Shoulds and Shouldn’ts

A tyrant, imagine, spares an innocent man from torture, but solely in order to reap good publicity. He does what morality demands, but not for the reasons that demand it. T.M. Scanlon’s new book Moral Dimensions plays on the two kinds of moral evaluation this case illustrates: the permissibility of an act, and the meaning of the way one does it. Half the book uses this distinction negatively, to refute two widespread claims about why some acts are impermissible. The other half tackles blame, as a response to an action’s meaning.

Scanlon has been noted above all for a comprehensive moral thesis he calls contractualism, which he elaborated in his long book of a decade ago, What We Owe to Each Other. There he formulated a general account of moral impermissibility. An act is wrong, he said, if it ‘would be disallowed by any set of principles for the general regulation of behaviour that no one could reasonably reject as a basis for informed, unforced general agreement’. This mouthful, Scanlon argued, explains which sorts of act are wrong and why they are wrong, and he devoted much of the book to drawing out its implications. Readers disputed whether his formula had real content, or whether, rather, he was fudging to get whatever conclusions he wanted. His new, shortish book contrasts with that earlier effort. Here, Scanlon’s comprehensive account of what we owe to each other plays no overt role. He sets out not to cover a unified topic, but to rethink a few different issues. These do tie in with a theme – the permissibility and meaning of actions – but chiefly, I take it, he picks topics on which he has new thoughts. The first half of the book, on permissibility and meaning, amounts to masterful and insightful philosophical housekeeping. The second half is revolutionary in the ways it tells us to think about blame.

First, then, moral permissibility. Even in a war that needs to be fought, some things are morally out of the question. Terror bombing is one of them: it is never permissible to aim to kill non-combatants with the aim of sapping enemy morale, even if that promises to reduce the total amount of killing. Other acts with equally horrific consequences, though, may be morally permissible in some circumstances. Bombing to damage the enemy war apparatus – ‘strategic’ bombing – may sometimes be permissible, even though one foresees that it will kill many innocent bystanders. The difference between terror bombing and strategic bombing lies not, it seems, in the deaths, but in the aims. The strategic bomber aims to destroy factories and the like and to kill enemy combatants. The terror bomber aims to kill innocent non-combatants. The results may be the same but the aims differ, and that is what makes the difference.

Or so it has seemed to many. Catholic moralists, drawing on Aquinas, have long elaborated this diagnosis as the doctrine of ‘double effect’, and a wide range of moral philosophers have found it plausible. The diagnosis raises a puzzle, however. What can be permissible or not, we might think, is what one does, not one’s aims in doing it. Sparing a man from torture is morally permissible – indeed, mandatory – however malign one’s aims may be. The Oxford philosopher W.D. Ross argued in 1930 that motives never bear on whether an act is right. He proposed using the word ‘act’ for the thing done and ‘action’ for the doing of it from a motive. Acts are right or not, he said, whereas actions are what can be morally good or bad. How, then, when bombing kills non-combatants, do the commander’s motives make a difference as to whether the act is morally permissible or not?

Scanlon agrees with Ross. The doctrine of double effect, he argues, is the wrong explanation of our judgments of terror bombing. The right explanation is specific to war. The general moral prohibition on killing has exceptions, and ‘just war’ is one of them. This exception, though, allows killing only when it can be expected to bring military advantage (and to be sure, not always then). Double effect has been used to explain moral judgments in a variety of other domains, such as medical care, but Scanlon argues that explanations of right and wrong are specific to each domain.

Even if we agree with Scanlon, as I do, that aims can’t bear on permissibility in a fundamental way, we might worry about the explanations he gives. Why does the narrow permission to kill in war have the shape it does? Scanlon is suspicious of overarching systems in moral theory, but we should want more than a list of rules and exceptions. He might have tried appealing to the master rationale he gave in What We Owe to Each Other, but in Moral Dimensions he doesn’t try to argue from first principles. In the chapters on permissibility and meaning, the main points are negative: double effect doesn’t work as an account of permissibility; neither does the disrespect involved in treating a person as a means to an end. As for what more general rationale might explain the moral rules in particular domains such as war or medicine, that is a question for elsewhere.

In the course of arguing his negative theses, Scanlon draws on a rich variety of examples to explore what is involved in living with intimates as well as strangers. The basic concept he invokes, drawn from his earlier book, is that of a reason to do something. Suppose a man has wronged me and I can now get back at him; I might ask myself whether what he did to me is, in itself, a reason to make him suffer in turn. The word ‘reason’ has a number of senses, but Scanlon treats the one illustrated in this example – ‘the standard normative sense’ – as basic. He defines it as ‘a consideration that counts in favour’. Questions of what we ought to do or what we have reason to do are termed ‘normative’: how to understand normative claims is a central topic of debate in philosophy. Whether, say, a past wrong in itself counts in favour of making the malefactor suffer isn’t a question that can be settled by scientific method. Whether you or I count it in favour is a matter of non-normative, psychological fact, but whether it really counts in favour is not. Scanlon thinks that normative questions have right and wrong answers, and that we have some power to judge these matters correctly.

The long final chapter on blame proceeds in this spirit. P.F. Strawson, a half-century ago, in his classic article ‘Freedom and Resentment’, insisted that ‘interpersonal relationships as we normally understand them’ by their very nature expose us to resentment and other ‘reactive attitudes’. Morality, he thought, deals in indignation, which is an impersonal analogue of resentment. Bernard Williams, in this vein, pictured morality as a ‘blame system’ fixated on guilt and resentment; like Nietzsche, he rejected this fixation. Scanlon works to vindicate morality and blame, but unlike both Strawson and Williams, he denies that resentment is crucial to either. His account of blame differs from others, as he says, ‘in placing emphasis on the expectations, intentions and other attitudes that constitute these relationships rather than on moral emotions such as resentment and indignation’.

What, then, does blame consist in? Scanlon starts with a vignette. I learn that a friend of mine, at a party, hearing a cruel joke told at my expense, laughed heartily and did nothing to defend me or to show offence. He even joined in, adding barbs of his own. True, he was tipsy and caught up in conviviality, and he now regrets what he did. Still, his behaviour has impaired our friendship. I will perhaps no longer value his company as I did. Blaming him, Scanlon maintains, consists in such responses as this: finding our relationship impaired and holding attitudes towards him accordingly. Blame is thus a response to the meaning of an action, its meaning for our relationship. Resentment may be involved, but even if I just feel sad, I may still be blaming my friend. Blame, then, consists in more than the disinterested response of finding his behaviour blameworthy. You too may think what he did blameworthy, but it won’t merit blame from you unless it impairs a relationship that you have with him yourself.

As well as friendship, Scanlon speaks of relationships within families and between neighbours. The ‘moral relationship’ in particular is one in which we stand with anyone who is ‘capable of understanding and responding to reasons’. Special relationships like friendship presuppose adequate moral relations, but the standards of these relationships aren’t just moral requirements applied to a special case. A relationship consists in many things: attitudes and dispositions, including intentions and expectations, and what one is inclined to recognise and respond to as reasons for behaving in particular ways. It consists also in hopes, feelings and ways of caring about someone. A normative ideal, Scanlon tells us, sets the standards for a relationship, and it is impaired ‘when one party, while standing in the relevant relation to another person, holds attitudes toward that person that are ruled out by the standards of that relationship’.

With a relationship such as friendship, we could presumably begin with the non-normative facts. Among other things, ‘being a friend involves being disposed to certain feelings: to take pleasure in the friend’s company, to hope that things go well for the friend and to take pleasure in their going well when they do.’ These facts don’t include such normative matters as whether, for instance, one ought to keep a friend’s confidence when he is plotting to harm someone else, but they do, nonetheless, have normative implications. In light of them, we can judge that the parties have reasons to treat each other in particular ways, and they can be right or wrong in these judgments. The parties themselves might be mistaken as to what reasons their relationship gives them. An abused wife, for instance, might agree with her husband that marriage gives her full reason to be subservient whatever he does, but as observers we would be right to disagree. (In saying this, I voice a normative judgment, expecting readers to agree.)

If the wife should then change her mind and stand up to her husband, he will blame her – though unjustifiably. What does his blame consist in, according to Scanlon? The husband takes the facts of marriage to give any wife reason to submit to her husband’s authority. When she violates these standards, he reconceives the relationship. Not that he will have a full, explicit doctrine of what their marriage now demands, but something will change in his sense of those demands. He may think he no longer has reason to feel warmly and indulgently towards her, as he used to feel at times when he wasn’t drunk and enraged. If he does, this is blame – unjustified though it is.

Scanlon works to explain why in many other cases blame would be justified. The wife, for instance, would always have had reason to blame her husband when he beat her, whether or not she actually did blame him. His actions violated the true normative ideal of marriage, and she had reason to adjust her feelings and ways of dealing with him accordingly. Like Strawson, Scanlon stresses the personal relationships that enrich our lives, and shows that they necessarily expose us to blame if we breach their standards.

A strength of Scanlon’s account is that it doesn’t make blame depend on an untenably sharp metaphysical distinction between what we are free to do and what we’re not. In particular, it doesn’t require that the ultimate causes of our action be indeterminate. I agree with Scanlon, as well as with Nietzsche and Williams, that such a pure metaphysical freedom is unintelligible, and isn’t required to ground our account of blame and morality. But is what Scanlon describes blame, as we know it and need it? In particular, can we elucidate blame without, like Strawson, basing our account on such feelings as resentment? I suspect that we can’t.

Scanlon’s account hinges on the notion of standards. One thing he says about friendships is that they can end blamelessly, as when friends just drift apart. There is no blame here: a friend who drifts apart from me hasn’t thereby violated the standards that govern our friendship. How are we to understand this? There are standards specifying how friends should treat each other and feel about each other, but what is the nature of this ‘should’? Is how friends ‘should’ treat each other simply the way they have most reason so to do? I may realise that drifting away was what my old friend had most reason to do, and if so, then, although I must reconsider my feelings towards him and how to act towards him, I don’t think of him as having violated the standards of friendship. Hence this isn’t blame. Standards, on this interpretation, would be what friends have reason to do because they are friends. They specify what it means to act wisely in light of being a friend.

Standards in the ordinary sense, though, are something more. They are what a person is held to – and we don’t hold each other to everything we have most reason so to do. Being your friend might give me most reason to doodle sketches of you fondly one day when I’m by myself, but this isn’t something to hold me to as a friend. What is special, then, about the standards I am held to? Scanlon’s examples and language suggest that by ‘standards’ he means something stronger than simply what the friendship gives me reason to do. If you tell me that my actions have breached the standards of friendship, you aren’t just criticising what I did as unwise in light of our friendship. You are complaining.

The standards of friendship are not merely counsels of wisdom, and we need to ask what makes the difference. Violating these standards seems a betrayal, at least in some small way. But what could that mean? A complaint of betrayal has a special emotional tone. A good friend misses my birthday, let’s suppose, to go to a concert. Even if I am convinced that he had insufficient reason to do what he did – the concert didn’t hold out much promise anyway, and he would have enjoyed himself more at my party – am I justified in thinking he has breached the standards of friendship? I might be in no doubt as to the non-normative facts of our relationship, and still ask myself this question. What I am then wondering, perhaps, is how to feel about what he did. If this is right, it suggests an answer to the question of what standards are in a relationship: they might be a matter of what sorts of conduct would warrant feelings of betrayal.

This brings us back to Strawson. It might not be possible to characterise the standards of a relationship like friendship, in the sense that Scanlon needs for his account, without appealing to the way one has reason to feel if a friend doesn’t meet those standards. Perhaps resentment isn’t quite the right term for the feelings that flavour a complaint of betrayal, but we may need some such emotional component. Scanlon renounces such punitive feelings in his account, but it may be these feelings that distinguish standards from counsels of wisdom in a relationship. Standards aren’t just a matter of how friends have reason to treat each other, but how they have reason to feel when the standards are breached.

Whether what Scanlon calls blame is blame as we know it is a significant question, but if it isn’t, we face a still greater one. Is what is missing in the account something we don’t need and would be better off without? Such punitive feelings as resentment are the root of much ill, and clearly we don’t need them when relationships are at their best. Nevertheless, Strawson thought, if we weren’t at least susceptible to such attitudes, we would lose much that was worth treasuring in our lives together. Whether he was right remains a vital issue in human affairs, from the most intimate to the most public.