctions:
    • على ("on") requires contact between figure and ground, while فوق ("above") does not. A substitution test is used to check this.
    • بين ("between") expresses inclusion relative to a plural ground, which the authors call distribution.
  • Ambiguity is the main source of difficulty. Prepositions such as بين, عند and وسط have spatial, temporal and abstract uses, and inclusion can be expressed by more than four different prepositions. Exact scores are not given in the available extract.

Interesting ideas

  • I1. A substitution test separates contact from non-contact readings: if "above" can be swapped for "on" without changing the meaning, the relation involves contact. This gives a cheap, testable criterion for a semantic feature.
  • I2. Spatial meaning is organised as a small hierarchy. Topological relations divide into support and inclusion, and projective relations divide into distance and orientation along three axes. This makes a compact target inventory instead of one label per preposition.
  • I3. The same preposition can be spatial, temporal or abstract, and the paper concludes that syntactic context is needed to tell these apart.
  • I4. The rules and lexical lists are built on 75% of the corpus and scored only on an untouched 25%. This keeps a hand-crafted rule system honest about generalisation.
  • I5. "Between" is modelled as inclusion relative to a plural or coordinated ground. Number and coordination therefore become part of the meaning of the relation.

For Mova

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

  • I1 (world): spatial relations could store an explicit contact flag, so that "the cup on the table" and "the lamp above the table" produce different world states that a reader can question.
  • I2 (world): the same hierarchy could serve as the ontology for locative frames, so that answers such as "inside", "next to" or "in front of" map onto typed relations.
  • I3 (frames): when labelling roles for obl dependents, the preposition sense could be decided from the noun class of its object (a place, a time expression or an abstract noun) rather than from the preposition alone.
  • I4 (learn): self-induced repair rules could be developed and induced on one slice of the data and judged only on a held-out slice, with negative controls.
  • I5 (tree): the parser could check whether the object of "between" is plural or has a conj structure, and flag parses where it is not as likely errors.

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 · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 works

🧭 Topics 1Research subjects and languages the work is about.Semitic languages · 3,066 works

Cite

Cite as Rita Hijazi, Amani Sabra and Moustafa Al-Hajj. 2018. Automatic Annotation of Locative and Directional Expressions in Arabic. arXiv:1805.06344.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.1805.06344 · arXiv 1805.06344

@misc{hijazi2018automatic,
  title = {{Automatic Annotation of Locative and Directional Expressions in Arabic}},
  author = {Rita Hijazi, Amani Sabra and Moustafa Al-Hajj},
  year = {2018},
  doi = {10.48550/arxiv.1805.06344},
  eprint = {1805.06344},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.1805.06344}
}

References

20 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Alhajj M, Mourad G. (2015). Extraction of reported speeches from Arabic Lebanese newspapers. IEEE The Fifth International Conference on Digital Information and Communication Technology and its Applications (DICTAP2015), Faculty of Engineering - Lebanese University, Lebanon, April 29 - May 1, 2015 Proceedings. find
  2. Alhajj M, Sabra A., (2018). Automatique Identification of Arabic Expressions Related to Future Events in Lenanon’s Economy. International Journal of Science and Research (IJSR), volume 7, Issue 4. find
  3. Asic T., (2004). La représentation cognitive du temps et de l'espace: étude pragmatique des données linguistiques en français et dans d'autres langues. Linguistique. Université Lumière - Lyon II. Français. find
  4. Atanassova I., (2012). Exploitation informatique des annotations sémantiques automatiques d’Excom pour la recherche d’informations et la navigation. Thèse de doctorat, Université Paris-Sorbonne. ↗
  5. Aurnague M. (1996). Les Noms de Localisation Interne : tentative de caractérisation sémantique à partir de données du basque et du français. Cahiers de Lexicologie, Centre National de la Recherche Scientifique, pp.159-192. find
  6. Aurnague M., Vieu L., Borillo A. (1997). La représentation formelle des concepts spatiaux dans la langue, in Langages et cognition spatiale, Masson, Paris, pp.69-102. find
  7. Aurnague M. (2004). Les structures de l’espace linguistique : regards croisés sur quelques constructions spatiales du basque et du français. Louvain/Paris, Peeters (Bibliothèque de l’Information Grammaticale, 56). find
  8. Berners-Lee T., Hendler J.A., Lassila O. (2001), The Semantic Web. Scientific American, 284(5), pp. 34-43. ↗
  9. Borillo A. (1988). Le lexique de l'espace : les noms et adjectifs de localisation interne, Cahiers de grammaire n°13, Université Toulouse-Le Mirail, pp. 3-22. find
  10. Borillo A. (1998). L'espace et son expression en français. L'essentiel, Ophrys. find
  11. Desclés J.-P. (2006a). Contextual exploration processing for discourse automatic annotations of texts. In FLAIRS 2006, Invited speaker, Melbourne, Floride. AAAI Press. find
  12. Herskovits A. (1986). Language and Spatial Cognition. An Interdisciplinary Study of Prepositions in English. Cambridge : Cambridge University Press. find
  13. Herskovits A. (1997). Language, spatial cognition, and vision. In O. Stock (ed.), Spatial and temporal reasoning, Dordrecht : Kluwer Academic Publishers, 155-202. ↗
  14. Jackendoff R. (1991). « Parts and boundaries », Cognition 41, 9-45. find
  15. Jackendoff R. (1996). « The Architecture of linguistic–spatial interface », in BLOOM P., PETERSON M.A., NADEL L. & GARRETT M.F. (eds), Language and Space, Cambridge (Mass.), The MIT Press, 2-32. find
  16. Kopecka A. (2004). Etude typologique de l’expression de l’espace ; Localisation et déplacement en français et en polonais. Thèse de doctorat, Université Lumière, Lyon2. find
  17. Talmy L. (2000). Toward a Cognitive Semantics, Cambridge (Mass.), The MIT Press. find
  18. Vandeloise C. (1986). L’espace en français : sémantique des prépositions spatiales, Paris : Seuil. find
  19. Vandeloise C. (1992). Les analyses de la préposition dans : Faits linguistiques et effets de méthodologie, Lexique 11, 15-40. find
  20. Vandeloise C. (1999). Quand dans quitte l’espace pour le temps. Approches sémantiques des prépositions, Revue de Sémantique et Pragmatique 6, 145-163. ،ةيملعلا بتكلا راد ،برعلا ملاك ةفرعم يف بهذلا روذش حرش ،) 2004 ( ماشه نبا ،يراصنلأا .توريب .اهلامعتسا يف ةاحنلا بهاذمو رجلا فورح .) 1988 ( نيسحلا دبع ،كرابم 1 ط ،توريب ،يدمحملا ةبتكم ،يفاولا وحنلا ،) 2007 ( سابع ، نسح find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/hijazi-2018-automatic-annotation-locative-directional-expressions/