Elamur Research⌕
☰
home / papers / 2022 / A Context-free Arabic Emoji Sentiment Lexicon (CF-Arab-ESL)

A Context-free Arabic Emoji Sentiment Lexicon (CF-Arab-ESL)

Paper · 2022 · Shatha Ali A. Hakami, Robert Hendley and Phillip Smith

Summary

Emoji are widely used as sentiment signals in sentiment analysis, but their reading varies from person to person, and cultural background is a major source of that variation. Lexicons built from other populations, such as the European Euro-ESL, may therefore not reflect how Arabic speakers perceive these symbols. The authors built CF-Arab-ESL, a lexicon of 1,069 unique emoji labelled for sentiment by 53 native Arabic speakers (men and women) who come from seven parts of the Arab world: Egypt, Sudan, Iraq, Yemen, the Gulf, the Levant and North Africa. The emoji were shown alone, with no surrounding text, so the labels record a default, context-free reading. Each annotator's reliability was checked with two tests: sensitivity (Recall) and consistency (Krippendorff's Alpha). The authors then examined the labels for cultural influences, including national affiliation, colour, animal and weather symbolism, and religion. The lexicon is freely released for research. It gives Arabic sentiment systems a culturally grounded starting point for emoji, and it can serve as a baseline for later comparison with in-context interpretations.

Key points

  • CF-Arab-ESL covers 1,069 unique emoji annotated by 53 native Arabic speakers of both genders, drawn from seven Arabic-speaking areas (Egypt, Sudan, Iraq, Yemen, the Gulf, the Levant and North Africa).
  • Annotation was deliberately context-free: participants judged each emoji's sentiment without any accompanying text.
  • Annotator reliability was assessed per participant through sensitivity (Recall) and consistency (Krippendorff's Alpha).
  • The cultural analysis covers national affiliation, colour, animals, weather and religious influence, building on the authors' earlier Arab-ESL comparison with Euro-ESL.
  • The resource is public on GitHub (ShathaHakami/Context-Free-Arabic-Emoji-Sentiment-Lexicon). The authors name two limitations: few annotators per region, and no comparison yet with in-context readings.

Interesting ideas

  • I1. The authors collect a context-free reading of each symbol first, so it can later be contrasted with its reading inside text; this separates a stable default from contextual shifts.
  • I2. Each annotator is screened with two complementary checks, one for recall against expected answers and one for self-consistency, rather than relying only on pooled inter-annotator agreement.
  • I3. Annotators are drawn from distinct regions, so systematic disagreement becomes a finding about culture rather than being averaged away as noise.
  • I4. Results are broken down by semantic category (colours, animals, weather, religious symbols) to show where cultural shifts concentrate.
  • I5. Building on the authors' earlier work, the paper treats emoji as having different roles: a true sentiment indicator, a multi-sentiment or ambiguous one, or a no-sentiment marker overshadowed by the text.

For Mova

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

  • I1 (frames): store default evaluative polarity for words and markers separately in the lexicon, and let prag record explicitly when context overrides it, e.g. as evidence for irony.
  • I2 (learn): score every witness parser and LLM teacher tier on both recall over control items and stability on repeated items before its silver votes are counted.
  • I3 (world): let value and connotation entries in compiled knowledge carry a community or dialect tag instead of a single global polarity, keeping divergence explainable.
  • I4 (learn): report evaluation errors by WordNet or FrameNet domain, so repair rules and teacher queries target the categories where readings diverge.
  • I5 (frames): classify evaluative markers by function before letting them contribute to speech-act or attitude frames, rather than adding their polarity blindly.

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 2The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 works

🧭 Topics 3Research subjects and languages the work is about.Sentiment & emotion · 972 worksSemitic languages · 2,908 worksGrammatical error correction & learner language · 156 works

Cite

Cite as Shatha Ali A. Hakami, Robert Hendley and Phillip Smith. 2022. A Context-free Arabic Emoji Sentiment Lexicon (CF-Arab-ESL). Proceedinsg of the 5th Workshop on Open-Source Arabic Corpora and Processing Tools with Shared Tasks on Qur'an QA and Fine-Grained Hate Speech Detection.

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

@inproceedings{hakami2022context,
  title = {{A Context-free Arabic Emoji Sentiment Lexicon (CF-Arab-ESL)}},
  author = {Shatha Ali A. Hakami, Robert Hendley and Phillip Smith},
  booktitle = {Proceedinsg of the 5th Workshop on Open-Source Arabic Corpora and Processing Tools with Shared Tasks on Qur'an QA and Fine-Grained Hate Speech Detection},
  year = {2022},
  url = {https://aclanthology.org/2022.osact-1.6/}
}

References

25 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Abdellaoui, H. and Zrigui, M. (2018). Using Tweets and Emojis to Build TEAD: an Arabic Dataset for Sentiment Analysis. Computación y Sistemas, 22:777 – 786. ↗
  2. Al-Azani, S. and El-Alfy, E.-S. (2018a). Emojis-based sentiment classification of arabic microblogs using deep recurrent neural networks. In 2018 International Conference on Computing Sciences and Engineering (ICCSE), pages 1–6. find
  3. Al-Azani, S. and El-Alfy, E.-S. M. (2018b). Combining emojis with arabic textual features for sentiment classification. In 2018 9th International Conference on Information and Communication Systems (ICICS), pages 139–144. find
  4. Al-Azani, S. and El-Alfy, E.-S. M. (2018c). Emoji-based sentiment analysis of arabic microblogs using machine learning. In 2018 21st Saudi Computer Society National Computer Conference (NCC), pages 1–6. find
  5. Al Rashdi, F. (2018). Functions of emojis in whatsapp interaction among omanis. Discourse, Context & Media, 26:117–126. ↗
  6. Al–Azani, S. and El–Alfy, E. M. (2018). Imbalanced sentiment polarity detection using emoji-based features and bagging ensemble. In 2018 1st International Conference on Computer Applications Information Security (ICCAIS), pages 1–5. find
  7. Badaro, G., Baly, R., Hajj, H., Habash, N., and El-Hajj, W. (2014). A large scale arabic sentiment lexicon for arabic opinion mining. In Proceedings of the EMNLP 2014 workshop on arabic natural language processing (ANLP), pages 165–173.
  8. Danesi, M. (2017). The semiotics of emoji: The rise of visual language in the age of the internet. Bloomsbury Publishing. find
  9. Donato, G. and Paggio, P. (2017). Investigating redundancy in emoji use: Study on a twitter based corpus. In Proceedings of the 8th Workshop on Computational Riordan, M. A. (2017). The communicative role of non-Approaches to Subjectivity, Sentiment and Social Media face emojis: Affect and disambiguation. Computers in 58 Human Behavior, 76:75–86. Analysis, pages 118–126. find
  10. Dresner, E. and Herring, S. C. (2010). Functions of the non-verbal in cmc: Emoticons and illocutionary force. Communication theory, 20(3):249–268. find
  11. Gao, B. and VanderLaan, D. P. (2020). Cultural influences on perceptions of emotions depicted in emojis. Cyberpsychology, Behavior, and Social Networking. ↗
  12. Good, I. J. (1965). The estimation of probabilities: An essay on modern bayesian methods, vol. 30. find
  13. Hakami, S. A. A., Hendley, R., and Smith, P. (2020). Emoji as sentiment indicators: An investigative case study in arabic text. In The Sixth International Conference on Human and Social Analytics, pages 26–32. IARIA. find
  14. Hakami, S. A. A., Hendley, R., and Smith, P. (2021). Arabic emoji sentiment lexicon (Arab-ESL): A comparison between Arabic and European emoji sentiment lexicons. In Proceedings of the Sixth Arabic Natural Language Processing Workshop, pages 60–71, Kyiv, Ukraine (Virtual). Association for Computational Linguistics.
  15. Husain, F. (2020). Osact4 shared task on offensive language detection: Intensive preprocessing-based approach. In Proceedings of the 4th Workshop on Open-Source Arabic Corpora and Processing Tools, with a Shared Task on Offensive Language Detection, pages 53–60, Marseille, France. European Language Resource Association. ↗
  16. Hussien, W. A., Tashtoush, Y. M., Al-Ayyoub, M., and Al-Kabi, M. N. (2016). Are emoticons good enough to train emotion classifiers of arabic tweets? In 2016 7th International Conference on Computer Science and Information Technology (CSIT), pages 1–6. find
  17. Kralj Novak, P., Smailović, J., Sluban, B., and Mozetič, I. (2015). Sentiment of emojis. PloS one, 10(12):e0144296. ↗
  18. Krippendorff, K. (2004). Measuring the reliability of qualitative text analysis data. Quality and quantity, 38:787– 800. ↗
  19. Mubarak, H., Hassan, S., and Chowdhury, S. A. (2022). Emojis as anchors to detect arabic offensive language and hate speech. CoRR, abs/2201.06723. ↗
  20. Nielsen, F. Å. (2011). A new anew: Evaluation of a word list for sentiment analysis in microblogs. arXiv preprint arXiv:1103.2903. ↗
  21. Njenga, K. (2018). Social media information security threats: Anthropomorphic emoji analysis on social engineering. In IT Convergence and Security 2017, pages 185–192. Springer. ↗
  22. Refaee, E. and Rieser, V. (2014). Can we read emotions from a smiley face? emoticon-based distant supervision for subjectivity and sentiment analysis of arabic twitter feeds. In 5th International Workshop on Emotion, Social Signals, Sentiment and Linked Open Data, LREC. find
  23. Shiha, M. and Ayvaz, S. (2017). The effects of emoji in sentiment analysis. Int. J. Comput. Electr. Eng.(IJCEE.), 9(1):360–369. ↗
  24. Skovholt, K., Grønning, A., and Kankaanranta, A. (2014). The communicative functions of emoticons in workplace e-mails::-. Journal of Computer-Mediated Communication, 19(4):780–797. ↗
  25. Su, L. T. (1994). The relevance of recall and precision in user evaluation. Journal of the American Society for Information Science, 45(3):207–217. ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/hakami-2022-context-free-arabic-emoji-sentiment/