---
title: "Metaphorical Expressions in Automatic Arabic — Alsiyat 2020"
description: "This paper asks whether current Arabic sentiment analysis tools can handle metaphorical language."
authors: ["Israa Alsiyat", "Scott Piao"]
year: 2020
venue: "Proceedings of the Twelfth Language Resources and Evaluation Conference"
source: "https://aclanthology.org/2020.lrec-1.604/"
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/alsiyat-2020-metaphorical-expressions-automatic-arabic-sentiment/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/alsiyat-2020-metaphorical-expressions-automatic-arabic-sentiment/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2020](https://md.research.elamur.ai/free/en/papers/2020/page/6/) / Metaphorical Expressions in Automatic Arabic Sentiment Analysis

# Metaphorical Expressions in Automatic Arabic Sentiment Analysis

**Authors:** [Israa Alsiyat](https://md.research.elamur.ai/free/en/authors/israa-alsiyat/), [Scott Piao](https://md.research.elamur.ai/free/en/authors/scott-piao/) · **Year:** 2020 · **Venue:** Proceedings of the Twelfth Language Resources and Evaluation Conference · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2020.lrec-1.604/) · [PDF](https://aclanthology.org/2020.lrec-1.604.pdf) · [OpenAlex](https://openalex.org/W3029754862)

**Citations:** 4 citations (OpenAlex, 2026-10-14) · [4](https://md.research.elamur.ai/free/en/papers/alsiyat-2020-metaphorical-expressions-automatic-arabic-sentiment/cited-by/) in our library, of the 140884 papers whose bibliographies we read

## Summary
This paper asks whether current Arabic sentiment analysis tools can handle metaphorical language. The authors argue that metaphor is especially common in Arabic for historical and cultural reasons. A metaphor's intended polarity often differs from the literal polarity of its words, which confuses systems that rely on surface cues.

To test this, they built a small, carefully chosen set of 40 opinionated sentences drawn from the LABR and HAAD book-review datasets. Half contain metaphors and half do not. They ran the set through Mazajak, a publicly available deep-learning sentiment tool with state-of-the-art results.

Mazajak performed noticeably worse on the metaphorical half. Manual error analysis showed that it labels sentences by the literal sentiment of individual words. For example, it read "on the injury" as negative and "most dangerous autobiography" as neutral. It succeeded only when the surface words already pointed the right way.

The authors propose a pipeline that first detects metaphor, recovers the hidden meaning, and only then classifies sentiment. They suggest Wilks-style co-occurrence statistics for detection. The work matters because it documents a resource gap in Arabic NLP.

## Key points
- Test set: 40 Arabic sentences (20 with metaphors, 20 without), selected from LABR (about 63,000 book reviews) and HAAD (2,389 aspect-annotated book reviews, a LABR subset).
- Tool evaluated: Mazajak, a web-based deep-learning Arabic sentiment classifier with strong results on SemEval 2017 and ASTD.
- Result: classification quality dropped clearly on sentences containing metaphors compared with literal ones.
- Typical errors:
  - "On the injury" (meaning "on point") was labelled negative.
  - "Capsules of advice" was labelled neutral, though the intended sense is positive.
  - "Licking your brain" (meaning confusion) was labelled neutral.
- Taxonomy: a simile keeps both the compared entity and the image. A metaphor drops one of them, giving implicit or explicit metaphor.

## Interesting ideas
- Build a minimal paired test set where one half contains the target phenomenon and the other half is a matched literal control, so any score gap can be attributed to that phenomenon. — For Mova (learn): Evaluation could pair each figurative test item with a literal control, a cheap negative-control design that shows whether a gain comes from real handling of the phenomenon.
- Diagnose surface reliance by checking whether errors cluster where literal word polarity disagrees with intended meaning; success only on agreeing cases reveals a shallow model. — For Mova (answer): Fable and tale QA tests could be split by literal/intended agreement to confirm the readers interpret meaning rather than matching keywords.
- Treat a weak statistical bond between a verb and its noun arguments (e.g., "shave" with "life") as a signal that the phrase is likely metaphorical. — For Mova (frames): A VerbNet selectional-restriction violation in a filled frame could raise an explainable "possible metaphor" flag before roles are taken literally.
- Route input through a metaphor gate: if a metaphor is detected, resolve its hidden sense first, otherwise use the normal path. — For Mova (frames): The prag crate could add a figurative-detection step next to irony handling, so frames are built from the resolved sense and the routing decision stays traceable.
- Define metaphor as a simile with either the real referent or the image omitted, distinguishing implicit and explicit types. — For Mova (world): When building entities, a missing tenor (as in "the lion shoots the enemy") could be recovered by linking the image back to a contextual referent instead of creating a literal lion.

*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: [Sentiment & emotion](https://md.research.elamur.ai/free/en/topics/sentiment-emotion/) · [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · [Idioms & multiword expressions](https://md.research.elamur.ai/free/en/topics/multiword-expressions/)

## Institutions

- [Lancaster University](https://md.research.elamur.ai/free/en/universities/lancaster-university/) — University, Lancaster, United Kingdom · 13 works · 1 069 citations

## Cite

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{alsiyat2020metaphorical,
  title = {{Metaphorical Expressions in Automatic Arabic Sentiment Analysis}},
  author = {Alsiyat, Israa and Piao, Scott},
  booktitle = {Proceedings of the Twelfth Language Resources and Evaluation Conference},
  year = {2020},
  url = {https://aclanthology.org/2020.lrec-1.604/}
}
```

- [← previous Integrating Graph-Based and Transition-Based Dependency Parsers in the Deep Contextualized Era](https://md.research.elamur.ai/free/en/papers/falenska-2020-integrating-graph-based-transition-based/)
- [next → Model-based Annotation of Coreference](https://md.research.elamur.ai/free/en/papers/aralikatte-2020-model-based-annotation-coreference/)
- [↑ back to papers of 2020](https://md.research.elamur.ai/free/en/papers/2020/page/6/)

## References

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

1. Abdul-Mageed, M., and Diab, M. T. (2012). AWATIF: A Multi-Genre Corpus for Modern Standard Arabic Subjectivity and Sentiment Analysis. In LREC (Vol. 515, pp. 3907-3914). [find](https://scholar.google.com/scholar?q=Abdul%2DMageed%2C+M%2E%2C+and+Diab%2C+M%2E+T%2E+%282012%29%2E+AWATIF%3A+A+Multi%2DGenre+Corpus+for+Modern+Standard+Arabic+Subjectivity+and+Sentiment+Analysis%2E+In+LREC+%28Vol%2E+515%2C+pp%2E+3907%2D3914%29%2E)
2. Al-Harrasi, A. N. K. (2001). Metaphor in (Arabic-into-English) translation with specific reference to metaphorical concepts and expressions in political discourse (Doctoral dissertation, University of Aston in Birmingham). [find](https://scholar.google.com/scholar?q=Al%2DHarrasi%2C+A%2E+N%2E+K%2E+%282001%29%2E+Metaphor+in+%28Arabic%2Dinto%2DEnglish%29+translation+with+specific+reference+to+metaphorical+concepts+and+expressions+in+political+discourse+%28Doctoral+dissertation%2C+University+of)
3. Alkhatib, M., and Shaalan, K. (2016). Natural language processing for Arabic metaphors: a conceptual approach. In International Conference on Advanced Intelligent Systems and Informatics, Springer, Cham. pp. 170-181. [find](https://scholar.google.com/scholar?q=Alkhatib%2C+M%2E%2C+and+Shaalan%2C+K%2E+%282016%29%2E+Natural+language+processing+for+Arabic+metaphors%3A+a+conceptual+approach%2E+In+International+Conference+on+Advanced+Intelligent+Systems+and+Informatics%2C+Springer%2C+Ch)
4. Alowisheq, A., Alhumoud, S., Altwairesh, N., and Albuhairi, T. (2016). Arabic sentiment analysis resources: a survey. In Proceedings of The Social Computing and Social Media: The 8th International Conference, SCSM 2016, Held as Part of HCI International 2016, Toronto, ON, Canada, July 17–22, 2016. Springer International Publishing. pp. 267-278. [find](https://scholar.google.com/scholar?q=Alowisheq%2C+A%2E%2C+Alhumoud%2C+S%2E%2C+Altwairesh%2C+N%2E%2C+and+Albuhairi%2C+T%2E+%282016%29%2E+Arabic+sentiment+analysis+resources%3A+a+survey%2E+In+Proceedings+of+The+Social+Computing+and+Social+Media%3A+The+8th+International+Con)
5. Al-Smadi, M., Qawasmeh, O., Talafha, B., and Quwaider, M. (2015). Human annotated arabic dataset of book reviews for aspect based sentiment analysis. In Proceedings of The 3rd International Conference 2015 on Future Internet of Things and Cloud, pp. 726-730. IEEE. [find](https://scholar.google.com/scholar?q=Al%2DSmadi%2C+M%2E%2C+Qawasmeh%2C+O%2E%2C+Talafha%2C+B%2E%2C+and+Quwaider%2C+M%2E+%282015%29%2E+Human+annotated+arabic+dataset+of+book+reviews+for+aspect+based+sentiment+analysis%2E+In+Proceedings+of+The+3rd+International+Conference)
6. Aly, M., and Atiya, A. (2013). Labr: A large scale arabic book reviews dataset. In Proceedings of The 51st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) (Vol. 2), pp. 494-498. [find](https://scholar.google.com/scholar?q=Aly%2C+M%2E%2C+and+Atiya%2C+A%2E+%282013%29%2E+Labr%3A+A+large+scale+arabic+book+reviews+dataset%2E+In+Proceedings+of+The+51st+Annual+Meeting+of+the+Association+for+Computational+Linguistics+%28Volume+2%3A+Short+Papers%29+%28Vol)
7. Beust, P., Ferrari, S., and Perlerin, V. (2003). NLP model and tools for detecting and interpreting metaphors in domain-specific corpora. In Proceedings of the Corpus Linguistics 2003 Conference, pp. 114-123. [find](https://scholar.google.com/scholar?q=Beust%2C+P%2E%2C+Ferrari%2C+S%2E%2C+and+Perlerin%2C+V%2E+%282003%29%2E+NLP+model+and+tools+for+detecting+and+interpreting+metaphors+in+domain%2Dspecific+corpora%2E+In+Proceedings+of+the+Corpus+Linguistics+2003+Conference%2C+pp%2E+)
8. Elarnaoty, M., AbdelRahman, S., and Fahmy, A. (2012). A machine learning approach for opinion holder extraction in Arabic language. arXiv preprint arXiv:1206.1011. [↗](https://arxiv.org/abs/1206.1011)
9. El-Masri, M., Altrabsheh, N., Mansour, H., and Ramsay, A. (2017). A web-based tool for Arabic sentiment analysis. Procedia Computer Science, 117, pp. 38-45. [find](https://scholar.google.com/scholar?q=El%2DMasri%2C+M%2E%2C+Altrabsheh%2C+N%2E%2C+Mansour%2C+H%2E%2C+and+Ramsay%2C+A%2E+%282017%29%2E+A+web%2Dbased+tool+for+Arabic+sentiment+analysis%2E+Procedia+Computer+Science%2C+117%2C+pp%2E+38%2D45%2E)
10. Farha, I. A., and Magdy, W. (2019). Mazajak: An Online Arabic Sentiment Analyser. In Proceedings of the Fourth Arabic Natural Language Processing Workshop, pp. 192-198. [find](https://scholar.google.com/scholar?q=Farha%2C+I%2E+A%2E%2C+and+Magdy%2C+W%2E+%282019%29%2E+Mazajak%3A+An+Online+Arabic+Sentiment+Analyser%2E+In+Proceedings+of+the+Fourth+Arabic+Natural+Language+Processing+Workshop%2C+pp%2E+192%2D198%2E)
11. Faycel, D. (2012). Food metaphors in Tunisian Arabic proverbs. Rice Working Papers in Linguistics, (Vol. 3). [find](https://scholar.google.com/scholar?q=Faycel%2C+D%2E+%282012%29%2E+Food+metaphors+in+Tunisian+Arabic+proverbs%2E+Rice+Working+Papers+in+Linguistics%2C+%28Vol%2E+3%29%2E)
12. Gholami, S., Montashery, I., and Khorrami, M. (2016). Metaphors in Arabic and English texts: A case study of translation of metaphors in the English versions of Al-Sahifa Al-Sajjadiyyah. Journal of Literature, Languages and Linguistics (Vol. 21), pp. 56-66. [find](https://scholar.google.com/scholar?q=Gholami%2C+S%2E%2C+Montashery%2C+I%2E%2C+and+Khorrami%2C+M%2E+%282016%29%2E+Metaphors+in+Arabic+and+English+texts%3A+A+case+study+of+translation+of+metaphors+in+the+English+versions+of+Al%2DSahifa+Al%2DSajjadiyyah%2E+Journal+of+Li)
13. Jakobson, R., and Halle, M. (1971). Fundamentals of language. The Hague: Mouton. [find](https://scholar.google.com/scholar?q=Jakobson%2C+R%2E%2C+and+Halle%2C+M%2E+%281971%29%2E+Fundamentals+of+language%2E+The+Hague%3A+Mouton%2E)
14. Lakoff, G., and Johnson, M. (1980). Conceptual metaphor in everyday language. The journal of Philosophy, 77(8), pp. 453-486. [find](https://scholar.google.com/scholar?q=Lakoff%2C+G%2E%2C+and+Johnson%2C+M%2E+%281980%29%2E+Conceptual+metaphor+in+everyday+language%2E+The+journal+of+Philosophy%2C+77%288%29%2C+pp%2E+453%2D486%2E)
15. Maamouri, M., Bies, A., Buckwalter, T., and Mekki, W. (2004). The penn arabic treebank: Building a large-scale annotated arabic corpus. In NEMLAR conference on Arabic language resources and tools (Vol. 27), pp. 466-467. [find](https://scholar.google.com/scholar?q=Maamouri%2C+M%2E%2C+Bies%2C+A%2E%2C+Buckwalter%2C+T%2E%2C+and+Mekki%2C+W%2E+%282004%29%2E+The+penn+arabic+treebank%3A+Building+a+large%2Dscale+annotated+arabic+corpus%2E+In+NEMLAR+conference+on+Arabic+language+resources+and+tools+%28Vol)
16. Raii, J. (2009). Metaphor in day-to-day Arabic speech: a conceptual approach. Tishreen University Journal for Research and Scientific Studies-Arts and Humanities Series, 31(1), pp. 175-193. [find](https://scholar.google.com/scholar?q=Raii%2C+J%2E+%282009%29%2E+Metaphor+in+day%2Dto%2Dday+Arabic+speech%3A+a+conceptual+approach%2E+Tishreen+University+Journal+for+Research+and+Scientific+Studies%2DArts+and+Humanities+Series%2C+31%281%29%2C+pp%2E+175%2D193%2E)
17. Refaee, E., and Rieser, V. (2014). An arabic twitter corpus for subjectivity and sentiment analysis. In Proceedings of The LREC 2014. pp. 2268-2273. [find](https://scholar.google.com/scholar?q=Refaee%2C+E%2E%2C+and+Rieser%2C+V%2E+%282014%29%2E+An+arabic+twitter+corpus+for+subjectivity+and+sentiment+analysis%2E+In+Proceedings+of+The+LREC+2014%2E+pp%2E+2268%2D2273%2E)
18. Rushdi‐Saleh, M., Martín‐Valdivia, M. T., Ureña‐López, L. A., and Perea‐Ortega, J. M. (2011). OCA: Opinion corpus for Arabic. Journal of the American Society for Information Science and Technology, 62(10), pp. 2045-2054. [find](https://scholar.google.com/scholar?q=Rushdi%E2%80%90Saleh%2C+M%2E%2C+Mart%C3%ADn%E2%80%90Valdivia%2C+M%2E+T%2E%2C+Ure%C3%B1a%E2%80%90L%C3%B3pez%2C+L%2E+A%2E%2C+and+Perea%E2%80%90Ortega%2C+J%2E+M%2E+%282011%29%2E+OCA%3A+Opinion+corpus+for+Arabic%2E+Journal+of+the+American+Society+for+Information+Science+and+Technology%2C+62)
19. Wilks, Y., Dalton, A., Allen, J., and Galescu, L. (2013). Automatic metaphor detection using large-scale lexical resources and conventional metaphor extraction. In Proceedings of the First Workshop on Metaphor in NLP, pp. 36-44. [find](https://scholar.google.com/scholar?q=Wilks%2C+Y%2E%2C+Dalton%2C+A%2E%2C+Allen%2C+J%2E%2C+and+Galescu%2C+L%2E+%282013%29%2E+Automatic+metaphor+detection+using+large%2Dscale+lexical+resources+and+conventional+metaphor+extraction%2E+In+Proceedings+of+the+First+Workshop+on)