Naoto Fujie

Peter Wolfendale’s computational Kantianism and minimalist Hegelianism are frequently described as translations of German Idealism into the idiom of computation and inferentialism. That description is incomplete. Wolfendale’s more consequential achievement is architectural: he reorganises Kantian objectivity and apperception, Hegelian conceptual revision, Sellarsian anti-givenness, Brandomian inferential normativity, and artificial-intelligence research as interdependent conditions of artificial rational agency. Object identification, representation, judgement, incompatibility, exception, self-modelling, and revision become functional constraints that can be compared with technical research. This article reconstructs that achievement through close engagement with The Revenge of Reason and Dialectics and Dialogue, while treating Reza Negarestani and Ray Brassier as internal comparison points within neorationalism. It then tests the architecture against research on joint-embedding world models, causal intervention, relational representation, value-equivalent planning, continual-learning plasticity, large-language-model reasoning, and mechanistic interpretability. The tests show genuine operational and explanatory gains, but they do not license an automatic inference from functional decomposability to ontological disclosure. The article develops that claim through three distinctions: operational architecture is not identical with genetic ontology; generated reasons are not necessarily faithful reports of causal computation; and model revision is not institutional correction. Embodied cognition, individuation theory, neural and destructive plasticity, distributed cognition, technical milieus, and external memory are introduced to specify how rational capacities arise across brain, body, space, public language, and infrastructure. The Amazon recruiting experiment provides a historical institutional case, while a carefully delimited LLM-assisted extension shows how summarisation, ranking, and generated explanations amplify the separation between technical revisability and affected-subject power. Five audit-philosophical constraints—Target Non-Exhaustibility, Decomposition–Genesis Non-Identity, Counterexample Externality, Institutional Return, and Directional Legitimacy—are proposed alongside a non-aggregative Human Capability Revision Vector. Digital systems may accelerate pragmatic inquiry, but such acceleration is truth-conducive and politically legitimate only when external counterevidence can defeat the system, embodied and social consequences remain visible, and objections can alter decisions, objectives, responsibilities, and repair.

Artificial intelligence turns the transcendental problem into a comparative inquiry into finite cognitive architectures. This article asks how materially realised systems stabilise objects, relevance, temporality, self-reference, and action without confusing architecture-relative access with exhaustive reality. It develops a generalized Kantian account organised around four claims. First, every finite cognitive architecture constitutes a typed operational world through selective reception, integration, memory, modelling, and action. Second, successful constitution does not entitle an architecture to identify its operational world with reality as such; the transcendental remainder names this non-entitlement rather than a hidden positive object. Third, recursive self-modification produces transcendental plasticity by altering conditions of access while preserving determinate organisation. Fourth, when artificial operational worlds acquire institutional authority, their non-exhaustiveness must be mediated by contestability, return, counter-audit, reversibility, and proportionate remedy. The argument integrates Kant, Sellars, Brandom, Simondon, Wolfendale, contemporary consciousness science, and architecture-specific analyses of language models, memory agents, embodied robotics, multi-agent systems, and institutional scoring. It neither attributes consciousness to present systems nor excludes it in principle. Instead, it supplies a disciplined framework for distinguishing functional indicators, phenomenal attribution, ontological progress, and legitimate governance.

This article asks how a reason for protecting human beings from technological sacrifice can be constructed after humanity has ceased to function as the metaphysical source, privileged measure, and final purpose of intelligence and evolution. The problem arises at the convergence of Reza Negarestani's account of intelligence as a socially and artificially elaborated project, Nick Land's account of technocapital as a nonpersonal process, formalist and network-state proposals for technologically selected governance, accelerationist appeals to growth and thermodynamic development, transhumanist projects of bodily redesign, and singularity narratives in which biological humanity becomes a transitional substrate. The article accepts the decentring of the human while rejecting the inference that decentring authorizes disposability. It develops Post-Anthropocentric Protective Enlightenment as a framework in which protection derives from embodied susceptibility to harm, asymmetric dependency, irreversibility, contestability, responsibility, and the preservation of future options. Its central norm, Non-Sacrificial Dialectics, holds that contradiction, failure, and injury can contribute to later knowledge without becoming retrospectively justified by that contribution. The framework is applied to neurotechnology, biological scoring, AI data labour, experimental burdens in low-resource settings, and the prospective closure of children's futures. It is then institutionalized through Ontological Score Legal Logic, Process-Logical AI Jurisprudence, Human Capability Biosecurity, and a vector-based FED-DLR directional audit. The resulting position preserves the nonhuman future of intelligence while denying any institution, corporation, state, artificial agent, or imagined future population a license to treat present persons as expendable developmental material.

Contemporary philosophy has significantly expanded the scope of philosophical inquiry through speculative realism, object-oriented ontology, posthumanism, process philosophy, science and technology studies, and critical AI research. However, relatively little attention has been given to how philosophical concepts become institutional infrastructures that organize classification, scoring, administrative intervention, and algorithmic governance. This paper argues that contemporary philosophy faces an institutional limitation: it lacks a systematic framework for evaluating how ontological and normative commitments are translated into operational systems. To address this problem, the paper develops Audit Philosophy, a framework that reconstructs contemporary philosophy through institutional analysis rather than replacing it. Building upon Kantian epistemic limits, Hegelian dialectics, and Marxian institutional critique, the framework introduces Free Evolution Dialectics and the Directional Legitimacy Rate (FED-DLR) as principles for evaluating whether institutions preserve future capacities for revision, contestation, repair, and re-entry. Rather than producing aggregate scores, the framework employs plural evaluative dimensions grounded in evidential proportionality, reciprocal revisability, intervention limits, and non-aggregation. The paper further analyzes how classifications, metrics, and AI-driven decision systems reshape social ontology through what is termed score-bureaucratic social ecosystems. It examines institutional mechanisms including ontological score inheritance, counter-audit rights, responsibility displacement, and institutional return, demonstrating that governance increasingly operates through classifications that regulate future opportunities rather than merely describing existing conditions. The central contribution of Audit Philosophy is to shift philosophical evaluation from questions of representational correctness toward the institutional consequences of concepts. Philosophical theories should therefore be assessed not only according to explanatory power or ontological sophistication but also according to the directions they authorize within technological and administrative systems. The paper concludes that philosophy requires an institutional criterion capable of evaluating whether concepts expand free evolutionary possibilities or reinforce closed systems of classification, control, and inherited constraint.

Contemporary philosophy has significantly expanded the scope of philosophical inquiry through speculative realism, object-oriented ontology, posthumanism, process philosophy, science and technology studies, and critical AI research. However, relatively little attention has been given to how philosophical concepts become institutional infrastructures that organize classification, scoring, administrative intervention, and algorithmic governance. This paper argues that contemporary philosophy faces an institutional limitation: it lacks a systematic framework for evaluating how ontological and normative commitments are translated into operational systems. To address this problem, the paper develops Audit Philosophy, a framework that reconstructs contemporary philosophy through institutional analysis rather than replacing it. Building upon Kantian epistemic limits, Hegelian dialectics, and Marxian institutional critique, the framework introduces Free Evolution Dialectics and the Directional Legitimacy Rate (FED-DLR) as principles for evaluating whether institutions preserve future capacities for revision, contestation, repair, and re-entry. Rather than producing aggregate scores, the framework employs plural evaluative dimensions grounded in evidential proportionality, reciprocal revisability, intervention limits, and non-aggregation. The paper further analyzes how classifications, metrics, and AI-driven decision systems reshape social ontology through what is termed score-bureaucratic social ecosystems. It examines institutional mechanisms including ontological score inheritance, counter-audit rights, responsibility displacement, and institutional return, demonstrating that governance increasingly operates through classifications that regulate future opportunities rather than merely describing existing conditions. The central contribution of Audit Philosophy is to shift philosophical evaluation from questions of representational correctness toward the institutional consequences of concepts. Philosophical theories should therefore be assessed not only according to explanatory power or ontological sophistication but also according to the directions they authorize within technological and administrative systems. The paper concludes that philosophy requires an institutional criterion capable of evaluating whether concepts expand free evolutionary possibilities or reinforce closed systems of classification, control, and inherited constraint.