The Civilizational State

In the [YOU] on AI Field Guide

The concept enters the cycle as the institutional counterpart to the epistemological argument about intellectual colonialism in AI ethics. If the Western AI ethics discourse embeds liberal-democratic assumptions without acknowledging them as such, the civilizational-state concept names the institutional target those assumptions are being applied to—and shows why the application distorts rather than illuminates. A state whose legitimation logic is performance-based will govern AI differently than one whose legitimation logic is procedural, and calling the difference authoritarianism rather than institutional difference is a category error that misrepresents both the state being analyzed and the governance alternatives available.

[YOU] on AI’s argument that AI is an amplifier whose quality depends on what is amplified connects to the civilizational-state concept at the level of institutional design: a state that understands its legitimacy as the quality of its governance outcomes has a structural incentive to amplify AI capabilities in the direction of those outcomes, and to suppress applications that damage them. The question of whether this incentive is realized in practice, and whether the state’s internal accountability mechanisms are adequate to prevent the same power from being used for less legitimate purposes, is the genuinely open question that the concept raises without resolving.

Origin

The civilizational-state concept was developed across Zheng’s scholarly career, most fully in The Chinese Communist Party as Organizational Emperor (2010) and subsequent works. Its intellectual roots lie in comparative political theory and in Zheng’s sustained engagement with the Chinese scholarly tradition, which understands statecraft (zhizheng) as a practice with its own logic, values, and institutional memory that cannot be reduced to any universal model of governance.

The concept has relatives in the scholarly literature on “state capacity” and in Samuel Huntington’s civilizational analysis, but Zheng’s specific contribution is to connect civilizational continuity to the internal logic of institutional legitimation. A state that loses the Mandate of Heaven in the classical formulation is not merely unpopular; it has failed the constitutive purpose of the governance project. Applying this logic to AI means that AI governance is not a regulatory add-on but a core test of the state’s institutional adequacy: if AI produces outcomes that damage social order, the civilizational state that fails to govern it has failed at its defining purpose.

The concept has gained new relevance with AI because the structural characteristics of advanced AI systems—highly concentrated capabilities, centralized control, diffused users—replicate, in technical form, the exact governance problem that Chinese statecraft has spent millennia developing tools to address: how to prevent the concentration of power from becoming tyrannical in the Confucian sense of losing the moral authority that justifies governance. The technical structure of AI and the institutional structure of the civilizational state are, on this reading, made for each other’s analysis.

Key Ideas

Legitimation logic and AI governance. A procedural state can satisfy its legitimation requirements by following the correct process regardless of outcome; a civilizational state cannot. This structural difference means that the civilizational state has a stronger and more immediate governance imperative when AI produces damaging outcomes—not because it is more controlling but because its institutional survival depends on demonstrating that it can identify and correct such outcomes. The performance test is built into the legitimation structure.

State capacity as governance prerequisite. The civilizational-state framework implies that governance capacity—the technical expertise, institutional infrastructure, and enforcement power to actually regulate AI systems—is not a convenience but a constitutional necessity. A state whose legitimacy depends on outcomes cannot satisfy that requirement through rules it lacks the capacity to enforce. This makes the EU AI Act’s gap between sophisticated principles and limited enforcement capacity, which Zheng identifies, a legitimation failure, not merely an administrative shortfall.

The co-evolutionary dynamic. The civilizational state does not merely regulate technology from outside; it co-evolves with it. The state’s institutional characteristics shape what AI becomes within the civilizational context, and AI’s capabilities reshape what the state can be. This co-evolutionary dynamic has historical precedents across Chinese history—hydraulic engineering, printing, industrialization—each of which produced forms of governance that could not have been derived from either the technology’s technical characteristics or the state’s pre-technological logic. The outcome of the AI co-evolution is one of the defining open questions of the century.

Limits and accountability gaps. Zheng is explicit that the civilizational state’s capacity advantage in AI governance comes with a reciprocal governance problem: the same concentration of power that enables effective state intervention in AI makes effective accountability of that intervention difficult. The performance-based legitimation logic depends on the population retaining the capacity to evaluate performance—which connects directly to the de-intellectualization concern: a state that governs AI in ways that atrophy the population’s evaluative capacity is undermining the very accountability mechanism that its own legitimation logic requires.