The rapidly growing field of cognitive neuroscience holds the promise of explaining the operations of the mind in terms of the physical operations of the brain. Some suggest that our emerging understanding of the physical causes of human (mis)behaviour will have a transformative effect on the law. Others argue that new neuroscience will provide only new details and that existing legal doctrine can accommodate whatever new information neuroscience will provide. We argue that neuroscience will probably have a transformative effect on the law, despite the fact that existing legal doctrine can, in principle, accommodate whatever neuroscience will tell us. New neuroscience will change the law, not by undermining its current assumptions, but by transforming people's moral intuitions about free will and responsibility. This change in moral outlook will result not from the discovery of crucial new facts or clever new arguments, but from a new appreciation of old arguments, bolstered by vivid new illustrations provided by cognitive neuroscience. We foresee, and recommend, a shift away from punishment aimed at retribution in favour of a more progressive, consequentialist approach to the criminal law.

The law takes a long-standing interest in the mind. In most criminal cases, a successful conviction requires the prosecution to establish not only that the defendant engaged in proscribed behaviour, but also that the misdeed in question was the product of mens rea, a 'guilty mind'. Narrowly interpreted, mens rea refers to the intention to commit a criminal act, but the term has a looser interpretation by which it refers to all mental states consistent with moral and/or legal blame. (A killing motivated by insane delusional beliefs may meet the requirements for mens rea in the first sense, but not the second.) (Goldstein et al 2003) Thus, for centuries, many legal issues have turned on the question: 'what was he thinking?'.

To answer this question, the law has often turned to science. Today, the newest kid on this particular scientific block is cognitive neuroscience, the study of the mind through the brain, which has gained prominence in part as a result of the advent of functional neuroimaging as a widely used tool for psychological research. Given the law's aforementioned concern for mental states, along with its preference for 'hard' evidence, it is no surprise that interest in the potential legal implications of cognitive neuroscience abounds. But does our emerging understanding of the mind as brain really have any deep implications for the law? This theme issue is a testament to the thought that it might. Some have argued, however, that new neuroscience contributes nothing more than new details and that existing legal principles can handle anything that neuroscience will throw our way in the foreseeable future (Morse 2004).

In our view, both of these positions are, in their respective ways, correct. Existing legal principles make virtually no assumptions about the neural bases of criminal behaviour, and as a result they can comfortably assimilate new neuroscience without much in the way of conceptual upheaval: new details, new sources of evidence, but nothing for which the law is fundamentally unprepared. We maintain, however, that our operative legal principles exist because they more or less adequately capture an intuitive sense of justice. In our view, neuroscience will challenge and ultimately reshape our intuitive sense(s) of justice. New neuroscience will affect the way we view the law, not by furnishing us with new ideas or arguments about the nature of human action, but by breathing new life into old ones. Cognitive neuroscience, by identifying the specific mechanisms responsible for behaviour, will vividly illustrate what until now could only be appreciated through esoteric theorizing: that there is something fishy about our ordinary conceptions of human action and responsibility, and that, as a result, the legal principles we have devised to reflect these conceptions may be flawed.

Our argument runs as follows: first, we draw a familiar distinction between the consequentialist justification for state punishment, according to which punishment is merely an instrument for promoting future social welfare, and the retributivist justification for punishment, according to which the principal aim of punishment is to give people what they deserve based on their past actions. We observe that the common-sense approach to moral and legal responsibility has consequentialist elements, but is largely retributivist. Unlike the consequentialist justification for punishment, the retributivist justification relies, either explicitly or implicitly, on a demanding — and some say overly demanding — conception of free will. We therefore consider the standard responses to the philosophical problem of free will (Watson 1982). 'Libertarians' (no relation to the political philosophy) and 'hard determinists' agree on `incompatibilism', the thesis that free will and determinism are incompatible, but they disagree about whether determinism is true, or near enough true to preclude free will. Libertarians believe that we have free will because determinism is false, and hard determinists believe that we lack free will because determinism is (approximately) true. 'Compatibilists', in contrast to libertarians and hard determinists, argue that free will and determinism are perfectly compatible.

We argue that current legal doctrine, although officially compatibilist, is ultimately grounded in intuitions that are incompatibilist and, more specifically, libertarian. In other words, the law says that it presupposes nothing more than a metaphysically modest notion of free will that is perfectly compatible with determinism. However, we argue that the law's intuitive support is ultimately grounded in a metaphysically overambitious, libertarian notion of free will that is threatened by determinism and, more pointedly, by forthcoming cognitive neuroscience. At present, the gap between what the law officially cares about and what people really care about is only revealed occasionally when vivid scientific information about the causes of criminal behaviour leads people to doubt certain individuals' capacity for moral and legal responsibility, despite the fact that this information is irrelevant according to the law's stated principles. We argue that new neuroscience will continue to highlight and widen this gap. That is, new neuroscience will undermine people's common sense, libertarian conception of free will and the retributivist thinking that depends on it, both of which have heretofore been shielded by the inaccessibility of sophisticated thinking about the mind and its neural basis.

The net effect of this influx of scientific information will be a rejection of free will as it is ordinarily conceived, with important ramifications for the law. As noted above, our criminal justice system is largely retributivist. We argue that retributivism, despite its unstable marriage to compatibilist philosophy in the letter of the law, ultimately depends on an intuitive, libertarian notion of free will that is undermined by science. Therefore, with the rejection of common-sense conceptions of free will comes the rejection of retributivism and an ensuing shift towards a consequentialist approach to punishment, i.e. one aimed at promoting future welfare rather than meting out just deserts. Because consequentialist approaches to punishment remain viable in the absence of common-sense free will, we need not give up on moral and legal responsibility. We argue further that the philosophical problem of free will arises out of a conflict between two cognitive subsystems that speak different 'languages': the 'folk psychology' system and the 'folk physics' system. Because we are inherently of two minds when it comes to the problem of free will, this problem will never find an intuitively satisfying solution. We can, however, recognize that free will, as conceptualized by the folk psychology system, is an illusion and structure our society accordingly by rejecting retributivist legal principles that derive their intuitive force from this illusion.

There are two standard justifications for legal punishment (Lacey 1988). According to the forward-looking, consequentialist theory, which emerges from the classical utilitarian tradition (Bentham 1982), punishment is justified by its future beneficial effects. Chief among them are the prevention of future crime through the deterrent effect of the law and the containment of dangerous individuals Few would deny that the deterrence of future crime and the protection of the public are legitimate justification; for punishment. The controversy surrounding con• sequentialist theories concerns their serviceability a; complete normative theories of punishment. Most theorist: find them inadequate in this regard (e.g. Hart 1968) and many argue that consequentialism fundamentally mis- characterizes the primary. justification for punishment which, these critics argue, is retribution (Kant 2002). A: a result, they claim, consequentialist theories justify intuitively unfair forms of punishment, if not in practice then in principle. One problem is that of Draconian penalties. It is possible, for example, that imposing the death penalty for parking violations would maximize aggregate welfare by reducing parking violations to near zero. But, retributivist/ claim, whether or not this is a good idea does not depend on the balance of costs and benefits. It is simply wrong to kill someone for double parking. A related problem is that of punishing the innocent. It is possible that, under certain. circumstances, falsely convicting an innocent person would have a salutary deterrent effect, enough to justify that person's suffering, etc. Critics also note that, so far as deterrence is concerned, it is the threat of punishment that is justified and not the punishment itself. Thus, consequentialism might justify letting murderers and rapists off the hook so long as their punishment could be convincingly faked.

The standard consequentialist response to these charges is that such concerns have no place in the real world. They say, for example, that the idea of imposing the death penalty for parking violations to make society an overall happier place is absurd. People everywhere would live in mortal fear of bureaucratic errors, and so on. Likewise, a legal system that deliberately convicted innocent people and/or secretly refrained from punishing guilty ones would require a kind of systematic deception that would lead inevitably to corruption and that could never survive in a free society. At this point critics retort that consequentialist theories, at best, get the right answers for the wrong reasons. It is wrong to punish innocent people, etc. because it is fundamentally unfair, not because it leads to bad consequences in practice. Such critics are certainly correct to point out that consequentialist theories fail to capture something central to common-sense intuitions about legitimate punishment.

The backward-looking, retributivist account does a better job of capturing these intuitions. Its fundamental principle is simple: in the absence of mitigating circumstances, people who engage in criminal behaviour deserve to be punished, and that is why we punish them. Some would explicate this theory in terms of criminals' forfeiting rights, others in terms of the rights of the victimized, whereas others would appeal to the violation of a hypothetical social contract, and so on. Retributivist theories come in many flavours, but these distinctions need not concern us here. What is important for our purposes is that retributivism captures the intuitive idea that we legitimately punish to give people what they deserve based on their past actions—in proportion to their 'internal wickedness', to use Kant's (2002) phrase—and not, primarily, to promote social welfare in the future. The retributivist perspective is widespread, both in the explicit views of legal theorists and implicitly in common sense. There are two primary motivations for questioning retributivist theory. The first, which will not concern us here, comes from a prior commitment to a broader consequentialist moral theory. The second comes from scepticism regarding the notion of desert, grounded in a broader scepticism about the possibility of free will in a deterministic or mechanistic world.

The problem of free will is old and has many formulations (Watson 1982). Here is one, drawing on a more detailed and exacting formulation by Peter Van Inwagen (1982): determinism is true if the world is such that its current state is completely determined by (i) the laws of physics and (ii) past states of the world. Intuitively, the idea is that a deterministic universe starts however it starts and then ticks along like clockwork from there. Given a set of prior conditions in the universe and a set of physical laws that completely govern the way the universe evolves, there is only one way that things can actually proceed.

Free will, it is often said, requires the ability do otherwise (an assumption that has been questioned; Frankfurt 1966). One cannot say, for example, that I have freely chosen soup over salad if forces beyond my control are sufficient to necessitate my choosing soup. But, the determinist argues, this is precisely what forces beyond your control do — always. You have no say whatsoever in the state of the universe before your birth; nor do you have any say about the laws of physics. However, if determinism is true, these two things together are sufficient to determine your choice of soup over salad. Thus, some say, if determinism is true, your sense of yourself and others as having free will is an illusion.

There are three standard responses to the problem of free will. The first, known as 'hard determinism', accepts the incompatibility of free will and determinism ('incompatibilism'), and asserts determinism, thus rejecting free will. The second response is libertarianism (again, no relation to the political philosophy), which accepts incompatibilism, but denies that determinism is true. This may seem like a promising approach. After all, has not modern physics shown us that the universe is indeterministic (Hughs 1992)?

The problem here is that the sort of indeterminism afforded by modern physics is not the sort the libertarian needs or desires. If it turns out that your ordering soup is completely determined by the laws of physics, the state of the universe 10 000 years ago, and the outcomes of myriad subatomic coin flips, your appetizer is no more freely chosen than before. Indeed, it is randomly chosen, which is no help to the libertarian. What about some other kind of indeterminism? What if, somewhere deep in the brain, there are mysterious events that operate independently of the ordinary laws of physics and that are somehow tied to the will of the brain's owner? In light of the available evidence, this is highly unlikely. Say what you will about the `hard problem' of consciousness (Shear 1999), there is not a shred of scientific evidence to support the existence of causally effective processes in the mind or brain that violate the laws of physics. In our opinion, any scientifically respectable discussion of free will requires the rejection of what Strawson (1962) famously called the 'panicky metaphysics' of libertarianism. ¹

Finally, we come to the dominant view among philosophers and legal theorists: compatibilism. Compatibilists concede that some notions of free will may require indefensible, panicky metaphysics, but maintain that the kinds of free will 'worth wanting', to use Dennett's (1984) phrase, are perfectly compatible with determinism. Compatibilist theories vary, but all compatibilists agree that free will is a perfectly natural, scientifically respectable phenomenon and part of the ordinary human condition. They also agree that free will can be undermined by various kinds of psychological deficit, e.g. mental illness or 'infancy'. Thus, according to this view, a freely willed action is one that is made using the right sort of psychology—rational, free of delusion, etc.

Compatibilists make some compelling arguments. After all, is it not obvious that we have free will? Could science plausibly deny the obvious fact that I am free to raise my hand at will? For many people, such simple observations make the reality of free will non-negotiable. But at the same time, many such people concede that determinism, or something like it, is a live possibility. And if free will is obviously real, but determinism is debatable, then the reality of free will must not hinge on the rejection of determinism. That is, free will and determinism must be compatible. Many compatibilists sceptically ask what would it mean to give up on free will. Were we to give it up, wouldn't we have to immediately reinvent it? Does not every decision involve an implicit commitment to the idea of free will? And how else would we distinguish between ordinary rational adults and other individuals, such as young children and the mentally ill, whose will—or whatever you want to call it—is clearly compromised? Free will, compatibilists argue, is here to stay, and the challenge for science is to figure out how exactly it works and not to peddle silly arguments that deny the undeniable (Dennett 2003).

The forward-looking–consequentialist approach to punishment works with all three responses to the problem of free will, including hard determinism. This is because consequentialists are not concerned with whether anyone is really innocent or guilty in some ultimate sense that might depend on people's having free will, but only with the likely effects of punishment. (Of course, one might wonder what it means for a hard determinist to justify any sort of choice. We will return to this issue in § 8.) The retributivist approach, by contrast, is plausibly regarded as requiring free will and the rejection of hard determinism. Retributivists want to know whether the defendant truly deserves to be punished. Assuming one can deserve to be punished only for actions that are freely willed, hard determinism' implies that no one really deserves to be punished. Thus, hard determinism combined with retributivism requires the elimination of all punishment, which does not seem reasonable. This leaves retributivists with two options: compatibilism and libertarianism. Libertarianism, for reasons given above, and despite its intuitive appeal, is scientifically suspect. At the very least, the law should not depend on it. It seems, then, that retributivism requires compatibilism. Accordingly, the standard legal account of punishment is compatibilist.
(Philosophical Transactions of the Royal Society London B (2004), pp.1775-8)