Elamur Research⌕
☰
home / papers / 2026 / From Categories to Decisions: A Framework for Attitudinal Analysis of Evaluative Language

From Categories to Decisions: A Framework for Attitudinal Analysis of Evaluative Language

Summary

Annotating attitude in evaluative language under Appraisal Theory is often inconsistent. The authors trace this to three sources. First, schemes usually tag only the evaluative wording and leave out who evaluates and what is evaluated. Second, the line between Affect and the other two Attitude subsystems has no operational test. Third, Judgement and Appreciation are hard to separate when the target is an abstract human product, such as a policy or report.

The paper proposes a unified annotation scheme. Every evaluative instance is marked as a triple: Appraiser (AS), Appraisal Element (AE) and Appraised (AT). The Affect boundary is redefined as two directional pathways. Emotions that become objectified drift toward Appreciation, and emotions that become socialised drift toward Judgement. For Judgement versus Appreciation, a decision procedure based on target and lexis types uses override conditions and substitution tests.

The procedures are illustrated with English COVID-19 news examples rather than evaluated quantitatively; agreement figures appear in a 2025 companion study. The goal is a replicable workflow in which annotator disagreements can be traced to specific decision steps.

Key points

  • The scheme encodes each evaluation as A = {AS, AT, AE}, which can be drawn as a tree with appraisal types and subcategories (e.g., Appreciation: Reaction; Affect: In/security).
  • Criterion 1 separates Affect from Appreciation by how salient the experiencer is, moving from individual feeling to collective emotion to an objectified property. Criterion 2 separates Affect from Judgement by asking whether a personal emotion or an evaluation of conduct is foregrounded.
  • A selection principle resolves overlapping evaluative readings (Thompson's "Russian doll" problem), based on how AT is realised and what kind of entity it is.
  • Examples come from a COVID-19 news corpus covering China Daily, New York Times, South China Morning Post, Straits Times and The Guardian (regions CN, HK, SG, UK, US).
  • Acknowledged limitations:
    • The scheme has been tested only on English media discourse.
    • The selection principle does not cover every configuration.
    • Substitution tests still require some analyst judgement.

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

Institutions

🔬Institut national de recherche en sciences et technologies du numériqueLe Chesnay-Rocquencourt · France55 works · 16 624 citations🎓Fudan UniversityShanghai · China45 works · 16 344 citations🎓Korea UniversitySeoul · South Korea4 works · 8 679 citations🎓Beihang UniversityBeijing · China18 works · 6 692 citations🎓Tilburg UniversityTilburg · Netherlands30 works · 5 198 citations🎓Brandeis UniversityWaltham · United States27 works · 4 999 citations🎓City University of Hong KongHong Kong · Hong Kong12 works · 2 138 citations

Also on the byline: Université de Franche-Comté · South China Normal University · Liaocheng University · China National Institute of Standardization.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/4sv645t72r82

@inproceedings{zeng2026categories,
  title = {{From Categories to Decisions: A Framework for Attitudinal Analysis of Evaluative Language}},
  author = {Zeng, Jiamei and Wang, Haitao and Bunt, Harry and Cao, Xinyu and Dong, Min and Hao, Tianyong and Lee, Kiyong and Pustejovsky, James and Romary, Laurent and Zong, Jianfang and Rey, François Claude and Cardey, Sylviane and Jia, Yangli and Liao, Shengqing and Fang, Alex Chengyu},
  booktitle = {Proceedings of the 22nd Joint ACL - ISO Workshop on Interoperable Semantic Annotation and Representation (ISA-22) @ LREC 2026},
  year = {2026},
  doi = {10.63317/4sv645t72r82},
  url = {https://doi.org/10.63317/4sv645t72r82}
}

References

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

  1. Bednarek, M. (2009). Language patterns and ATTITUDE. Functions of Language, 16(2):165– 192. find
  2. Fuoli, M. (2018). A stepwise method annotating APPRAISAL. Functions Language, 25(2):229–258. for of Martin, J.R. (2000). Beyond exchange: APPRAISAL systems in English. In S. Hunston & G. Thompson (Eds.), Evaluation in text: Authorial stance and the construction of discourse. Oxford University Press,… find
  3. Martin, J.R. & White, P.R.R. (2005). The language of evaluation: Appraisal in English. Palgrave Macmillan. find
  4. Read, J. & Carroll, J. (2012). Annotating expressions of Appraisal in English. Language Resources and Evaluation, 46(3):421–447. find
  5. Starfield, S., Paltridge, B., McMurtrie, R., Holbrook, A., Bourke, S., Fairbairn, H., Kiley, M., & Lovat, T. (2015). Understanding the language of evaluation in examiners' reports on doctoral theses. Linguistics and Education, 31:130–144. find
  6. Taboada, M. & Carretero, M. (2012). Contrastive analyses of evaluation in text: Key issues in the design of an annotation system for attitude applicable to consumer reviews in English and Spanish. Linguistics and the Human Sciences, 6(1–3):275–295. find
  7. Thompson, G. (2014). AFFECT and emotion, target-value mismatches, and Russian dolls: Refining the APPRAISAL model. In G. Thompson & L. Alba-Juez (Eds.), Evaluation in context. John Benjamins, pp. 47–66. find
  8. White, P.R.R. (n.d.). An introductory tour through appraisal theory. Retrieved from http://www.grammatics.com/appraisal/ ↗
  9. White, P.R.R. (2004). Subjectivity, evaluation and point of view in media discourse. In C. Coffin, A. Hewings, & K. O’Halloran (Eds.), Applying English grammar: Corpus and functional approaches. Hodder Arnold, pp. 229–246. find
  10. White, P.R.R. (2012). Exploring the axiological workings of ‘reporter voice’ news stories: Attribution and attitudinal positioning. Discourse, Context & Media, 1(2–3):57–67. find
  11. Zeng, J., Wang, H., Bunt, H., Cao, X., Cardey, S., Dong, M., Hao, T., Jia, Y., Lee, K., Liao, S., Pustejovsky, J., Rey, F.C., Romary, L., Zong, J., & Fang, A.C. (2025). Enhanced evaluative language annotation through refined theoretical framework and workflow. In H. Bunt (Ed.), Proceedings of the 21… find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zeng-2026-categories-decisions-framework-attitudinal-analysis/