Freedom & Determinism Are Incompatible
Here, I explain why free will is incompatible with determinism (Peter van Inwagen is right).*
(This is based on the first paper that I got into Philosophical Review, the top-ranked philosophy journal. It was a little technical, which I thought would help me to get into Phil Review.)
1. Background: Three Views About Free Will
What does it mean to have free will? Well, you need two things: (i) there must be alternative possibilities that are genuinely open to you, and (ii) you must be in control of your actions (you determine which possibilities are realized).
What is determinism? Basically, the view that there is exactly one future evolution of the universe that is consistent with the present state of the world and all of the laws of nature. Or: the past and the laws determine a unique future.
Traditionally, people thought determinism was incompatible with free will due to condition (i) – i.e., if determinism is true then there aren’t really any alternative possibilities. So the traditional views are that we lack free will since determinism is true, or that determinism is false since we have free will.
Surprisingly, though, most philosophers who have thought about the issue in modern times have held a third view, which practically no non-philosopher would ever entertain: the view that determinism and free will are perfectly compatible. I.e., even if everything is predetermined, we can still often have a choice about what happens. This is called Compatibilism.
How do they explain this seemingly incoherent view? Well, they might say something like this: “To say that a person can do A just means that if they tried to do A, then they would succeed. Now, it’s perfectly compatible with determinism that there should be two incompatible courses of action such that for each, if you tried to do it, you would succeed. Thus determinism is compatible with your having multiple things you can do.” Or maybe something like this: “A free action isn’t an uncaused action. It's an action that is caused by your own beliefs & values, rather than by external forces.” Or something like that. (This isn’t a post about the compatibilists’ views, though, so I won’t explain them more precisely.)
2. More Background: PvI’s Argument
Peter van Inwagen (PvI) has famously defended the “Consequence Argument” against compatibilism. Here’s a slightly modified version of it.
Terminology
Define “Np” (where p is any proposition) to mean “no one has any choice about the fact that p”. PvI also explains this as meaning: “p, and no one can render p false”. He goes on to explain that “rendering p false” means performing an action such that your performance of it is sufficient for ~p. This is important.
Example: Intuitively, I have a choice about the fact that the apple in my kitchen is uneaten, because I can eat that apple, which would suffice for “the apple is uneaten” to be false.
Let P0 be a complete and correct description of the state the universe was in 14 billion years ago (just after the Big Bang).
Let L be a complete and correct statement of all the laws of nature.
Let P be a correct description of anything happening now or in the future.
Determinism holds that P0 and L together logically entail P.
Assumptions
Intuitively, it seems that no one has any choice about the state the universe was in 14 billion years ago. Also, no one has any choice about what the laws of nature are. Finally, it seems that these two inference rules are valid:
Rule A: If you have Np and Nq (where p and q are any propositions), you can infer N(p & q).
Rule B: If you have Np, and p logically entails q, you can infer Nq.
Deduction
So here’s a deduction:
1. NP0. (Premise.)
2. NL. (Premise.)
3. N(P0 & L). (From 1, 2; rule A.)
4. Assume determinism is true. Then (P0 & L) entails P.
5. NP. (From 3, 4; rule B.)
This shows that if determinism is true, then no one has any choice about anything that is happening or will ever happen. So the compatibilists are wrong.
3. Objection
I thought about this issue off and on for several years. Eventually, I thought of a counterexample to rule A (then I learned that Thomas McKay and David Johnson had scooped me, but I still published about it anyway).
Let’s say I have a device that can shoot particles into a certain box. I never in fact activate the device, but I could activate it, in which case it would shoot one particle. Suppose that if I activated it, it would be quantum mechanically random which half of the box the particle would wind up in, but it would definitely end up in either the left half or the right half. Then the following are true, on PvI’s definitions:
1*. I have no choice about the fact that (no particle enters the left half of the box).
2*. I have no choice about the fact that (no particle enters the right half of the box).
3*. I have a choice about (no particle enters the box).
This is because if I shoot a particle into the box, this would be sufficient for a particle to enter the box, but it wouldn’t be sufficient for a particle to enter the left half, nor would it be sufficient for a particle to enter the right half. Letting p = “no particle enters the left half” and q = “no particle enters the right half”, we have
Np
Nq
~N(p & q),
which is a counterexample to rule A.
4. A Fix
So van Inwagen is wrong, and freedom is compatible with determinism? No. We can fix up PvI’s argument. Redefine “Np” to mean “No matter what I do, p”. In other words: “p, and for each action, A, that I can perform, if I were to perform A, p would definitely still be the case.”
The premises of the main argument are still plausible: the universe was in a certain state in 14 billion B.C., and that will continue to be true no matter what I do. Also, the laws of nature will continue to be what they are no matter what I do. E.g., I can eat an apple, and if I eat an apple, the past will definitely still be what it was, and the laws of nature will definitely still be as they are. I can go for a walk, and if I do that, the past and the laws of nature will definitely still be as they are. And so on, for every action I can do.
The two inference rules are also now valid. If, for any action I can do, p would definitely be the case, and for any action that I can do, q would definitely be the case, then for any action I can do, (p & q) would definitely be the case.
5. Illustrations
Okay, here are a couple of illustrations that make the argument clearer.
First illustration
Let’s say that a student comes to me near the end of the semester and asks, “Hey Mr. Huemer, how can I get an A in this class?” (students often do stuff like that, including calling professors “Mister”). Suppose I say: “Well, in order to get an A in the class, you would have to have gotten at least an 87% average on the first 4 tests. But in fact, you only got 75%.”
Now, what can the student infer from my answer? Obviously, the student could validly infer “I can’t get an A in Huemer’s class.”
(Note: Don't say he could get an A by convincing me to change the grading standards. That's rejecting the premise. You are to assume that what I said is true (in order to get an A, he would have to have gotten 87%, and he only got 75%), then work out what follows from that.)
Second illustration
A scientist is testifying before Congress about global warming. Congresswoman Alexandria Ocasio-Cortez asks him, “Hey, science dude, how can we prevent any further global warming?”
The scientist responds: “Well, after you release CO2, it takes about 50 years for it to reach the upper atmosphere, where it has its maximum effect. So in order to avoid any more warming, we would have to have stopped CO2 emissions 50 years ago. As you know, we did no such thing.”
Here, even AOC could work out the consequence: we cannot prevent further global warming.
(Note: Don't say we could avoid further warming through geoengineering. Again, that's just denying the premise. The question isn't whether the scientist's statement is true. The question is what follows, if the statement is true.)
Lesson
What these examples illustrate: If, in order for you to do A, something would have to have happened in the past that did not happen, then you cannot now do A.
This supports the premise “NP0” in the above argument with my “no matter what I do” interpretation. It also makes for a more direct argument for incompatibilism:
If determinism is true, then in order for me to do anything different from what I’m actually going to do, things would have to have happened in the past that did not happen. Therefore, if determinism is true, then I cannot do anything different from what I’m actually going to do.
Hence, determinism is incompatible with my having alternative possibilities, and so incompatible with free will.
* Based on my “Van Inwagen’s Consequence Argument,” Philosophical Review 109 (2000): 524-43.