--- title: "A Syntactic Lexicon for Arabic Verbs — Loukil 2010" description: "Such resources are expensive to build, and each grammar framework tends to use its own format. 2 references, cited by 2 papers here." authors: ["Noureddine Loukil", "Kais Haddar", "Abdelmajid Benhamadou"] year: 2010 venue: "Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC'10)" source: "https://aclanthology.org/L10-1601/" licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)" licence_url: "https://creativecommons.org/licenses/by-nc/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/loukil-2010-syntactic-lexicon-arabic-verbs/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/loukil-2010-syntactic-lexicon-arabic-verbs/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2010](https://md.research.elamur.ai/free/en/papers/2010/page/2/) / A Syntactic Lexicon for Arabic Verbs # A Syntactic Lexicon for Arabic Verbs **Authors:** Noureddine Loukil, [Kais Haddar](https://md.research.elamur.ai/free/en/authors/kais-haddar/) and Abdelmajid Benhamadou · **Year:** 2010 · **Venue:** Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC'10) · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/) **Links:** [Publisher's page](https://aclanthology.org/L10-1601/) · [PDF](https://aclanthology.org/L10-1601.pdf) · [OpenAlex](https://openalex.org/W13202936) **Citations:** 6 citations (OpenAlex, 2026-10-14) · 2 in our library, of the 140,884 papers whose bibliographies we read ## Summary The paper tackles the lack of reusable, shareable syntactic lexicons for Arabic. Such resources are expensive to build, and each grammar framework tends to use its own format. The authors encode Arabic verb syntax in the Syntax extension of the ISO Lexical Markup Framework (LMF). The model rests on a verbal type hierarchy (triliteral vs. quadriliteral, sound vs. defective, voice, mood) and on subcategorization frames that list ordered arguments together with constraints such as which prepositional particle may introduce a complement. The lexicon was built semi-automatically in the Lexus editor and holds 2500 verb lemmas. To show the LMF data is practically usable, they wrote a rule-based converter that maps LMF entries into Type Description Language (TDL). The output is loaded into a pre-existing Arabic HPSG grammar implemented on the LKB platform. On a 205-sentence test set, the generated lexicon parsed at least as well as a hand-written one. This suggests a standards-based lexicon can drive a deep grammar without loss of quality, while staying exchangeable between projects. ## Key points - The lexicon covers 2500 Arabic verb lemmas, defines 17 subcategorization frames (SFs) and averages 2.7 SFs per verb. Entries were edited in Lexus, which checks LMF compatibility and outputs XML. - An Arabic verb takes at most three required complements. A complement is either a direct NP or is introduced by a particle from a closed set (min, ila, bi, ala, an, fi, lem). A per-verb restriction property lists the allowed particles; for example, taharraka allows only ala, bi, min, lem. - Every verb carries two sets of frames: one for verbal sentences (VSO, the more common order) and one for nominal sentences (SVO). - The LMF→TDL conversion uses 32 rules. Each rule has an IN path into the LMF data and an OUT TDL feature structure; for example, PartOfSpeech maps to MAJ. A syntactic behaviour id becomes inheritance from a TDL type with the same name. - Evaluation used the LKB HPSG grammar of Boukedi et al. on 205 transliterated sentences drawn from a 781-word vocabulary: - **Manual lexicon:** 175 sentences got one parse (about 85%), 25 got none and 5 got two. - **Generated lexicon:** 190 got one parse, 17 got none and 8 got two. - The authors call this a slight improvement in one passage, yet elsewhere say no clear improvement is visible. They blame the remaining failures on missing grammar rules, for instance for relatives and coordination. ## Interesting ideas - Frame patterns are stored once as named syntactic behaviours and shared by many verbs, so a verb entry only points to the frames it accepts and the redundancy lives in the type system. - Format conversion is written as declarative rule pairs: a source path and a target feature path. Changing the mapping then means editing rules rather than code, which keeps the converter transparent and easy to extend. - Each verb carries an explicit list of the prepositions allowed to introduce its complements, instead of a generic "takes a PP" flag. - The same verb gets separate frame sets for different clause word orders, so constructional variation is modelled directly rather than left to the parser. - The generated resource is tested by swapping it into a fixed grammar and comparing against a hand-built resource on the same sentences. Results are counted by number of analyses (0, 1, 2), which separates coverage gaps from ambiguity. ## For Mova How these ideas could serve Mova, our small open language model — its files, stages and tests: - **I1** (world): verb-class knowledge could keep VerbNet-like frame patterns as shared, named objects that lemmas reference by id, so a frame bug is fixed once for all its verbs. - **I2** (tree): UD dialect and treebank conversion could be expressed as such inspectable path-to-path rules, giving every converted tree a traceable rule for each change. - **I3** (frames): mapping obl/nmod dependents to semantic roles could check a per-verb preposition whitelist, which helps separate arguments from adjuncts and explains rejected readings. - **I4** (frames): inverted, fronted or passive clause patterns could have their own frame variants linked to one predicate, keeping role assignment explicit across constructions. - **I5** (learn): silver or induced lexicons could be evaluated by A/B replacement in a frozen pipeline, reporting no-parse and multi-parse counts as distinct failure signals. *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: [Linguistics & languages](https://md.research.elamur.ai/free/en/shelves/linguistics/) Topics: [🐪 Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) ## Institutions - [University of Sfax](https://md.research.elamur.ai/free/en/universities/university-of-sfax/) — University, Sfax, Tunisia · 3 works · 17 citations ## Cite Cite as: Noureddine Loukil, Kais Haddar and Abdelmajid Benhamadou. 2010. A Syntactic Lexicon for Arabic Verbs. Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC'10). The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{loukil2010syntactic, title = {{A Syntactic Lexicon for Arabic Verbs}}, author = {Noureddine Loukil, Kais Haddar and Abdelmajid Benhamadou}, booktitle = {Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC'10)}, year = {2010}, doi = {10.63317/56hwn2hwsztn}, url = {https://doi.org/10.63317/56hwn2hwsztn} } ``` - [← previous Towards Semi-Automated Annotation for Prepositional Phrase Attachment](https://md.research.elamur.ai/free/en/papers/rosenthal-2010-semi-automated-annotation-prepositional-phrase/) - [next → Morphological Analysis Can Improve a CCG Parser for English](https://md.research.elamur.ai/free/en/papers/honnibal-2010-morphological-analysis-can-improve-ccg/) - [↑ back to papers of 2010](https://md.research.elamur.ai/free/en/papers/2010/page/2/) ## References 2 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. G. Francopoulo. 2005. Lexical markup framework (lmf = iso 24-613). Technical report, INRIA Loria. [our card](https://md.research.elamur.ai/free/en/papers/francopoulo-2006-lexical-markup-framework-lmf/) 2. H-U. Krieger and U. Schafer. 1994. Tdl: A type description language for hpsg. part 2: User guide. Technical report, German’s research center for artificial intelligence. [find](https://scholar.google.com/scholar?q=H%2DU%2E+Krieger+and+U%2E+Schafer%2E+1994%2E+Tdl%3A+A+type+description+language+for+hpsg%2E+part+2%3A+User+guide%2E+Technical+report%2C+German%E2%80%99s+research+center+for+artificial+intelligence%2E)