Can ‘Downward Causation’ Save Free Will?

Abstract

Recently, Trenton Merricks has defended a libertarian view of human freedom. He claims that human persons have downward causal control of their constituent parts, and that downward causal control of this sort is sufficient for free will. In this paper I examine Merricks’s defense of free will, and argue that it is unsuccessful. I show that having downward causal control is not sufficient for for free will. In an Appendix I also argue that Merricks’s defense of free will, together with assumptions implicit in his broader ontology, commit him to the implausible conclusion that determinism is incompatible with the existence of human persons.

Appendix: Determinism and Human Persons—A Surprising Conclusion

In this Appendix I argue that (M2) together with several assumptions implicit in Merricks’s broader ontology commit him to the Surprising Conclusion that determinism is incompatible with the existence of human persons. Of course, Merricks is no stranger to surprising conclusions, as we shall see, and so perhaps he would be content with this result. Nevertheless, I shall try to explain why I (and I suspect many others) would want to resist it. Moreover, if the Surprising Conclusion is as implausible as I claim, then to avoid this conclusion Merricks must either abandon his defense of free will or revisit some of the assumptions that lead to his broader ontological views.

Merricks is an eliminativist about inanimate macrophysical objects (e.g. baseballs); he denies that such objects exist. He argues at length that if these objects did exist, they would not cause things not also caused by their constituent atoms working in concert. Therefore, such objects, were they to exist, would be causally redundant; they would overdetermine effects also caused by their proper microphysical parts. But Merricks thinks we should resist this sort of systematic overdetermination, and so given that a plausible criterion for existence is that an entity have causal powers, he concludes we should eliminate inanimate macrophysical objects from our ontology. This is a (very) brief sketch of what Merricks calls the Overdetermination Argument. Although Merricks thinks this argument eliminates inanimate macrophysical objects, he does not think it eliminates human persons. Recall that on his view human persons have non-redundant causal powers in virtue of which we cause things not caused by our constituent atoms working in concert. Human persons are therefore neither epiphenomenal nor causally redundant. We therefore survive the Overdetermination Argument.

Merricks’s broader assumptions together with (M2) commit him to the Surprising Conclusion. Here is (an unfortunately rather cumbersome) statement of the argument for this claim. It takes the form of conditional proof: assume determinism is true and derive the conclusion that human persons do not exist.

• 1. Determinism is true: every human action is a strict consequence of events in the distant past and the laws of nature (assumption for conditional proof).

• 2. If determinism is true, no one has a choice about what actions one performs (assumption: free will and determinism are incompatible).

• 3. So no one has a choice about what actions one performs and hence no choice about what one’s constituent atoms do or are like (from 1 and 2).

• 4. If no one has choice about what one’s atoms do or are like, then no one has downward causal control of one’s atoms (equivalent to M2).

• 5. So no one has downward causal control of one’s atoms (from 3 and 4).

• 6. If no one has downward causal control of one’s atoms, then no one has non-redundant causal powers (assumption).

• 7. So no human person has non-redundant causal powers (from 5 and 6).

• 8. Whatever lacks non-redundant causal powers does not exist (consequence of the Overdetermination Argument).

• 9. Therefore, no human persons exist (from 7 and 8).

The key steps in this argument are (2), (4), (6) and (8). Of these, Merricks is explicitly committed to (2) and (4) and, I think, implicitly to (6) and (8). Premise (2) is incompatibilism. Merricks is explicitly committed to (2) because he is a libertarian, an incompatibilist who believes we have free will. Premise (4) is equivalent to (M2), which as we have seen is an essential component of his response to the bottom-up threat. Turning to (6), although Merricks is not explicitly committed to this premise, nevertheless (6) is plausible given his broader commitments. If human persons lack downward control—if, for example, my decisions are not causally responsible for my bodily movements—it is hard to see how we could ever cause anything that was not also caused by our constituent atoms working in concert. Finally, premise (8) seems to be a consequence of the Overdetermination Argument. Objects that do not cause things not also caused by their constituent atoms do not exist. If there were an object that was either epiphenomenal or regularly and non-accidentally causally redundant, then we would have to revisit the Overdetermination Argument against inanimate macrophysical objects. I conclude then that Merricks is indeed committed to (2), (4), (6) and (8) in which case he is committed to the Surprising Conclusion mentioned above: if determinism is true, no human persons exist.

The Surprising Conclusion, or SC for short, strikes me as not only surprising but highly implausible as well, and I suspect I am not alone in thinking this. It seems doubtful that anyone would abandon the belief in their own existence should it be discovered that determinism is true. The empirical discovery of determinism just doesn’t seem to be the sort of thing that could undermine the claim that human persons exist. If determinism were known to be true, most of us would not conclude that we do not exist but rather that some mistake has been made in the argument for SC. I think this fact casts at least some doubt on SC. Of course, perhaps our refusal to abandon our belief in our own existence is just stubbornness on our part, but I doubt it. Determinism is, in principle at least, an empirically verifiable thesis. So if SC is true, the existence of human persons is in principle empirically falsifiable. But the thought that scientists could one day make a discovery the implication of which is that human persons do not exist borders on the absurd.⁹

In any event, we have already been given reason to think that SC is false. Earlier I defended the claim that determinism is compatible with a human agent’s exerting downward control over her atoms, and clearly if this claim is true, determinism is compatible with the existence of human persons.