This article is a "Best of the Blog" repeat
For centuries philosophers have been debating the existence of Free Will. One of the curious features of this debate is how lopsided it is. Over the millennia thousands, if not tens of thousands, of arguments in defense of the proposition that we free willed entities have been put forth. By contrast, there have been a mere handful of largely timid counter-arguments. In any other debate, one would suppose that the mere fact of overwhelming consensus would render the debate into nothing more than an academic curiosity with, at most, a small group of devotees. When one reads between the lines, one gets a sense that even though the majority of people fervently insist that we have free will there is, never the less, a deep and abiding fear that we may not. I believe that it is this fear or, more precisely, these fears that I believe are being argued against.
The reason that I use the plural of fear to describe the situation is that it seems that there are a multitude of concerns that people have with regards to the possibility that we may not be free agents. The philosopher Daniel Dennett addresses a number of these fears in his book Elbow Room: The Varieties of Free Will Worth Wanting (such as the fear of loss of self, the fear of nihilism, the fear of "sphexishness" and so on) before going on to argue his own stance that we are free enough. It is not my intention, in this essay, to make an argument either for or against free will — except to note that I have yet to encounter a definition of the term that didn't result in either infinite regress, question begging terminations, an assertion of randomness, an equivocation, or which was, simply put, incoherent — but rather to address one of the more common fears which is often stated as an argument for the existence of free will.
The basic form of the argument is that if we did not have free will, we would not be able to hold people accountable for their actions therefore, since we do hold people accountable, we must have freedom of will. It does not take a trained logician to see that the argument is faulty. The fundamental problem of the argument is that it commits an inversion of the Naturalistic Fallacy. Whereas the common form of the naturalistic fallacy conflates is with ought ("since survival of the fittest is the rule of nature, we should let the poor starve to death"), this particular form confuses ought with is. Essentially, the argument assumes that if we can not hold people accountable for their actions, society would collapse. Since this would be a bad thing, we must be able to hold them accountable. This requires the existence of free will. Ergo, people have free will.
The primary problem with this line of deduction is the assumption that the universe must necessarily conform itself to our desires. Clearly, the universe is under no necessary obligation to do so. I could just as easily say that suffering would be bad therefore there must be no suffering in the world. Unless one is willing to entertain the ad hoc supposition that all suffering is an illusions (and I'll leave it to the reader to ponder what possible distinction can be made between real misery and imaginary misery), the evidence of the universe flatly contradicts the assumption. If you don't believe that suffering is a good example (perhaps you are a Christian Scientist), I'm sure you can think of any number of examples of how the universe ought to be but isn't (at the very least, the universe ought not deceive us into believing that we suffer). Even if we ought to be able to hold people accountable for their actions, we have no guarantee that we will be able to logically do so.
Fortunately, the second fault of the argument is the presumption that freedom of will is necessary for the existence of the rule of law. Dispelling this view is going to be more challenging because we have a number of intuitions regarding the nature of accountability that stand in the way of seeing why free will is an unnecessary component to the issue.
The first concept that I need to address is the concept of responsibility. In philosophical terms, responsibilities pertain to certain types of causal relationships. Causal relationships, in turn, can be described in terms of cause and effect. A given cause is considered to be proximate to an effect if there is an immediate correlation between it and the effect. An ultimate cause, by contrast, sits at the end of a change of proximate causes.
One way to illustrate this is to consider an avalanche. The proximate cause to an avalanche might be a boulder that started rolling down a bill. To proximate cause to the boulder rolling down the hill might be a high gust of wind pushing against it. And so forth and so on. Trying to reach the ultimate cause of the avalanche might well lead us back through issues of climatology, geology, planetary formation, stellar genesis and so on all the way back to the Big Bang and, perhaps, even beyond that. This is a crucial difference between proximate causes and ultimate causes. When talking about proximate causes, we don't need to extend our investigations beyond the mundane bounds of the everyday world that we live in. When the subject turns to ultimate causes, though, we invariably enter a realm dominated by questions of philosophy, theology and cosmology. It is for this reason that ultimate causation, as interesting as the topic may be to us on an intellectual level, rarely enters into our concerns when it comes to the pragmatic issue of how to deal with a given causal arrangement.
Let us consider the issue of avalanches. Avalanches are a problem. It is in our interests to either prevent avalanches from occurring or, failing that, to prevent people from being harmed by avalanches. When grappling with this issue, we do not turn to cosmologist for their insights into the deep ontological questions of what initial conditions at the origin of time (if time, indeed, has an origin point) lead to the existence of a universe where avalanches occur. Doing so would not only pose a set of problems that, in all likelihood, may well be insurmountable but, indeed, even if we could answer the question of what ultimately causes avalanches, doing so would give us no insights in how to deal with the immediate issue. In order to make headway on that problem, we have to restrict the domain of our inquiries to the proximate and the near proximate.
Let us say that we have determined that loose patches of snow tend to lead to avalanches (a proximate cause leading to a proximate effect). We may decide that setting off explosives to clear out loose snow would help forestall future avalanches — that is, using a proximate solution to anticipate a different proximate effect. We may post warning signs in areas where avalanches are likely in the hopes of preventing people from being in the path of an avalanche. Note that if we do post such signs, we only need to concern ourselves with the immediate fact that warning signs are a deterrent to people being in a danger area. We don't need to worry about deep issue of nature vs. nurture or whether any given person is ultimately free to obey or disobey the sign. Our interests are pragmatic and do not need to be concerned with deep philosophical questions.
We apply proximate standards and solutions to the vast majority of problems that face us. Questions of free will are not proximate questions. The question of free will invariably plunges into deep ontological considerations of whether or not we are ultimately free agents. Such discussions may well be fascinating but why ought we attempt to apply a question of ultimate causality to the proximate concerns inherent in crime and punishment?
Let us first ask ourselves why we want to be able to hold others legally and socially responsible for their actions. I will propose that our reasons fall into four broad concerns: quarantine, rehabilitation, deterrence and retribution. Quarantine is the principle that a person who has demonstrated that they are a danger to the rest of society must be sequestered away from society in order to prevent them from causing further harm. Rehabilitation is the theory that punishing someone for antisocial acts can lead to a state of moral and social improvement such that they may be rehabilitated back into society as an improved individual. The theory of deterrence suggests that by punishing one person for unacceptable behavior will have the effect of discouraging others from following similar courses of action. Finally, retribution is the desire to cause harm to those who have harmed others. It is the retributive stance that says that convicts be made to suffer for their crimes in order to pay for what they've done. It is, essentially, a moral theory based on the notion that one's actions should be balanced by an equivalent set of consequences. We must now ask ourselves if a theory of free will is necessary to accomplish these goals.
Most proponents of free will argue that an individual must be a free moral agent in order to be held to any sort of accountability for their actions. It is not enough that we can demonstrate that a given individual stabbed another individual to death. If that person can not be shown to be ultimately responsibility for their actions, there is no logical basis by which we can charge them, try them, or sentence them for their actions. We would not, after all, hold someone accountable for murder if it could be proven that a mad scientist had seized control of their brains and made them do horrible things against their will, now would we? As such, how can we hold a person responsible for their actions if they were the product of bad genes or an abusive environment or, ultimately, just a puppet dancing under the controlling strings of uncountable deterministic permutations leading back to the origin of the universe? It is argued that a presumption of free will is critical to the existence of any system of justice and that it's absence negates the very concept of just action. Without free will we can't accomplish our goals because the very framework of law is a sham.
Is this true, though? Let us consider our first goal: quarantine. Leaving aside the question of free will, for a moment, why do we want to quarantine criminals in the first place? The answer is self-evidently that if didn't quarantine those who have proven themselves to be a danger we would fear that that their continued freedom would pose a threat to the rest of us. We would not want someone who has already raped several women (or even one woman!) to walk the streets for fear that he would continue to rape more women. A burglar might continue to rob. A murder might murder again. It is worth noting that all of these concerns are rooted in the expectation that people are not utterly free agents. We act as though people are, in fact, biased by their previous experiences and prone to repeat them. It would seem that a pristine theory of free will would allow us to assume that one day a person could choose to be a murderer and the next, just as easily, not to murder anyone else. Indeed, a pristine theory of free will would suggest that any one of us are just as free to go on a crime spree as anyone else and that there's no good reason to treat those who have already done so with especial caution. I will not pursue this line of argument, however, because I am certain that I will be accused of erecting strawmen if I do. Never the less, it is worth noting that even the staunchest advocates of free will are willing to concede that we do have experiential constraints upon our actions – Mother Theresa did not have the same moral latitude in her choices as Charlie Manson did.
I would ask, rather, to consider an agent that indisputably lacks free will. Let us suppose that it is discovered that a given car has a faulty brake line. We know, using proximate logic, that allowing that car on the road would likely lead to accident, injury and possibly death. It is, quite literally, a dangerous vehicle. What moral justification do we need to obtain in order to ban it from the roads? Must we prove that the car wants to be a killer? Do we need to demonstrate that the car wasn't a product of its environment and manufacture? Must we tread lightly around the fear that there are things beyond the scope of the car that made it what it is? Do we have to be able to morally condemn the car for what it has become? No, of course we don't have to do any of these things. We may well choose to ponder why the car became as it is but the critical consideration of how to actually deal with it only needs to consider that it is, in fact, as it is. It is sufficient to demonstrate that the car is a danger and that we have every reason to think that putting it on the roads would be to put others at risk.
It is obvious that a dangerous machine, like a defective car, lacks any sort of will (free or otherwise) but the freedom of the machine has no bearing on the decisions we make in dealing with it. The danger it poses is enough, in and of itself, to motivate us to act to mitigate the danger. Why do we presume, then, that we must we be paralyzed in the face of the possibility that a given individual may not have ultimate moral responsibility for their actions? If a person has demonstrated that they are a threat to others, this is all the cause we need to segregate them from others.
Ah, but what of our mind controlling mad scientist? Doesn't this say that we should imprison the poor person that he's enslaved with his mind control device? If not, how does this different from being a slave to a deterministic universe? I would answer that my saying that, yes, the very first thing that we should do is, in fact, detain him so that he won't keep killing people while under the nefarious influence of a madman. It would be reckless to allow someone to go around stabbing people just because it was under someone else's influence. Once we have successfully detained him, we would want to break the control that the scientist had over him. It would not be until that point that we could afford to release him. This brings us to the question of rehabilitation.
It is important to note that the concept of rehabilitation is a fairly modern invention. In ancient times, the primary methods of justice were punitive. Often, punishment was simply tantamount to torture or a form of state sponsored vengeance whose primary function was to short circuit the endless cycles of vendetta that vigilante justice tends to promote. Never the less, the theory that punishment can be used to rehabilitate criminals has not only gained currency in the modern world but it is based upon the realization that any successful efforts to rehabilitate a criminal would represent a net gain to society since the rehabilitated criminal would, in principle, would be able to contribute to the social good.
Do we need a theory of free will in order to hope that a given person could be rehabilitated into a productive member of society? No. To demonstrate this we will, once again, utilize the example of an agent that is not considered to have free will: non-sapient animals. Very few philosophers would contend that dogs are free agents, for instance. While we may say that a given dog is a "bad dog" for making a mess on the carpet, we aren't using the term as a moral evaluation of the animal's conduct. We understand that dogs do not have a well developed sense of agency. When we correct a dog, the only thing that we are relying on is that the dog will be able to remember the rebuke and will associate the rebuke with the undesirable act and will, with repeated corrections and the proper application of positive enforcements, become trained not to defecate indoors.
Training is the cornerstone of any theory of rehabilitation. All that is required is that the agent in question is sufficiently intelligent to respond to the training and sufficiently motivated to take the lessons to heart. The amount of intelligence required is actually fairly small. If one is tempted to assign free will to dogs, we can shift the considerations to cats, mice, or even worms if faced with a particularly aggressive skepticism. So long as an agent has some degree of memory and some aversion to corrective stimuli, it can be trained to avoid a given course of action. Humans, of course, are significantly more complex than dogs, cats, et al. This complicates efforts towards rehabilitation but it also allows for more sophisticated sorts of association.
Whether rehabilitative treatments work or not is a question that I will not attempt to address except to note that the data is complex, controversial and, all too often, equivocal. It does appear that certain social deviations are more prone to efforts at correction than others (pedophilia being a prime example of a condition that does not seem to respond to rehabilitative efforts, thus throwing a counterexample to the proposition that we may freely choose to morally improve ourselves) but in no case are we required to believe that a person must be a free agent in order to assume that said person could, in theory, be trained to be a better person. Let us now examine deterrence.
The theory of deterrence essentially amounts to a kind of preemptive rehabilitation at a distance. The idea behind the theory is that having a sufficiently harsh punishment for a given crime, people who might be tempted to commit the crime will think twice. With any hope, they will decide that the potential gains of committing the crime will be outweighed by the potential costs.
We like to believe that crime doesn't pay. Unfortunately, the truth of the matter is that crime often does pay. A person who deals drugs, for instance, can often make tens of thousands of dollars a month while his law abiding compatriots are forced to subsist on a fraction of that wealth. It is delusional to think that all criminals are caught. At best, we can hope that most criminals are caught. Even in the cases where a criminal is apprehended, his apprehension does not negate the harm he's done. We may be pleased when a murderer is taken off the streets but that can only offer so much comfort to the families of his victims.
By making the cost of a crime high enough, we hope to prevent crime thus reducing the number of individuals that get away with crimes unapprehended as well as preventing acts of victimization before they happen. In order for a theory of deterrence to work, we need a fairly sophisticated sort of agent. It is not sufficient that an agent can learn to avoid a course of action as a result of his own experiences; we must have an agent intelligent enough to anticipate the consequences of an action before the fact (what is sometimes called a Popperian agent). Fortunately, human beings are such agents.
Consider a room with a sign that says "DANGER: VX NERVE GAS, DO NOT ENTER". It is rather unlikely that you've been exposed to nerve gas before (otherwise you'd probably be dead). Never the less, so long as you understand the sign (e.g., you can read English) and appreciate the consequences, you are unlikely to enter the room. Once again, our capacity to perform proximate evaluations allows us to realize that if we open the room, we might die. Since most people are adverse to dying, most people would, therefore, choose to leave the room undisturbed.
The only thing that a theory of deterrence requires a handful of critical ingredients:
- That a given set of actions will result in a predetermined set of consequences
- The presence of agents who are aware of the consequences
- An inherent aversion to the consequence in a majority of the agents
- The ability for the agents to associate their potential actions to the consequences
- That the consequences are sufficiently steep to offset any benefits of performing the actions
Given that efforts towards deterrence are intended to inhibit our choices, the contention that we are required to be free agents in order to make a theory of deterrence work is defeated by self-contradiction. Even more critically, the only criteria that matters in order to advance a system of deterrence is whether or not the efforts to deter others from the set of proscribed actions does, in fact, succeed in its goal. A system of deterrence that doesn't work should be rejected. One that does work may well be embraced (so long as it doesn't conflict with other moral goals). In this context, the issue of free will is simply a non-sequitur.
Finally, we come to retribution. Retribution is not the same as generic punishment. Because we are social animals, any sort of quarantine will be automatically punitive (which is one reason that solitary confinement is often used as a supplementary punishment in prisons), punitive actions are typically considered necessary to foster criminal rehabilitation, and efforts at deterrence would be toothless unless we actually carried out the threat of punishment. Given this, retribution must be considered actions that go above and beyond these goals. Retributive justice is punishment for the sake of punishment (although it is often phrased in terms of obtaining a moral balance). Typically it is based on the notion that a criminal deserves his fate and must be made to suffer because he freely chose his course of action.
Here, at last, we have a cause that seems to logically appeal to a sense of free will. If we do not believe that being are ultimately responsible for their own actions, it becomes more difficult to justify a course of action that is based solely on a desire to inflict harm.
The first question I would raise to this contention is to ask whether or not this is a desirably goal for a civilized society. While I would agree that there is something inside of us that wants to hurt those who have hurt us (and where does this innate desire stand with respect to our own supposed free will — but I digress), should this be something that we want to do as a culture? Historically we have been moving away from doing so. As an example of this, the crime for theft, in most of the civilized world, is not amputation but incarceration. I don't doubt that there are still people (especially the victims of robbery) who might long for the old days when thieves would lose their hands but, as a society, we consider such a punishment excessive. The notion that the punishment should fit the crime and that it should serve to promote social order is central to modern theories of justice. Vengeance may well be sweet but it isn't considered to be civilized.
It does behoove us to recognize that the meaning of civilization has changed dramatically over time. The Romans, certainly, would have had little difficulty with the idea that we should torture convicts (and even made a sport out of it). Who is to say that our own moral qualms represent the high water mark of civilization? Who is to say that some future society won't find our notion of a balance between a crime and a punishment to be a hopelessly quaint invention and an artifact of our place in history? If this is the case, must we appeal to a theory of free will in order to justify a belief in retribution? Perhaps not.
In order to consider the question of retribution we must first ask ourselves why retribution would be desirable. We can not appeal to its rehabilitative effects and we can not appeal to its deterrent effects since doing so would mean that we were either talking about a system of rehabilitation or a system of deterrence. Punishment for the sake of punishment must provide its own justification for its necessity. Perhaps we can appeal to human nature. Maybe there's something inside of us that needs to see villains suffer. If that's the case, though, an appeal to free will is the last thing we want. Instead, we would be better served to actively deny our own free will — we indulge in excess punishment because we have no choice! Perhaps we should punish criminals beyond the strict limits of necessity in order to make the victims of criminalization feel better. Maybe there's a sort of moral catharsis that we are denying to something who has, for instance, been raped by our prissy restraints on torture and mutilation. If that's the case, though, then our reason isn't punishment for the sake of punishment but, rather, punishment for the sake of emotional recompensation (this would likewise apply to any theory of punishment that appealed to a desire for moral balance). Surely if that's the case, any supposition of free will is unnecessary; all that is required is an agent that requires this means of punitive balance. Again, supposing free will actually complicates the issue because we could suppose that a victim has the freedom to simply choose not to need any sort of sanctioned retribution.
It is possible that I am being unimaginative. I must consider the possibility that there is some reason that would justify punishment for the sake of punishment and that such a reason would demand a theory of free will to justify it. What then? If free will proved to be a chimera, would the very notion of law crash to our feet? No. The only thing that would be lost is the principle of retribution. The principle would have to be dismissed precisely because it was unjustified by the facts of the universe. In its place, we would still have the option of having legal institutions based on the principles of quarantine, rehabilitation, deterrence and, perhaps, emotional recompensation. Since all of these allow (and even demand) punishment, criminals would continue to be punished for their crimes regardless of whether or not we considered them to have any sort of ultimate responsibility for their actions. Likewise, the degree of punishment we could apply may well be very broad — even the ultimate punishment of death could, in principle, be made to fit the demands of these goals. All we would lose is the justification to apply an arbitrarily harsh punishment for any given crime. I would contend that this is a very small loss.
Have I dispensed with free will, then? No. My goal has not been to address the reality or unreality of free will. The question of whether or not we have free agency is beyond the scope of this essay and, I dare say, probably beyond the ability of anyone to resolve conclusively. The only thing that I have hoped to accomplish is the banishment of a particular fear regarding the subject of free will. We rightly desire the stability of our societies and the rule of law (anarchists to the contrary) is a critical component of that stability. If our systems of crime and punishment did, in fact, require freedom of will then we would have cause to fear the possibility that free will might only be a philosopher's dream. Fortunately, our punitive institutions are not such fragile things. Whether our actions are in our genes, our upbringing, or in some sort of ephemeral soul, the legal foundations of our society will persist.