X-Bias: Exploring Gender Bias in Argumentation
Gender bias is still a widespread phenomenon. According to a report issued by the United Nations Development Programme (2023, p. 7), data collected in 80 countries over two periods (2010 – 2014 and 2017 – 2022) showed that almost 90% of people have one or more biases against women when it comes to politics, education, economy, or physical integrity. This high number underscores the importance of studying the reasons behind gender bias, its forms, and how it manifests across different contexts. The X-Bias project aims to explore, experimentally assess, and explain gender bias in a specific domain: argumentation. People exchange reasons for or against a claim in various situations, and studying the argumentative structures that are used, as well as their pragmatic and cognitive underpinnings, contributes to a better understanding of biases related to gender. In this project, gender is considered a relevant variable for the argument perception and its effects on the speaker’s image.
The project combines research on gender in language and argumentation theory with an experimental methodology. Gender has been investigated in numerous fields outside of and within linguistic contexts, such as medicine (e.g., Hamberg, 2008), law (e.g., Patton & Smith, 2017), academia (e.g., Skov, 2020), politics (e.g., Paxton et al., 2007), persuasion research (e.g., Galasso & Nannicini, 2025), psycholinguistics (e.g., Gygax, Elmiger, et al., 2019), and Large Language Models (henceforth LLMs, e.g., Urchs et al., 2024). Yet, research on gender in argumentation theory, in a narrow sense, especially the perception and evaluation of argument types, is still scarce (but see e.g., Ciurria & Altamimi, 2014; Hundleby, 2012; Rooney, 1995; Yap, 2020, for some examples). The role of gender in argumentation has been discussed neither by classical theories (e.g., Anscombre & Ducrot, 1983; Perelman & Olbrechts-Tyteca, 1958; Toulmin, 1958; van Eemeren & Grootendorst, 1992)nor by most of the contemporary heirs, as the latest handbook on argumentation theory shows (Aikin et al., 2025).
The X-Bias project fills this gap and contributes to a better understanding of the occurrences, perceptions, evaluations, and possible reasons for gender bias in argumentation. It provides a systematic empirical assessment of argument types such as the appeal to pity, “You should find the defendants not guilty because they have already suffered enough.”, uttered by a male/female speaker, investigating the audience’s view on it. The planned experiments isolate gender and argument types as variables, and control for context (topics, setting, etc.) and information about the speaker (age, profession, etc.). The results from the studies will be compared to how LLMs evaluate the same argument types given a set of rules based on definitions from argumentation theory. Taken together, it informs how gender affects argumentation, as well as the way different types of (non-)fallacious arguments are evaluated by humans and machines. The project comprises three strands:
Strand 1, Gender Bias and (Non-) Fallacious Arguments: This strand investigates whether specific types of non-fallacious (1a) (e.g., appeals to expert opinion, appeals to emotion, etc.) and fallacious arguments (1b) (e.g., appeals to authority, appeals to pity, etc.) are evaluated differently based on the speaker’s gender.
Strand 2, Gender Bias and Speaker Ethos: The focus lies on the role of pre-discursive and discursive ethos (i.e., the speaker’s image before and during speech) in the perception of arguments uttered by speakers of different genders. The aim is to assess whether gender affects the speaker’s image and the perception of arguments.
Strand 3, Gender Bias in Large Language Models: The insights obtained from Strands 1 and 2 will be compared with the evaluations of the same argument types through LLMs. The goal is twofold: replicate the experimental studies in LLMs, i.e., analyse and compare humans’ and LLMs’ performance given the same tasks.
References
Aikin, S. F., Casey, J. P., & Stevens, K. (Eds). (2025). The Routledge Handbook of Argumentation Theory. Routledge. https://www.routledge.com/The-Routledge-Handbook-of-Argumentation-Theory/Aikin-Casey-Stevens/p/book/9781041128410
Anscombre, J.-C., & Ducrot, O. (1983). L’argumentation dans la langue. Mardaga.
Ciurria, M., & Altamimi, K. (2014). Argumentum ad Verecundiam: New Gender-based Criteria for Appeals to Authority. Argumentation, 28(4), 437–452. https://doi.org/10.1007/s10503-014-9328-0
Galasso, V., & Nannicini, T. (2025). Persuasion and gender: Experimental evidence from two political campaigns. Public Choice, 203(1–2), 183–204. https://doi.org/10.1007/s11127-024-01192-y
Gygax, P. M., Elmiger, D., Zufferey, S., Garnham, A., Sczesny, S., von Stockhausen, L., Braun, F., & Oakhill, J. (2019). A Language Index of Grammatical Gender Dimensions to Study the Impact of Grammatical Gender on the Way We Perceive Women and Men. Frontiers in Psychology, 10, 1604. https://doi.org/10.3389/fpsyg.2019.01604
Hamberg, K. (2008). Gender Bias in Medicine. Women’s Health, 4(3), 237–243. https://doi.org/10.2217/17455057.4.3.237
Hundleby, C. (2012). Feminist Empiricism. In S. N. Hesse-Biber (Ed.), Handbook of Feminist Research: Theory and Praxis (2nd edn, pp. 28–45). SAGE Publications Inc.
Patton, D., & Smith, J. L. (2017). Lawyer, Interrupted: Gender Bias in Oral Arguments at the US Supreme Court. Journal of Law and Courts, 5(2), 337–361. https://doi.org/10.1086/692611
Paxton, P., Kunovich, S., & Hughes, M. M. (2007). Gender in Politics. Annual Review of Sociology, 33(1), 263–284. https://doi.org/10.1146/annurev.soc.33.040406.131651
Perelman, C., & Olbrechts-Tyteca, L. (1958). Traité de l’argumentation. La Nouvelle Rhétorique. Presses Universitaires de France.
Rooney, P. (1995). Rationality and the Politics of Gender Difference. Metaphilosophy, 26(1/2), 22–45.
Skov, T. (2020). Unconscious Gender Bias in Academia: Scarcity of Empirical Evidence. Societies, 10(2), 31. https://doi.org/10.3390/soc10020031
Toulmin, S. E. (1958). The Uses of Argument. Cambridge University Press.
UNESCO IRCAI. (2024). Challenging systematic prejudices: An investigation into gender bias in Large Language Models (pp. 1–21). United Nations Educational, Scientific and Cultural Organization.
United Nations Development Programme. (2023). Breaking Down Gender Biases: Shifting Social Norms Towards Gender Equality: 2023 Gender Social Norms Index (pp. 1–40). United Nations. https://doi.org/10.18356/9789210028028
Urchs, S., Thurner, V., Aßenmacher, M., Heumann, C., & Thiemichen, S. (2024). How Prevalent is Gender Bias in ChatGPT? -- Exploring German and English ChatGPT Responses (No. arXiv:2310.03031). arXiv. http://arxiv.org/abs/2310.03031
van Eemeren, F. H., & Grootendorst, R. (1992). Argumentation, Communication, Fallacies. A Pragma-Dialectical Perspective. Lawrence Erlbaum Associates.
Yap, A. (2020). Argumentation, Adversariality, and Social Norms. Metaphilosophy, 51(5), 747–765. https://doi.org/10.1111/meta.12458