Free Will

� Author for correspondence ( [email protected]).

One contribution of 16 to a Theme Issue ‘Law and the brain’.

Phil. Trans. R. Soc. Lond. B (2004) 359, 1775–1785 1775 doi:10.1098/rstb.2004.1546

Published online 26 November 2004

For the law, neuroscience changes nothing and everything

Joshua Greene � and Jonathan Cohen

Department of Psychology, Center for the Study of Brain, Mind, and Behavior, Princeton University, Princeton, NJ 08544, USA

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 assump-

tions, 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.

Keywords: law; brain; morality; free will; punishment; retributivism

1. INTRODUCTION The law takes a long-standing interest in the mind. In most

criminal cases, a successful conviction requires the pros-

ecution 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 delu-

sional 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 con-

tributes 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 crimi-

nal behaviour, and as a result they can comfortably

assimilate new neuroscience without much in the way of

conceptual upheaval: new details, new sources of evi-

dence, 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, neu-

roscience will challenge and ultimately reshape our intuit-

ive 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 con-

ceptions 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

# 2004 The Royal Society

1776 J. Greene and J. Cohen Neuroscience changes nothing and everything

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 prob-

lem of free will (Watson 1982). ‘Libertarians’ (no relation

to the political philosophy) and ‘hard determinists’ agree

on ‘incompatibilism’, the thesis that free will and determin-

ism 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 deter-

minists, 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 meta-

physically 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 compati-

bilist 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 con-

sequentialist approach to punishment, i.e. one aimed at

promoting future welfare rather than meting out just

deserts. Because consequentialist approaches to punish-

ment 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

Phil. Trans. R. Soc. Lond. B (2004)

retributivist legal principles that derive their intuitive force

from this illusion.

2. TWO THEORIES OF PUNISHMENT: CONSEQUENTIALISM AND RETRIBUTIVISIM

There are two standard justifications for legal punishment

(Lacey 1988). According to the forward-looking, con-

sequentialist theory, which emerges from the classical utili-

tarian 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 justifications

for punishment. The controversy surrounding con-

sequentialist theories concerns their serviceability as

complete normative theories of punishment. Most theorists

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). As

a result, they claim, consequentialist theories justify intuit-

ively 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, retributivists

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 per-

son’s suffering, etc. Critics also note that, so far as deter-

rence is concerned, it is the threat of punishment that is

justified and not the punishment itself. Thus, con-

sequentialism might justify letting murderers and rapists

off the hook so long as their punishment could be convin-

cingly 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 pen-

alty for parking violations to make society an overall hap-

pier 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 rea-

sons. 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 bet-

ter job of capturing these intuitions. Its fundamental prin-

ciple 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

Neuroscience changes nothing and everything J. Greene and J. Cohen 1777

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 wel-

fare 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 con-

sequentialist moral theory. The second comes from scepti-

cism regarding the notion of desert, grounded in a broader

scepticism about the possibility of free will in a determinis-

tic or mechanistic world.

3. FREE WILL AND RETRIBUTIVISM 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 deter-

ministic 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 uni-

verse 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 (‘incompa-

tibilism’), 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 cho-

sen than before. Indeed, it is randomly chosen, which is no

Phil. Trans. R. Soc. Lond. B (2004)

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 evi-

dence, 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 meta-

physics’ of libertarianism. 1

Finally, we come to the dominant view among philoso-

phers and legal theorists: compatibilism. Compatibilists

concede that some notions of free will may require indefen-

sible, 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 obvi-

ously 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 men-

tally 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 pun-

ishment works with all three responses to the problem of

free will, including hard determinism. This is because con-

sequentialists 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 x 8.) The retributivist approach, by contrast, is plausibly regarded as requiring

free will and the rejection of hard determinism. Retributi-

vists 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,

1778 J. Greene and J. Cohen Neuroscience changes nothing and everything

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 rea-

sons given above, and despite its intuitive appeal, is scien-

tifically 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.

4. NEUROSCIENCE CHANGES NOTHING The title of a recent paper by Stephen Morse (2004), ‘New

neuroscience, old problems’, aptly summarizes many a sea-

soned legal thinker’s response to the suggestion that brain

research will revolutionize the law. The law has been deal-

ing with issues of criminal responsibility for a long time,

Morse argues that there is nothing on the neuroscientific

horizon that it cannot handle.

The reason that the law is immune to such threats is that

it makes no assumptions that neuroscience, or any science,

is likely to challenge. The law assumes that people have a

general capacity for rational choice. That is, people have

beliefs and desires and are capable of producing behaviour

that serves their desires in light of their beliefs. The law

acknowledges that our capacity for rational choice is far

from perfect (Kahneman & Tversky 2000), requiring only

that the people it deems legally responsible have a general

capacity for rational behaviour.

Thus, questions about who is or is not responsible in the

eyes of the law have and will continue to turn on questions

about rationality. This approach was first codified in the

M’Naghten standard according to which a defence on the

ground of insanity requires proof that the defendant

laboured under ‘a defect of reason, from disease of the

mind’ (Goldstein 1967). Not all standards developed and

applied since M’Naghten explicitly mention the need to

demonstrate the defendant’s diminished rationality (e.g.

the Durham standard; Goldstein 1967), but it is generally

agreed that a legal excuse requires a demonstration that the

defendant ‘lacked a general capacity for rationality’ (Gold-

stein et al. 2003). Thus, the argument goes, new science

can help us figure out who was or was not rational at the

scene of the crime, much as it has in the past, but new

science will not justify any fundamental change in the law’s

approach to responsibility unless it shows that people in

general fail to meet the law’s very minimal requirements for

rationality. Science shows no sign of doing this, and thus

the basic precepts of legal responsibility stand firm. As for

neuroscience more specifically, this discipline seems

especially unlikely to undermine our faith in general mini-

mal rationality. If any sciences have an outside chance of

demonstrating that our behaviour is thoroughly irrational

or arational it is the ones that study behaviour directly

rather than its proximate physical causes in the brain. The

law, this argument continues, does not care if people have

‘free will’ in any deep metaphysical sense that might be

threatened by determinism. It only cares that people in

general are minimally rational. So long as this appears to be

the case, it can go on regarding people as free (compatibi-

lism) and holding ordinary people responsible for their mis-

deeds while making exceptions for those who fail to meet

the requirements of general rationality.

Phil. Trans. R. Soc. Lond. B (2004)

In light of this, one might wonder what all the fuss is

about. If the law assumes nothing more than general mini-

mal rationality, and neuroscience does nothing to under-

mine this assumption, then why would anyone even think

that neuroscience poses some sort of threat to legal doc-

trines of criminal responsibility? It sounds like this is just a

simple mistake, and that is precisely what Morse contends.

He calls this mistake ‘the fundamental psycholegal error’

which is ‘to believe that causation, especially abnormal

causation, is per se an excusing condition’ (Morse 2004, p.

180). In other words, if you think that neuro-

scientific information about the causes of human action, or

some particular human’s action, can, by itself, make for a

legitimate legal excuse, you just do not understand the law.

Every action is caused by brain events, and describing those

events and affirming their causal efficacy is of no legal inter-

est in and of itself. Morse continues, ‘[The psycholegal

error] leads people to try to create a new excuse every time

an allegedly valid new ‘‘ syndrome ’’ is discovered that is

thought to play a role in behaviour. But syndromes and

other causes do not have excusing force unless they suffi-

ciently diminish rationality in the context in question’

(Morse 2004, p. 180).

In our opinion, Morse and like-minded theorists are

absolutely correct about the relationship between current

legal doctrine and any forthcoming neuroscientific results.

For the law, as written, neuroscience changes nothing. The

law provides a coherent framework for the assessment of

criminal responsibility that is not threatened by anything

neuroscience is likely to throw at it. But, we maintain, the

law nevertheless stands on shakier ground than the forego-

ing would suggest. The legitimacy of the law itself depends

on its adequately reflecting the moral intuitions and com-

mitments of society. If neuroscience can change those

intuitions, then neuroscience can change the law.

As it happens, this is a possibility that Morse explicitly

acknowledges. However, he believes that such develop-

ments would require radical new ideas that we can scarcely

imagine at this time, e.g. a new solution to the mind–body

problem. We disagree. The seeds of discontent are already

sown in common-sense legal thought. In our opinion, the

‘fundamental psycholegal error’ is not so much an error as

a reflection of the gap between what the law officially cares

about and what people really care about. In modern crimi-

nal law, there has been a long tense marriage of con-

venience between compatibilist legal principles and

libertarian moral intuitions. New neuroscience, we argue,

will probably render this marriage unworkable.

5. WHAT REALLY MATTERS FOR RESPONSIBILITY? MATERIALIST THEORY, DUALIST INTUITIONS AND

THE ‘BOYS FROM BRAZIL’ PROBLEM According to the law, the central question in a case of puta-

tive diminished responsibility is whether the accused was

sufficiently rational at the time of the misdeed in question.

We believe, however, that this is not what most people

really care about, and that for them diminished rationality

is just a presumed correlate of something deeper. It seems

that what many people really want to know is: was it really

him? This question usually comes in the form of a disjunc-

tion, depending on how the excuse is constructed: was it

him, or was it his upbringing ? Was it him, or was it his genes?

Neuroscience changes nothing and everything J. Greene and J. Cohen 1779

Was it him, or was it his circumstances? Was it him, or was it

his brain? But what most people do not understand, despite

the fact that naturalistic philosophers and scientists have

been saying it for centuries, is that there is no ‘him’ inde-

pendent of these other things. (Or, to be a bit more accom-

modating to the supernaturally inclined, there is no ‘him’

independent of these things that shows any sign of affecting

anything in the physical world, including his behaviour.)

Most people’s view of the mind is implicitly dualist and

libertarian and not materialist and compatibilist. Dualism, for

our purposes, is the view that mind and brain are separate,

interacting, entities. 2 Dualism fits naturally with liber-

tarianism because a mind distinct from the body is pre-

cisely the sort of non-physical source of free will that

libertarianism requires. Materialism, by contrast, is the

view that all events, including the operations of the mind,

are ultimately operations of matter that obeys the laws of

physics. It is hard to imagine a belief in free will that is

materialist but not compatibilist, given that ordinary mat-

ter does not seem capable of supplying the non-physical

processes that libertarianism requires.

Many people, particularly those who are religious, are

explicitly dualist libertarians (again, not in the political

sense). However, in our estimation, even people who do or

would readily endorse a thoroughly material account of

human action and its causes have dualist, libertarian intui-

tions. This goes not only for educated people in general,

but for experts in mental health and criminal behaviour.

Consider, for example, the following remarks from Jona-

than Pincus, an expert on criminal behaviour and the brain.

When a composer conceives a symphony, the only way he or

she can present it to the public is through an orchestra. . . If the

performance is poor, the fault could lie with the composer’s

conception, or the orchestra, or both. . . Will is expressed by the

brain. Violence can be the result of volition only, but if a brain is

damaged, brain failure must be at least partly to blame.

(Pincus 2001, p. 128)

To our untutored intuitions, this is a perfectly sensible

analogy, but it is ultimately grounded in a kind of dualism

that is scientifically untenable. It is not as if there is you, the

composer, and then your brain, the orchestra. You are your

brain, and your brain is the composer and the orchestra all

rolled together. There is no little man, no ‘homunculus’, in

the brain that is the real you behind the mass of neuronal

instrumentation. Scientifically minded philosophers have

been saying this ad nauseum (Dennett 1991), and we will

not belabour the point. Moreover, we suspect that if you

were to ask Dr Pincus whether he thinks there is a little con-

ductor directing his brain’s activity from within or beyond

he would adamantly deny that this is the case. At the same

time, though, he is comfortable comparing a brain-

damaged criminal to a healthy conductor saddled with an

unhealthy orchestra. This sort of doublethink is not

uncommon. As we will argue in x 7, when it comes to moral responsibility in a physical world, we are all of two minds.

A recent article by Laurence Steinberg and Elizabeth

Scott (Steinberg & Scott 2003), experts respectively on

adolescent developmental psychology and juvenile law,

illustrates the same point. They argue that adolescents do

not meet the law’s general requirements for rationality and

that therefore they should be considered less than fully

responsible for their actions and, more specifically,

Phil. Trans. R. Soc. Lond. B (2004)

unsuitable candidates for the death penalty. Their main

argument is sound, but they cannot resist embellishing it

with a bit of superfluous neuroscience.

Most of the developmental research on cognitive and psycho-

social functioning in adolescence measures behaviors, self-

perceptions, or attitudes, but mounting evidence suggests that

at least some of the differences between adults and adolescents

have neuropsychological and neurobiological underpinnings.

(Steinberg & Scott 2003, p. 5)

Some of the differences? Unless some form of dualism is

correct, every mental difference and every difference in

behavioural tendency is a function of some kind of differ-

ence in the brain. But here it is implicitly suggested that

things like ‘behaviours, self-perceptions, or attitudes’ may

be grounded in something other than the brain. In sum-

ming up their case, Steinberg and Scott look towards the

future.

Especially needed are studies that link developmental changes

in decision making to changes in brain structure and

function. . . In our view, however, there is sufficient indirect

suggestive evidence of age differences in capacities that are rel-

evant to criminal blameworthiness to support the position that

youths who commit crimes should be punished more leniently

then their adult counterparts.

(Steinberg & Scott 2003, p. 9)

This gets the order of evidence backwards. If what the

law ultimately cares about is whether adolescents can

behave rationally, then it is evidence concerning adolescent

behaviour that is directly relevant. Studying the adolescent

brain is a highly indirect way of figuring out whether adoles-

cents in general are rational. Indeed, the only way we neu-

roscientists can tell if a brain structure is important for

rational judgement is to see if its activity or damage is corre-

lated with (ir)rational behaviour. 3

If everyone agrees that what the law ultimately cares

about is the capacity for rational behaviour, then why are

Steinberg and Scott so optimistic about neuroscientific evi-

dence that is only indirectly relevant? The reason, we sug-

gest, is that they are appealing not to a legal argument, but

to a moral intuition. So far as the law is concerned, infor-

mation about the physical processes that give rise to bad

behaviour is irrelevant. But to people who implicitly believe

that real decision-making takes place in the mind, not in

the brain, demonstrating that there is a brain basis for ado-

lescents’ misdeeds allows us to blame adolescents’ brains

instead of the adolescents themselves.

The fact that people are tempted to attach great moral or

legal significance to neuroscientific information that,

according to the letter of the law, should not matter, sug-

gests that what the law cares about and what people care

about do not necessarily coincide. To make this point in a

more general way, we offer the following thought experi-

ment, which we call ‘The Boys from Brazil problem’. It is an

extension of an argument that has made the rounds in

philosophical discussions of free will and responsibility

(Rosen 2002).

In the film The Boys from Brazil, members of the Nazi old

guard have regrouped in South America after the war.

Their plan is to bring their beloved führer back to life by

raising children genetically identical to Hitler (courtesy of

some salvaged DNA) in environments that mimic that of

Hitler’s upbringing. For example, Hitler’s father died while

1780 J. Greene and J. Cohen Neuroscience changes nothing and everything

young Adolph was still a boy, and so each Hitler clone’s

surrogate father is killed at just the right time, and so on,

and so forth.

This is obviously a fantasy, but the idea that one could, in

principle, produce a person with a particular personality

and behavioural profile through tight genetic and environ-

mental control is plausible. Let us suppose, then, that a

group of scientists has managed to create an individual—

call him ‘Mr Puppet’—who, by design, engages in some

kind of criminal behaviour: say, a murder during a drug

deal gone bad. The defence calls to the stand the project’s

lead scientist: ‘Please tell us about your relationship to

Mr Puppet. . .’

It is very simple, really. I designed him. I carefully selected

every gene in his body and carefully scripted every significant

event in his life so that he would become precisely what he is

today. I selected his mother knowing that she would let him cry

for hours and hours before picking him up. I carefully selected

each of his relatives, teachers, friends, enemies, etc. and told

them exactly what to say to him and how to treat him. Things

generally went as planned, but not always. For example, the

angry letters written to his dead father were not supposed to

appear until he was fourteen, but by the end of his thirteenth

year he had already written four of them. In retrospect I think

this was because of a handful of substitutions I made to his

eighth chromosome. At any rate, my plans for him succeeded,

as they have for 95% of the people I’ve designed. I assure you

that the accused deserves none of the credit.

What to do with Mr Puppet? Insofar as we believe this

testimony, we are inclined to think that Mr Puppet cannot

be held fully responsible for his crimes, if he can be held

responsible for them at all. He is, perhaps, a man to be

feared, and we would not want to return him to the streets.

But given the fact that forces beyond his control played a

dominant role in causing him to commit these crimes, it is

hard to think of him as anything more than a pawn.

But what does the law say about Mr Puppet? The law

asks whether or not he was rational at the time of his mis-

deeds, and as far as we know he was. For all we know, he is

psychologically indistinguishable from the prototypical

guilty criminal, and therefore fully responsible in the eyes

of the law. But, intuitively, this is not fair.

Thus, it seems that the law’s exclusive interest in ration-

ality misses something intuitively important. In our opi-

nion, rationality is just a presumed correlate of what most

people really care about. What people really want to know

is if the accused, as opposed to something else, is respon-

sible for the crime, where that ‘something else’ could be the

accused’s brain, genes or environment. The question of

someone’s ultimate responsibility seems to turn, intuit-

ively, on a question of internal versus external determi-

nation. Mr Puppet ought not be held responsible for his

actions because forces beyond his control played a domi-

nant role in the production of his behaviour. Of course, the

scientists did not have complete control—after all, they had

a 5% failure rate—but that does not seem to be enough to

restore Mr Puppet’s free will, at least not entirely. Yes, he is

as rational as other criminals, and, yes, it was his desires

and beliefs that produced his actions. But those beliefs and

desires were rigged by external forces, and that is why,

intuitively, he deserves our pity more than our moral con-

demnation. 4

Phil. Trans. R. Soc. Lond. B (2004)

The story of Mr. Puppet raises an important question:

what is the difference between Mr Puppet and anyone else

accused of a crime? After all, we have little reason to doubt

that (i) the state of the universe 10 000 years ago, (ii) the

laws of physics, and (iii) the outcomes of random quantum

mechanical events are together sufficient to determine

everything that happens nowadays, including our own

actions. These things are all clearly beyond our control. So

what is the real difference between us and Mr Puppet? One

obvious difference is that Mr Puppet is the victim of a dia-

bolical plot whereas most people, we presume, are not. But

does this matter? The thought that Mr Puppet is not fully

responsible depends on the idea that his actions were exter-

nally determined. Forces beyond his control constrained

his personality to the point that it was ‘no surprise’ that he

would behave badly. But the fact that these forces are con-

nected to the desires and intentions of evil scientists is

really irrelevant, is it not? What matters is only that these

forces are beyond Mr Puppet’s control, that they’re not

really his. The fact that someone could deliberately harness

these forces to reliably design criminals is an indication of

the strength of these forces, but the fact that these forces

are being guided by other minds rather than simply operat-

ing on their own seems irrelevant, so far as Mr Puppet’s

freedom and responsibility are concerned.

Thus, it seems that, in a very real sense, we are all pup-

pets. The combined effects of genes and environment

determine all of our actions. Mr Puppet is exceptional only

in that the intentions of other humans lie behind his genes

and environment. But, so long as his genes and environ-

ment are intrinsically comparable to those of ordinary

people, this does not really matter. We are no more free

than he is.

What all of this illustrates is that the ‘fundamental psy-

cholegal error’ is grounded in a powerful moral intuition

that the law and allied compatibilist philosophies try to

sweep under the rug. The foregoing suggests that people

regard actions only as fully free when those actions are seen

as robust against determination by external forces. But if

determinism (or determinism plus quantum mechanics) is

true, then no actions are truly free because forces beyond

our control are always sufficient to determine behaviour.

Thus, intuitive free will is libertarian, not compatibilist.

That is, it requires the rejection of determinism and an

implicit commitment to some kind of magical mental caus-

ation. 5

Naturalistic philosophers and scientists have known for a

long time that magical mental causation is a non-starter.

But this realization is the result of philosophical reflection

about the nature of the universe and its governance by

physical law. Philosophical reflection, however, is not the

only way to see the problems with libertarian accounts of

free will. Indeed, we argue that neuroscience can help

people appreciate the mechanical nature of human action

in a way that bypasses complicated arguments.

6. NEUROSCIENCE AND THE TRANSPARENT BOTTLENECK

We have argued that, contrary to legal and philosophical

orthodoxy, determinism really does threaten free will and

responsibility as we intuitively understand them. It is just

that most of us, including most philosophers and legal

Neuroscience changes nothing and everything J. Greene and J. Cohen 1781

theorists, have yet to appreciate it. This controversial opi-

nion amounts to an empirical prediction that may or may

not hold: as more and more scientific facts come in, provid-

ing increasingly vivid illustrations of what the human mind

is really like, more and more people will develop moral

intuitions that are at odds with our current social practices

(see Robert Wright (1994) for similar thoughts).

Neuroscience has a special role to play in this process for

the following reason. As long as the mind remains a black

box, there will always be a donkey on which to pin dualist

and libertarian intuitions. For a long time, philosophical

arguments have persuaded some people that human action

has purely mechanical causes, but not everyone cares for

philosophical arguments. Arguments are nice, but physical

demonstrations are far more compelling. What neuro-

science does, and will continue to do at an accelerated

pace, is elucidate the ‘when’, ‘where’ and ‘how’ of the

mechanical processes that cause behaviour. It is one thing

to deny that human decision-making is purely mechanical

when your opponent offers only a general, philosophical

argument. It is quite another to hold your ground when

your opponent can make detailed predictions about how

these mechanical processes work, complete with images of

the brain structures involved and equations that describe

their function. 6

Thus, neuroscience holds the promise of turning the

black box of the mind into a transparent bottleneck. There

are many causes that impinge on behaviour, but all of

them—from the genes you inherited, to the pain in your

lower back, to the advice your grandmother gave you when

you were six—must exert their influence through the brain.

Thus, your brain serves as a bottleneck for all the forces

spread throughout the universe of your past that affect who

you are and what you do. Moreover, this bottleneck con-

tains the events that are, intuitively, most critical for moral

and legal responsibility, and we may soon be able to

observe them closely.

At some time in the future we may have extremely high-

resolution scanners that can simultaneously track the

neural activity and connectivity of every neuron in a human

brain, along with computers and software that can analyse

and organize these data. Imagine, for example, watching a

film of your brain choosing between soup and salad. The

analysis software highlights the neurons pushing for soup in

red and the neurons pushing for salad in blue. You zoom in

and slow down the film, allowing yourself to trace the

cause-and-effect relationships between individual neu-

rons—the mind’s clockwork revealed in arbitrary detail.

You find the tipping-point moment at which the blue neu-

rons in your prefrontal cortex out-fire the red neurons, seiz-

ing control of your pre-motor cortex and causing you to

say, ‘I will have the salad, please’.

At some further point this sort of brainware may be very

widespread, with a high-resolution brain scanner in every

classroom. People may grow up completely used to the idea

that every decision is a thoroughly mechanical process, the

outcome of which is completely determined by the results

of prior mechanical processes. What will such people think

as they sit in their jury boxes? Suppose a man has killed his

wife in a jealous rage. Will jurors of the future wonder whe-

ther the defendant acted in that moment of his own free will?

Will they wonder if it was really him who killed his wife

rather than his uncontrollable anger? Will they ask whether

Phil. Trans. R. Soc. Lond. B (2004)

he could have done otherwise? Whether he really deserves to be

punished, or if he is just a victim of unfortunate circum-

stances? We submit that these questions, which seem so

important today, will lose their grip in an age when the

mechanical nature of human decision-making is fully

appreciated. The law will continue to punish misdeeds, as

it must for practical reasons, but the idea of distinguishing

the truly, deeply guilty from those who are merely victims

of neuronal circumstances will, we submit, seem pointless.

At least in our more reflective moments. Our intuitive

sense of free will runs quite deep, and it is possible that we

will never be able to fully talk ourselves out of it. Next we

consider the psychological origins of the problem of free

will.

7. FOLK PSYCHOLOGY AND FOLK PHYSICS COLLIDE: A COGNITIVE ACCOUNT OF THE PROBLEM OF ATTRIBUTIVE FREE WILL

Could the problem of free will just melt away? This ques-

tion begs another: why do we have the problem of free will

in the first place? Why does the idea of a deterministic uni-

verse seem to contradict something important in our con-

ception of human action? A promising answer to this

question is offered by Daniel Wegner in The illusion of con-

scious will (Wegner 2002). In short, Wegner argues, we feel

as if we are uncaused causers, and therefore granted a

degree of independence from the deterministic flow of the

universe, because we are unaware of the deterministic pro-

cesses that operate in our own heads. Our actions appear to

be caused by our mental states, but not by physical states of

our brains, and so we imagine that we are metaphysically

special, that we are non-physical causes of physical events.

This belief in our specialness is likely to meet the same fate

as other similarly narcissistic beliefs that we have cherished

in our past: that the Earth lies at the centre of the universe,

that humans are unrelated to other species, that all of our

behaviour is consciously determined, etc. Each of these

beliefs has been replaced by a scientific and humbling

understanding of our place in the physical universe, and

there is no reason to believe that the case will be any differ-

ent for our sense of free will. (For similar thoughts, see

Wright (1994) on Darwin’s clandestine views about free

will and responsibility.)

We believe that Wegner’s account of the problem of free

will is essentially correct, although we disagree strongly

with his conclusions concerning its (lack of) practical moral

implications (see below). In this section we pick up on and

extend one strand in Wegner’s argument (Wegner 2002,

pp. 15–28). Wegner’s primary aim is to explain, in psycho-

logical terms, why we attribute free will to ourselves, why

we feel free from the inside. Our aim in this section is to

explain, in psychological terms, why we insist on attribu-

ting free will to others—and why scientifically minded philo-

sophers, despite persistent efforts, have managed to talk

almost no one out of this practice. The findings we review

serve as examples of how psychological and neuroscientific

data are beginning to characterize the mechanisms that

underlie our sense of free will, how these mechanisms can

lead us to assume free will is operating when it is not, and

how a scientific understanding of these mechanisms can

serve to dismantle our commitment to the idea of free will.

1782 J. Greene and J. Cohen Neuroscience changes nothing and everything

Looking out at the world, it appears to contain two fun-

damentally different kinds of entity. On the one hand, there

are ordinary objects that appear to obey the ordinary laws

of physics: things like rocks and puddles of water and

blocks of wood. These things do not get up and move

around on their own. They are, in a word, inanimate. On

the other hand, there are things that seem to operate by

some kind of magic. Humans and other animals, so long as

they are alive, can move about at will, in apparent defiance

of the physical laws that govern ordinary matter. Because

things like rocks and puddles, on the one hand, and mice

and humans, on the other, behave in such radically differ-

ent ways, it makes sense, from an evolutionary perspective,

that creatures would evolve separate cognitive systems for

processing information about each of these classes of

objects (Pinker 1997). There is a good deal of evidence to

suggest that this is precisely how our minds work.

A line of research beginning with Fritz Heider illustrates

this point. Heider and Simmel (Heider & Simmel 1944)

created a film involving three simple geometric shapes that

move about in various ways. For example, a big triangle

chases a little circle around the screen, bumping into it.

The little circle repeatedly moves away, and a little triangle

repeatedly moves in between the circle and the big triangle.

When normal people watch this movie they cannot help

but view it in social terms (Heberlein & Adolphs 2004).

They see the big triangle as trying to harm the little circle,

and the little triangle as trying to protect the little circle; and

they see the little circle as afraid and the big triangle as frus-

trated. Some people even spontaneously report that the big

triangle is a bully. In other words, simple patterns of move-

ment trigger in people’s minds a cascade of complex social

inferences. People not only see these shapes as ‘alive’. They

see beliefs, desires, intentions, emotions, personality traits

and even moral blameworthiness. It appears that this kind

of inference is automatic (Scholl & Tremoulet 2000). Of

course, you, the observer, know that it is only a film, and a

very simple one at that, but you nevertheless cannot help

but see these events in social, even moral, terms.

That is, unless you have damage to your amygdala, a

subcortical brain structure that is important for social cog-

nition (Adolphs 1999). Andrea Heberlein tested a patient

with rare bilateral amygdala damage using Heider’s film

and found that this patient, unlike normal people,

described what she saw in completely asocial terms, despite

that fact that her visual and verbal abilities are not compro-

mised by her brain damage. Somehow, this patient is blind

to the ‘human’ drama that normal people cannot help but

see in these events (Heberlein & Adolphs 2004).

The sort of thinking that is engaged when normal people

view the Heider–Simmel film is sometimes known as

‘folk psychology’ (Fodor 1987), ‘the intentional stance’

(Dennett 1987) or ‘theory of mind’, (Premack &

Woodruff 1978). There is a fair amount of evidence

(including the work described above) suggesting that

humans have a set of cognitive subsystems that are specia-

lized for processing information about intentional agents

(Saxe et al. 2004). At the same time, there is evidence to

suggest that humans and other animals also have sub-

systems specialized for ‘folk physics’, an intuitive sense of

how ordinary matter behaves. One compelling piece of evi-

dence for the claim that normal humans have subsystems

specialized for folk physics comes from studies of people

Phil. Trans. R. Soc. Lond. B (2004)

with autism spectrum disorder. These individuals are parti-

cularly bad at solving problems that require ‘folk psy-

chology’, but they do very well with problems related to how

physical objects (e.g. the parts of machine) behave, i.e. ‘folk

physics’ (Baron Cohen 2000). Another piece of evidence for

a ‘folk physics’ system comes from discrepancies between

people’s physical intuitions and the way the world actually

works. People say, for example, that a ball shot out of a

curved tube resting on a flat surface will continue to follow a

curved path outside the tube when in fact it will follow a

straight path (McCloskey et al. 1980). The fact that people’s

physical intuitions are slightly, but systematically, out of

step with reality suggests that the mind brings a fair amount

of implicit theory to the perception of physical objects.

Thus, it is at least plausible that we possess distinguish-

able cognitive systems for making sense of the behaviour of

objects in the world. These systems seem to have two fun-

damentally different ‘ontologies’. The folk physics system

deals with chunks of matter that move around without pur-

poses of their own according to the laws of intuitive phys-

ics, whereas the folk psychology system deals with unseen

features of minds: beliefs, desires, intentions, etc. But

what, to our minds, is a mind? We suggest that a crucial

feature, if not the defining feature, of a mind (intuitively

understood) is that it is an uncaused causer (Scholl &

Tremoulet 2000). Minds animate material bodies, allow-

ing them to move without any apparent physical cause and

in pursuit of goals. Moreover, we reserve certain social atti-

tudes for things that have minds. For example, we do not

resent the rain for ruining our picnic, but we would resent a

person who hosed our picnic (Strawson 1962), and we

resent picnic-hosers considerably more when we perceive

that their actions are intentional. Thus, it seems that folk

psychology is the gateway to moral evaluation. To see

something as morally blameworthy or praiseworthy (even if

it is just a moving square), one has to first see it as ‘some-

one’, that is, as having a mind.

With all of this in the background, one can see how the

problem of attributive free will arises. To see something as

a responsible moral agent, one must first see it as having a

mind. But, intuitively, a mind is, among other things, an

uncaused causer. Consequently, when something is seen as

a mere physical entity operating in accordance with deter-

ministic physical laws, it ceases to be seen, intuitively, as a

mind. Consequently, it is seen as an object unworthy of

moral praise or blame. (Note that we are not claiming that

people automatically attribute moral agency to anything

that appears to be an uncaused causer. Rather, our claim is

that seeing something as an uncaused causer is a necessary

but not sufficient condition for seeing something as a moral

agent.)

After thousands of years of our thinking of one another

as uncaused causers, science comes along and tells us

that there is no such thing—that all causes, with the

possible exception of the Big Bang, are caused causes

(determinism). This creates a problem. When we look at

people as physical systems, we cannot see them as any more

blameworthy or praiseworthy than bricks. But when we

perceive people using our intuitive, folk psychology we can-

not avoid attributing moral blame and praise.

So, philosophers who would honour both our scientific

knowledge and our social instincts try to reconcile these

two competing outlooks, but the result is never completely

Neuroscience changes nothing and everything J. Greene and J. Cohen 1783

satisfying, and the debate wears on. Philosophers who can-

not let go of the idea of uncaused causes defend libertarian-

ism, and thus opt for scientifically dubious, ‘panicky

metaphysics’. Hard determinists, by contrast, embrace the

conclusions of modern science, and concede what others

will not: that many of our dearly held social practices are

based on an illusion. The remaining majority, the compati-

bilists, try to talk themselves into a compromise. But the

compromise is fragile. When the physical details of human

action are made vivid, folk psychology loses its grip, just as

folk physics loses its grip when the morally significant

details are emphasized. The problem of free will and deter-

minism will never find an intuitively satisfying solution

because it arises out of a conflict between two distinct cog-

nitive subsystems that speak different cognitive ‘languages’

and that may ultimately be incapable of negotiation.

8. FREE WILL, RESPONSIBILITY AND CONSEQUENTIALISM

Even if there is no intuitively satisfying solution to the prob-

lem of free will, it does not follow that there is no correct

view of the matter. Ours is as follows: when it comes to the

issue of free will itself, hard determinism is mostly correct.

Free will, as we ordinarily understand it, is an illusion.

However, it does not follow from the fact that free will is an

illusion that there is no legitimate place for responsibility.

Recall from x 2 that there are two general justifications for holding people legally responsible for their actions. The

retributive justification, by which the goal of punishment is

to give people what they really deserve, does depend on this

dubious notion of free will. However, the consequentialist

approach does not require a belief in free will at all. As con-

sequentialists, we can hold people responsible for crimes

simply because doing so has, on balance, beneficial effects

through deterrence, containment, etc. It is sometimes said

that if we do not believe in free will then we cannot legiti-

mately punish anyone and that society must dissolve into

anarchy. In a less hysterical vein, Daniel Wegner argues

that free will, while illusory, is a necessary fiction for the

maintenance of our social structure (Wegner 2002, ch. 9).

We disagree. There are perfectly good, forward-looking

justifications for punishing criminals that do not depend on

metaphysical fictions. (Wegner’s observations may apply

best to the personal sphere: see below.)

The vindication of responsibility in the absence of free

will means that there is more than a grain of truth in com-

patibilism. The consequentialist approach to responsibility

generates a derivative notion of free will that we can

embrace (Smart 1961). In the name of producing better

consequences, we will want to make several distinctions

among various actions and agents. To begin, we will want

to distinguish the various classes of people who cannot be

deterred by the law from those who can. That is, we will

recognize many of the ‘diminished capacity’ excuses that

the law currently recognizes such as infancy and insanity.

We will also recognize familiar justifications such those

associated with crimes committed under duress (e.g. threat

of death). If we like, then, we can say that the actions of

rational people operating free from duress, etc. are free

actions, and that such people are exercising their free will.

At this point, compatibilists such as Daniel Dennett may

claim victory: ‘what more could one want from free will?’.

Phil. Trans. R. Soc. Lond. B (2004)

In a word: retributivism. We have argued that common-

sense retributivism really does depend on a notion of free

will that is scientifically suspect. Intuitively, we want to

punish those people who truly deserve it, but whenever the

causes of someone’s bad behaviour are made sufficiently

vivid, we no longer see that person as truly deserving of

punishment. This insight is expressed by the old French

proverb: ‘to know all is to forgive all’. It is also expressed in

the teachings of religious figures, such as Jesus and Bud-

dha, who preach a message of universal compassion. Neu-

roscience can make this message more compelling by

vividly illustrating the mechanical nature of human action.

Our penal system is highly counter-productive from a

consequentialist perspective, especially in the USA, and yet

it remains in place because retributivist principles have a

powerful moral and political appeal (Lacey 1988; Tonry

2004). It is possible, however, that neuroscience will

change these moral intuitions by undermining the intuitive,

libertarian conceptions of free will on which retributivism

depends.

As advocates of consequentialist legal reform, it behoves

us to briefly respond to the three standard criticisms levied

against consequentialist theories of punishment. First, it is

claimed that consequentialism would justify extreme over-

punishing. As noted above, it is possible in principle that

the goal of deterrence would justify punishing parking vio-

lations with the death penalty or framing innocent people

to make examples of them. Here, the standard response is

adequate. The idea that such practices could, in the real

world, make society happier on balance is absurd. Second,

it is claimed that consequentialism justifies extreme under-

punishment. In response to some versions of this objection,

our response is the same as above. Deceptive practices such

as a policy of faking punishment cannot survive in a free

society, and a free society is required for the pursuit of most

consequentialist ends. In other cases consequentialism may

advocate more lenient punishments for people who, intuit-

ively, deserve worse. Here, we maintain that a deeper

understanding of human action and human nature will lead

people—more of them, at any rate—to abandon these retri-

butivist intuitions. Our response is much the same to the

third and most general criticism of consequentialist punish-

ment, which is that even when consequentialism gets the

punishment policy right, it does so for the wrong reasons.

These supposedly right reasons are reasons that we reject,

however intuitive and natural they may feel. They are, we

maintain, grounded in a metaphysical view of human

action that is scientifically dubious and therefore an unfit

basis for public policy in a pluralistic society.

Finally, as defenders of hard determinism and a con-

sequentialist approach to responsibility, we should briefly

address some standard concerns about the rejection of free

will and conceptions of responsibility that depend on it.

First, does not the fact that you can raise your hand ‘at will’

prove that free will is real? Not in the sense that matters. As

Daniel Wegner (2002) has argued, our first-person sense of

ourselves as having free will may be a systematic illusion.

And from a third-person perspective, we simply do not

assume that anyone who exhibits voluntary control over

his body is free in the relevant sense, as in the case of

Mr Puppet.

A more serious challenge is the claim that our commit-

ments to free will and retributivism are simply inescapable

1784 J. Greene and J. Cohen Neuroscience changes nothing and everything

for all practical purposes. Regarding free will, one might

wonder whether one can so much as make a decision with-

out implicitly assuming that one is free to choose among

one’s apparent options. Regarding responsibility and pun-

ishment, one might wonder if it is humanly possible to deny

our retributive impulses (Strawson 1962; Pettit 2002).

This challenge is bolstered by recent work in the beha-

vioural sciences suggesting that an intuitive sense of fair-

ness runs deep in our primate lineage (Brosnan & De Waal

2003) and that an adaptive tendency towards retributive

punishment may have been a crucial development in the

biological and cultural evolution of human sociality (Fehr

& Gachter 2002; Boyd et al. 2003; Bowles & Gintis 2004).

Recent neuroscientific findings have added further support

to this view, suggesting that the impulse to exact punish-

ment may be driven by phylogentically old mechanisms in

the brain (Sanfey et al. 2003). These mechanisms may be

an efficient and perhaps essential, device for maintaining

social stability. If retributivism runs that deep and is that

useful, one might wonder whether we have any serious

hope of, or reason for, getting rid of it. Have we any real

choice but to see one another as free agents who deserve to

be rewarded and punished for our past behaviours?

We offer the following analogy: modern physics tells us

that space is curved. Nevertheless, it may be impossible for

us to see the world as anything other than flatly Euclidean

in our day-to-day lives. And there are, no doubt, deep evol-

utionary explanations for our Euclidean tendencies. Does

it then follow that we are forever bound by our innate

Euclidean psychology? The answer depends on the domain

of life in question. In navigating the aisles of the grocery

store, an intuitive, Euclidean representation of space is not

only adequate, but probably inevitable. However, when we

are, for example, planning the launch of a spacecraft, we

can and should make use of relativistic physical principles

that are less intuitive but more accurate. In other words, a

Euclidean perspective is not necessary for all practical pur-

poses, and the same may be true for our implicit commit-

ment to free will and retributivism. For most day-to-day

purposes it may be pointless or impossible to view ourselves

or others in this detached sort of way. But—and this is the

crucial point—it may not be pointless or impossible to

adopt this perspective when one is deciding what the crimi-

nal law should be or whether a given defendant should be

put to death for his crimes. These may be special situations,

analogous to those routinely encountered by ‘rocket scien-

tists’, in which the counter-intuitive truth that we legiti-

mately ignore most of the time can and should be

acknowledged.

Finally, there is the worry that to reject free will is to ren-

der all of life pointless: why would you bother with any-

thing if it has all long since been determined? The answer is

that you will bother because you are a human, and that is

what humans do. Even if you decide, as part of a little intel-

lectual exercise, that you are going to sit around and do

nothing because you have concluded that you have no free

will, you are eventually going to get up and make yourself a

sandwich. And if you do not, you have got bigger problems

than philosophy can fix.

Phil. Trans. R. Soc. Lond. B (2004)

9. CONCLUSION Neuroscience is unlikely to tell us anything that will chal-

lenge the law’s stated assumptions. However, we maintain

that advances in neuroscience are likely to change the

way people think about human action and criminal

responsibility by vividly illustrating lessons that some

people appreciated long ago. Free will as we ordinarily

understand it is an illusion generated by our cognitive

architecture. Retributivist notions of criminal responsi-

bility ultimately depend on this illusion, and, if we are

lucky, they will give way to consequentialist ones, thus

radically transforming our approach to criminal justice.

At this time, the law deals firmly but mercifully with

individuals whose behaviour is obviously the product of

forces that are ultimately beyond their control. Some day,

the law may treat all convicted criminals this way. That is,

humanely.

The authors thank Stephen Morse, Andrea Heberlein, Aaron Schurger, Jennifer Kessler and Simon Keller for their input.

ENDNOTES 1 Of course, scientific respectability is not everyone’s first priority.

However, the law in most Western states is a public institution

designed to function in a society that respects a wide range of religious

and otherwise metaphysical beliefs. The law cannot function in this

way if it presupposes controversial and unverifiable metaphysical facts

about the nature of human action, or anything else. Thus, the law

must restrict itself to the class of intersubjectively verifiable facts, i.e.

the facts recognized by science, broadly construed. This practice need

not derive from a conviction that the scientifically verifiable facts are

necessarily the only facts, but merely from a recognition that verifiable

or scientific facts are the only facts upon which public institutions in a

pluralistic society can effectively rely. 2 There are some forms of dualism according to which the mind and

body, although distinct, do not interact, making it impossible for the

mind to have any observable effects on the brain or anything else in the

physical world. These versions of dualism do not concern us here. For

the purposes of this paper, we are happy to allow the metaphysical

claim that souls or aspects of minds may exist independently of the

physical body. Our concern is specifically with interactionist versions

of dualism according to which non-physical mental entities have

observable physical effects. We believe that science has rendered such

views untenable and that the law, insofar as it is a public institution

designed to serve a pluralistic society, must not rely on beliefs that are

scientifically suspect (see previous endnote). 3 It is conceivable that rationality could someday be redefined in

neurocognitive rather than behavioural terms, much as water has been

redefined in terms of its chemical composition. Were that to happen,

neuroscientific evidence could then be construed as more direct than

behavioural evidence. But Steinberg and Scott’s argument appears to

make use of a conventional, behavioural definition of rationality and

not a neurocognitive redefinition. 4 This is not to say that we could not describe Mr Puppet in such a

way that our intuitions about him would change. Our point is only

that, when the details are laid bare, it is very hard to see him as morally

responsible. 5 Compatibilist philosophers such as Daniel Dennett (2003) might

object that the story of Mr Puppet is nothing but a misleading

‘intuition pump’. Indeed, this is what Dennett says about a similar

case of Alfred Mele’s (1995). We believe that our case is importantly

different from Mele’s. Dennett and Mele imagine two women who are

psychologically identical: Ann is a typical, good person, whereas Beth

has been brainwashed to be just like Ann. Dennett argues, against

Mele, that if you take seriously the claim that these two are psychologi-

cally identical and properly imagine that Beth is as rational, open-

minded, etc. as Ann, you will come to see that the two are equally free.

We agree with Dennett that Ann and Beth are comparable and that

Neuroscience changes nothing and everything J. Greene and J. Cohen 1785

Mele’s intuition falters when the details are fleshed out. But does the

same hold for the intuition provoked by Mr Puppet’s story? It seems to

us that the more one knows about Mr Puppet and his life the less

inclined one is to see him as truly responsible for his actions and our

punishing him as a worthy end in itself. We can agree with Dennett

that there is a sense in which Mr Puppet is free. Our point is merely

that there is a legitimate sense in which he, like all of us, is not free and

that this sense matters for the law. 6 We do not wish to imply that neuroscience will inevitably put us in a

position to predict any given action based on a neurological examin-

ation. Rather, our suggestion is simply that neuroscience will eventu-

ally advance to the point at which the mechanistic nature of human

decision-making is sufficiently apparent to undermine the force of

dualist/libertarian intuitions.

REFERENCES Adolphs, R. 1999 Social cognition and the human brain. Trends Cogn. Sci. 3, 469–479.

Baron Cohen, S. 2000 Autism: deficits in folk psychology exist alongside superiority in folk physics. In Understanding other minds: perspectives from autism and developmental cognitive neuroscience (ed. S. Baron Cohen, H. Tager Flusberg & D. Cohen), pp. 78–82. New York: Oxford University Press.

Bentham, J. 1982 An introduction to the principles of morals and legislation. London: Methuen.

Bowles, S. & Gintis, H. 2004 The evolution of strong recip- rocity: cooperation in heterogeneous populations. Theor. Popul. Biol. 65, 17–28.

Boyd, R., Gintis, H., Bowles, S. & Richerson, P. J. 2003 The evolution of altruistic punishment. Proc. Natl Acad. Sci. USA 100, 3531–3535.

Brosnan, S. F. & De Waal, F. B. 2003 Monkeys reject unequal pay. Nature 425, 297–299.

Dennett, D. C. 1984 Elbow room: the varieties of free will worth wanting. Cambridge, MA: MIT Press.

Dennett, D. C. 1987 The intentional stance. Cambridge, MA: MIT Press.

Dennett, D. C. 1991 Consciousness explained. Boston, MA: Little Brown and Co.

Dennett, D. C. 2003 Freedom evolves. New York: Viking. Fehr, E. & Gachter, S. 2002 Altruistic punishment in humans. Nature 415, 137–140.

Fodor, J. A. 1987 Psychosemantics: the problem of meaning in the philosophy of mind. Cambridge, MA: MIT Press.

Frankfurt, H. 1966 Alternate possibilities and moral responsi- bility. J. Philosophy 66, 829–839.

Goldstein, A. M., Morse, S. J. & Shapiro, D. L. 2003 Evalu- ation of criminal responsibility. In Forensic psychology. vol. 11 (ed. A. M. Goldstein), pp. 381–406. New York: Wiley.

Goldstein, A. S. 1967 The insanity defense. New Haven, CT: Yale University Press.

Hart, H. L. A. 1968 Punishment and responsibility. Oxford Univeristy Press.

Heberlein, A. S. & Adolphs, R. 2004 Impaired spontaneous anthropomorphizing despite intact perception and social knowledge. Proc. Natl Acad. Sci. USA 101, 7487–7491.

Heider, F. & Simmel, M. 1944 An experimental study of apparent behavior. Am. J. Psychol. 57, 243–259.

Phil. Trans. R. Soc. Lond. B (2004)

Hughs, R. I. G. 1992 The structure and interpretation of quantum

mechanics. Cambridge, MA: Havard University Press. Kahneman, D. & Tversky, A. (eds) 2000 Choices, values, and frames. Cambridge University Press.

Kant, I. 2002 The philosophy of law: an exposition of the funda-

mental principles of jurisprudence as the science of right. Union, NJ: Lawbook Exchange.

Lacey, N. 1988 State punishment: political principles and com-

munity values. London and New York: Routledge & Kegan Paul.

McCloskey, M., Caramazza, A. & Green, B. 1980 Curvi-

linear motion in the absence of external forces: naive

beliefs about the motion of objects. Science 210, 1139–

1141. Mele, A. 1995 Autonomous agents: from self-control to autonomy.

Oxford University Press. Morse, S. J. 2004 New neuroscience, old problems. In Neu- roscience and the law: brain, mind, and the scales of justice (ed.

B. Garland), pp. 157–198. New York: Dana Press. Pettit, P. 2002 The capacity to have done otherwise. Rules, rea- sons, and norms: selected essays. Oxford University Press.

Pincus, J. H. 2001 Base instincts: what makes killers kill? New

York: Norton. Pinker, S. 1997 How the mind works. New York: Norton. Premack, D. & Woodruff, G. 1978 Does the chimpanzee have

a theory of mind? Behav. Brain Sci. 4, 515–526. Rosen, G. 2002 The case for incompatibilism. Philosophy

Phenomenol. Res. 64, 699–706. Sanfey, A. G., Rilling, J. K., Aronson, J. A., Nystrom, L. E. &

Cohen, J. D. 2003 The neural basis of economic decision-

making in the ultimatum game. Science 300, 1755–1758. Saxe, R., Carey, S. & Kanwisher, N. 2004 Understanding

other minds: liking developmental psychology and func-

tional neuroimaging. A. Rev. Psychol. 55, 87–124. Scholl, B. J. & Tremoulet, P. D. 2000 Perceptual causality and

animacy. Trends Cogn. Sci. 4, 299–309. Shear, J. (ed.) 1999 Explaining consciousness: the hard

problem. Cambridge, MA: MIT Press. Smart, J. J. C. 1961 Free will, praise, and blame. Mind 70,

291–306. Steinberg, L. & Scott, E. S. 2003 Less guilty by reason of ado-

lescence: developmental immaturity, diminished responsi-

bility, and the juvenile death penalty. Am. Psychol. 58,

1009–1018. Strawson, P. F. 1962 Freedom and resentment. Proc. Br.

Acad. xlviii, 1–25. Tonry, M. 2004 Thinking about crime: sense and sensibility in

American penal culture. New York: Oxford University

Press. Van Inwagen, P. 1982 The incompatibility of free will and

determinism. In Free will (ed. G. Watson), pp. 46–58. New

York: Oxford University Press. Watson, G. (ed.) 1982 Free will. New York: Oxford University

Press. Wegner, D. M. 2002 The illusion of conscious will. Cambridge, MA: MIT Press.

Wright, R. 1994 The moral animal: evolutionary psychology and

everyday life. New York: Pantheon.