Is Determinism Self-Refuting?

Determinism is usually considered to be the claim that every event is necessitated by the laws of nature (or whatever) in conjunction with the causal history leading up to that event. Many people consider whether or not determinism is true to be an open question. Some folks think it’s a contingent thesis, obtaining in some but not all possible worlds, while others take it to be necessarily true or false, like many positions in metaphysics. In this post, the modal status of determinism will be briefly touched upon at the end. First, what I want to know about determinism is whether or not it is false. Here’s an argument concluding that it is:

  1. We should refrain from believing falsehoods about determinism.
  2. Whatever should be done, can be done (Ought Implies Can).
  3. If determinism is true, then whatever can be done, is done.
  4. At least one person believes that determinism is false.
  5. We can refrain from believing falsehoods about determinism. (1, 2)
  6. If determinism is true, then we refrain from believing falsehoods about determinism. (3,5)
  7. If determinism is true then it is true that determinism is false. (6,4)
  8. It’s true that determinism is false. (7)

The argument isn’t very straightforward. Premise two is a formulation of “ought implies can” which is a principle that states that if S ought to bring about some state of affairs P, then S can bring about the state of affairs P. Premise three is a consequence of determinism. If determinism is true, then everything that you can do, you do at some point in your causal history. In effect, it’s a denial of the notion that we have alternative possibilities available to us when deciding to act. Premise four is the claim that at least one person believes that determinism is false. The fourth premise is an empirical claim in need of empirical justification, which is readily available in the form of books, papers, and talks from philosophers who accept indeterminism. From the first two premises, it follows that we can refrain from believing falsehoods about determinism, because ought implies can plus the epistemic norm expressed by premise one jointly entail that we can do what we ought to do, and in this case that is refrain from believing things that are false about determinism. From three and five it follows that we (at least somebody) refrains from believing falsehoods about determinism, because we can refrain from believing falsehoods about determinism, which, combined with determinism, entails that we do refrain from believing falsehoods about determinism (what we ought to do we can do, and what we can do we actually do). From six and four it follows that given the truth of determinism, at least one person’s belief that determinism is false turns out to be true, which is because that person ought to refrain from believing falsehoods about determinism, and that person can refrain from believing falsehoods about determinism, so that person actually does refrain from believing falsehoods about determinism. The transition is from ought to refrain from P to can refrain from P to actually does refrain from P, where the first transition is due to ought implies can, and the second transition is due to determinism. Finally, the conclusion is that it’s true that determinism is false, which follows from seven since determinism implies its own falsehood given the other premises.

This is quite a fishy argument in my opinion. I’m utterly unconvinced that it is sound, yet I cannot find anything blatantly wrong with it. The argument most likely fails somewhere, but I cannot think of anywhere in particular that sticks out as being clearly implausible/false. If I had to take a stab at critiquing the argument, I would question the application of “ought implies can” to belief formation. The application of that principle to doxastic practices assumes that belief formation is a subset of human ability. Perhaps belief formation is not within the realm of our control, so it does not qualify as an ability.

Another issue is with formulating determinism as the thesis what whatever a person can do, that person actually does. Perhaps the “can” in this formulation is implicitly indexed to our actual world such that it is a redundant formulation of determinism, and a broader interpretation of “can” that involves nearby possible worlds is in order. I’m thinking of something to the effect of a person can do X even if they do not do X in the actual world, but in some possible world, but in that world that person is determined to do X. However, I’m unsure how this formulation of determinism avoids the argument above, as it would require loading different background conditions into the analyses of what a person can do at a particular world. Since there are different background conditions, “ought implies can” wouldn’t apply to the actual world. Given S’s background history in the actual world, S ought to P, but S can only P in a nearby possible world with a different causal history and set of epistemic responsibilities, thus making “ought implies can” as formulated by the argument irrelevant to this discussion.

One last way to possibly resist the conclusion is to deny that ought necessarily implies can. Perhaps there are impossible demands that our norms place on us. If moral obligations/norms are not always such that we can fulfill them, then perhaps the same holds for epistemic obligations/norms.

A last question that can be asked is about determinism’s modal status given this argument’s soundness. Some may think that if the principles that are conjoined with determinism are themselves necessarily true/false, and they entail that determinism is necessarily false, then determinism has the status of necessity, be it necessary truth or falsity. But this argument proves no such thing. There is a contingent premise in the argument, and there are (probably) some possible worlds where nobody believes that determinism is false, which means that if a philosopher in those worlds ran this argument, they would conclude that determinism is true, so it follows that determinism’s truth value is contingent on which possible world one examines it in.

What I find interesting about the argument is that there is no obvious, knockdown refutation at hand (as far as I can tell). All of the ways around the argument involve substantive philosophical commitments that require independent argument that will itself probably carry further substantive philosophical commitments. So, the argument is a bit of a challenge to those who entertain the possibility of determinism being true (as I do).

The argument can be found here.


  • Guestiguess

    Is the first premise stating that there is a kind of non-moral obligation?

    • I suppose it can be called an epistemic obligation to refrain from believing falsehoods. It does seem presupposed by the debate in the literature on free will, save for those who think even if free will doesn’t exist, we should believe that it does for social benefits and will power purposes.

  • Mind-Forged Manacles

    I’ve a more general question about “ought implies can”. How should this be understood? I can think of 2 equally plausible things it could mean, and they are compatible so far as I can tell:

    1) X must be at least morally permissible to be a moral obligation. That is, moral permissibility is a necessary (but not sufficient) condition for moral obligation.

    2) A mix of 2 types of modality. X must at least be metaphysically possible in order to be an ought.

    These 2 potential meanings of “ought implies can” seen sensible & compatible to me, I’m just not sure which one isn often meant by that phrase. 😅😅