Fairness and Free Will

(Originally published at Talking Philosophy (2013) as two posts.)

PART ONE

The concept of what is “fair” causes trouble all the way through moral and political philosophy, and even in the free will debate. Here is an argument against the existence of free will, based on a notion of fairness:

P1. We have free will only if we sometimes deserve to be praised or blamed for our actions.
P2. We do not deserve to be praised or blamed for our actions in circumstances where such praise or blame is unfair.
P3. Praise or blame for our actions is unfair unless we are causally responsible for our relevant actions all the way down.
P4. We are never causally responsible for our actions all the way down.
C1. Praise or blame for our actions is always unfair. (From P3. and P4.)
C2. We never deserve to be praised or blamed for our actions. (From P2. and C1.)
C3. ~(We sometimes deserve to be praised or blamed for our actions. (From C2.)
C. We do not have free will. (From P1. and C3.)

You can question whether the step at C3. is really needed, and you might want to rephrase something in the above to make sure this is all able to be represented in your favourite system of propositional calculus. However, at the end of the day you’ll find that it’s a deductively valid argument. So the only question is whether you accept its premises – if you do, you are intellectually committed to accepting the conclusion.

You’ll come across arguments that seem to go something like this (either expressly or implicitly), whether in the formal literature on the subject or in informal discussions on the internet and elsewhere. So, should we deny the existence of free will, based on something like this argument? (By all means, offer your improved version of it.)

Before we get to that, note a couple of points. The argument does not proceed straight from: “The world is deterministic” to “We don’t have free will.” It doesn’t even move, in a more sophisticated way, straight from: “The world is a mix of determinism and randomness” to “We don’t have free will.” That seems to be an advantage. It does not simply beg the question against theories that are compatibilist about determinism and free will (or about “a mix of determinism and randomness” and free will). Rather, it offers richer (purported) insights into how we understand free will, and if we accept these (purported) insights we might then be intellectually compelled to reject compatibilist approaches.

Conversely, note that the argument depends very much on a philosopher’s conception of free will. That may also be a strength. Philosophers are better placed than most to understand how free will has been conceived of in philosophical debate over the years and centuries. For example, they may be more familiar with the literature than scientists. If the argument shows that we lack free will, as thus understood, that is an important outcome.

Nonetheless, there may be other conceptions of free will that are not caught by the argument. What if it turns out that, when ordinary people who are not philosophers claim to have free will, they actually mean (and convey to each other) something that is rather vague and messy? Perhaps one of its primary components is merely a denial of fatalism. Or perhaps it is a denial of both fatalism and any related doctrines that involve our experienced choices being bypassed or overridden, so as to lack causal efficacy. For example, epiphenomenalism would have such an effect, even if it’s not a theory that the folk are specifically familiar with. If, as things turn out, that is really what ordinary people are trying to convey when they use the (perhaps inapposite) term “free will”, then they might not accept P1, at least not without a lot of caveats, qualifications, etc. For them, P1. might miss the point.

That said, P1. does sound rather plausible (doesn’t it?), so there is much to explore if somebody wants to reject it while also claiming that human beings have free will in some interesting sense. It seems to me that there is a lot more to be said here, but I’ll leave it for another time.

P4. is a crucial premise. It is basically a way of denying that we have libertarian free will, though again there is more to be said. Note, however, that the argument as a whole is not an argument against libertarian positions, as libertarians are likely to deny P4. Against them, I am assuming that P4. is true, and I’d support it with arguments about the world being a mix of determinism and randomness in a way that seems to preclude, if not free will (after all, there are accounts of free will that are compatible with this as far as it goes), at least ideas of ultimate self-determination and the like. The argument, then, would need to be supplemented with some further argument for P4. before it could persuade a free-will libertarian. As the argument stands, without that supplementation, it begs the question against libertarian positions on the free will question.

The drift of the argument, then, is that we cannot reject ideas of ultimate self-determination and the like while also maintaining the existence of anything that ought to be called “free will”.

So compatibilist ideas are not being rejected in P4., taken by itself. The argument is supposed to persuade compatibilists to become incompatibilists – the question is not begged against compatibilism.

Where, then, does the argument go wrong? Perhaps some compatibilists will argue that P1. is beside the point or even false, because it is not getting at what is really bugging the folk when they talk about free will (as evidenced in cultural history and contemporary popular culture). In that case, they may criticise much of the contemporary discussion of free will by analytic philosophers for having lost contact with what was bugging ordinary people in the first place. I feel some sympathy for this, as, perhaps, is evident over here, but much work would need to be done before we could be confident about this approach one way or the other.

I’ll set that aside. Meanwhile, I think an equally troubling problem arises with P3. Frankly, I see no reason at all to accept this premise. Nonetheless, it appears to have intuitive appeal for many people and to enjoy some popularity. In my next post, I’ll focus squarely on P3., asking whether we should accept it and the conception of fairness that it implies.

PART TWO

In my previous post, I considered an argument against free will (let’s call it “the fairness argument”) along these lines:

P1. We have free will only if we sometimes deserve to be praised or blamed for our actions.
P2. We do not deserve to be praised or blamed for our actions in circumstances where such praise or blame is unfair.
P3. Praise or blame for our actions is unfair unless we are causally responsible for our relevant actions all the way down.
P4. We are never causally responsible for our actions all the way down.
C1. Praise or blame for our actions is always unfair. (From P3. and P4.)
C2. We never deserve to be praised or blamed for our actions. (From P2. and C1.)
C3. ~(We sometimes deserve to be praised or blamed for our actions. (From C2.)
C. We do not have free will. (From P1. and C3.)

As I indicated in the earlier post, you can quibble with this formulation if you like; however, the fairness argument, properly formalised is deductively valid.

I accepted P4. for the sake of argument, and in any event I think P4. is actually true. Beyond that, I engaged in a certain amount of fencing, some of which, I think, casts (serious) doubt on P1. Whatever you think of either P4. or P1., however, there seems to be a real problem with P3., and this turns on the nature of “fairness”. I actually see no reason at all to accept P3. – it does not strike me as intuitively compelling, or even appealing, or as something that could be given any decisive intellectual support. To be fair(!), though, something like it appears, at least from my reading and interactions, to be rather popular.

Part of the problem with P3. is that there’s a mystery about what “praise” and “blame” really amount to. Perhaps on some conceptions of these things (error theories about praise and blame), sentences that praise or blame are always false. To keep this simple, let’s stick with praise. What if, when we praise someone for an act, we are stating (perhaps among other things) that the act complied with some objectively binding standard? If no such objectively binding standards exist, it follows that we are always saying something false when we praise somebody.

Again, what if, when we praise someone for an act, we are saying (perhaps among other things) that the person is causally responsible, all the way down, for a good act? If no one is ever causally responsible for an act all the way down, it follows, again, that we are always saying something false whenever we praise somebody.

But what if, when we praise someone for an act, we simply mean that the act is a good one in the sense of one that has such properties as to tend, relatively efficiently, to bring about the sorts of consequences favoured by the people involved in the conversation? In that case, we might often say something that is simply true. There might also be non-cognitive content to our praise, such as an expression of approval, but that content cannot be true or false – and there is going to be an interesting question about how such non-cognitive content can be unfair. Perhaps it can be, but we’d need to explore some arguments to see whether this makes sense.

I suspect, meanwhile, that praise involves both more and less than any than this. E.g., when I praise someone for an act I might be saying not just that the act is good (in the sense discussed above) but that the person is good. I.e., the person’s performance of the act has provided me with evidence that she possesses certain dispositions of character (courage, kindness, honesty, or whatever they may be) such as to be a desirable person to have around: such as to tend to act in ways favoured by me and the other people involved in the conversation, etc., etc.

If this is what expressions of praise really amount to, and if something analogous applies to dispraise or blame, then there is nothing necessarily unfair about praising someone for an act for which she was not causally responsible all the way down. Indeed, the fact that the action flowed from the dispositional structure of her character might support my words of praise or blame. The action did not happen at random, but was, to some extent, caused by the person’s character (even if this also had causes).

There is a huge body of academic literature on the words “praise” and “blame”, and what they mean, but at this stage someone who wants to run the fairness argument is already in trouble. P3. depends on a highly controversial idea, perhaps far detached from the thoughts of the folk, of what it is to praise or blame people. Or so it seems.

That’s troublesome enough, but P3. also depends on a controversial idea of fairness. The idea actually seems rather vague. Its essence seems to be an absence of bias, favouritism, patronage, nepotism, hostility, “bad vibes”, etc., in situations where, first, we are allocating/withholding benefits, rights, penalties, etc., and, second, the situation is such that exercising bias, favouritism, and so on, is somehow socially inappropriate or “bad”.

So Alice is not acting unfairly if she favours Bella as her lover rather than Clarice, even if Clarice’s good qualities might exceed Bella’s from some supposedly objective viewpoint. In a situation like this, bias, favouritism, idiosyncratic feelings of liking or attraction, “good vibes’, etc., are permitted (or so we usually assume), and fairness does not even enter into the picture.

In certain other situations, we think that bias, favouritism, etc., are not appropriate, and these are the situations where questions about fairness arise. But what does “fairness” then require? Well, the requirements will vary from situation to situation, as will the justifications that support them. In some cases, the requirements and the justifications will be deeply contested. For example, we tend (don’t we?) to think it fair that a person who is on trial in the criminal courts, or who is being sued in the civil courts, be given an accurate idea of what is alleged against her before she has to answer it. She should not be denied this because of bias, favouritism, hostility toward her from the judge, or the like.

Again, we tend to think that parents should give their various children roughly equal opportunities for happiness and success in life over time, and that any blatant lapse from this is “unfair”. We don’t want a mother or father, or the combination, to show blatant favouritism to a particular child. However, a parent may normally show bias, favour, etc., toward her own child, vis-à-vis other people’s children.

In some familiar cases, things get more complicated. What if I am working out what rates of pay to give my employees? There may be a problem if I do this based on personal bias, favouritism, patronage, whim, nepotism, etc. But it does not follow that I should pay them all equally. Fine, so how should I pay them? Should I pay my employees on the basis of their respective developed skills; on the basis of the responsibility that they have willingly taken on within my enterprise; on the basis of their average or daily productivity, as individuals, compared to similar employees in the enterprise (which may bring in issues of diligence, industriousness, etc.); on the basis of what employees with similar skills, records, etc., are likely to be paid by other enterprises (within the same labour market); or some mixture of all this (in which case, how do I measure and weight these things?); or something else? These questions weigh heavily on the minds of wages negotiators, industrial arbitration tribunals, etc., and they often develop pay fixation principles that are of at least some local assistance.

So… if I praise you as having performed a good act, or as being a generally good (or morally virtuous) person, or for having certain good dispositions of character, or having made a good judgment, or anything of the kind, am I under any obligation to be fair? Well, perhaps these important judgments should not be made on the basis of whether or not I like you (or have a family connection with you, or some such thing). They should, perhaps, be based on general criteria that I would apply to others, irrespective of personal feelings, familial loyalties, or the like. But that does not tell me what criteria I should actually use!

I take it that a claim that I am, first, in a situation where fairness is relevant (I should not exercise bias, favouritism, etc.), and, second, that I should use certain specific criteria (not others) in handing out benefits and rewards, will require something like a utilitarian justification. Of course, in many circumstances there is much conventional wisdom that may be worth deferring to about when fairness is (and is not) relevant, and about what criteria should be used to make judgments and to grant benefits, apply penalties, etc. There may be some merit in not trying to review these from scratch, using explicit utilitarian criteria. Either way, to say that I acted fairly is more or less to say that I applied the criteria that were relevant (whatever they were), in a situation that called for them (i.e., in a situation where I was not entitled to act on bias, favouritism, etc.), and (if this has to be said separately) without distortion from my personal feelings toward an affected person, etc.

Nothing at all follows from this that I must praise or blame people only if I find virtue or fault with them all the way down. People can genuinely make mistakes, act badly, show poor judgment, evidence a vicious character – or the opposites of any or all of these – without being causally responsible for their actions, judgments, characters, etc., all the way down. When we appraise them, we act fairly if we apply appropriate standards to the facts in evidence, without being biased by whom we like or dislike, the wish we could help or hinder the individual, etc.

Whatever, exactly, the ideas of praising and blaming really amount to, it is not at all obvious that they can be done fairly only if the people being appraised were responsible all the way down for their actions, judgments, and characters. Accordingly, P3. is not an attractive premise at all… and hence the whole argument is in trouble.

I’d love to see a similar argument that does not (in the view of its author) fall prey to this problem.