TB: Huemer's Proof of Free Will (Huemer)

Core Thesis

Huemer argues that free will must exist because the epistemic axioms required to pursue rational knowledge are incompatible with determinism. John finds the argument formally interesting but structurally circular — the key premise ("you should believe the truth") already presupposes the freedom to do otherwise.

Key Takeaways

Huemer's argument

Certain axioms (e.g., "you should believe the truth") must be assumed to pursue knowledge via reason at all.
These axioms are incompatible with determinism: if determinism is true, the concept of should has no meaning — you will only ever do what you are determined to do.
Therefore, determinism is false and free will exists (QED).

John's critiques

"Should" is doing a lot of work: The premise "you should believe the truth" presupposes the very freedom it claims to prove. It's circular.
Incompleteness analogy: We should only believe what we have rational evidence for, not simply "believe the truth" as a raw imperative. This moves closer to an incompleteness problem.
Step 5 breaks something: The argument appears to prove you can have no false beliefs in a deterministic world — which is clearly wrong. Something has gone wrong in the logical structure.
Probabilities: What if there is a certain probability of free will? The binary framing may be the source of the problem.

Technical diagnosis of the failure

Premise 3 ("if determinism is true, then if S can do A, S does A") is either false or makes the argument circular. Under the practical-ability reading of "can," P3 is simply wrong: determinism says what you will do is fixed by prior causes, not that you do everything you are physically able to do. Under the nomological-possibility reading of "can" (could have done otherwise), P3 is true but then "ought implies can" (Premise 2) already assumes libertarian free will — the argument is circular.
There is no third reading of "can" that makes both P2 and P3 true without circularity. This is the load-bearing flaw.

The Epicurus comparison

Epicurus's version is more honest: "The man who says all things come to pass by necessity cannot criticize the man who denies it." This is a pragmatic self-defeat — the determinist cannot consistently deploy normative vocabulary — not a formal proof.
Huemer's error was converting the intuition into a formal argument, which forced him to write down Premise 3 explicitly, where it collapses. The sketch survives; the formalization does not.

What makes it interesting

The argument "holds whether or not determinism is in fact false" — it proves determinism is false from within the deterministic framework by showing the framework is self-defeating. This is the genuinely interesting move. But John's meta-argument observation turns this against the proof: insensitivity to the actual truth of determinism is a sign the logical machinery is broken, not a feature.

Mental Models

Inversion — applying inversion to the argument reveals the circularity more clearly

Source

Michael Huemer, Fake Nous (Substack), 2026 Original article