Moral Responsibility and Rational Agency: An exchange between Dennett, Waller and Clark
As I mentioned in a recent post, I'm not aware of any satisfactory arguments against Kant's demonstration that the freedom of the will can neither be proved nor disproved by pure reason. Kant puts forward that argument in his Critique of Pure Reason. Yet, in his Critique of Practical Reason, he makes a pragmatic argument for belief in free will: We need to believe in free will because it is a necessary condition for moral responsibility.
Kant's arguments are not just about free will. They're also about the existence of God and the immortality of the soul. Yet, these days, the latter two are not seen as practically required for morality. We can have moral responsibility without eternal souls or Divine judgment. But can we have it without free will?
Kant was a deontologist--he believed that morality requires duty and dignity, and not just behavior calculated to maximize some quantity of happiness, pleasure or goodness. Yet, even those pursuing other approaches to ethics (e.g., consequentialism and virtue ethics) have generally stuck by Kant's guns: Free will is necessary for moral responsibility.
Compatibilism is the attempt to rescue free will (and thus moral responsibility) from the jaws of determinism: If human behavior is determined by the laws of nature, then what room can there be for freedom? Compatibilists like Dan Dennett argue that even in a deterministic universe, we can have free will and therefore moral responsibility, and seemingly along consequentialist lines.
At this point, the definition of "free will" must be considered. For the sake of this argument, and following Michael McKenna's SEP article, let's start with the following definition of "freewill": the unique ability of persons to exercise control over their conduct in the fullest manner necessary for moral responsibility. Given that definition, free will is clearly necessary for moral responsibility. The question, then, is whether or not compatibilists can justify belief in this sort of free will.
In a recent book, Against Moral Responsibility, Bruce Waller tries to undermine this whole approach. He argues that compatibilism successfully justifies belief in free will, but that it fails to justify belief in moral responsibility. I have not read the book yet, but I've read his exchange with Dennett and Tom Clark at naturalism.org, as well as his, Tom's and other people's comments about it in a discussion thread at Russell Blackford's blog.
I've tried to contribute to that discussion, too, but I need to flesh out my point of view in a little more detail.
There are a number of issues to untangle. First, it's not clear what Waller means by "free will," if it does not entail moral responsibility. Waller's idea is that the sort of free will worth wanting--the compatibilist notion of free will that Dennett espouses--does not entail moral responsibility. I think Dennett disagrees, sticking to a definition of "free will" more or less in line with McKenna's suggestion. I'm not sure what definition of "free will" Waller is advocating. What he has to show is that the compatibilist view of free will is not what Dennett thinks it is. I'm not sure if he has done that, so I'll leave the point aside for now.
Second, according to Neil Levy (in a comment at Russell's blog), Waller's claim is that moral responsibility is incompatible with naturalism. Neil says naturalism is a red herring. Compatibilists try to reconcile free will and determinism, and naturalism does not entail determinism. (Neil says that most physicists actually reject determinism, but I'm not so sure about that. Bell's Theorem, though commonly thought to put the last nail in the coffin of determinism, is perhaps more commonly accepted as a refutation of the locality thesis. Determinism is still alive and well in physics, I think. But that's not essential here.) However, even a non-deterministic view of nature would not obviously make any more room for moral responsibility than that offered by a deterministic one. Even if quantum indeterminacy scales up in the neurological process of generating choices, the making of the choice itself must have natural causes, or else naturalism is false. So it's not clear to me that naturalism is a red herring. Compatibilists argue that free will (the kind that is necessary for moral responsibility) is possible in the natural world, be that world deterministic or not. The fact that there might be some indeterminacy involved in the exercise of will does not open the door for more moral responsibility, so far as I can see. That is the point that compatibilists like Dennett insist on, anyway.
Another issue to untangle involves the role of consequentialism in this debate. Dennett is arguing for a strictly consequentialist justification for moral responsibility, including the notion of 'just deserts.' Clark responds by accusing Dennett of being inconsistent. He says Dennett cannot stick to his consequentialist guns and still adopt the deontological notion of 'just deserts.' I'm not sure if that is true, though there does seem to be something inconsistent about Dennett's position.
I think Dennett's approach is generally cogent, and Waller's and Clark's is fatally flawed. Before explaining my thoughts on Dennett, I'll point out the weaknesses in his opposition. Besides the fact that it's not clear (to me, not having read his book) what sort of free will Waller believes in, Waller's argument seems to be self-defeating. He appeals to fairness in his argument against moral responsibility: We are not libertarian agents, he says; therefore, we are not "truly responsible" for our actions; therefore, we cannot truly deserve punishment. Of course, punishment can be justified on consequentialist grounds, Waller says; but if we are aware that people are not truly responsible for their actions, we will (hopefully) do a better job of curbing our retributive impulses. It's just not fair, he says, to make somebody pay for their crimes, unless you're doing it to create a better future.
By appealing to fairness, Waller is assuming some moral responsibility to be fair. He might defend this by saying that fairness is justifiable on consequentialist grounds. But then there is the assumption that we have a moral responsibility to be consequentialists. If there is no moral responsibility, then we are not accountable for our lack of fairness. This seems like an undesirable outcome.
What is perhaps more devastating to Waller's argument is this: the notion of fairness implies that all people deserve equal treatment. But how can one person deserve more, less or the same as anyone else, if there is no such thing as accountability in the first place? Without accountability, there is no dignity. Without dignity, no justice. Without justice, no fairness.
In Neil Levy's comment, he says that Waller does not treat the issue of fairness at all in his book. If that is true, then it is disappointing. It is hard to see how Waller's argument can stand on these feet.
Of course, the fact that Waller's argument has weaknesses does not mean that Dennett is correct. While Waller's position looks potentially incoherent and without clear support, Dennett's position is also problematic. As Clark pointed out, Dennett seems to be inconsistent, though I'm not sure it is for the reason Clark suggests.
Dennett's argument against Waller goes like this:
Moral responsibility is justified by appealing to social contract theory. To say that we are rational agents (persons with free will, capable of making rational choices for which we are accountable) is to say that we are competent to enter into contracts. Moral responsibility is tacitly accepted by anyone who actively participates in society as a rational agent. If we break a contract, we are accountable. Thus, we deserve the punishment stipulated by the contract. Thus, we can punish people just for breaking a contract, and not because we think that punishing them will increase the overall good in the world. Dennett says this is ultimately consequentialist, because the entire system of moral responsibility is justifiable on consequentialist grounds. We are increasing the overall good, Dennett says, even though we are acting on our retributivist impulses. We have found a way to "direct [our retributivist desires] down justifiable channels," he says (with his own italics).
It seems that Dennett is justifying a non-consequentialist (or perhaps semi-consequentialist) approach to ethics on consequentialist grounds. It is okay to be a deontologist to some extent, Dennett says, because doing so increases the overall good. So Dennett has (perhaps) found a pragmatic justification for limited deontology: moral responsibility and retributive justice are justifiable as necessary aspects of social contracts.
The question I raised over at Russell's blog is: If Dennett's argument works, and he has successfully offered a pragmatic justification for belief in moral responsibility, then why hasn't he also offered a pragmatic justification for belief in contra-causal free will?
Dennett wants us to believe in rational agency, but not in the ability to act contra-causally--that is, the ability to act independently of those forces which caused us to act in the first place. But if we not acting in some way independently of those causes, then how could we be accountable? Doesn't the notion of accountability entail some kind of independence from other causes? I think it might.
It's important here to recall Waller's objection to the idea that people really are libertarian agents. Waller says that we do not have agency of the sort that justifies moral responsibility. Dennett says we do. Does that mean Dennett thinks we are libertarian agents? He says no, but I'm not convinced. I don't see why his appeal to contracts can justify moral accountability, but not contra-causal free will.
Some people might say that contra-causal free will is just going too far. Like the ideas of God and the soul, contra-causal free will has no place in a rational discussion of natural phenomena. Okay, true, but we are not talking about natural phenomena, exactly. What we are talking about supervenes on natural phenomena, but this is social phenomena, and therefore deserving of a unique language. Dennett has argued for this at length, of course.
Dennett says that rational agency is not a physical fact. It does not cash out in natural kind terms. It is the product of what he calls taking the intentional stance. We regard ourselves as agents capable of acting intentionally in the world, and this sustains our ability to make and respond to contracts. If we take a nature's-eye-view of the situation (adopting what Dennett calls "the physical stance"), there is no moral responsibility. But, Dennett would continue (or should continue, I think, if he wants to be consistent), we still act as if there were moral responsibility--and in so doing, we create it. We are not living a lie. The fact is that, by acting as if we have it, we (tacitly) grant it to each other and ourselves. And by virtue of that act, we actually do have it. It is real, not an illusion, but not reducible to physical terms. This is why I suggested, in my first comment at Russell's blog, that Dennett might see Waller's rejection of moral responsibility as a rejection of intentionality itself.
Yet, if this works as a justification for moral responsibility, then it seems our social creation of moral responsibility rests on the belief that we can act independently of natural causes. We do, Dennett could argue, act as if we act responsibly--as not merely the playing out of natural causes. But acting as if we are somehow independent of nature does not make us actually independent of nature, does it?
Well, if rational agency itself is the product of taking the intentional stance, and if it thereby cannot be reduced to the physical stance, then doesn't rational agency exist in some way independently of physical causes and effects? Aren't rational acts, therefore, contra-causal?
If adopting the intentional stance is enough to make rational agency real, then why isn't contra-causal free will also real?
I'm not sure what Dennett would say to this question. Perhaps what he should say is that agency is part of a causal order which supervenes on, but which is not reducible to, natural law. Contra-causal free will does not contradict any laws of nature. It does not require abandoning determinism. It just entails a sort of causality which supervenes on, but which is not definable in terms of, physical causality.
In sum, Dennett may have motivated (1) a consequentialist justification for (limited) deontological ethics and (2) a practical belief in (a compatibilistic conception of) contra-causal free will. Dennett might be a lot closer to Kant than he thought.
Kant's arguments are not just about free will. They're also about the existence of God and the immortality of the soul. Yet, these days, the latter two are not seen as practically required for morality. We can have moral responsibility without eternal souls or Divine judgment. But can we have it without free will?
Kant was a deontologist--he believed that morality requires duty and dignity, and not just behavior calculated to maximize some quantity of happiness, pleasure or goodness. Yet, even those pursuing other approaches to ethics (e.g., consequentialism and virtue ethics) have generally stuck by Kant's guns: Free will is necessary for moral responsibility.
Compatibilism is the attempt to rescue free will (and thus moral responsibility) from the jaws of determinism: If human behavior is determined by the laws of nature, then what room can there be for freedom? Compatibilists like Dan Dennett argue that even in a deterministic universe, we can have free will and therefore moral responsibility, and seemingly along consequentialist lines.
At this point, the definition of "free will" must be considered. For the sake of this argument, and following Michael McKenna's SEP article, let's start with the following definition of "freewill": the unique ability of persons to exercise control over their conduct in the fullest manner necessary for moral responsibility. Given that definition, free will is clearly necessary for moral responsibility. The question, then, is whether or not compatibilists can justify belief in this sort of free will.
In a recent book, Against Moral Responsibility, Bruce Waller tries to undermine this whole approach. He argues that compatibilism successfully justifies belief in free will, but that it fails to justify belief in moral responsibility. I have not read the book yet, but I've read his exchange with Dennett and Tom Clark at naturalism.org, as well as his, Tom's and other people's comments about it in a discussion thread at Russell Blackford's blog.
I've tried to contribute to that discussion, too, but I need to flesh out my point of view in a little more detail.
There are a number of issues to untangle. First, it's not clear what Waller means by "free will," if it does not entail moral responsibility. Waller's idea is that the sort of free will worth wanting--the compatibilist notion of free will that Dennett espouses--does not entail moral responsibility. I think Dennett disagrees, sticking to a definition of "free will" more or less in line with McKenna's suggestion. I'm not sure what definition of "free will" Waller is advocating. What he has to show is that the compatibilist view of free will is not what Dennett thinks it is. I'm not sure if he has done that, so I'll leave the point aside for now.
Second, according to Neil Levy (in a comment at Russell's blog), Waller's claim is that moral responsibility is incompatible with naturalism. Neil says naturalism is a red herring. Compatibilists try to reconcile free will and determinism, and naturalism does not entail determinism. (Neil says that most physicists actually reject determinism, but I'm not so sure about that. Bell's Theorem, though commonly thought to put the last nail in the coffin of determinism, is perhaps more commonly accepted as a refutation of the locality thesis. Determinism is still alive and well in physics, I think. But that's not essential here.) However, even a non-deterministic view of nature would not obviously make any more room for moral responsibility than that offered by a deterministic one. Even if quantum indeterminacy scales up in the neurological process of generating choices, the making of the choice itself must have natural causes, or else naturalism is false. So it's not clear to me that naturalism is a red herring. Compatibilists argue that free will (the kind that is necessary for moral responsibility) is possible in the natural world, be that world deterministic or not. The fact that there might be some indeterminacy involved in the exercise of will does not open the door for more moral responsibility, so far as I can see. That is the point that compatibilists like Dennett insist on, anyway.
Another issue to untangle involves the role of consequentialism in this debate. Dennett is arguing for a strictly consequentialist justification for moral responsibility, including the notion of 'just deserts.' Clark responds by accusing Dennett of being inconsistent. He says Dennett cannot stick to his consequentialist guns and still adopt the deontological notion of 'just deserts.' I'm not sure if that is true, though there does seem to be something inconsistent about Dennett's position.
I think Dennett's approach is generally cogent, and Waller's and Clark's is fatally flawed. Before explaining my thoughts on Dennett, I'll point out the weaknesses in his opposition. Besides the fact that it's not clear (to me, not having read his book) what sort of free will Waller believes in, Waller's argument seems to be self-defeating. He appeals to fairness in his argument against moral responsibility: We are not libertarian agents, he says; therefore, we are not "truly responsible" for our actions; therefore, we cannot truly deserve punishment. Of course, punishment can be justified on consequentialist grounds, Waller says; but if we are aware that people are not truly responsible for their actions, we will (hopefully) do a better job of curbing our retributive impulses. It's just not fair, he says, to make somebody pay for their crimes, unless you're doing it to create a better future.
By appealing to fairness, Waller is assuming some moral responsibility to be fair. He might defend this by saying that fairness is justifiable on consequentialist grounds. But then there is the assumption that we have a moral responsibility to be consequentialists. If there is no moral responsibility, then we are not accountable for our lack of fairness. This seems like an undesirable outcome.
What is perhaps more devastating to Waller's argument is this: the notion of fairness implies that all people deserve equal treatment. But how can one person deserve more, less or the same as anyone else, if there is no such thing as accountability in the first place? Without accountability, there is no dignity. Without dignity, no justice. Without justice, no fairness.
In Neil Levy's comment, he says that Waller does not treat the issue of fairness at all in his book. If that is true, then it is disappointing. It is hard to see how Waller's argument can stand on these feet.
Of course, the fact that Waller's argument has weaknesses does not mean that Dennett is correct. While Waller's position looks potentially incoherent and without clear support, Dennett's position is also problematic. As Clark pointed out, Dennett seems to be inconsistent, though I'm not sure it is for the reason Clark suggests.
Dennett's argument against Waller goes like this:
Moral responsibility is justified by appealing to social contract theory. To say that we are rational agents (persons with free will, capable of making rational choices for which we are accountable) is to say that we are competent to enter into contracts. Moral responsibility is tacitly accepted by anyone who actively participates in society as a rational agent. If we break a contract, we are accountable. Thus, we deserve the punishment stipulated by the contract. Thus, we can punish people just for breaking a contract, and not because we think that punishing them will increase the overall good in the world. Dennett says this is ultimately consequentialist, because the entire system of moral responsibility is justifiable on consequentialist grounds. We are increasing the overall good, Dennett says, even though we are acting on our retributivist impulses. We have found a way to "direct [our retributivist desires] down justifiable channels," he says (with his own italics).
It seems that Dennett is justifying a non-consequentialist (or perhaps semi-consequentialist) approach to ethics on consequentialist grounds. It is okay to be a deontologist to some extent, Dennett says, because doing so increases the overall good. So Dennett has (perhaps) found a pragmatic justification for limited deontology: moral responsibility and retributive justice are justifiable as necessary aspects of social contracts.
The question I raised over at Russell's blog is: If Dennett's argument works, and he has successfully offered a pragmatic justification for belief in moral responsibility, then why hasn't he also offered a pragmatic justification for belief in contra-causal free will?
Dennett wants us to believe in rational agency, but not in the ability to act contra-causally--that is, the ability to act independently of those forces which caused us to act in the first place. But if we not acting in some way independently of those causes, then how could we be accountable? Doesn't the notion of accountability entail some kind of independence from other causes? I think it might.
It's important here to recall Waller's objection to the idea that people really are libertarian agents. Waller says that we do not have agency of the sort that justifies moral responsibility. Dennett says we do. Does that mean Dennett thinks we are libertarian agents? He says no, but I'm not convinced. I don't see why his appeal to contracts can justify moral accountability, but not contra-causal free will.
Some people might say that contra-causal free will is just going too far. Like the ideas of God and the soul, contra-causal free will has no place in a rational discussion of natural phenomena. Okay, true, but we are not talking about natural phenomena, exactly. What we are talking about supervenes on natural phenomena, but this is social phenomena, and therefore deserving of a unique language. Dennett has argued for this at length, of course.
Dennett says that rational agency is not a physical fact. It does not cash out in natural kind terms. It is the product of what he calls taking the intentional stance. We regard ourselves as agents capable of acting intentionally in the world, and this sustains our ability to make and respond to contracts. If we take a nature's-eye-view of the situation (adopting what Dennett calls "the physical stance"), there is no moral responsibility. But, Dennett would continue (or should continue, I think, if he wants to be consistent), we still act as if there were moral responsibility--and in so doing, we create it. We are not living a lie. The fact is that, by acting as if we have it, we (tacitly) grant it to each other and ourselves. And by virtue of that act, we actually do have it. It is real, not an illusion, but not reducible to physical terms. This is why I suggested, in my first comment at Russell's blog, that Dennett might see Waller's rejection of moral responsibility as a rejection of intentionality itself.
Yet, if this works as a justification for moral responsibility, then it seems our social creation of moral responsibility rests on the belief that we can act independently of natural causes. We do, Dennett could argue, act as if we act responsibly--as not merely the playing out of natural causes. But acting as if we are somehow independent of nature does not make us actually independent of nature, does it?
Well, if rational agency itself is the product of taking the intentional stance, and if it thereby cannot be reduced to the physical stance, then doesn't rational agency exist in some way independently of physical causes and effects? Aren't rational acts, therefore, contra-causal?
If adopting the intentional stance is enough to make rational agency real, then why isn't contra-causal free will also real?
I'm not sure what Dennett would say to this question. Perhaps what he should say is that agency is part of a causal order which supervenes on, but which is not reducible to, natural law. Contra-causal free will does not contradict any laws of nature. It does not require abandoning determinism. It just entails a sort of causality which supervenes on, but which is not definable in terms of, physical causality.
In sum, Dennett may have motivated (1) a consequentialist justification for (limited) deontological ethics and (2) a practical belief in (a compatibilistic conception of) contra-causal free will. Dennett might be a lot closer to Kant than he thought.
Comments