This review was commissioned originally by the Financial Times. But they did not like it, and it was not published in this form. However a version more suitable for the general reader (and perhaps slightly less irritated in tone) was published eventually, and I have added this for comparison, at the end. Only connoisseurs of rhetoric need bother reading both versions. For more reservations about Parfit’s treatment of expressivism, see my paper ‘All Souls Night’ on this website under recent papers.

Derek Parfit On What Matters, Vols I & II. Oxford, Oxford University Press, 1,440 pages, £30.00.

Together weighing two kilos or just under five pounds, these two colossal volumes represent many years of work by one of the most influential moral philosophers of our time. They have been long in the making, and the subject of much pre-publication discussion on the web and elsewhere. Indeed, the first volume starts with an extensive introduction by one colleague, while the second volume includes four quite lengthy essays by commentators on other parts of the work, as well as replies by Parfit to those commentaries. And in the acknowledgements Parfit lists, by my count, some 260 other philosophers whom he claims to have helped him. Oxford University Press is especially to be congratulated for being able to price the result at just over 2p a page, an astonishing figure when the norm for academic books is probably nearer 20p a page. Other Oxford authors must hope that no element of cross-subsidy has taken place.

So is this, as Peter Singer hailed in the TLS, the most significant contribution to moral philosophy since 1874, when Henry Sidgwick sculpted his own great tombstone, The Methods of Ethics? Or is it a long voyage down a stagnant backwater? Like most work on moral philosophy, it is divided between two distinct areas. There are theories within ethics, telling us what our values should be, or what are the contours of our rights and duties. This is known as first-order moral philosophy. Then there are second-order theories, telling us something about the status of first-order pronouncements. In this area, often called metaethics, notions such as objectivity, knowledge, truth, proof, and reason are used to debate the nature of first-order claims. If I pronounce, for example, that vanity is a sin, could my remark count as objective and perhaps true, or even known to be true, by the light of reason? This is Parfit’s view, rationalism. Or am I more in the business of expressing an attitude, or encouraging a sentiment of disapprobation of vanity, voicing a stance rather than describing a fact?

This alternative, (‘Humeanism’) is the view held by philosophers from Augustine (‘in the pull of the will and of love appears the worth of everything to be sought or avoided, to be thought of greater or less value’) to Hume, Adam Smith, and Wittgenstein. It is also the view implicit in all the fascinating work on actual decision-making that has exploded in recent years, with writers such as Antonio Damasio, Jonathan Haidt, Jesse Prinz, Joshua Green, Pat Churchland, and others showing, in illuminating detail, how we actually work. But Parfit wishes to uproot and stamp out Humeanism.

‘We are the animals that can both understand and respond to reasons’, he says the first sentence of the book, launching his attempt to demonstrate that Humeans cannot do justice to this fact. The stick he uses to beat what he condescendingly calls ‘these people’ is that reasons are ‘object-given’, that is, they exist in virtue of the properties of the things said to give reasons. So far so good: that there is a bull in a field might be a nice solid fact, and one that gives some of us a reason to stick close to the perimeter. But now, the argument continues, the reason, being object-given, would exist whether we are aware of it or not, or whether we respond to it or not. This contrasts with what he thinks is said by ‘subject-given’ views in which reasons exist only in the light of our desires. And as the work unfolds the objectivity and independence of reasons from mere human desires and preferences are ever more firmly asserted, with Humeans and others banished from any kind of commerce with these ‘object-given’ reasons, Parfit’s own private hunting preserve. Thus after four hundred pages, Parfit roundly forbids Humeans even from saying that when we are forming desires or plans, our standpoint would be improved if we knew more of the relevant facts about the environment, such as it’s not being a bull but a cow. Silly old Hume.

In fact, Humeans must say that there are no reasons for anything—nothing matters. They are rank nihilists! Nicely illustrating how to combine poverty of imagination with vulgarity of tone, one of the commentators included here, Allen Wood, describes them as ‘either radically defective specimens of humanity who are incapable of feeling respect for anyone or anything, or else every time they do feel it they commit themselves to contradicting their own metaethical theories’. Golly.

Vice Chancellors bent on finding excuses to close philosophy departments must be rubbing their hands if not one of Parfit’s 260 helpers smelled a rat in all this. Philosophers do say funny things, but none that I can call to mind has ever denied that we respond to facts about objects, such as the bull in the field, when we decide what to do. Nor have they doubted that if we get those facts wrong, our decisions and desires are likely to be worse. What Humeans have said is that to take the bull’s presence as a reason for sticking to the edge of the field is indeed to go beyond merely perceiving it, that it will require a particular profile of concerns, fears, and desires, and that this profile is not simply given by anything like our capacity for such things as mathematics and logic. This is the point of Hume’s famously provocative remark that reason is and ought only to be the slave of the passions. What has gone wrong is that Parfit’s strategy of erecting a binary opposition between ‘object-given’ and ‘subject-given’ theories is completely ludicrous. Any sensible theorist has both elements, working in harmony. So the Titanic hits its iceberg before leaving port, although, if one may abuse the metaphor, it hits plenty more before the end of the voyage.

All that Hume holds is that our passions are part of whatever mental state is revealed by our taking something as mattering to us. Far from implying that there are no reasons for anything or that nothing matters this is the only plausible account of why we think that there are reasons for things, and find that things do matter. Hume never bars himself from using the word ‘reasonable’ as a term of praise, and indeed peppers all his works with it, talking happily of reasonable precautions, demands, policies, traits and feelings.

It’s not as if Parfit has an alternative account that is much help. In his view reasons play roughly the role of Plato’s Form of the Good, and are in many ways just as elusive. Reasons are not part of the natural, causally interlocking world. We do not perceive them or respond to them in anything like the way we gain sensory information about our physical environment. Parfit compares our knowledge of them to our knowledge of mathematics, forgetting Frege’s insight that numerals start life as adjectives describing the empirical magnitudes of collections, and forgetting as well that it is quite easy to describe why we might be interested in those magnitudes. But in Parfit’s account, reasons are kept within a tight circle of evaluative terms (good, right, obligation), linked up in eternal verities whose intelligible connection with anything outside the circle, such as actual human decision-making, has to be left utterly mysterious. Parfit frequently presents himself as having an ‘account’ of ethical truth, but since the account simply consists of restating value judgments in terms drawn from the tight little circle, it is not an account, and it is not unavailable to Humeans.

When he turns from this shipwreck to first-order ethics, Parfit’s aim is to find a reconciliation between two philosophies that are often opposed: utilitarianism and Kantianism. This has also been the aim of many other philosophers, notably J. S. Mill, and R. M. Hare. From the Kantian tradition Parfit draws the idea of principles that could be universally willed. From distinguished modern followers of Kant, such as Rawls, and especially Thomas Scanlon, he draws the idea of principles that nobody could reasonably reject. Such abstract formulae need a great deal of filling out, so from the utilitarian tradition he draws the idea of principle whose universal acceptance would make things go best. Putting all these together we get that ‘an act is wrong just when such acts are disallowed by some principle that is optimific, uniquely universally willable, and not reasonably rejectable’. There are scholarly questions, some of which are pursued by the four commentators, of the extent to which this formula is a true offspring of either Kant or utilitarians, such as Mill. But the fidelity to both traditions, or to previous reconciling attempts, is not Parfit’s prime interest. Instead, with relentless, indeed obsessive, concentration he steers his principle through such urgent questions as whether we ought to send a lifeboat that can only make one trip to a rock where it can pick up five people rather than to a rock where there is only one, or whether a fat man might reasonably object to being pushed off a bridge to stop a trolley hurtling towards five others.

The image, then, is of a unique principle from which we can deduce which actions are wrong, thereby revealing the one true morality. A strange aspect of this approach is that it is entirely modeled on the judicial problem of coming to a verdict: was this something that it was permissible to do? But legal verdicts matter: they have consequences attached to them. Yet Parfit has no explanation why the moral verdict, and the scholastic apparatus necessary to deduce it, similarly matter to anyone. Suppose someone says, ‘OK, I did wrong. So what?’ Set all the forces that move people to zero, and people do not move. Moral emotions, such as pride, guilt and shame must be recruited to add some motivational pushes, but then we are back in the world of Hume and Smith, and the rationalism supposed to get us there has been nothing but a mirage, a fifth wheel.

The classical traditions in moral philosophy, and the great philosophers who followed it, see the subject very differently. In the Aristotelian and Ciceronian view, what matters is the character of the agent, and the virtues that make it up. Yet ‘character’ is a word that does not appear in the index to either of these volumes, presumably because it has no more to do with the rationalist aim of proving theorems about eternal reasons than do emotion or desire. Indeed, it is contestable whether a good character should need to make room for much of a notion of a ‘principle’ at all, let alone a deduction of judicial verdicts from such things. A well-tuned sense of shame or necessity, and with it a well-tuned sensitivity to the needs of others, go a long way before any principles loom into view. A sense of what will do and what will not, exercised on individual real, messy, human cases, and refined through education, experience, imagination, and sympathy, might never result in any urge to write everything down into a complete code. Any principles that might in some way summarize or assist the work of practical reasoning are likely to be provisional, liable to exceptions and qualifications without end, and to require interpretation and tact in their application.

Parfit is of a different temperament. ‘It would be a tragedy’ he tells us on page 2, ‘if there is no single true morality’. Well, outside the charmed walls of All Souls College, there actually are tragedies. Often the messy pluralities of conflicting moral demands—one might have said, the conflicting demands on human life itself—are part of the cause. Inside the charmed walls I fear that the tragedy is more like that of Ajax slaying sheep, or perhaps it is the comedy of Don Quixote tilting at windmills.

The version actually published was this:

Derek Parfit On What Matters, Vols I & II. Oxford, Oxford University Press, 1,440 pages, £30.00.

Together weighing two kilos, or just under five pounds, these two colossal volumes represent many years of work by one of the most influential moral philosophers of our time. They have been long in the making: it is over twenty five years since Parfit impressed the world with Reasons and Persons, a work that made his reputation, and firmly ensconced him behind the walls of All Souls. The intervening years have seen very little more. But Parfit has not entirely secluded himself: in the acknowledgements he lists, by my count, some 260 other philosophers whom he claims to have helped him. Oxford University Press is especially to be congratulated for being able to price the result at just over 2p a page, an astonishing figure when the norm for academic books is probably nearer 20p a page.

Like most work on moral philosophy, Parfit’s book it is divided between two distinct areas. There are theories within ethics, telling us what our values should be, or what are the contours of our rights and duties. These are theories in what is known as first-order moral philosophy. Its aim has often been to reduce the teeming plurality of rights and duties, obligations and benefits to some kind of order. At the limit there might be either a small number of principles, or even one unique principle, from which everything else could be derived. Hence we find suggestions such as the Golden Rule, John Stuart Mill’s principle of Maximizing Utility, or Kant’s Categorical Imperative. But we also find writers like Isaiah Berlin, or Bernard Williams, who mistrust all this tidiness, and insist instead on the irreducible plurality of virtues, or the inevitability of insoluble dilemmas as different obligations conflict and jar against each other. Classical tragedy is especially concerned with such conflicts and their insoluble nature.

The other branch of the subject consists of second-order theories, telling us something about the status of first-order pronouncements. In this area, often called meta-ethics, notions such as objectivity, knowledge, truth, proof, and reason are used to debate the nature of first-order claims. If I pronounce, for example, that vanity is a sin, could my remark count as objective and perhaps true, or even known to be true, by the light of reason? This is Parfit’s view, rationalism. Or am I more in the business of expressing an attitude, or encouraging a sentiment of disapprobation of vanity, voicing a stance rather than describing a fact?

This is the view held by philosophers from Augustine (‘in the pull of the will and of love appears the worth of everything to be sought or avoided, to be thought of greater or less value’) to Hobbes, Hume, and Adam Smith. It is also the view implicit in a huge amount of fascinating work on actual decision-making that draws on cognitive science, neurophysiology, evolutionary psychology and other disciplines. This has exploded in recent years, with writers such as Antonio Damasio, Jonathan Haidt, Jesse Prinz, Joshua Green, Pat Churchland, and others showing, in illuminating detail, how we actually work.

The battle between rationalism and its more emotional competitor may sound to be only of academic interest. But it spills into the real world. Rationalism more readily consorts with absolutism, with untroubled conviction that our own moral views are uniquely correct, so that other cultures that do not share them are defective, sunk in unreason, irrational, and perhaps after all best governed by us. It is the view of a mandarin class, and in international politics, a colonial or imperial view. Its particular horror is the ‘relativism’ that it associates with the alternative: the idea that our morality is just the haphazard creature of our particular culture, upbringing, emotional makeup, or prejudice.

Out in the world of politics, perhaps the most dangerous people of all are those self-righteously confident that reason alone determines their courses, but whose actual motivations are made up from the turbulent stew of their own emotional natures. Tyrants and democratic politicians alike claim the mantle of reason, when in actuality ambition, narcissism, vanity and lack of imagination propel their courses. One can think of cases in recent British history.

Parfit is an unashamed rationalist. ‘We are the animals that can both understand and respond to reasons’, he says the first sentence of the book, launching his attempt to demonstrate that those who side with Hobbes or Hume cannot do justice to this fact. The stick he uses to beat what he condescendingly calls ‘these people’ is that reasons are ‘object-given’, that is, they exist in virtue of the properties of the things said to give reasons. So far so good: that there is a bull in a field might be a nice solid fact, and one that gives some of us a reason to detour around it. But now, the argument continues, the reason, being object-given, would exist whether we are aware of it or not, or whether we respond to it or not. Parfit contrast this with ‘subject-given’ views in which reasons exist only in the light of our desires, as Augustine said. As the work unfolds the objectivity and independence of reasons from mere human desires and preferences are ever more firmly asserted, with Hume and others banished from any kind of commerce with these ‘object-given’ reasons, Parfit’s own private hunting preserve.

It’s a very idiosyncratic way of drawing the battle lines, so much so that Vice-Chancellors bent on finding excuses to close philosophy departments must be rubbing their hands if not one of Parfit’s 260 helpers smelled a rat in it. Philosophers do say funny things, but none that I can call to mind has ever denied that we respond to facts about objects, such as the bull in the field, when we decide what to do. Nor have they doubted that if we get those facts wrong, our decisions and desires are likely to be worse. If it wasn’t a bull but a cow, the arduous detour was unnecessary. What Hume and others have said is that to take the bull’s presence as a reason for going around the field is indeed to go beyond merely perceiving it, that doing so will depend upon some profile of fear and desire, and that this profile is not simply given by anything like our capacity for such things as mathematics and logic. This is the point of Hume’s famously provocative remark that reason is and ought only to be the slave of the passions.

Parfit does notice that Hume happily uses the word ‘reasonable’ as a term of praise, and indeed peppers all his works with it, talking happily of reasonable precautions, demands, policies, traits and feelings. He uncharitably supposes that Hume constantly forgot his own theory. But there is no forgetfulness and no inconsistency. The section of his Treatise in which the remark comes is called ‘Of the Influencing Motives of the Will’. It is about the explanation of choice and action, not about praise or blame. When he turns to those Hume can indeed happily go on to commend all kinds of things as reasonable or to criticize them as unreasonable. A person who fills with rage when overtaken on the motorway is unreasonable. But it is his passions and temperament that are at fault, not his awareness of the road nor his capacity for logic.

When he turns from this disastrous engagement with the Humean tradition to first-order ethics, Parfit is on less shaky ground. His aim is to find a reconciliation between two philosophies that are often opposed: one that talks of costs and benefits, utilitarianism, and one that talks of rigid duties and principles, Kantianism. Such reconciliation has also been the aim of many other philosophers, notably John Stuart Mill, and R. M. Hare. From the Kantian tradition Parfit draws the idea of principles that could be universally willed. From distinguished modern followers of Kant, such as Rawls, and especially Thomas Scanlon, he draws the idea of principles that nobody could reasonably reject. As Hegel noticed shortly after Kant wrote, such abstract formulae need a great deal of filling out, so from the utilitarian tradition Parfit draws the idea of principle whose universal acceptance would make things go best. Putting all these together we get that ‘an act is wrong just when such acts are disallowed by some principle that is optimific, uniquely universally willable, and not reasonably rejectable’.

The image, then, is of a unique principle from which we can deduce which actions are wrong, thereby revealing the one true morality. Parfit is assiduous, obsessive even, in pitting his principle against an exhausting variety of thought experiments designed to test which answer it is right to give in various circumstances. Yet a strange aspect of this approach is that it is entirely modeled on the judicial problem of coming to a verdict: was this something that it was permissible to do? But legal verdicts matter: they have consequences attached to them. Yet these volumes offer no explanation why the moral verdict, and the scholastic apparatus necessary to deduce it, similarly matter to anyone. Suppose someone says, ‘OK, you can deduce from your principle that I did wrong. So what?’ Moral emotions, such as a sense of honour, self-respect, pride, guilt, or shame must be recruited to add some motivational pushes, but then we are back in the world of Hume and Smith, and the rationalism supposed to go beyond them has been nothing but a mirage, a fifth wheel.

The classical traditions in moral philosophy, and the great philosophers who followed it, see the subject very differently. In the Aristotelian and Ciceronian view, what matters is the character of the agent, and the virtues that make it up. Yet ‘character’ is a word that does not appear in the index to either of these volumes, presumably because it has no more to do with the rationalist aim of proving theorems about eternal reasons than do emotion or desire.

It is actually contestable to what extent a virtuous character will be structured by hard-and-fast principles. A well-tuned sense of shame or necessity, and with it a well-tuned sensitivity to the needs of others, go a long way before any such principles loom into view. A sense of what will do and what will not, exercised on individual real, messy, human cases, and refined through education, experience, imagination, and sympathy, might never result in any urge to codify everything. Any principles that might in some way summarize or assist the work of practical reasoning are likely to be provisional, liable to exceptions and qualifications without end, and to require interpretation, judgment and tact in their application.

Parfit is of a different temperament. ‘It would be a tragedy’ he tells us on page 2, ‘if there is no single true morality’. Well, as tragedies go, this one seems quite supportable. Often the messy pluralities of conflicting moral demands—one might have said, the conflicting demands on human life itself—are part of the cause. But none of that implies that ‘anything goes’. Human life imposes demands on all of us. When people fall short, it may be our contingent, culturally formed natures that make us feel aversion to them, but the aversion is real enough. And we endlessly discuss and modify and rearrange the pieces on the moral board: in our own time attitudes to homosexuality, equality, race, gender, and childhood have all changed for the better. Others, such as our attitudes to greed and wealth may have gone into a trough, but perhaps there are welcome signs of recovery.