Moral Desert & Dennett & Caruso

The problem of 'moral desert' is closely linked to regret minimization. 

It is useful to have a notion of 'deserved outcome' alongside one of 'actual outcome'. In some contexts, it makes better sense to minimise regret with respect to the former not the latter.

Why? 

Well, as Ecclesiastes says- For in much wisdom is much grief: and he that increaseth knowledge increaseth sorrow.

The regret minimizing course is to invest in 'non thetic awareness', Bataille's 'non-knowledge', agnotology, or any thing else which moves society to reliance upon open markets or zero knowledge proofs.

Moral desert then becomes a problem of 'mutual information' which in turn can be looked at as one of 'Brown clustering' which in turn could enable better 'directed information'. There is a technical reason why 'mutual information' must be 'ontologically dysphoric'- i.e. arise only by something like 'Cohen forcing'- if we evolve on an uncertain fitness landscape.

Obviously, there is a canonical way of saying this but, till South Park does 'the Good Place', I don't know what that is.

Meanwhile, here is Caruso & Denett's comedy double act- 

Caruso: [Dan,] you have famously argued that freedom evolves and that humans, alone among the animals, have evolved minds that give us free will and moral responsibility.

But, if a trait has evolved, then some members of the species won't have it or will have it in a different way. In that case nothing involving universal morality could be predicated of it. 

I, on the other hand, have argued that what we do and the way we are is ultimately the result of factors beyond our control, and that because of this we are never morally responsible for our actions, in a particular but pervasive sense – the sense that would make us truly deserving of blame and praise, punishment and reward.

The fact that something can't be truly predicated doesn't mean it can't be usefully predicated- i.e feature in pragmatics. We can get along just fine without 'true' Justice. But we do need a more or less independent, protocol bound, juristic process to get along. Thus, it is enough that there be 'natural' or 'canonical' Schelling focal points for coordination games to proceed in an orderly fashion as if there were a 'true' public signal giving rise to a robust correlated equilibrium.

While these two views appear to be at odds with each other, one of the things I would like to explore in this conversation is how far apart we actually are. I suspect that we may have more in common than some think – but I could be wrong. To begin, can you explain what you mean by ‘free will’ and why you think humans alone have it?

Dennett: A key word in understanding our differences is ‘control’. [Gregg,] you say ‘the way we are is ultimately the result of factors beyond our control’ and that is true of only those unfortunates who have not been able to become autonomous agents during their childhood upbringing. There really are people, with mental disabilities, who are not able to control themselves, but normal people can manage under all but the most extreme circumstances, and this difference is both morally important and obvious, once you divorce the idea of control from the idea of causation. Your past does not control you; for it to control you, it would have to be able to monitor feedback about your behaviour and adjust its interventions – which is nonsense.

In fact, if your past is roughly normal, it contains the causal chains that turned you into an autonomous, self-controlling agent. Lucky you. You weren’t responsible for becoming an autonomous agent, but since you are one, it is entirely appropriate for the rest of us to hold you responsible for your deeds under all but the most dire circumstances. As [the American country singer] Ricky Skaggs once put it: ‘I can’t control the wind, but I can adjust the sails.’ To suppose that some further condition should be met in order for you or anyone else to be ‘truly deserving’ is to ignore or deny the manifest difference in abilities for self-control that we can observe and measure readily. In other words, the rationale or justification for excusing someone, holding them not deserving of criticism or punishment, is their deficit in this competence. We don’t try to reason with bears or babies or lunatics because they aren’t able to respond appropriately. Why do we reason with people? Why do we try to convince them of conclusions about free will or science or causation or anything else? Because we think – for good reason – that in general people are reasonable, are moved by reasons, can adjust their behaviour and goals in the light of reasons presented to them. There is something indirectly self-refuting in arguing that people are not moved by reasons! And that is the key to the kind of self-control which we are justified in treating as our threshold for true desert.

We would long ago have abandoned reasoning if our aim was to change others' behaviour because, experience has shown us, only economic incentives, or juristic penalties, can do that. However, being able to give reasons for our actions has a signalling or screening function. 

Caruso: I don’t disagree with you that there are important differences between agents who have the kind of rational control you highlight and those who lack it. Such a distinction is undeniable. A normal adult who is responsive to reasons differs in significant ways from one who is suffering from psychopathy, Alzheimer’s or severe mental illness.

This is merely a difference of degree. Furthermore, 'normal adults' may also lack capacity of a similar kind by reason of lack of knowledge or means or opportunity or motive.

I have no issue, then, with acknowledging various degrees of ‘control’ or ‘autonomy’ – in fact, I think you and other compatibilists have done a great job highlighting these differences. My disagreement has more to do with the conditions required for what I call ‘basic desert’ moral responsibility. As a free-will skeptic, I maintain that the kind of control and reasons-responsiveness you point to, though important, is not enough to ground basic-desert moral responsibility – the kind of responsibility that would make us truly deserving of blame and praise, punishment and reward in a purely backward-looking sense.

Nothing is 'grounded'. It is either useful of useless. In any case, grounding could always be in something like Bataille's 'base materialism' or involve weird 'zero-point energy', so why bother with it?

Consider, for example, the various justifications one could give for punishing wrongdoers. One justification, the one that dominates our legal system, is to say that they deserve it.

Nonsense! The legal system only punishes illegal acts. It takes no cognisance whatsoever of all manner of wrong-doing- e.g. sticking pins in a voodoo doll with the intent of causing excruciating pain to me.

This retributive justification for punishment maintains that punishment of a wrongdoer is justified for the reason that he/she deserves something bad to happen to them just because they have knowingly done wrong.

But this is not the theory which underpins our legal system. Legal positivism has no trace of it. 

Such a justification is purely backward-looking.

Rubbish! Consider the famous case of Regina vs Shivpuri. The letter of the law said that intent- which is forward looking- is enough. This was upheld. Shivpuri, who studied Law at S.O.A.S, thought he couldn't be convicted because it was 'impossible attempt'- he hadn't brought any drugs into the country, rather he had proof that he had brought in some thing which looked like drugs but was  harmless. The Court still thought he had 'intent' and so he was sent to jail.

For the retributivist, it is the basic desert attached to the criminal’s immoral action alone that provides the justification for punishment.

Not necessarily. The retributivist may just get off on punishing people, or seeing them punished, and suggests that this desert attaches itself to him with respect to criminals but not law-abiding people.

This means that the retributivist position is not reducible to consequentialist considerations that try to maximise good outcomes in the future, nor in justifying punishment does it appeal to wider goods such as the safety of society or the moral improvement of those being punished.

But, it could easily do so by assigning a value to the pleasure taken in another's punishment, or by postulating a 'demonstration effect'. 

I contend that retributive punishment is never justified since agents lack the kind of free will and basic-desert moral responsibility needed to ground it.

You could also contend that retributive punishment is never justified since the Nicaraguan horcrux of the neighbor's cat said so. The epistemic value of both statements is equal. They are wholly arbitrary.

While we may be sensitive to reasons, and this may give us the kind of voluntary control you mention, the particular reasons that move us, along with the psychological predispositions, likes and dislikes, and other constitutive factors that make us who we are, themselves are ultimately the result of factors beyond our control.

But not beyond the control of the Nicaraguan horcrux of the neighbor's cat.

And this remains true whether those factors include determinism, indeterminism, chance, or luck.

Which is why no situation where we have caused a person to be unjustly punished was our fault or represented a moral failure.  All is as willed by that Nicaraguan horcrux.

This is not to say that there are not other conceptions of responsibility that can be reconciled with determinism, chance or luck.

Nor is it to say anything at all. 

Nor is it to deny that there may be good forward-looking reasons for maintaining certain systems of punishment and reward. For instance, free-will skeptics typically point out that the impositions of sanctions serve purposes other than punishment of the guilty: it can also be justified by its role in incapacitating, rehabilitating and deterring offenders.

But only if the Nicaraguan horcrux of the neighbor's cat wills it so.  Otherwise the free will sceptic isn't really making any sort of argument any more that a parrot, or a wind up toy can make an argument.

My question, then, is whether the kind of desert you have in mind is enough to justify retributive punishment?

Free will, exercised in good faith, is its own justification. However a justification offered by some one acting in bad faith, our under the control of another, is no justification at all. Why? To be worthy of evaluation, a justification must have a concrete model- i.e. someone, with free will and in good faith must actually believe it to be a justification. 

If not, then it becomes harder to understand what, if anything, our disagreement truly amounts to since forward-looking justifications of punishment are perfectly consistent with the denial of free will and basic-desert moral responsibility.

No justification is consistent with the denial of free-will because it can not certify itself as a good faith proposition.

And if you are willing to reject retributivism, as I think you might be, then I’m curious to know exactly what you mean by ‘desert’ – since it’s debatable whether talk of giving agents their just deserts makes any sense devoid of its backward-looking, retributive connotations.


Dennett: You grant that the distinction I make between people who are autonomous and those who are not (because of various limits on their abilities to control themselves) is important, but then say that it is not enough for ‘the kind of desert’ that would ‘justify retributive punishment’. I too reject retributivism. It’s a hopeless muddle, and so is any doctrine of free will that aspires to justify it.

A muddle, like a ball of string which has gotten knotted up, may be very difficult to untangle. Indeed, it may be a hopeless task for most people. But, if the thing was important enough, an expert in knot theory might be able to do so. 

An innocent man may give a hopelessly tangled account of how he came to appear, circumstantially speaking, indisputably guilty. Yet, Perry Mason can untangle that muddle. 

It is enough that a person with freewill, acting in good faith, has a, however muddled, justification of a certain sort, for that justification to have a consistent depiction as a set of juristic doctrines. If this were not the case, then no juristic proceeding could be both positivist and Fair. It must be the case that the game has been rigged in advance.

But that doesn’t mean there is no ‘backward-looking’ justification of punishment.
It’s quite straightforward. On Monday you make me a promise, which I accept in good faith, and rely on when I adjust my own activities. On Friday, I discover you have broken your promise, with no excuse (what counts as an excuse has been well-explored, so I will take that on without further notice). I blame you for this. My blaming you is of course backward-looking: ‘But you promised me!’

It is also retributive.  The person upbraided in this way loses face. That is itself a punishment.

Autonomous people are justly held responsible for what they did because all of us depend on being able to count on them.

So are slaves or agents or heteronomous people or animals trained for specific tasks. Autonomy neither increases nor reduces responsibility. What matters is whether others had a right to count on them. 

It is for this reason that among their responsibilities is preserving their status as autonomous agents, guarding against the usurpation or manipulation of their own powers of discernment and decision.

There is no such responsibility unless we have a legal right, arising from a contract, or statute or customary law, enforceable against them. 

Thus if Dennett gets drunk and is duped by an Iyengar, I have no recourse against him. He has no responsibility towards me with respect to abstaining from alcohol or not conversing with cunning Iyengars.

So we can blame them for being duped, for getting drunk, etc.

No we can't- at least, if we wish to preserve our reputation for sanity. 

When we blame them, we are not just diagnosing them, or categorising them; we are holding them deserving of negative consequences.

No. We are displaying our stupidity and childishness. 

If this isn’t ‘basic desert’ then so much the worse for basic desert. What is it supposed to add to this kind of desert?

'Basic desert' could mean the set of shadow prices associated with a particular solution to Society's Kantarovich-Monge problem. If this solution is 'Muth Rational' and coincides with the outcome of a repeated game, then a system of positive law could enforce it. We could say 'Basic desert' can have a concrete model as a Moral Economy. This enables us to dismiss stupid or childish claims of the sort Dennett has put forward. It gets rid of the nuisance that would be occasioned by everybody holding everybody else accountable for things which they are not legally accountable at all.

The fact is – and I invite you to consider whether it is a fact – that autonomous people understand that they will be held to account and have tacitly accepted this as a condition for their maintaining their freedom in the political sense.

But, heteronomous people understand this even better as do animals trained for a specific economic purpose. 

Autonomous people have the right of exit. If they don't wish to use that right, they may still use their own countervailing power to hold the other to account if they make a nuisance of themselves. 

I take this to be all the grounds we need for justifying the imposition of negative consequences (under all the usual conditions).

No. You are assuming that autonomous people have already tacitly accepted the thing which you want to impose. Why bother? It would be a waste of resources. We have all tacitly accepted that we need to breathe in and then breathe out. There is no need to justify the imposition of negative consequences for failure to breathe in and then breathe out.

The difference between the madman who is physically restrained and removed to quarantine for the sake of public safety, and the deserving culprit who is similarly restrained and then punished, is large, and it is a key feature of any defensible system of government.

The differences between madmen are large as are the differences between culprits. So are the differences between the stars and flowers. So what? Such assertions have no epistemic value. 

A defensible system of Government is one that can defend itself against actual threats. Philosophers are too stupid and ignorant to defend anything.

The culprit has the kind of desert that warrants punishment (but not ‘retributive’ punishment, whatever that is).

There is no way to prevent a culprit who is being punished from being retributively punished in the eyes of his victims.

There is no incompatibility between determinism and self-control
As I have argued before, we can see this rationale in a simpler domain of human activity: sport. The penalty kicks and red cards of soccer, the penalty box of ice hockey, the ejection of players for flagrant fouls, etc, all make sense; the games they enable would not survive without them. The punishment (consider the etymology of ‘penalty’) is relatively mild because ‘it’s only a game’, but if the transgression is serious enough, large fines can be assessed, or banishment from the game, and, of course, criminal prosecution for assault or cheating also lurks in the wings. Free-will skeptics should consider if they would abolish all these rules because the players don’t have real free will. And if they would grant a special exemption for such penalties in sport, what principle would they cite for not extending the same policies to the much more important game of life?
You also say ‘the particular reasons that move us, along with the psychological predispositions, likes and dislikes, and other constitutive factors that make us who we are, themselves are ultimately the result of factors beyond our control’. So what? The point I think you are missing is that autonomy is something one grows into, and this is indeed a process that is initially entirely beyond one’s control, but as one matures, and learns, one begins to be able to control more and more of one’s activities, choices, thoughts, attitudes, etc.

But this is equally true of kids- who play way more hockey and football and so on than elderly academics- as well as slaves or fatalists or people under the spell of a hypontist. Autonomy adds nothing and takes away nothing from what is simply an adaptation to the fitness landscape. 

Yes, a great deal of luck is involved, but then a great deal of luck is involved in just being born, in being alive. We human beings are well designed to take advantage of the luck we encounter, and to overcome or deflect or undo the bad luck we encounter, to the point where we are held responsible for not taking foolish chances (for instance) that might lead to our losing control. There is no incompatibility between determinism and self-control.

Obviously, 'self-control' is only meaningful if there is some deterministic process whereby the self can be controlled. If the thing is wholly non-deterministic, self-control might still exist but- with our current level of mathematical and biological knowledge-  that self would be something truly rich and strange.

Caruso: Well, I’m glad to know that you reject retributivism along with ‘any doctrine of free will that aspires to justify it’. This point of agreement is significant since it entails that major elements of the criminal justice system are unjustified.

Is Dennett a judge? Has he helped Society reduce Crime? What about Caruso? If neither of them have achieved prominence in the field of Law or Criminology, how can any point of agreement between them entail anything at all?