The philosopher Stuart Hampshire served in British military intelligence during the Second World War. When we were colleagues at Princeton he told me about the following incident, which must have taken place shortly after the Normandy landings. The French Resistance had captured an important collaborator, who was thought to have information that would be useful to the Allies. Hampshire was sent to interrogate him. When he arrived, the head of the Resistance unit told Hampshire he could question the man, but that when he was through they were going to shoot him: that’s what they always did with these people. He then left Hampshire alone with the prisoner. The man said immediately that he would tell Hampshire nothing unless Hampshire guaranteed he would be turned over to the British. Hampshire replied that he could not give such a guarantee. So the man told him nothing before he was shot by the French.
Another philosopher, when I told him this story, commented drily that what it showed was that Hampshire was a very poor choice for the assignment. But I tell it here not in order to determine whether Hampshire did the right thing in failing to lie to the prisoner in these circumstances. I offer it as a real-life example of the force of a certain type of immediate moral reaction. Even those who think that Hampshire should, for powerful instrumental reasons, have made a false promise of life to this man facing certain death can feel the force of the barrier that presented itself to Hampshire. It is an example of the sort of moral gut reaction that figures prominently in the recent literature of empirical moral psychology. I assume that a scan of Hampshire’s brain at the time would have revealed heightened activity in the ventromedial prefrontal cortex.
Much intellectual effort has gone into the delineation of the protective boundaries around people that ordinary morality says we must not cross. Usually the examples designed to call forth our intuitions are more artificial than this one – as in the famous trolley problem. But the phenomenon is real, and an inescapable part of human morality. I am interested in the question of how to decide what authority to give to these moral judgments, or perceptions, or intuitions – what kind of thinking can lead us either to affirm them as correct and fundamental, or to detach from them so that we come to regard them as mere appearances without practical validity – or alternatively perhaps to step back from them but nevertheless allow them some influence in our lives that is not fundamental but is derived from other values. This problem has been around for a long time, and much of what I say about it will be familiar. But recent discussion prompts another look.
It is a question of moral epistemology: not the kind of epistemological question posed when we consider how to respond to a general scepticism about morality, or about value, but an epistemological question internal to moral thought. There is a venerable tradition of scepticism about whether any moral judgments, or the intuitions that support them, can be regarded as correct or incorrect, rather than as mere feelings of a special kind that we express in the language of morality. I am not going to enter into that larger debate here. I will proceed on the assumption that it makes sense to try to discover what is really right and wrong, and that moral intuitions provide prima facie evidence in this inquiry. The problem I want to discuss arises because, for some of our most powerful intuitions, there are various possible explanations, both moral and causal, that would, if correct, undermine their claim to fundamental authority – the claim that those convictions should be taken at face value as perceptions of the moral truth. Challenges of this kind present us with the task of finding a way to conduct ourselves that is consistent with the best understanding of ourselves from outside – as biological, psychological, social or historical products.
The question has broad legal and political importance because in liberal constitutional regimes many of the rights and protections of the individual against the exercise of collective power appear initially as intuitive boundaries of this type. Freedom of religion, freedom of thought and expression, freedom of association, sexual and reproductive freedoms, protections of privacy, prohibitions of torture and cruel punishment are all supported and in part identified by an immediate sense of what may and may not be done to people, a constraint that precedes cost-benefit calculations.
Even though it is possible to construct more or less plausible consequentialist justifications – justifications in terms of long-term costs and benefits – for strict legal rules embodying such protections, that is not the moral aspect under which they immediately present themselves. The violation of an individual right seems wrong in itself, and not merely as the transgression of a socially valuable strict general rule. The question is whether this is an illusion – a natural illusion built into our moral psychology. Though Hampshire’s uncrossable boundary arose in the context of an individual decision, it feels similar to the boundary which bars the state from employing torture to get information, even against its enemies and for reasons of national security. And as we have seen in our own time, the bar against torture is not uncontested.
Ihave a lot of sympathy with the classic intuitionist W.D. Ross when he says: ‘To ask us to give up at the bidding of a theory our actual apprehension of what is right and what is wrong seems like asking people to repudiate their actual experience of beauty, at the bidding of a theory which says “only that which satisfies such and such conditions can be beautiful.”’ Still, some move into theory seems unavoidable in response to the deep disagreements that repeatedly emerge in discussion of this subject.
John Rawls gave the name ‘reflective equilibrium’ to the process of putting one’s moral thoughts in order by testing general principles against considered judgments about particular cases, and adjusting both until they fit more or less comfortably together. The process does not treat particular judgments as unrevisable givens, or general principles as self-evident axioms, so it need not be conservative: it can lead to radical revision of some of the considered judgments with which one begins. But it must take intuitive value judgments as starting points, and in order to dismiss some of those judgments as mistaken, it must rely on others – just as we must rely on perceptual evidence when dismissing some perceptual appearances as illusions. I think there is no alternative to this method for pursuing answers to moral questions in which we can maintain some confidence, even in the face of disagreement.
While the process is structurally similar to that of testing empirical hypotheses in the natural or social sciences against observational evidence, there is a crucial difference. In the scientific case we understand our perceptual observations to be the result of causal interaction with the world we are investigating. Even though we haven’t solved the mind-body problem, and don’t understand how the brain produces conscious experience, we have a rough conception of ourselves as organisms causally embedded in the world whose constituents and laws we are trying to discover; so how things seem from our perceptual point of view clearly provides data for such an investigation. However, in the moral case we do not take our evaluative intuitions to be the result of causal interaction with the moral domain, and it is not clear what other kind of embeddedness or access to the moral truth moral judgment might involve. As is often observed, moral judgment has this in common with logical and mathematical judgment, which are also not the result of our causal interaction with the realms of logic and mathematics. Knowledge arrived at just by thinking is mysterious. But I am going to leave aside the large and difficult question of how this is possible, and concentrate on more specific issues.
Here are the familiar features of ordinary moral thought that give rise to our problem. We evaluate many different kinds of thing, but important among them are states of affairs or outcomes, on the one hand, and actions or policies, on the other. To evaluate states of affairs we use the concepts of good and bad, better and worse. To evaluate actions we use in addition the concepts of right and wrong. The classical problem is whether there is an independent aspect of morality governing the rightness and wrongness of acts and policies – either of individuals or of institutions – or whether the only truly fundamental values are good and bad, so that standards of right and wrong must be explained instrumentally, by identifying the types of actions and policies that lead to good and bad outcomes. The latter possibility was given the name ‘consequentialism’ by Elizabeth Anscombe, and its best-known version is utilitarianism. The opposite view, that the right is at least in some respects independent of the good, doesn’t have a name, but the principles that it identifies are usually called ‘deontological’ – an ugly word, but we seem to be stuck with it. Deontological principles say that whether an act is morally permitted, prohibited or required often depends not on the goodness or badness of its overall consequences but on intrinsic features of the act itself. In a case like Hampshire’s, the calculation of probable consequences is clearly in favour of lying to the prisoner, so if it would be wrong to do so, that would have to be for some other reason.
The coexistence in ordinary moral thought and discourse of these two types of evaluation is the starting point for moral reflection that might lead us to a new equilibrium. If we take both consequences and deontology seriously as guides to conduct, they will naturally present us often with the sense of a dilemma. When a deontological prohibition – against killing the innocent, against breach of promise, against betrayal – blocks an act that would prevent a greater evil or produce a great good, we are likely still to feel the force of the reason to promote the good, which makes it tempting to violate the prohibition. I believe the sense of moral conflict in these cases arises naturally, and is not just a philosophical artefact.
One possible response to such a dilemma is simply to take the consequentialist reasons as decisive, so that they override the deontological intuition, which is construed as a form of moral squeamishness or a groundless taboo. Yet without more explanation this all-consequentialist response would be arbitrary. The problem arises because we have both types of moral intuition, and if we are going to let one type override the other, it could just as easily go in the opposite direction: we could decide that the principle that one should always do what will have the best overall consequences is shown to be wrong by the obvious impermissibility of using particular means like torture to achieve good ends. The conclusion would be that to allow good and bad outcomes always to determine right and wrong is an illusion, perhaps an illusion about what is demanded by rationality.
Merely pitting the intuitions against each other results in a stand-off. To decide which of them to credit and which, if any, to throw out we must do more. For each side of the dilemma, we must consider the pros and cons – both how to make the best sense of these types of judgments and how most convincingly to undermine their authority.
On the positive side, it is easiest to explain the meaning of consequentialist values, as in utilitarianism. Each of us has direct access to the goodness or badness of certain things in our own lives, such as pleasure and pain, freedom and coercion, survival and death. Once we accept the crucial judgment that such things are objectively good and bad, the imagination allows us to extend these values to similar features in the lives of others. There seems no reason to weigh a given quantity of pleasure or pain differently depending on who undergoes it. So the obvious way to compare the value of two states of affairs is to add up the good or bad in the lives of all those involved, and see which has the higher net balance. That then is the one we should prefer and, if possible, bring about. Good is preferable to bad and better is preferable to worse. Whether or not they are correct, consequentialist value judgments are unmysterious.
The positive interpretation of deontological values is less transparent, but I believe we can give them a definite sense. When we think in this way about how to value others, it is their status as autonomous beings independent of us that is central, not their susceptibility to pleasure and pain, or other good or bad things that might happen to them. Deontological requirements govern our direct interaction with each other person; they determine how we may treat him, rather than what we should want to happen to him. As with consequentialist values, this is an extension to everyone of a value whose importance we can recognise in our own case. The basic idea is that we must regard each person, including ourselves, as immune from subjection to others – the centre of a morally protected sphere of individual autonomy that can be granted equally to everyone. Each of us respects this autonomy in others most fundamentally not as a good that we should promote but as a boundary that we must not violate. We must not violate it even to prevent more violations of the same kind by others – hence the prohibition against torture even to get information about a planned terrorist attack. This idea of inviolability seems clear even if there are significant uncertainties in the precise shape of the deontological boundaries. There is room for indeterminacy and ongoing reflection here as elsewhere. So far this is merely an interpretation of deontology as an intelligible way of valuing people – not a proof of its correctness. But in that respect it is on the same footing as the interpretation I have given of consequentialism.
According to this account, our deontological convictions that certain things may not be done to people are just the subjective experience of running up against objective boundaries of inviolability that define a systematic form of value. Hampshire evidently felt that to manipulate the prisoner by falsely promising him his life was something he just could not do, even to an enemy. It is phenomenologically more complex than the feelings of impartial sympathy for the pleasures and pains of others that would motivate an obedience to utilitarian requirements. But it is intelligible as a response to a different kind of value that humans possess.
Thus far there is something positive to be said for both types of intuition, consequentialist and deontological. However, this is not the end of the story. I said that it would be necessary to examine the pros and cons for both types of intuition, and the most interesting arguments are those that claim to undermine the authority of some of our intuitive starting points – especially the deontological ones. I shall describe two.
The first is a theory owed to David Hume, which seeks not to discredit deontological requirements but to show that they are not morally basic, because they can be explained in terms of other values. The theory is called rule-consequentialism, or in one of its more specific versions, rule-utilitarianism. It purports to vindicate many of our deontological intuitions – about property, contract, promise, political obligation and various individual rights – by attributing them to our having internalised certain conventions with strict rules that serve the general interest.
The theory is presented by Hume in his account of what he calls the ‘artificial’ virtues, and its point is that these moral requirements do not have a foundation independent of the general good of society. They seem to need an independent foundation only because the general good of society requires strict rules, such as the rules of property and contract, that can serve their consequentialist purpose only if they are followed even when their violation in an individual case would produce more good than harm. The utility of the rule in creating security and predictability demands that utility not be considered in deciding whether to adhere to it in each individual case. Participants in these conventions who have internalised the rules experience the wrongness of stealing and promise-breaking as simple and intrinsic; but those requirements do not wear their true meaning on their sleeves. Their conventionality, and their consequentialist foundations, are hidden, and this is an aid to their effectiveness.
Although I believe it cannot give a fully adequate account of rights, Hume’s theory is a brilliant contribution to moral philosophy. It offers to explain deontological requirements in a way that also opens the possibility of revision of those requirements if the good of society would be better served by modification of the governing conventions and rules. I believe rule-consequentialism is at least part of the truth about rights and deontological obligations. This is strongly supported by the fact that societies that have abandoned the protection of individual rights in order to leave their hands free for the more effective pursuit of the general good have some of the most dismal records with respect to its actual achievement.
Rule-consequentialism doesn’t seek to eliminate deontology but to show that it is not morally fundamental. But a second, more negative criticism of anti-consequentialist moral intuitions has recently become prominent. It is found in the writings of certain psychologists, notably Daniel Kahneman, Jonathan Haidt and Joshua Greene, who have turned their attention to the psychological analysis of moral judgment and motivation, and also to their neurophysiological, evolutionary or sociological underpinnings. The approach has also been taken up by non-psychologists like Peter Singer and Cass Sunstein. Like Hume, it ascribes to deontological rules some form of social utility, but it is often disposed to find that utility more pronounced in the past, when our hunter-gatherer ancestors lived in small groups and evolutionary forces produced the then functional dispositions that we find operating in us to this day. Those dispositions strongly inhibit direct interpersonal aggression and violation of interpersonal agreements, and therefore contribute to peaceful coexistence and co-operation. They are efficient because they don’t require knowledge of long-term consequences, but respond only to the immediate character of the action, and engage the emotions rather than reasoning.
In the end, the account is partly vindicating, like Hume’s, but partly debunking. It does not set out to discover the moral foundation for our intuitions. Instead it asks us to take a detached stance towards ourselves and our moral responses – to try to understand ourselves from outside, so to speak. We then discover, it is claimed, that some of our most confident moral judgments are the product not of reason but of emotionally charged instinct, created by natural selection. Only after explaining those responses scientifically can we decide which of them we should continue to allow to influence us and which we should instead dismiss as emotional illusions.
The conclusion of these psychological critics is that while the evolutionary legacy of deontological morality retains some usefulness as a set of heuristics for identifying the right thing to do much of the time – or, to put it in Kahneman’s terms, as an example of ‘thinking fast’ as opposed to the less efficient but more accurate ‘thinking slow’ – nevertheless the only moral standards that are valid in their own right are the rational standards of consequentialism, more specifically some form of utilitarianism. That is the standard that underpins what is useful in the deontological aspects of intuitive morality, and it is also the standard that justifies overriding deontological intuitions when a slower calculation of costs and benefits reveals that they would lead us to do more harm than good.
Even with this alleged scientific support, however, I believe it is too early to declare victory for consequentialism.
The detached biological stance towards ourselves is now a cultural commonplace. People are accustomed to thinking of their psychological dispositions as the product of natural selection, and of their minds and motives as significantly not under their conscious rational control. But this merely poses the problem of distinguishing between moral appearance and moral reality, and does not yet solve it. Psychologically reductive or debunking accounts of our moral intuitions are not self-validating; they are contributions to the process of reflective equilibrium, and we have to decide whether they are more plausible than the judgments they propose to replace.
The problem of appearance and reality is at the heart of philosophy, and it is found in every branch of the subject. It arises when we step outside ourselves temporarily, and consider the way things seem to us in some respect as a psychological fact – a fact about a certain type of creature in the world. The question then is whether the best explanation of this psychological fact – employing the forms of biological, neurophysiological, psychological and historical understanding available to us – is compatible with our continuing to affirm that things are as they appear to be from that perspective. We can pose this question whether the appearance is a sensory perception, a memory, a mathematical certainty, an aesthetic judgment, or a moral conviction.
However, it is important that when we take this step outside ourselves, the inside point of view that we are examining does not disappear. We cannot completely withdraw from our own point of view and observe ourselves as if we were someone else. Even when we take up the external point of view it is still ours. In the cases we are discussing, both deontological intuitions about the wrongness of murder or betrayal and consequentialist intuitions about the goodness of saving more lives rather than fewer continue to offer prima facie grounds for moral belief, and we have to decide whether what appears to the external view will justify us in disregarding some of them.
In arriving at such a judgment, we simply have to make use of moral intuitions – otherwise we could not draw any moral conclusions. So the question will be whether the external view – scientific, historical or sociological – of our moral responses convincingly weakens the authority of some more than others. Since the responses themselves participate in this contest, the outcome is not automatic. For the kinds of individual rights we are considering, I believe the outcome of the external challenge remains just as problematic as the original moral dilemma.
When psychological and neurophysiological data and speculative evolutionary explanations are brought to bear on these judgments, reactions typically diverge, in ways that I suspect depend heavily on the reflector’s antecedent predisposition for or against a purely consequentialist outlook. Some people just find utilitarianism the only rationally intelligible form of moral justification; others find it equally obvious that morality has another dimension. For the former, the fact that deontological judgments are arrived at quickly, without deliberation, and that they are associated with emotion, both phenomenologically and neurophysiologically, is a reason to discredit them as instances of moral knowledge, especially since they sometimes require us to do what is clearly irrational, namely choose the greater evil over the lesser. The hypothesis about their evolutionary origin provides an alternative explanation of their emotional force.
But for anti-consequentialists, the immediacy and emotional force of reactions against murder, torture, betrayal and so forth are not surprising, since they are responses to the immediate moral character of our interaction with another person, rather than to broader consequences. Of course your ventromedial prefrontal cortex will squirm when you think about murdering or torturing someone! Anti-consequentialists also point to another psychological fact: that impartial benevolence, the motive that is supposed to ground utilitarianism, is far too weak in most human beings to support obedience to its demanding moral requirements; which makes it unsuitable as the sole basis for human morality. As for the evolutionary speculations, they hardly count as independent empirical data: they are really driven by the consequentialist moral theory, and can be taken with a grain of salt.
In other words, nothing in this complex set of arguments and counterarguments forces one to abandon either the consequentialist or the deontological position. The main point, as Selim Berker has pointed out, is that disagreements over how to respond to information about the psychology and neurophysiology of moral judgment are themselves moral disagreements. It is certainly legitimate to introduce these findings into the process of reflective equilibrium, but in the end, it is we who have to decide whether they should undermine our confidence in the validity of the deontological judgments they are supposed to explain away. And in this decision, those intuitions themselves play a role. They don’t automatically withdraw from the scene in response to an MRI reading. It is not a question of whether the outside view will cause the challenged intuitions to become emotionally weaker. It is a question of whether it convinces us to regard them, whatever their emotional vividness, as mere appearances rather than recognitions of moral reality. (After all, optical illusions don’t go away as features of visual experience when we learn by measurement that they are illusions.) When we regard ourselves and our moral psychology from outside, we still have to decide whether the outside view requires us to withdraw our assent from what is presented as true from the inside. And the inside view is a participant in that choice.
I think it is clear from the history of this debate and the tenacity of the two sides that there are two quite different reflective equilibria to be found here, one of which preserves a significant deontological component in morality and one of which is significantly revisionary.
Isuggest that in light of this stand-off it is unrealistic for either side to think of itself as having refuted the other. Both are viable moral outlooks. So let us ask a slightly different question. Would a move to the revisionary position count as moral progress? That is how it is usually presented, by contrast with the conservative unwillingness to abandon a visceral attachment to basic individual rights, seen as a legacy from the past.
We know that human morality changes over time, and it seems unlikely that it will ever reach a final steady state, any more than science will. Instead of saying that some form of consequentialism is the only rational moral outlook, moral revisionists might be understood as offering us a step forward in this journey of moral development. One might think of it, mutatis mutandis, on the analogy of a later scientific theory not simply refuting, but subsuming and replacing an earlier one, in a way that preserves and explains many of its results while revising others. The difference is that the moral theories are not alternative descriptions of the external world, but normative alternatives.
Some examples of moral reform can only be seen as outright refutations of the outlooks they seek to overthrow. This is true of the contemporary revolution in views about homosexuality, which has overridden a very powerful taboo in the name of individual liberty and human happiness, and seems to be on the way to dissolving the emotionally freighted sense of pollution that has for so long sustained the taboo. On the other hand some proposals for moral progress are simply false, sometimes atrociously so. Both Nazism and Bolshevism were justified, after all, under the colour of moral progress. But there are other examples that pose the issue of progress in a different way, as an evaluative comparison between alternative conceptions of the same moral domain.
One example that can be seen in this way, and that is closely related to the issue before us, is that of property rights. The radical hope that private property rights could be abolished in favour of common ownership has turned out to be a destructive dream. But I find utterly convincing Hume’s account of property not as a basic moral right but as a convention sustained by its indispensable contribution to the collective interest of society. And so long as it provides security of possession, succession and exchange, permitting capital accumulation, economic planning and co-operation over the long term, the system of property rules and rights can also serve other ends, such as distributive justice. This instrumental, anti-Lockean conception of property rights has some currency, but it is resisted by a powerful libertarian strain in Western morality, which continues to have large political influence, especially in the US. So I think the conventionalist, largely consequentialist conception of property rights, making them a vehicle for social justice, is best presented as a call for moral progress – a call to subsume the prevailing conception in an expanded one. Reform in the morality of property would mean that property rights widely came to be seen as based not on individual liberty but on the collective good. Even though individual liberty is an important value, and should be protected in the exercise of those property rights that we hold under the collectively valuable property conventions, it should not be seen as the foundation of those rights and should not determine their content.
This is an example of how a moral outlook that has considerable intuitive support might be replaced by another that is deemed superior. In my view, even though a conception of property rights based largely on liberty and self-ownership embodies significant values, its replacement by a conventionalist conception that includes some protection of liberty would be a clear example of progress.
But would it be moral progress if we came to see all deontological boundaries not as fundamental to morality, but at best as rough guides worthy of respect only to the extent that their relatively strict protection actually serves the welfare of society or humanity as a whole? I think that depends on whether it would be progress to simplify the moral point of view.
In some sense the moral point of view requires putting oneself in everyone else’s shoes and taking the separate point of view of each individual into account in deciding what to do. The question is: how?
The appeal of the consequentialist way of valuing people impartially cannot be denied. It may seem that there is no way to take the points of view of all individuals into account without merging them into a single ocean of benefits and harms, which is then valued as a whole. Yet there is another way. In deontological morality each individual faces us separately, and the inviolability of the individual facing us dominates the rival claims of those we could help by sacrificing him to the general welfare. In a sense the inviolability of that individual stands for the inviolability of all. I think we feel this even when, in a typical dilemma, we put ourselves simultaneously in the place of several potential victims of these two different kinds.
This individualised moral respect is something that morality can guarantee to everyone equally: the same entitlement to be treated in certain ways, the same status, the same limits or boundaries. It determines the character of our relations with one another. Each person with whom I interact presents me with the same stubborn and impenetrable moral surface that I present to him: there are certain things that may not be done by either of us to the other. And this ‘moral minimum’ is an expression, in the structure of morality, of the demand that moral consideration should respect each individual separately. That is not the case if morality in principle permits the complete subordination of one person’s interests to the greater interests of others. As Frances Kamm has emphasised, such an inviolable status is denied to everyone by any consequentialist system, whatever outcomes it values.
The question I am asking is whether, looking at ourselves from outside, we should come to view our attachment to rights and deontology as an unnecessarily cluttered moral outlook, which grossly magnifies the claims of the person facing us, and limits our rationality. Would it be progress if we no longer took the individual-centred deontological outlook and the intuitions stemming from it as fundamental moral guides?
I believe something would be lost. No doubt human morality will evolve, and perhaps it will move in a more consequentialist direction. But the completely different way of valuing individuals, by treating each of them decently come what may, and demanding such treatment for ourselves, is a vital part of our lives. Most important, it is a distinctive way of thinking about how to relate to one another, the source of our constantly developing interpretations of people’s equality of status as the bearers of human rights. Without it the advantages of membership in the moral community would be seriously diminished – not quantitatively but qualitatively. Those who disagree will see me as simply begging the question – but perhaps that is unavoidable.