Andrea Rocci on “From polyphony to polylogue: problems and opportunities of the representation of concessions and counterarguments in large societal discussions”
Recently, Nature published a piece in its comments section (Chen et al. 2026), whose title — Does AI already have human-level intelligence? The evidence is clear — is both bold and in tune with the spirit of the times. While the title alludes to the evidence supporting the standpoint that it does, most of the piece consists in evoking arguments against this standpoint and countering them, as in “They lack agency. It is true that present-day LLMs do not form independent goals or initiate action unprompted, as humans do. […] But intelligence does not require autonomy.” (Chen et al. 2026: 39) — thus instantiating patterns of concession and counterargument (CO(u)N patterns). An integrated analytical approach to these patterns, currently in development within the SNF funded project ReCOUNT, makes the object of this talk, which builds on previous work by myself and others on the significance of counterarguments and of concessive moves for the analytical reconstruction and evaluation of argumentative discourse in context (Rocci 2021, Rocci & Lucchini 2025, Palmieri 2025, Rocci et al. 2020).
In the first part of the talk, I sketch a working definition for a multi-dimensional notion of concession, in terms of polyphony and partial commitment, briefly positioning it vis-à-vis the dialectical, rhetorical and grammatical traditions, as well as the analysis of counterarguments in a defeasible reasoning context (Pollock 1987). I argue for a sequential view of counterargument, as an argument interacting with an extant argument (a necessary but difficult notion) and I hypothesize that, in a dialectical defeasible reasoning setting, all counterargument inherently presupposes an act of concession (resuming, re-presenting and — to some extent — recognizing the reasons of the other). In the second part of the talk, armed with a barebones analytical approach for CO(u)N patterns, I look at how individual discursive argumentative contributions (such as the Nature piece by Chen et al. 2026) placed in the context of a vast argumentative polylogue (Lewinski & Aakhus 2023), such as the multifarious debate on AI, introject the polylogue into the discourse through the polyphony (Nølke 2017) of CO(u)N patterns. Looking closely at how it happens, I contend, is important to evaluate how the contribution assumes and discharges the burden of proof, or — adopting the perspective of Lewinski and Aakhus (2023) — how reasonably it contributes to extending the polylogue’s disagreement space. The exemplification will be drawn from Chen et al. (2026) and from the response to that piece by Quattrociocchi et al. (2026).
Andrea Rocci (Università della Svizzera italiana)
REFERENCES
Chen, E. K., Belkin, M., Bergen, L., & Danks, D. (2026). Does AI already have human-level intelligence? The evidence is clear. Nature, 650.
Lewinski, M., & Aakhus, M. A. (2023). Argumentation in complex communication: Managing disagreement in a polylogue. Cambridge University Press.
Nølke, H. (2017). Linguistic Polyphony: The Scandinavian Approach: ScaPoLine. BRILL.
Palmieri, R. (2025). Reasons for trust. The (counter-) argumentative dynamics of image-repair strategies. Journal of Pragmatics, 240, 142–163. https://doi.org/10.1016/j.pragma.2025.02.014
Pollock, J. L. (1987). Defeasible Reasoning. Cognitive Science, 11(4), 481–518. https://doi.org/10.1207/s15516709cog1104_4
Quattrociocchi, W., Capraro, V., & Marcus, G. (2026). Statistical approximation is not general intelligence (Qjrhs_v1). PsyArXiv. https://osf.io/preprints/psyarxiv/qjrhs_v1/
Rocci, A. (2021). Diagramming Counterarguments: At the Interface Between Discourse Structure and Argumentation Structure. In R. Boogaart, H. Jansen, & M. van Leeuwen (Eds.), The Language of Argumentation (Vol. 36, pp. 143–166). Springer International Publishing. https://doi.org/10.1007/978-3-030-52907-9_8
Rocci, A., & Lucchini, C. (2025). Diagramming the Enthymematic Structure of Counterarguments. An Introduction to IAMT Diagrams. Argumentation et Analyse du Discours, (34), Article 34. https://doi.org/10.4000/13q0e
Rocci, A., Greco, S., Schär, R., Convertini, J., Perret-Clermont, A.-N., & Iannaccone, A. (2020). The significance of the adversative connectives aber, mais, ma (‘but’) as indicators in young children’s argumentation. Journal of Argumentation in Context, 9(1), 69–94. https://doi.org/10.1075/jaic.00008.roc
FURTHER INFORMATION
To join the session on Teams, please get in touch with Maria Grazia Rossi at mgrazia.rossi@fcsh.unl.pt or Giulia Terzian at giuliaterzian@fcsh.unl.pt for the details.
This event is part of the ArgLab Research Colloquium organised by Maria Grazia Rossi, Giulia Terzian and Alberto Oya at the Laboratory of Argumentation, Cognition and Language of the NOVA Institute of Philosophy. For any inquiries, please contact Maria Grazia, Giulia or Alberto.