Agonistic Pluralism
The constructive alternative to both deliberative democracy (which seeks rational consensus) and antagonistic politics (which collapses into enmity). Agonistic pluralism treats political life as the ongoing struggle among plural, legitimate, incompatible visions of the good. The goal is not to arrive at agreement but to sustain institutional spaces where disagreement can be expressed, contested, and provisionally settled — always with the understanding that settlements are temporary, excluded positions will return, and the struggle is democracy's beating heart rather than a deficiency to be overcome. Adversaries recognize each other's legitimacy; they do not seek each other's elimination. They fight, but within rules that preserve the conditions of continued fighting.
In the AI Story
The Buyl et al. 2025 study of LLM ideology demonstrated what agonistic pluralism looks like in the material infrastructure of AI. The researchers found that different large language models from different geopolitical regions reflect systematically different ideological positions. Their explicitly Mouffean conclusion: rather than pursuing ideological neutrality (which is a hegemonic operation disguised as procedural fairness), regulation should preserve this diversity as a feature. 'The strong ideological diversity shown across publicly available, powerful LLMs would even be considered healthy under Mouffe's democratic model of pluralistic agonism.'
Applied to the AI discourse, agonistic pluralism rejects both the triumphalist dismissal of critics and the silent middle's balanced absorption of all positions. It demands institutional spaces where Han's grief, the builder's exhilaration, the displaced worker's refusal, and the developing child's vulnerability are represented as legitimate political positions with institutional standing — not as emotional reactions to be accommodated but as contending visions of what the AI transition should produce.
The model requires specific institutional conditions: structures through which affected populations can organize; mechanisms for contestation that are not merely consultative; recognition that the outcomes of contestation are provisional; and cultural practices that treat disagreement as valuable rather than as a problem to be managed. These conditions are under-developed in current AI governance, which relies heavily on expert advisory structures and industry self-regulation.
Critics charge that agonistic pluralism offers no resource for settling specific disputes and that some form of shared rational framework must underlie legitimate contestation. Mouffe's response — that the shared framework is the commitment to democratic form itself, not to any substantive outcome — reframes the objection but does not fully dissolve it.
Origin
Developed most fully in The Democratic Paradox (2000) and Agonistics: Thinking the World Politically (2013). The framework emerged from Mouffe's engagement with and critique of Habermasian deliberative democracy — her argument that the Habermasian ideal of rational consensus misunderstands the nature of political life by treating passion, conflict, and power as deficiencies to be overcome.
Key Ideas
Adversaries, not consensus. The goal is sustained productive contestation, not agreement.
Institutional form. Agonism requires specific institutions — not mere tolerance.
Provisional settlements. Every political outcome remains open to future contestation.
Passion is political. Democratic life is constituted by passionate commitment, not by the bracketing of passion in favor of reason.
Appears in the Orange Pill Cycle
Further reading
- Chantal Mouffe, Agonistics: Thinking the World Politically (Verso, 2013)
- Chantal Mouffe, The Democratic Paradox (Verso, 2000)
- Maarten Buyl et al., 'Large Language Models Reflect the Ideology of their Creators' (Nature Machine Intelligence, 2025)