The Doctrine of Elements examined the materials of cognition, what the architecture supplies and what it receives, how it organizes experience, and where it necessarily fails. The Doctrine of Method asks: given what we now understand about what cognition is, how should we operate? What discipline, what canon, what architecture of inquiry is appropriate to a finite, self-referential, deterministic, prediction- making system that cannot fully know itself?

Kant frames this as the plan for an edifice whose materials have been examined. Our framing is different: there is no final edifice to build. There is a practice to cultivate, a stance to take toward the ongoing, fallible operation of the cognitive architecture.

Chapter I. The Discipline of Pure Reason

Discipline, for Kant, is negative: a restraint that prevents reason from overreaching. Under the reconstruction, this is preserved but transformed. The architecture will overreach, that is its nature. Discipline is not a switch that turns off the overreach. It is the cultivated capacity to notice when overreach is happening, and to hold the system’s confident outputs with appropriate epistemic weight given where on the convergence continuum the reasoning sits.

Discipline in the Sphere of Dogmatism

Kant warns against applying the mathematical method to philosophy. Mathematical cognition constructs its concepts in pure intuition, giving each concept an immediately verifiable correlate. Philosophical cognition works through concepts whose correlates are not similarly constructible.

Under the reconstruction, this distinction maps onto the convergence continuum. Mathematics operates at the maximally constrained end, where contextual rigidity forces near-immediate agreement among competent reasoners. Philosophy, when it concerns human cognition and its limits, operates in a middle range: there are real constraints (neuroscience, phenomenology, the structure of inference) but agreement is harder to force. And traditional metaphysics operates at the unconstrained end, where no evidence can discipline inference and posteriors never converge.

The discipline required is: recognize where on the continuum you are operating. Do not import the expected certitude of the constrained domain into the unconstrained one. The dogmatic mistake is to treat loosely-constrained reasoning as if it delivered mathematical certainty.

Discipline in Polemics

Kant observes that in purely speculative matters, no side can be definitively refuted because no evidence can decide. Defending dogmatic claims by attacking opposing ones produces endless polemics without resolution.

The reconstruction confirms this sharply. In the unfalsifiable domain, all positions are projections of the architecture. No amount of argumentation within that domain will force convergence, because the mechanism for convergence (shared evidence) is absent. What critique offers is not a method for winning speculative debates but the capacity to recognize that they cannot be won, and to redirect effort toward domains where it can produce shared knowledge.

Discipline in Hypothesis

A hypothesis, in the legitimate sense, is a provisional model held with appropriate uncertainty while evidence is gathered to support or refute it. In Bayesian terms: a hypothesis is a live element of the hypothesis space with non-negligible probability, awaiting the evidence that will raise or lower that probability.

The illegitimate use of hypothesis is to posit entities or mechanisms that cannot in principle be evidentially disciplined. A hypothesis about the soul’s immortality cannot be tested, no evidence could raise or lower its probability, and so it does not function as a hypothesis in the proper sense. It is a commitment disguised as a conjecture.

Discipline in Proofs

Kant demands that every proof in transcendental philosophy proceed from premises that are themselves secure. The reconstruction replaces this with: every inference must acknowledge the evidence base it rests on and the degree of confidence that base warrants.

Proof in the strict sense belongs to the constrained end of the continuum (mathematics). Demonstration from evidence belongs to the middle range (science). Argumentation from shared architectural features of cognition belongs to a further zone (this project). Claims to necessity or universality beyond what the architecture and the evidence can support are the characteristic overreach.

Chapter II. The Canon of Pure Reason

A canon, for Kant, is the sum of the a priori principles of the correct use of cognitive faculties. The Dialectic showed that pure reason has no canon for speculative use, its speculative ventures produce only illusion. But reason has a canon for its practical use: morality, the regulation of action, the pursuit of the ends of life.

The Ultimate Aims

Kant identifies three problems as the ultimate aims of reason: the freedom of the will, the immortality of the soul, and the existence of God. The speculative interest in these questions is, he says, small, proving them adds little to empirical knowledge. Their real importance is practical.

Under the reconstruction, these three questions correspond to the three structural features of the Dialectic:

• Freedom is the functionally necessary self-model of the agent the architectural requirement that the subject present itself as capable of choice in order to act at all. Its “truth” is neither metaphysical nor illusory but structural: it is what the system must do to be a system.

• Immortality is what the self-model, lacking access to its own termination, naturally projects. The architecture cannot model its own non-existence; the projection of continuation is a default output. Whether anything corresponds to it is undetermined.

• God is the coherence-drive projected as the ground of all possibility. Permanently undetermined as to reference, permanently present as architectural tendency.

For Kant, these practical postulates support moral life. For the reconstruction, they are structural features to be recognized and worked with, not truths to be demonstrated. We do not need to resolve the question of freedom to act freely; we act freely because the architecture requires it. We do not need to establish God’s existence to orient toward completeness; the coherence-drive will orient us regardless.

Opinion, Knowledge, and Belief

Kant distinguishes three degrees of holding-true:

• Opinion: consciously insufficient, subjectively and objectively.
• Belief: subjectively sufficient but objectively insufficient held without shared evidence to validate.
• Knowledge: both subjectively and objectively sufficient.

Under the reconstruction, these correspond to regions of the Bayesian credence space with different relations to shared evidence:

• Opinion: a credence held with acknowledged uncertainty, where the posterior is known to be weak and no commitment is warranted.
• Belief: a credence that feels certain but is not validated by evidence that would force convergence. Private conviction without intersubjective test.
• Knowledge: a credence supported by evidence strong enough to force convergence among differently-prior-ed reasoners. This is what falsifiable science produces.

The crucial reconstruction: knowledge is not certainty. Even the strongest convergence in falsifiable domains is probabilistic. The difference between knowledge and belief is not the presence of certainty versus its absence, but the presence of evidence that disciplines inference versus its absence. Knowledge is calibrated belief; belief without calibration is opinion pretending to be knowledge.

The practice of discipline is, in part, the practice of honest calibration: knowing when one has knowledge, when one has belief, when one has mere opinion. The architecture does not natively make these distinctions, it tends to feel confident regardless. Calibration is learned.

Chapter III. The Architectonic of Pure Reason

Architectonic, for Kant, is the art of constructing a system of cognition. Reason demands systematic unity, not mere aggregation. Knowledge that is not systematically organized is pre-scientific.

Under the reconstruction, the demand for systematic unity is the coherence-drive of the hierarchical generative model. Higher cortical levels demand consistency across lower levels; longer temporal horizons demand coherence across shorter ones; the most abstract levels seek to unify everything under a single organizing idea. This is what the architecture does. It is useful, it produces theories, unified explanations, fruitful generalizations.

But the demand for systematic unity is not a discovery about reality. It is a tendency of the architecture that sometimes tracks real structure (when the world is composable, as LeCun observed) and sometimes overreaches (when it demands totalities that exceed what any architecture could hold). The architectonic is a practice of system-building that knows its own tendencies, that takes systematic unity as a useful orientation without mistaking it for a guarantee.

Our own reconstruction exemplifies this. It proposes a system that organizes what is known about cognition, inference, and their limits. It does not claim this system is the final edifice. It is a provisional articulation, subject to revision as the science develops and as errors are exposed. The architectonic is always open-ended, because the architecture that produces it is itself plastic and finite.

Chapter IV. The History of Pure Reason

Kant ends the Critique with a brief sketch of the history of philosophy under three headings: the object of cognition (sensualist vs. intellectualist), the origin of pure cognition (empiricist vs. noologist), and the method (naturalistic vs. scientific, dogmatic vs. sceptical vs. critical).

Under the reconstruction, the history of philosophy is the history of different emphases on different parts of the cognitive architecture:

• Sensualists (Epicurus, Locke, the empiricists) emphasized the likelihood, the sensory evidence side of the posterior. They were right that knowledge begins with experience and wrong to the extent that they denied the architectural priors that make experience possible.

• Intellectualists (Plato, Leibniz, the rationalists) emphasized the prior, the architectural contribution cognition supplies from itself. They were right that priors are required and wrong to the extent that they thought the priors were transcendentally universal rather than biologically produced.

• Kant’s critical path was the first to recognize that both are needed and that the prior/likelihood distinction is structural, not developmental. He was right in outline and wrong in detail, wrong in specifics (the table of categories, the transcendental unity of apperception, the possibility of dispelling the paralogisms) and wrong in his diagnosis of what the prior is (he thought it transcendental; we now know it to be biological).

• The Bayesian / neuroscientific path extends Kant’s critical move: it preserves the structural insight while grounding it in the architecture we can actually study. The prior becomes the evolutionary / developmental shaping of neural tissue. The likelihood becomes the sensory signal crossing the Markov blanket. The synthesis becomes the ongoing Bayesian updating of the generative model. The categories become biologically primitive and biologically constructed features of that model. The Dialectic’s illusions become structural tendencies of finite self-referential systems.

The history of pure reason is, on this reading, the slow recognition of what cognition is: not a transcendental faculty with universal structure, but a biological process in specific tissue, shaped by evolution, producing its own limits, and generating both science and illusion from the same mechanism.

---

Closing Note

The Critique of Pure Reason opened with the image of metaphysics as a queen fallen into anarchy, endlessly contested, producing no cumulative knowledge, yet unable to be abandoned because the questions arise naturally from the structure of human thought.

The reconstruction preserves the diagnosis and recasts the prescription. Metaphysics in the traditional sense cannot produce shared knowledge because it operates in domains where no evidence can force posterior convergence. The reasons different thinkers produce mutually exclusive metaphysical systems are structural: they are running the same architecture past the domain where evidence can arbitrate. The “endless contests” are what that architecture does when the evidence runs out.

What remains is a critical practice:

• Ground inference in the architecture that actually produces it.
• Calibrate confidence to the availability of disciplining evidence.
• Recognize structural tendencies of the architecture (paralogism, antinomy, ideal) as unavoidable features, not dissolvable errors.
• Hold the self-model lightly, knowing it is produced downstream of the mechanism.
• Hold the world-model lightly, knowing it is one among many models our architecture permits.
• Remember that the subject’s felt freedom is functionally necessary even within a deterministic world-model, and that the incompleteness of self-knowledge is what makes subjectivity possible.
• Treat “necessity” and “universality” as the limit behavior of convergence under constraint, not as transcendental absolutes.
• Acknowledge that the most important questions (Is there a ground? A freedom? A self that persists?) may be structurally undeterminable, and that this undeterminability is itself a feature of what we are.

This is not Kant’s tribunal delivering verdicts, nor is it a replacement philosophy that claims his mantle. It is a naturalization of the critical project: an attempt to state clearly what the cognitive object is and does, grounded in the sciences we now have, open to the revisions they will produce.

The critical path is still open. It is wider now than Kant could have known.