JJC Smart: “Indeterminism does not confer freedom on us: I would feel that my freedom was impaired if I thought that a quantum mechanical trigger in my brain might cause me to leap into the garden and eat a slug.” (2003: 63)
Libertarian free will (FW) cannot be defended solely by looking at the implications of quantum mechanics (QM). However, the last century threw a lot of empirical evidence onto us, the implications of which for FW are not yet understood. I shall define FW as the ability to choose out of my own will between genuine alternatives. Alas I require the actions of an agent to be causae sui (CS) i.e. their own cause, or, for an agent to have the ability to originate actions. To achieve such a definition from QM implications, one ought to first look at what these implications are and establish that they can affect us. I shall then examine the trade-off between randomness and freedom and finally look for CS among their implications. I will finish by offering a partial solution to the problem. This post assumes a logical positivist point of view and ignores epistemic concerns over QM. For simplicity, it will furthermore employ a counterfactual account of causation.
LaPlace’s definition of Determinism implies that if in a nearby possible world all events up to some time T were the same, then all events after T must be the same as well. This suggests that any intellect sufficiently powerful to comprehend all events before T would be able to extrapolate all events after T as well. This is false.
According to the postulates of QM, knowledge of events within a system up to T, ceteris paribus, does not entail knowledge of events afterwards. This is because the laws of events are probabilistic at a microscopic level. Suppose you look out the window. You will find that you can see both yourself as well as the outside world. This is because some 4% of light is reflected, while the rest goes through the glass. When looking at light as particles, one says that of 1 hundred photons, 4 will be reflected. When just looking at a single photon however, one cannot say whether it will be reflected or not, only that there is a chance (4%) that it will be.
This is entailed by Heisenberg’s indeterminacy principle. We do not just lack the necessary knowledge to avoid uncertainty, there are no hidden variables, and although Heisenberg is not irrefutable, it is “difficult to see how its essence […] can ever be relinquished”. It seems rather that science will evolve in ways that further circumcise Laplace. Any Laplacian demon is impossible and “essentially illusionary“ (Margenau : 718). In essence, the claim that the same past will bring about the same future is incommensurable with reality. Heisenberg’s indeterminacy principle allows for genuine chance. Therefore there are true alternatives. (Margenau : 722)
What follows is that the same past can bring about different futures. Either a photon is reflected or it is not. Physicists ironically call this “Degrees of Freedom”, or the number of “independent ways in which a system can change in the course of its motion.” (Polkinghorne : 96). Whether this freedom is the freedom we seek I shall discuss later, for now it is important to note that there are only two possible futures, each of which has a certain likelihood of occurring.
This now suggests that quantum randomness does not imply unpredictability. There are only so many metaphysically possible outcomes of an experiment. On scale this effect becomes apparent: If the reflection experiment is repeated a billion times, we have a very accurate prediction of the outcome. Therefore even QM, although indeterministic, is not unpredictable and we must resist a reading of the theory that suggests that there are no laws governing it. In essence, acts are still “consequences of the laws of nature and events in the remote past” (van Inwagen, 1983: 16 & 56).
Now let us examine the effect of QM on human behaviour: It must be one of the premises of a libertarian argument from QM that mental events are in some way reducible to the laws of QM, otherwise the libertarian must not appeal to them.
One may argue however that we are not like photons. We are macroscopic, i.e. not governed by the laws of QM, more complex and our metabolisms better predictable. Our physiology operates on scale. For a single action, for example raising my arm, uncountable processes need to occur within me. Even if some of these were on a quantum level, uncountably many would let the effect of not accurately knowing the outcome disappear. For a single photon there is chance, but for a billion we know what will happen. If I drop a plate, only one thing can happen, and all quantum fluctuations that make up the plate will not save it from breaking. Humans are made up by quantum states, but too intricate to be explained this way.
Margenau disagrees with me: He argues that there are visible ingressions of quantum randomness into the macroscopic world. All neurochemistry is in principle reducible to QM, hence if only one of the uncountable delicate processes were governed by QM, then its outcome would be undetermined like for a single photon. If now that outcome were to affect another macroscopic processes then QM would have entrance into the biological realm. (Margenau : 720) Whether this really is the case is for neuroscientists to research. For now, let us assume such entry exists.
Now we must face the libertarian challenge of showing that this randomness can be linked to freedom. To do that we must 1) show that there are genuine alternatives as we have just done, and 2) that we can bring about some CS to bring about an uncaused choice between these alternatives.
Though, are these CS really necessary now that indeterminism is demonstrated? Let me illustrate why randomness is insufficient: Marcel chooses between an apple and an orange. Through neurophysiological processes that are influenced by QM, there is a 96% chance of picking the orange, and it is impossible to say a priori which he will pick. Surely we would not want to say that he is free to choose between them for this would logically imply that photons are free also, but they are not. To say that agents that do things according to some probability are free entails that all processes in the universe that happen according to some probability are free. There is no causal difference between Marcel and a photon. In fact one may argue that I am begging the question just by stating that Marcel chooses in the thought experiment.
Smart agrees with me by saying that if we allow mere chance to govern our actions, then we lose ownership over them. Also, it would defeat one of the ulterior aims of FW, namely moral responsibility, since any morally good action would have been brought about by chance. (Smart, 2003: 62). Still, O’Conner would differ here: He argues that if QM can be attached to behaviour, then either we are undetermined or free; and it would make little sense to go for the former if we could just as well preserve freedom (O’Conner : 282). I disagree deeply with O’Conner here, since just choosing a theory because we like it better seems arbitrary and incoherent from a logical positivist perspective.
There is a great flaw in my reasoning though! I assume that photons and Marcel are causally indistinguishable, yet many argue that we must treat agents differently from natural phenomena (Smart, 2004; O’Conner, 2003; Searle, 1984; Margenau, 1967). Alas, there is another way out of this problem: Let me come back to my original definition of FW, namely the ability to choose out of my own will between genuine alternatives. Let us assume a two-stage model of FW, where on the first ‘free’ stage causeless alternatives need to be present, and on the second ‘will’ stage an agent chooses out of her own will. We have already shown alternatives, so now we only need to account for the own will. Which leads us back to the deeper problem of CS and agent causation. To say that there are special attributes to agents that do not apply to photons would mean that Marcel can be free without having to concede to photons being free. This is a very appealing solution.
The problems of agent causation shall be preserved for another essay. Suffice it to say though, that QM to the best of my knowledge, does not allow for CS, Honderich even calls the very idea just “confused” (Honderich : 196); and even if they were observable in physics, one would again face the same challenge as before of having to explain why CS in agents would allow for freedom and not in other phenomena in the cosmos just like with randomness in photons and agents. In essence, what QM is able to do for the libertarian is to show is that there are genuine alternatives, which according to the two stage model, is already half the solution. QM hence removes “an essential obstacle” (Margenau : 719) to a libertarian account of FW. “Human freedom involves more than chance: […] choice. But it needs the chance” (Margenau : 725) Still, the main problem remains unsolved.