From Lexicon to Prediction
A Protocol for Transformer Fine-Tuning and Multilingual Validation of a Psycho-Forensic Linguistic Framework for Suicide-Related Discourse Detection
Keywords:
suicide discourse; psycho-forensic linguistics; multilingual NLP; XLM-RoBERTa; cultural calibrationAbstract
Suicide prevention requires attention not only to clinically recognized crises but also to communicative signals that may precede or accompany them. Language offers an observable behavioural trace through which distress can be studied outside formal clinical encounters, although linguistic evidence must not be treated as a diagnosis. The Psycho-Forensic Linguistic Surveillance Initiative (PFLSI) has progressed from culturally situated Pakistani suicide notes to corpus-scale English analysis, theoretical integration, and a proposed multilingual Urdu-Punjabi corpus. Series Paper 5 extends that programme by specifying a protocol for transformer-based modelling and multilingual validation of suicide-related discourse in English, Urdu, Punjabi, Roman Urdu, and Urdu-English code-switched communication. Because the proposed Multilingual Suicide Discourse Corpus (M-SDC) has not yet been completed, this article does not report empirical model results. Instead, it prespecifies the classification task, corpus-matching principles, annotation procedures, baseline and transformer models, culturally informed feature streams, validation regimes, evaluation metrics, ablation strategy, explainability procedures, and ethical boundaries to be implemented after corpus completion. The protocol compares corpus-linguistic and classical machine-learning baselines with multilingual BERT and XLM-RoBERTa, while testing whether structured linguistic and culturally situated features add information beyond contextual representations. Evaluation prioritizes discrimination, calibration, source-held-out generalization, cross-linguistic and cross-script transfer, subgroup stability, and interpretability. The central argument is that multilingual suicide-discourse modelling should be evaluated not by predictive performance alone but by linguistic validity, cultural calibration, source generalization, transparent error analysis, and strict separation between computational risk indication and clinical judgement.
Downloads
References
Ahmad, A. I. S. (2026). Language of influence: A corpus-based lexical analysis of psychological power strategies in Robert Greene's The Laws of Human Nature. In Proceedings of the 2nd Riphah International Conference on Language, Literature, and Culture.
Al-Mosaiwi, M., & Johnstone, T. (2018). In an absolute state: Elevated use of absolutist words is a marker specific to anxiety, depression, and suicidal ideation. Clinical Psychological Science, 6(4), 529–542.
Bali, K., Sharma, J., Choudhury, M., & Vyas, Y. (2014). I am borrowing ya mixing? An analysis of English-Hindi code mixing in Facebook. In Proceedings of the First Workshop on Computational Approaches to Code Switching (pp. 116–126).
Beck, A. T. (1963). Thinking and depression: I. Idiosyncratic content and cognitive distortions. Archives of General Psychiatry, 9(4), 324–333.
Beck, A. T., Weissman, A., Lester, D., & Trexler, L. (1974). The measurement of pessimism: The Hopelessness Scale. Journal of Consulting and Clinical Psychology, 42(6), 861–865.
Bialer, A., Izmaylov, D., Segal, A., Tsur, O., Levi-Belz, Y., & Gal, K. (2022). Detecting suicide risk in online counseling services: A study in a low-resource language. In Proceedings of the 29th International Conference on Computational Linguistics (pp. 4241–4250). International Committee on Computational Linguistics.
Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20(1), 37–46.
Conneau, A., Khandelwal, K., Goyal, N., Chaudhary, V., Wenzek, G., Guzmán, F., Grave, E., Ott, M., Zettlemoyer, L., & Stoyanov, V. (2020). Unsupervised cross-lingual representation learning at scale. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (pp. 8440–8451). Association for Computational Linguistics. https://doi.org/10.18653/v1/2020.acl-main.747
Coppersmith, G., Dredze, M., & Harman, C. (2014). Quantifying mental health signals in Twitter. In Proceedings of the Workshop on Computational Linguistics and Clinical Psychology (pp. 51–60).
De Choudhury, M., Counts, S., & Horvitz, E. (2013). Predicting postpartum changes in emotion and behavior via social media. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 3267–3276).
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (pp. 4171–4186). Association for Computational Linguistics. https://doi.org/10.18653/v1/N19-1423
Gkotsis, G., Oellrich, A., Hubbard, T., Dobson, R., Liakata, M., Velupillai, S., & Dutta, R. (2017). Characterisation of mental health conditions in social media using informed deep learning. Scientific Reports, 7, 45141.
Guo, C., Pleiss, G., Sun, Y., & Weinberger, K. Q. (2017). On calibration of modern neural networks. In Proceedings of the 34th International Conference on Machine Learning (pp. 1321–1330). PMLR.
Ishfaq, A. (2025). Ethnolinguistic identity and cultural memory in the Dawoodi community. Annual Methodological Archive Research Review, 3(6), 185–208.
Ishfaq, A. (2026). Thinking through language: Linguistic foundation and advanced academic writing.
Ishfaq, A., & Bhatti, A. M. (2019). Language shift and imminent language death: A diachronic study of Dawoodi. Harf-o-Sukhan, 3, 1–14.
Ishfaq, A., & Bhatti, A. M. (2020). Lexical attrition and generational language competence in Dawoodi speakers. Harf-o-Sukhan, 4, 4–18.
Ishfaq, A., & Bhatti, A. M. (2021). From pidgin to creole to collapse: The evolutionary trajectory of Dawoodi language. Jahan-e-Tahqeeq, 4.
Ishfaq, A., & Bhatti, A. M. (2022). Linguistic hegemony and the silencing of Dawoodi: Power, stigma, and structural marginalization. Jahan-e-Tahqeeq, 5(4), 48–60.
Ishfaq, A., & Bhatti, A. M. (2023). Code-switching, borrowing, and linguistic dilution: Contact-induced change in Dawoodi. Jahan-e-Tahqeeq, 6(3), 577–592.
Ishfaq, A., & Bhatti, A. M. (2024). From heritage to liability: Language attitudes and identity reconstruction as drivers of obsolescence in the Dawoodi language. Al-Mahdi Research Journal (MRJ), 5(3), 1303–1336.
Ishfaq, A., & Sultan, N. (2024). Identification of different methodologies for treatment of autism in Urdu-speaking adolescents: An investigative report. Contemporary Journal of Social Science Review, 2(4), 1611–1618.
Ishfaq, A., & Sultan, N. (2025a). Artificial intelligence and human language cognition: An integrative comparative framework. Critical Review of Social Sciences and Humanities, 5(2), 57–71.
Ishfaq, A., & Sultan, N. (2025b). Cognitive control and executive function in advanced second-language writing. Sareer-a-Khama, 4(4).
Ishfaq, A., & Sultan, N. (2025c). Language of stress: A corpus-based study to detect early signs of suicide through lexical choice. Journal of Applied Linguistics and TESOL (JALT), 9(1), 862–880. https://doi.org/10.63878/jalt2310
Ishfaq, A., & Sultan, N. (2025d). Narrative and meaning in Surah Yūsuf: A critical hermeneutic analysis. AL-HAYAT Research Journal (AHRJ), 2(4), 11–23. https://doi.org/10.5281/zenodo.17091179
Ishfaq, A., & Sultan, N. (2025e). Trauma, resilience, and narrative healing: A psycho-hermeneutic reading of Surah Yūsuf. AL-JAMEI Research Journal, 3(1), 229–239.
Ishfaq, A., & Sultan, N. (2026a). Modeling meaning in generative AI: A corpus-assisted discourse analysis of coherence, framing, and persuasion. Pakistan Journal of Social Science Review, 5(3), 1147–1164.
Ishfaq, A., & Sultan, N. (2026b). Voices across scripts: A protocol for a multilingual Urdu–Punjabi suicide discourse corpus and the cultural calibration of psycho-forensic linguistic detection. Critical Review of Social Sciences and Humanities, 6(1), 43–56.
Ishfaq, A., Ahmad, S., & Sultan, N. (2025). Decoding despair: A multidisciplinary psycho-forensic linguistic approach to suicide notes. Journal of Psychology, Health and Social Challenges, 3(2), 83–89.
Ishfaq, A., Azim, M. U. (2025). Phono-semantics and translation: A cross-linguistic study of Urdu and Punjabi ideophones. International Research Journal of Arts, Humanities and Social Sciences, 2(3).
Ishfaq, A., Malik, A. H., & Sultan, N. (2025). Developing trauma-sensitive pedagogical practices for resilient learning in academia: A multidisciplinary approach of psycholinguistics and ELT. Al Aasar, 2(1), 171–189.
Ishfaq, A., Sultan, N., & Bhatti, A. M. (2026). The lexicon of crisis: A psycho-forensic linguistic architecture for real-time suicide risk detection integrating big data, corpus evidence, and theoretical frameworks. Annual Methodological Archive Research Review, 4(6), 1–27.
Ishfaq, A., Sultan, N., & Healy, B. (2025). Turn-taking, politeness, and identity: A conversational study of Speak Your Heart. Journal of Applied Linguistics and TESOL (JALT), 8(3), 1567–1581. https://doi.org/10.63878/jalt1146
Ishfaq, A., Sultan, N., Hassan, N., Aleem, F., & Maldonado, M. G. (2022). Elif Shafak's Forty Rules of Love: Contextual variation in adjectives. Turkish Online Journal of Qualitative Inquiry, 13(1), 2029–2036.
Izmaylov, D., Segal, A., Gal, K., Grimland, M., & Levi-Belz, Y. (2023). Combining psychological theory with language models for suicide risk detection. In Findings of the Association for Computational Linguistics: EACL 2023 (pp. 2430–2438). Association for Computational Linguistics. https://doi.org/10.18653/v1/2023.findings-eacl.184
Jain, S., & Wallace, B. C. (2019). Attention is not explanation. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (pp. 3543–3556). Association for Computational Linguistics. https://doi.org/10.18653/v1/N19-1357
Joiner, T. E. (2005). Why people die by suicide. Harvard University Press.
Khan, I. A., Khaled, F., & Ishfaq, A. (2025). Operation Bunyan un Marsoos: A critical analysis of human rights compliance: A study of the operation's adherence to human rights law and international humanitarian law. Dialogue Social Science Review (DSSR), 3(6), 173–184.
Khattak, A., Asghar, M. Z., Saeed, A., Hameed, I. A., Ahmad, S. A., & Hassan, M. (2021). A survey on sentiment analysis in Urdu: A resource-poor language. Egyptian Informatics Journal, 22(1), 53–74.
Kövecses, Z. (2015). Where metaphors come from: Reconsidering context in metaphor. Oxford University Press.
Lakoff, G., & Johnson, M. (1980). Metaphors we live by. University of Chicago Press.
Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33(1), 159–174.
Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. In Advances in Neural Information Processing Systems 30.
Menon, R. (2024). The narrative-crisis model: Understanding suicidal ideation through personal storytelling. Journal of Mental Health and Narratives, 12(3), 193–208.
Molina, G., AlGhamdi, F., Ghoneim, M., Hawwari, A., Rey-Villamizar, N., Diab, M., & Solorio, T. (2016). Overview for the second shared task on language identification in code-switched data. In Proceedings of the Second Workshop on Computational Approaches to Code Switching (pp. 40–49).
O'Connor, R. C. (2011). Towards an integrated motivational-volitional model of suicidal behaviour. In R. C. O'Connor, S. Platt, & S. Gordon (Eds.), International handbook of suicide prevention (pp. 181–198). Wiley-Blackwell.
Pennebaker, J. W. (2011). The secret life of pronouns: What our words say about us. Bloomsbury Press.
Rayson, P., Archer, D., Piao, S., & McEnery, T. (2004). The UCREL Semantic Analysis System. In Proceedings of the LREC Workshop on Beyond Named Entity Recognition.
Shneidman, E. S. (1993). Suicide as psychache: A clinical approach to self-destructive behavior. Jason Aronson.
Solorio, T., Blair, E., Maharjan, S., Bethard, S., Diab, M., Ghoneim, M., Hawwari, A., AlGhamdi, F., Hirschberg, J., Chang, A., & Fung, P. (2014). Overview for the first shared task on language identification in code-switched data. In Proceedings of the First Workshop on Computational Approaches to Code Switching (pp. 62–72).
Sultan, N., & Ishfaq, A. (2026). Instagram captions as identity performance: A multimodal discourse analysis. AL-HAYAT Research Journal (AHRJ), 3(2), 9–21.
Teddlie, C., & Yu, F. (2007). Mixed methods sampling: A typology with examples. Journal of Mixed Methods Research, 1(1), 77–100.
World Health Organization. (2023). Suicide worldwide in 2019: Global health estimates. WHO Press.
Downloads
Published
Data Availability Statement
It is a proposal paper; hence no data is collected.
Issue
Section
Categories
License
Copyright (c) 2026 Afshan Ishfaq, Nida Sultan (Author)
This work is licensed under a Creative Commons Attribution 4.0 International License.
You are free to:
- Share — copy and redistribute the material in any medium or format for any purpose, even commercially.
- Adapt — remix, transform, and build upon the material for any purpose, even commercially.
- The licensor cannot revoke these freedoms as long as you follow the license terms.
Under the following terms:
- Attribution — You must give appropriate credit , provide a link to the license, and indicate if changes were made . You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
- No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
