Friday, December 27, 2024

Four Objections to Free Will Skepticism

I examine four objections against the no free will view (NF) and argue that they pose no threat to the view.

Objection 1: NF is a contradictory view.

Michael Huemer gives five arguments for the incoherence of determinism.[*1] Of the five I find three to be independent (i.e., addressing the three addresses the five), so here are my responses to those three:

First argument - There is no point in arguing for determinism.

If determinism is true, then we cannot believe other than we do.[*2] So there’s no point in arguing for determinism, because those who disagree with determinism must disagree with it, and thus cannot change their minds, and those who agree with determinism must agree to it, and thus do not need to be persuaded.

But it’s obviously false that those who agree with determinism must agree to it as if their belief in determinism is magically zapped into their brain. It may be precisely an argument for determinism that caused their agreement with determinism.

Those who argue for determinism might do so because they believe spreading truth makes the world a better place. Or they may not have persuasion in mind at all, and simply argue for determinism because they find doing so compelling. The point in arguing for determinism, in that case, is to do what you find yourself compelled to do, which is, in an oversimplified way, why anyone does anything.

Second argument - There is no point in deliberating without free will.

Huemer says:

“When we reason about free will and determinism, we are deliberating about what to believe. But deliberation presupposes the existence of alternatives that one has control over. It makes no sense to deliberate about whether to do A if you don’t think you have any choice about it.”[*3]

This is false; it makes perfect sense to deliberate even when you know that whatever choice you end up making, you couldn’t help but make that choice. You have to live with your choices, free or not. So it’s in your best interest to deliberate as well as you can so you make the best choices you can make. It would be silly to pick the first option that comes to mind before weighing the pros and cons because “I was going to do that anyway.” You know that will lead to poor choices, which is exactly why you don't adopt that way of thinking.

Because you are a rational being, you have no choice but to be sensitive to reasons. So you cannot help but reason as you do, weighing the reasons to φ or not-φ for any action φ. You inevitably choose based on how the reasons weigh up in your mind. 

So I do not see my deliberating as an expression of my free will but as an expression of my attempt (or my brain’s attempt) to make the best choice I can (or it can).

Third argument - Determinism entails a contradiction.

I’ve adapted the following argument[*4] for clarity, but the core idea remains: by assuming determinism and incompatibilism, along with some common sense premises, we can show a contradiction:

1) If S should do A, then S can do A. (Premise 1 - ‘Ought implies can’)
 
2) S should believe what’s true. (Premise 2 - from common sense)
 
3) Therefore, S can believe what’s true. (From 1,2)
 
4) S cannot believe other than how S does. (Premise 3 - assume determinism and incompatibilism)
 
5) S believes in free will. (Premise 4)
 
6) Therefore, S cannot believe what’s true. (From 4,5)

This contradiction threatens the coherence of NF. Premise 4 is guaranteed to be true with respect to some S. Premise 2 cannot, by my lights, be coherently rejected. The hard determinist (which includes the hard incompatibilist) cannot deny Premise 3. So that forces the hard incompatibilist to deny Premise 1 and to reject ‘ought implies can’, which brings us to our next objection to NF. 

I have independent reasons for rejecting ‘ought implies can’, so my rejection of the principle here is not ad hoc.

Objection 2: NF entails that 'Ought implies can' is false.

While it’s true that NF, as I defend it, entails that ‘ought implies can’ is false, this is not the unacceptable consequence some might think.

Amy Karofsky, in her defense of necessitarianism, responds to the incoherence objection herself, and in doing so rejects ‘ought implies can’:

According to the necessitarian, terms like ought and should are used to prescribe a future course of action. The speaker uses them to encourage, influence, affect, or recommend that a person takes that course. But that does not mean that there are two, genuinely possible courses that the person might take; instead, the speaker’s encouragement is merely another factor in the situation.[*5]

I’m not advocating for necessitarianism here. But I agree that normative terms like ‘ought’ and ‘should’ need not imply alternative possibilities.

Regardless of whether we have alternative possibilities, we need language for prescribing actions, and words like ‘ought’ and ‘should’ are part of that language. When faced with mutually exclusive options we must choose one option at the expense of all other options, and therefore, if we wish to avoid arbitrary choices (and we very much do; our survival depends on it), we must have a method for non-arbitrary option selection. That method, as mentioned above, involves adding up reasons for choosing each option and selecting the option with the weightiest reasons. Reasons are often consequential in nature: what are the outcomes of each option and how do they compare?

If ‘shoulds’ play this prescribing role based on reasons, then when we think that someone should have done something, we are prescribing an action we can imagine that person taking, and we prescribe the action based on the reasons apparent to us. It’s easy to say “My dad should have invested in Microsoft,” because we see the superior outcome had he done so. We imagine placing ourselves in Dad’s shoes a few decades ago, equipped with our future knowledge of stock prices, and making choices on the basis of the reasons available to us.

If that sounds like cheating, then it should, because it is. Whenever we judge someone’s actions and say they should have done differently, we are always making that judgment on the basis of the reasons available to us and not necessarily on the basis of the reasons available to the person we are judging at the time they made their choice. Even if the reasons available to us were available to the person at the time the choice is made, our sensitivity to those reasons and the relative strength of those reasons is different for us than for the person we are judging.

Given our ability to imagine Dad investing in Microsoft, does it follow that he could have invested in Microsoft? Not at all. Dad did not have access to the reasons, or sensitivity to reasons, that would have made it possible for him to do so.

Moreover, we imagine impossible things all the time, like in the case of consuming fiction. We can imagine Harry Potter well enough to follow the story and immerse ourselves into its world. But if you pull at any threads then the logic of the world quickly unravels and plot holes emerge. Magic is physically impossible, and yet we can imagine it. We can imagine Harry Potter doing otherwise (say, casting the Killing Curse on Bellatrix instead of sparing her), but Harry Potter cannot do otherwise as he does not exist.

So reasons give us the ought, and the ease of imagining someone making choices, even when those choices are impossible, gives us the (mistaken) can.

'Ought implies can' is correct in the following sense: ought implies blameworthiness in the sense of critical blame, i.e. beholding someone’s lack of quality. If someone ought to φ, but fails to φ, then we criticize this person’s lack of understanding of the reasons we see why they should have φ'd.

Compatibilists want to say that Dad could have invested in Microsoft in the sense that had he tried to do so, he would have succeeded.[*6] But that pushes the problem back a step: Of course Dad could not have tried doing so given his ignorance.

The compatibilists are right in the following sense: given the fact that Dad could not have tried not because of constraints outside of his character, but due to constraints related to his character, Dad’s failure to try says something about his character, namely his ignorance.

Denying ‘ought implies can’ entails the following: We ought to do things we cannot do. For many people that sounds strange, even unintelligible. If I cannot do something, then I cannot be blamed for not doing it. And yet, if I ought to do something, then I can be blamed for failing to do it. So if I ought to do it, I can do it. We can show the logic in the following way:

1) If I ought to φ, then not φ-ing is a failure on my part. (Premise 1 - common sense)
 
2) I ought to φ. (Premise 2 - necessarily true in some cases if there are moral facts)
 
3) Therefore, not φ-ing is a failure on my part. (Modus ponens, from 1,2)
 
4) If I cannot φ, then not φ-ing is not a failure on my part. (Premise 3 - common sense)
 
5) Therefore, I can φ. (Modus tollens, from 3,4)

Given the common sense nature of Premises 1 and 3, it looks like the proponent of NF who denies ‘ought implies can’ is forced to deny Premise 2, which amounts to denying the existence of moral facts. This brings us to Objection 3 against NF.

Objection 3: NF entails there are no moral facts.

On the contrary, moral facts fit nicely on NF. Denying moral facts wreaks havoc on morality, so the objection goes. If there is no truth to the statement “murder is wrong”, then someone can be perfectly within their epistemic rights and yet commit murder. And yet, as I myself would argue, one cannot commit murder while being perfectly within their epistemic rights.

Given what I’ve said about reasons, it’s no surprise that I accept rationalism about moral facts, which says that moral facts are a species of epistemic fact. Moral facts are facts about what we should do. Epistemic facts are facts about what we should believe, or about what’s rational to believe. We can apply epistemic facts to moral contexts when behaviors are grounded in beliefs. If an action depends on an objectively irrational belief, then that action is objectively irrational.

Therefore I must deny one of the ‘common sense’ premises, and I deny Premise 1, that oughts entail blameworthiness, which leads to our last Consequence. In short, I accept critical blameworthiness but I reject moral blameworthiness.

Objection 4: NF cannot make sense of praise and blame.

On the contrary, NF can make perfect sense of our common sense attitudes of praise and blame. If we can think of praise as in ‘singing the praises of’ something, then we can think of blame as in beholding a lack of quality in something, and this kind of bemoaning the badness of something is what I mean by critical blame.

Saying that I ought to φ is to say that the reasons for φ-ing outweigh the reasons for not φ-ing. So failing to φ is to fail to have access to, grasp, understand, or have the proper sensitivity to the reasons for φ-ing. But whether one succeeds or fails to have access to, grasp, understand, or have the proper sensitivity to reasons is not up to us, and so we cannot be credited or blamed for this success or failure. We can acknowledge the quality, or lack thereof, of a person in some respect while simultaneously acknowledging that such quality is not chosen by the person.

So on one kind of blame – critical blame – we can appropriately praise or criticize the qualities of persons, such as the quality of being rational or irrational. I suggest that our common attitudes of praise and blame often manifest exactly as this kind of critical blame. But on the blame we care about in this context – moral blame – we cannot appropriately praise or blame persons for succeeding or failing to be rational, as such success or failure is not up to the person.


*1 - Michael Huemer, Knowledge, Reality, and Value (Self-published, 2021), 192-193.
*2 - I accept hard incompatibilism, which says free will is impossible on either determinism or indeterminism. If we take free will to require a kind of self-determination, then both determinism and indeterminism preclude self-determination, so the hard incompatibilist argues, and so we can refer to both as ‘determinism.’
*3 - Huemer, Knowledge, 192.
*4 - Huemer, Knowledge, 193.
*5 - Amy Karofsky, A Case for Necessitarianism (New York: Routledge, 2022), 152-153.
*6 - Robert Kane, Four Views on Free Will (Malden: Blackwell Publishing, 2007), 12.
*7 - For an example of this kind of defense of moral facts from epistemic facts, see Terence Cuneo, The Normative Web (New York: Oxford University Press, 2007), 224.

No comments:

Post a Comment