The Majesty of Reason
In this paper I contemplate two phenomena that have impressed theorists concerned with the domain of reasons and of what is now called ‘normativity’. One is the much-discussed ‘externality’ of reasons. According to this, reasons are just there, anyway. They exist whether or not agents take any notice of them. They do not only exist in the light of contingent desires or mere inclinations. They are ‘external’ not ‘internal’. They bear on us, even when through ignorance or wickedness we take no notice of them. They thus very conspicuously shine the lights of objectivity, and independence, and even necessity. By basking in this light, ethics is rescued from the slough of sentiment and preference, and regains the dignity denied to it by theorists such as Hobbes or Hume, Williams, Gibbard or myself. Hence, many contemporary philosophers compete to stress and to extol the external nature of reasons, their shining objectivity (Broome 2004; Dancy 2000; Nagel 1970; Parfit 1997; Raz 1975, 1978, 2003; Schafer-Landau 2003; Wallace 1999).
The other phenomenon is that of the inescapable ‘normativity’ of means-ends reasoning. Here the irrationality of intending an end but failing to intend the means is a different shining beacon. It is that of pure practical reason in operation: an indisputable norm, again showing a sublime indifference to whatever weaknesses people actually have, and ideally fitted to provide a Trojan horse for inserting rationality into practical life. If the means-end principle is both unmistakably practical and yet the darling child of rationality itself, then other principles of consistency or of humanity, or of universalizing the maxims of our action, can perhaps follow through the breach in the Humean citadel that it has spearheaded. And so we get the dazzling prospect that if people who choose badly are choosing against reason, then this can be seen to be a special and grave defect. It would locate the kind of fault they are indulging. It would give us, the people of reason, a special lever with which to dislodge their vices. Being able to corral knaves and villains in a compound reserved for those who trespass against reason and rationality therefore represents definite progress.
It is sad to have to spoil the party, but I fear that these apotheoses of reason contain much less than meets the eye. Ethics is given no new light, nor is its armoury in the least strengthened, nor is its status beyond anything dreamed of by Hume remotely established, by these contemporary ethusiasms. In fact, the massive amount of work that has gone into the coronation of reason has been almost entirely misdirected.
II. Moving the Mind
Clearly we should not start by being deluded by the noun, thinking of Reason as a kind of magical faculty or structure. We should start with the relation. Reasons are reasons for something: the primary datum is relational. The field of the relation is less clear, or rather, more diffuse. Propositions are reasons for propositions, facts are reasons for intentions and desires, some intentions are reasons for others. Actions have reasons and one action may be another person’s reason for a different action. But corresponding to each of these and other relations there is a potential movement of the mind, a movement guided by the first mental state, and issuing in the second, when the reason is accepted or operative. So when we talk in the abstract of one proposition being a reason for another, or a fact being a reason for a norm or decision, the field is one of abstract representations corresponding to potential movements of a mind so guided. The movement in question might be one from one cognitive state or a true belief state to another: this is when we talk of theoretical reason. Or it might be one from the apprehension of a fact about a situation to an action or a desire or the formation of a motive or intention: this is when we talk of practical reason. We can also include movements that have action itself as its terminus, if we wish. A movement might also be that from a plan or an intention, or the supposition that a policy is settled, coupled with belief about the means that are open, to the postulation or adoption of a strategy for realizing the intention. A particular movement of the mind might therefore consist in Sally noticing that there is a mouse under the chair and inferring that the cat is somewhere around, or noticing that her scratching the blackboard is distressing Molly and then, guided by this thought, supposing she should stop.
The notion of guidance is intended to suggest the difference between taking one thing as a reason for another, and being subject to some kind of free association in which one thing leads to another, but not by a process of reasoning. It is the difference between thinking that a restaurant is expensive, and for that reason going elsewhere, and finding (either consciously or unconsciously) that the restaurant reminds you of an evening long ago with your mother-in-law, and finding yourself going elsewhere. Since the relationship between reasoning and causes of behaviour is puzzling, it may be hard to say in what the difference consists, and there will certainly be cases that are neither clearly one nor the other. Psychologists in Newcastle found that in weeks when a picture of eyes looking at subjects added as a kind of banner headline to a poster indicating suggested prices for coffee, in a communal coffee room, the amount stumped up in the honesty box went up by nearly three times compared with weeks when a neutral image of flowers was substituted (Bateson, Nettle & Roberts 2006). Is this a case of free association between a cue suggestive of being watched, leading to an unconscious fear of exposure, or is it a case of unconscious reasoning, triggered by that same cue, from the possibility of being watched to the same fear? For the purposes of this paper we can choose either way. I shall mostly be concerned with conscious sensitivity to the starting point and to its tendency to steer the movement of mind in question, rather than with subliminal or subdoxastic forces which may or may not be counted as giving reasons.
By an abstract representation I mean simply that we can leave out mention of actual agents and their actual states of mind, and contemplate the guidance purely in the abstract, considered as a relation between truths, or one between truths and possible intentions or desires. We can say, for instance, that the fact that an action is distressing people is a reason for desisting, or the fact that there is a dead mouse under the chair is a reason for inferring that there is a cat around. It is often difficult to frame such abstract relations without a ceteris paribus clause, since the particular circumstances of particular cases may nullify the reason. If you keep a pet mink a dead mouse may not be a reason for inferring the presence of a cat, and if someone has no business being distressed because what you are doing is harmless, there may perhaps be no reason to desist. Nevertheless the abstract generalization may be a useful general guide, even when it is liable to exceptions.
Some philosophers insist that the actual and potential movements of the mind must start from genuine cognitions, or even facts. I shall mostly defer to this usage, in which a false belief or a misapprehension does not provide a reason for anything. We might say that it does not provide a real reason for anything, although unfortunately some people take it to do so. I do not entirely like the stipulation, since it forces us to say that people who through little or no fault of their own misapprehended the facts and inferred or acted accordingly, had no good reason, or no real reason for what they did. And that sounds harsh, for they may not have been at all irrational, after all. They certainly had their reasons for what they did, and they may have acted well in the light of them. The general who is misinformed by a normally reliable source about the disposition of enemy troops, but who then plans well accordingly, is only unfairly accused of having had no reason for what he did, unlike the one who has proper information and then thoroughly botches his plans. The second might be court-martialled for acting irrationally or for no reason, but surely not the first. Similarly the victim of hallucination taking himself to perceive a rat in the drawer acts reasonably enough in then shutting it, although his movement of mind does not start with apprehension of a fact. However, nothing important hinges on this stipulation in what follows and we could in these cases follow the course of saying that there were after all ‘factive’ reasons in play: not the fact of the enemy troops being thus-and-so, nor the fact of the rat’s proximity, for these were not facts, but the fact that the informant reported as he did, or the fact that it looked as if there were a rat there. Similarly in the Newcastle case we cannot say that the subjects reasoned from the fact of being watched, but we might choose to say that they reasoned from the fact that there is a possibility of being watched. As an aside it is interesting that an exaggerated version of the thought that only facts make reasons plays an important role in Spinoza, for whom reason had to avoid inputs from experience (perceptions), since the senses provided confused or misinterpreted ideas, and also since we are ‘passive’ in the face of them, and this sullies the free exercise of reason (Spinoza 2001, 3.1.3; Bennett 1984, 324—8). But we can confine our field to genuine apprehendings, true representations, without any such crippling restriction.
Of course aims and intentions as well as apprehensions provide reasons, and introduce another need for care in our scorekeeping. Sally’s reason for scratching on the blackboard may be to annoy Molly. If we say, as no doubt we should, that this was a bad reason, what we say is unfortunately ambiguous between negatively evaluating Sally’s intention, and negatively evaluating the means she adopted to realize it. We might want to do this if, for instance, we approve of Sally’s mischief, but Molly was unfortunately out of earshot. It is important to distinguish these since they impute quite different faults to Sally. It would matter, for instance, if we are wondering whether to employ Sally to annoy Molly in the future.
When we say that the field of reason is that of movements of the mind, we must be include failures to move as the kind of thing which excite verdicts of reasonable or the reverse. Gordon may be unreasonable in ignoring Jack’s interventions, or failing to pick up Molly’s signs of distress. This is just an instance of the way in which more generally we criticize failures to act as well as positive actions.
So what are we saying about the actual or potential movement of the mind? We say that p is a reason for q or that the fact of x is a reason for doing y when we think it is good to infer q from p, or to be moved towards doing y upon apprehending x. By invoking the relation we commend or endorse the kind of guidance of the mind that it indicates. Molly’s distress is a therefore a reason for Sally to stop scratching the blackboard. Movement from apprehension of that distress to her stopping would be a good movement of Sally’s mind. It would be good even if in fact Sally does not know about Molly’s distress, or does not care a jot about it, or is actively enjoying it. It would be good even if Sally could not implement it, perhaps because of some kind of ingrained insensitivity or some equally ingrained and immovable determination to ignore or humiliate Molly. The reason for Sally to stop is just there anyway—it is Molly’s distress—and Sally, or a slightly improved version of Sally can apprehend it by normal perception. But this does not imply that she needs nothing more to apprehend it as a reason for stopping. She may or may not be guided by it. If she is, we say she is being reasonable, by way of commending her.
III. A Blind Alley
An agent’s blindness or malformation may prevent a good reason for a movement from being his reason for doing anything. When we describe his reason, we are simply producing a fact about him and the explanation of his states—the way his mind was guided. It has nothing to do with how it would have been good for him to be guided. When agents not only are not moved but cannot be moved in the right direction, they are still liable to criticism, and this is the sense, and the only sense, in which reasons are external. This in turn means that there is absolutely no need to follow Bernard Williams’s regrettable move of making the contingent profiles of actual concern of an agent determine what is to be said in the context of evaluation. This is so even if, like Williams, we expand the domain of an agent’s actual concerns to include an idealized set of concerns, the ones to which they could deliberate in ways they themselves approve. Thus when Williams considers an agent who is a confirmed wife-beater and who has not got sufficient internal resources to deliberate to a better way of being, he finds it difficult to judge that there is good reason for him to stop (Williams 1995, p. 191). I say instead that there is no difficulty here. There is indeed good reason—excellent reason—for him to stop. He sees no reason to stop, and perhaps his mind is too corrupt or impoverished ever to be guided in that way, or even for him to comprehend improvement in this respect. Nevertheless, it would be better if he did. Some may be optimistic enough to suppose that all human beings have enough resources within them to come to adopt, as their reason for acting, anything which actually is a reason for acting. It is a nice, pious, hope, but our language and our thoughts are far from presupposing that at the outset. There need be no optimistic assumption that any agent can be moved by any reason.
With this understood, this whole debate between “internalists” and “externalists” in the theory of reason collapses. Externalists were right that reasons are just there anyway, for the starting points of guidances of the mind are there anyway—i.e. regardless of whether particular people notice them, or could bring themselves to move in good directions because of them. But this is an entirely hollow victory, for internalists remain right that it is only in the light of the contingent ways we are that we can instance movements, and just as obviously it is only in the light of the contingent ways we are that we commend and endorse them. So the phenomenon is of no interest to the debate between Humeans in the theory of motivation and value, and others.
IV. Kinds of Guidance
We should notice that it is the kind of guidance that we are commending, not its endpoint nor its consequences. In principle a good journey can take you to a bad place, or a bad journey to a good place, but the movements are good or bad for what they are rather than where they take you. It may be good that Henry returned to the house (because he found that the cat was locked in the bathroom) even if he had no good reason for doing so, but he decided to do so because of some neurosis, or some nefarious design. There was in fact a good movement of the mind that could have been made—from the thought that the cat might be locked in, to a decision to return to the house—but it was not Henry’s. Conversely it may be a pity that Sally came to believe that the cat was around, because her project was to decapitate it, although her reason for believing it was the perfectly good one that there was a dead mouse under the table. It may be good that Cedric brought Sally flowers on her birthday, even if his reason for doing so was the bad one that it would exacerbate her allergies. You can move in a bad way but get to truths, and to doing the right thing or the fortunate thing. And conversely, you can move well, but be moving to falsehoods, and to doing the wrong thing or the unfortunate thing, although we should accept that there are some destinations so bad that nothing could count as a good journey ending up with them: an intention to commit genocide, for instance.
A distinction that we need not dwell upon holds between movements that are in some sense deliberate, those with which the agent himself is comfortable, or which he endorses, and movements that the agent either does not consciously know about, or might wish away. This is close to Gibbard’s distinction between accepting a norm and being in the grip of one (Gibbard 1990). Thus someone in the grip of a fetish or a compulsion might be said not to have a reason for doing what he does, but only find himself caused to do it, as if by some outside force. But he could equally be said to have had a reason for acting as he did. Plato’s Leontius, who had a shameful thing about recently executed corpses, found himself sufficiently gripped by the consideration that there were corpses to be seen to go and see them (Plato 2008, IV, 439e). He may have felt as if he were doing so ‘almost’ against his own will. But we can properly say that he had his reason for going. What we will not say is that the fact that there are recently executed corpses somewhere is actually a reason for going and looking, unless we wish to commend the process, for instance to medical students.
In the case of theoretical reason, our sense of how truths relate to each other give us our standards for good or bad movements. Of course, it is not entirely easy to describe the relations behind these standards. But we know the general pattern. The premise, p makes q more probable, or q provides the best explanation of p or the simplest or only plausible explanation of p. The gold standard, of course, is that p could not be true without q being true, but few movements of the mind are guided by relations that meet the gold standard, except in logic lecture rooms and mathematics classes. In most cases we have to settle for less, or, if we use the modal term it may be because we are operating under a tacit contextual assumption that some possibilities are too outlandish or irrelevant to take into consideration (Lewis 1996). When we settle for less, we may only want to say that in the circumstances p was a good enough reason for assuming q, and here the circumstances may determine not only the probability of q being false, but the gravity of getting it wrong and the cost of investigating further. It may only correspond to a good movement of the mind if nothing much hangs on it. This touches upon the relation between alethic standards and pragmatic ones, an area in which there is a clear difference between the ‘right’ kind of reason, for believing something, and the ‘wrong’ kind of reason, such as the advantage in doing so. A similar distinction arises in practical reasoning, where it hinges on the difference between a reason for admiring something, which is on account of the way it merits or deserves admiration, and an extraneous or ‘wrong’ kind of reason, such as strategic or political reasons for doing the same. The difference lies in the kind of movement of the mind in question. If we are egging someone on to admire something because it will be politic to do so, we are not commending the kind of movement that takes in only the relevant properties the thing possesses—those we take to be indicators of merit— and is guided to admiration on their account. It is if, but only if, we were prepared to commend this kind of movement we would say that the thing merits or deserves admiration. But if we see advantage in admiring it, for instance in becoming one of the club or sneaking a financial return, we are only hoping for a particular endpoint, and the only movement of the mind that is commended is one that takes account of the advantage and sets about gaining it. In the alethic case there are deep issues here, going to the heart of pragmatism, about the connections between success in action on the one hand and a general cognitive ability to represent the world on the other (Blackburn 2005). Fortunately, however, they do not concern us in this paper.
In the case of practical reason, the widest standards are those for evaluation in general. In saying that Molly’s being in distress is a reason for Sally to desist I commend or endorse or express appoval of the movement of mind in which Sally takes in Molly’s distress and as a result desists. This is entering an ethical judgement. I will have my own reasons for it: I hold that things go better if people are guided like this. If I go further and say that it is a decisive reason or a compulsory reason for Sally to desist, then I do not merely commend the movement, but insist upon it or regard it as compulsory, and stand ready to censure Sally if she fails to move the the appropriate way.
There are obviously many ways for our minds to fall short. We may be twitchy or lethargic: too quick or too slow to form beliefs, attitudes, or emotions. Our movement from apprehension of fact to flaming anger might be very regrettable. We are hot-tempered and therefore unreasonable. We might not be moved to alter our opinions by the well-judged comments of others. We are pig-headed and therefore unreasonable. There are many, many, ways of going wrong, and none of us avoid all of them all of the time.
In standard cases of succumbing to temptation, we can be described, albeit unhelpfully, as being unreasonable. Seeing the situation as it is, and judging which action is best to perform given how things stand, and then doing the other thing, is the standard case of succumbing to a temptation, and it will be generally (although not always) be an instance of a bad or inferior movement of the mind, that is, a case of being unreasonable. Perhaps this is what theorists are after when they talk of akrasia as involving irrationality, a claim which otherwise seems orphaned from any sensible train of thought.
If life were simple the virtue of reason would simply be a matter of moving well, one dyadic relation at a time. But of course it is not. Many considerations clamour for attention; many movements which would otherwise be good are nullified or outweighed by others. The phronimos or person of judgement and practical reason needs not only sensitivity to reasons one at a time, but a capacity to amalgamate them, weigh them and prioritise them. The better he does this, the more reasonable we allow him to be.
Since movements of the mind, in the generous sense we have given ourselves, occupy so much of the territory of ethics, it should be little surprise that Scanlon’s project of ‘buck passing’, or seeing talk of good and bad, right and wrong, obligation and trespass, as verdicts entered in terms of ‘reasons’, might be feasible (Scanlon 1998). Nor is it surprising that Michael Smith can urge the sovereignty of the ideally rational self, since this will just be the self whose mind moves exactly as it should. But of course, the takeover is merely nominal. For all we are given are moves within the ethical. We are not provided any independent methodology, or independent underwriting of the ethical as a domain. The suggestion we have been following out tells us nothing about the authority of these verdicts on good or bad practical movements of the mind. It merely uses the judgments themselves. If, for instance, we were troubled by objectivity before, we will be troubled by it after. For in spite of any contrary appearance, this talk of reasons imports no new standards and no new buttress for whatever standards we deploy.
V. Unkindness to Animals
Derek Parfit writes that ‘Other animals can be motivated by desires and beliefs. Only we can understand and respond to reasons’ (Parfit 1997, p. 127). But we now see that this is not so. That there is a snake in the path is an excellent reason for me to step aside. But it is also an excellent reason for my dog to step aside, and the dog may do so every bit as quickly and nimbly as me. Animals indeed differ from us: we can bring to mind wider ranges of considerations than they can, and perform complex estimates of which reasons to prioritize. But on this occasion, the dog responds to the reason with an alacrity more than matching my own. Parfit had earlier said, correctly, that ‘reasons for acting are facts that count in favour of some act’ (p. 121), and on this occasion the dog understood and responded smartly to one of those, the fact that counted in favour of jumping aside, just as I did.
Parfit probably did not intend to deny, as he actually did, that the dog responds to reasons. The tenor of his discussion, as of many others, is that we ourselves are not just responding to the presence of the snake, but to some ‘normative feature’ of the snake or in other words a further evaluative or deontic fact about the situation of its being in the path, a nimbus of normativity beatifying the union between the presence of the snake and a subsequent sidestep, a radiance in which Parfit, but not the dog, can bask. Needless to say, this is pure fantasy. The position of the snake and its posture are quite sufficient to set one’s legs racing. We do not need to respond to anything more or anything different; indeed, since speed is probably of the essence, we need not to do so. There is no time for extra processing. And since stepping aside is highly appropriate, this is a good movement of the mind, and equally so for the dog. We could, if we wish, give some meaning to saying that we, but not the dog, see the snake as a reason to jump aside. If this is to mean more than that we are disposed so to act, a property we share with the dog, it must be along the lines of our satisfaction with the movement, or willingnness to endorse it and recommend it for similar occasions, or in other words our own positive valuation of our own conduct. This is all that separates us from the dog. If we put them in the negative and said rather that we do not regret the movement or feel ashamed of it, or inclined to apologise for it, then once again the dog and we are on all fours, since it too feels no regret or tendency to apologise. Perhaps a young mongoose would feel some proto-version of these emotions, were its sidestep derided for cowardice by its mongoose mentors. But not the dog, and we share everything essential to walking with equal safety through the forest. Hence, there is nothing about our thoughts conducted in terms of reasons that affords any evidence at all for Parfit’s speciesist intuitionism.
Do we gain anything by subscribing to the thesis that if an agent has a reason to do something and is properly aware that the reason obtains, then they must be motivated to do it ‘on pain of irrationality’ (Wallace 1999, p. 218, citing many others)? First, notice that it is not very apparent how severe this pain is: Sally and her mischievous ilk can evidently put up with it quite contentedly. If we want to improve Sally by threatening pain, it had better be of a different sort. And most people find it sufficient to call children like Sally naughty, insensitive, mischievous, careless, callous, or even wicked, while after all it was the pain the snake might cause rather than any other imagined pain that explained our sidestep. We thought ‘Oh heavens, it might bite me’ rather than ‘Oh heavens, how horrid to feel irrational’. Is the invocation of irrationality an improvement, or, as it appears, a mere gesture, an impoverishment that washes out the interesting textures or particular contours of individual cases of vice and virtue? We would of course like Sally’s mind to move in better ways. We would like her to take Molly’s distress to guide her more reliably, and in the reverse way than it evidently does at present. We have familiar devices of persuasion and argument. How would you like it if Molly did the same to you? we might ask. Perhaps Sally does not mind the sound of fingernails on the blackboard, but Molly can reciprocate by playing her bagpipes, which annoys Sally just as much. Sally wouldn’t like it at all. We hope that thinking about that will motivate her to stop. But it may not. She can gamble on the kind and forgiving Molly not playing her bagpipes, or gamble on her parents stopping her if she does. Or, she can expect Molly to play her bagpipes, and be getting her own strike in first. Or, she can usually beat Molly in a fight. Or, she knows she may have to pay for her fun later, but still finds it irresistible to be naughty now. So we might try rubbing Sally’s nose in Molly’s distress, hoping to activate empathy or pity, and thence remorse and a better frame of mind. But perhaps we fail. It was, after all, the prospect of Molly’s distress that excited Sally’s mischief in the first place.
So we go away shaking our heads. Sally seems incorrigible. She lacks respect for the law (and for Molly). She doesn’t have her heart in the right place. But doesn’t she have her head in the right place? On the face of it Sally’s understanding is impeccable. She knows exactly what she is doing, and why she is doing it (this does not mean that she understands whichever psychological facts underlie sibling rivalry). Suggesting that it is her head at fault now looks simply like a deformation professionelle that afflicts moral philosophers, rather than an open road to new proofs of Sally’s wrongness, or new therapies for bringing her back to the straight and narrow. It is in this vein that Bernard Williams scoffed at that ignis fatuus of moral philosophy ‘the argument that will stop them in their tracks when they come to take you away’.
VI. The authority of reason
A problem area which my proposal clears up nicely is that of the ‘authority’ of reason, a problem some writers have found in ‘Humean’ proposals about motivation and desire. In an influential paper on this theme, Warren Quinn urged that there is a basic issue between rationalists such as himself, and “subjectivists” or “noncognitivists” (Quinn, 1995). Although I disown the labels, he clearly has in mind expressivist and in general naturalistic approaches to ethics of the kind that I favour. He writes that:
The basic issue here is more fundamental: whether pro-and con-attitudes conceived as functional states that dispose us to act have any power to rationalize those acts.
He points out that bizarre, pointless functional states (such as a disposition to switch on any radio that I find not to be on), do not ‘give me even a prima facie reason to turn on radios’. The disposition may explain how I am, but by itself it cannot make any resulting act of turning on a random radio sensible. And after rejecting any attempt to invoke higher-order states, such as pro or con-attitudes to the having of this first order disposition, to help with this problem, he concludes that in themselves dispositions such as tendencies to try to obtain things or to feel pained by things do not ‘rationalize’ choices. Even choices of means to given ends are not rationalized unless the ends themselves are, and only a genuine cognition of the objects of choice as ‘good’ could do that. Parfit enthusiastically takes the same line (Parfit 1997, p. 128).
It seems strange to say that a movement towards, say, eating a proferred piece of pie is not ‘rationalized’ by my occurrent hunger, so we need to take a closer look at this line of thought. First of all, which movement of the mind is in question? One proposal would be that it is from an awareness of a desire to a tendency to satisfy the desire. But that is not the typical case. When acting on a desire we are not typically self-reflective, taking a fact about ourselves as our starting point. Rather, we take in a fact about our situation, and our desires are functional states manifested in the relationship between the fact we apprehend, and the tendency towards action which results. As the desire for food, hunger is manifested in the way in which a tendency to take the pie issues from and is guided by an awareness that it is being proferred. Does the desire, then, “rationalize” the tendency? It explains it, in whichever way dispositions may be said to explain their manifestations. But Quinn is indeed right that it does not by itself show that the movement of mind is either good or bad, admirable or despicable, and so does not fund evaluative talk in terms of reason or rationality. That is not its job. However, all that shows is that Quinn’s demand that desire should validate or rationalize choice was entirely misplaced. To enter on the enterprise of arguing that a movement of the mind was a good one or a ‘rational’ one is a different business. To do this one has to step back, and sees if one can fit the movement into whichever practices in the area one endorses, or at least shares or understands or accepts as immune to criticism. The compulsive, oddball desire, such as the addiction to turning silent radios on, is pointless, and potentially costly and irritating. So we are far from inclined to endorse the movement of the mind from awareness of a silent radio to the motivation to turn it on, that manifests the compulsion.
Quinn may have thought that if particular desires cannot rationalize themselves, then nothing in our conative dispositions taken as a whole, could do so either: the picture is that the Humean world is one with ‘normativity’ bleached out of it. This would be a dangerous form of argument, whose weakness is more familiar from discussions of coherentism and foundationalism as they apply to cognitive states. While many writers accept that a belief cannot validate itself, they tend to suppose that its membership of a sufficiently coherent set may do so. Or, if other things than beliefs are allowed into the justificatory pool, they may include things like processes and actions, such as the engagement of perceptual processes in causal interaction with the world, or the experiences resulting from such engagement. If this begins to paint a satisfactory picture of cognitive justification, which it had better do because it is really the only game in town, then a parallel story can do a parallel job for practical dispositions, first invoking a whole matrix of surrounding dispositions, and then potentially invoking experience of the way those dispositions stand the test of time, as they are tried out in human practice. These together provide the only tribunal that a single desire could ever face. In other words, although we can stand apart from any particular desire or disposition, and consider the good of it in the light of other desires and dispositions, taken as a whole, there is no process of standing back from all of them at once, any more than there is in the case of belief. Someone with Quinn’s orientation might try urging that so long as this is ‘just us’, it can only tells us what we actually value, but not what is of value. But little is gained by denigrating the only methods we use, or could use. Insisting upon a wholesale cleavage between ‘fact’ and ‘value’ at this point would not so much be protecting the autonomy of the normative world, as making it on the one hand immune to awareness, and on the other hand of no conceivable interest. It is in fact only philosophers’ illusions, not valuations and norms, that are bleached out of the Humean world.
VII. Open Questions
Nevertheless, the contemporary enthusiasm for reasons suggests that in many minds, the substitution of the sovereignty of the good by the sovereignty of reasons is to be not just a change of idiom, but a change of regime. It is to open the way to a new dawn of philosophy, a new dispensation, and new philosophical territory to occupy and explore. It is important therefore to consider the view that by moving onto the territory of reason we are, actually, moving. I suggest that the only remaining temptation to think this arises because of the possibility of an ‘open question’ akin to Moore’s famous open question about goodness. However, in this application this question opens not between goodness and some natural property, but between reason and goodness. Thus if everything I have said is true, a critic may complain, how can there be the open and difficult question of whether it is always reasonable to be good? How can there be an issue, for instance, of whether reason might sometimes demand a sacrifice of goodness, in favour of such competing candidates as self-interest? How could we so much as worry whether reason stands on the side of prudence and self-interest, or on the side of justice or benevolence or the common good?
The question is very real, and fertilizes the idea of reason as a particular kind of authority, a self-standing normative structure magnificent enough to be used to measure and assay even the claims of virtue themselves. But I want to explain this open question differently.
For since ‘reasonable’ and its clan are general terms of commendation, like other such terms they can take on a particular cast. The can be confined to commendation within a subset of possible dimensions. This happens whenever we talk of ‘good for (the economy, the crops) or ‘good from (the point of view of the banks, the farmers)’, and in the same way we talk of reasons of state, economic reasons, reasons of health, personal reasons, or strategic reasons. Consider Machiavelli’s notorious claim in chapter XVIII of The Prince:
Therefore it is unnecessary for a prince to have all the good qualities I have enumerated, but it is very necessary to appear to have them. And I shall dare to say this also, that to have them and always to observe them is injurious, and that to appear to have them is useful; to appear merciful, faithful, humane, religious, upright, and to be so, but with a mind so framed that should you require not to be so, you may be able and know how to change to the opposite.
I do not think Machiavelli says in so many terms that it is often reasonable for a prince to be unmerciful, inhumane, faithless and the rest. But he says something that looks as if it implies it, namely that it is often absolutely necessary for him to be so, and that if fails to be so when the occasion requires it he will not achieve power or will be destroyed if he does. The Prince, we might say, sometimes has overwhelming reason to behave cruelly and inhumanely, treacherously and in bad faith. In short, he must behave badly. The dimension within which the commendation is given is simply that of self-interest, stability, or survival, and Machiavelli notoriously thinks that when these compete with conventional goodness not only do they win in mens’ actual conduct, but that it is necessary they should do so. Here what the Prince has most reason to do is not what is best: the movement of mind that is commended may be crafty, deceitful, treacherous and inhumane. He has to be these things (while appearing not to be) in order to survive.
All this is in accord with our proposal. The point is that the crafty and strategic movement of mind is indeed commended. It may not be being commended in conventional terms—that is why Machiavelli prompted such shock and gained his dark reputation—but it is commendation within what he regarded as the most important dimensions of statecraft, namely survival and success. And if we think about it this is always so. Whenever anyone describes a potential conflict between reason and virtue, what we find is that reasons are restricted to within a dimension, and the question is whether wider, more humane, virtues of justice or benevolence need curtailing because of the insistent demands of that dimension.
So we can open the question whether it is always reasonable to be good, not because reason is an autonomous lawgiver at some unspecified distance from the good, whose injunctions have their own authority, yet ones that may conflict with the injunctions of virtue or obligation. We open it, for instance, when we explicitly or implicitly worry about the old and uneasy conflict between self-interest and the other-regarding virtues. In an ideal world, perhaps, we could commend each without ever ranking them, for they march in step. But in the real world, and in spite of the optimism of some classical philosophers, any coincidence between them is a fragile business; servants of the world are not necessarily good trustees of their own interests, and indeed it is a political achievement to bring them into anything resembling an alignment. Machiavelli thought that in the Italy of his time, no such alignment obtained; hence, reasons of state had to trump better-known virtues, and the ideal prince had better be aware of that ugly fact.
We may also have other misalignments in mind. We might ask whether it is always reasonable to be just, having in mind potential conflicts between the common good, and one or another principle of justice. Here we probably confine the dimension of commendation indicated by ‘reason’ to consequential or utilitarian ones, and worry about the sacrifice of principle that they seem capable of requiring. But to solve this clash, if it can be solved, we cannot appeal to the autonomous court of reason. We can only walk around our own moral and ethical thought, and then campaign for whatever resolution appeals to us.
VIII. Means and Ends
We now turn to the much-discussed issue of means-ends reasoning, which is so frequently paraded as a prize specimen of ‘practical rationality’, a normative constraint of almost divine authority, and even a Trojan horse to insert into the citadel of naturalism. If Humean naturalism cannot even account for the majesty of this norm, then it is indeed in trouble.
It may be worth remarking that Kant did not think of it like that. Kant thought it is analytic that that if we will the end we will what is known to be the only means to it:
In the volition of an object as my effect, my causality as acting cause, that is, the use of means, is already thought, and the imperative extracts the concept of actions necessary to this end merely from the concept of a volition of this end …when I know that only by such an action can the proposed effect take place it is an analytic proposition that if I fully will the effect I also will the action requisite to it; for it is one and the same thing to represent something as an effect possible by me in a certain way and to represent myself as acting in this way with respect to it. (Kant 1785)
One can see why he might have thought that if we consider the problem of interpretation offered by the agent who might at first sight seem to intend (which I shall use as synonymous with ‘will’) an end, yet shows little or no inclination to adopt what he knows to be necessary means. It is at least plausible that we cannot be sure where he stands on the issue. Does he really intend to meet me for golf, if he said he would, but has not bothered to collect his clubs or put gas in his car? Perhaps he said so, but if he is comfortably resting in front of the TV as the necessary time ticks away, interpretation falters. Kant only says that if we fully will the end we intend the means, and that seems about right. His intention may be half-hearted; or his knowledge of the necessary means may be insufficiently robust, as when he knows that the time at which he might have got himself to the course has gone, but “hopes that something might turn up”. What is clear is that we cannot rely on him; we do not know where he stands on the project of playing golf, and perhaps he does not either, and probably there is no determinate reality about where he does stand. Socially he is a thundering nuisance, since on the basis of his apparently sincere say-so, we turn up, only to find he is not going to be there.
A norm of action is something to which we can conform, or fail to conform. But if Kant is right then there is a difficulty about failing to conform to the ‘norm’ of means-ends rationality. It cannot be done. There is however a cluster of very closely related norms, and indeed our errant golf partner exhibits what it is to fail to conform to them. He is a nuisance, as already noticed. He is incapable of following through on apparent commitments, for communicating an intention on which the audience is likely to rely is normally undertaking a commitment. He is weak-willed, in the sense properly made prominent by Richard Holton, and that builds on Michael Bratman’s pathbreaking discussions of the virtue of diachronic consistency in aims (Bratman 1988, Holton 1999). All that is sufficiently serious, and after all, it has been known to be important for a long time: ‘No man, having put his hand to the plough, and turning back, is fit for the Kingdom of God’. But what remains unclear is whether there is a more specific ‘norm’ of means-ends rationality, against which he has trespassed, or against which anyone can trespass. It is here that Kant’s doctrine stands in the way.
If there is such a more specific norm, then it requires careful formulation. Let us consider only situations in which it is known, sufficiently vividly or wholeheartedly, that means M is the only means to end E. Suppose we try:
If a person intends the end E, then they ought to intend means M.
We then meet problems of ‘factual detachment’, made prominent in deontic logic by the paradox of gentle murder (Forrester’s paradox), and recently resurrected by John Broome and Joseph Raz. The original paradox, we may recall goes:
If you murder someone, then you ought to murder them gently
You murder someone
Hence: you ought to murder them gently
Hence: you ought to murder them.
And the problem is how to interpret the first premise so that the conclusion does not follow, firstly by a simple application of modus ponens, and secondly by the principle that if a complex ought to occur then its consituents ought to occur. For a gentle murder is also a murder.
Before continuing, it will be as well to remember some points drummed into us in elementary formal logic. An argument takes us from premises to conclusions, not from beliefs to beliefs. If we talk of conclusions of arguments being ‘detached’, this does not imply that they are accepted or that it is a good idea to believe them. That is only so if it is a good idea to accept the premises, and the very fact that they imply the conclusion may count against that. Secondly, within an argument, a conclusion may be detached, but remain under an assumption. Detachment is not the same as discharging all the assumptions still in play. There is no limit, for instance, to the assumptions in play under which premises of a modus ponens may be assumed, and its conclusion appropriately detached—but still remaining under assumption. The importance of remembering these distinctions will shortly appear.
A second point to remember is that the auxiliary in the consequent, the ‘ought’ of deliberation, is not on its face an ‘ought’ of ethics. It is not in ‘if we want to get to Blackwells, we ought to go down the Turl’. In fact, everyday English idiom is quite happy to substitute other modal auxiliaries with more flavour of necessity and less of obligation: ‘To get to Blackwells you must/have to/should/had better/ go down the Turl’, and we can quite equally substitute a conditional prescription: ‘To get to Blackwells, go down the Turl’. Things are clearer if we generally reserve ‘ought’ for cases where there is genuinely a moral or evaluative element. The point to keep hold of is that we are advising a couse of action in the context of the assumption of an end to be achieved.
A final preliminary warning is that we should notice something treacherous about our habit of introducing apparent reference to states of mind, such as desires or intentions, into the antecedent of such conditionals. In the context of deliberation, the most the conditional can easily be heard to mean is that if we are to achieve the end, we have to intend the means; that is, in a normal world in which the end is to be achieved, such-and-such is the plan to adopt. In the context of deliberation ‘if we want him to come we have to write a letter’, ‘if we would like a good time we had better not go to Torremolinos’, or ‘if we wish to get home tonight we had better leave now’ would normally be taken to have as antecedents not states of mind, but their satisfaction: we could equally or better have put it by saying that ‘if he is to come, …’, ‘if we are to have a good time…’ or ‘if we are to get home, …’. The reference to wants, intentions or wishes is, in my view, an incidental way of indicating why we are interested in planning for those outcomes, rather than an integral way of specifying the condition in question itself. There is no inference, no movement of the mind, from the recognition of a state of mind itself to a demand or plan, but only an inference from the presumption that an end is to be achieved, to proposing a plan for achieving it. Such auxiliary mention of intentions, wants, or wishes, may also get into the consequents of conditionals. I might say ‘if you are to do the washing-up, you will want to wear an apron’ when I suppose that (a) you are to do the washing up (b) you do not want to do it and (c) you do not and will not want to wear an apron either. The conditional does not induce contradiction, because the mention of a want is incidental to its real content, which is to recommend that if you are to do the washing up, wear an apron.
A popular suggestion, which I used to accept, is that in the paradox of gentle murder the detachment is invalid because the first premise should be interpreted in terms of a wide scope ‘ought’. We only have:
It ought to be that (either you do not murder anybody, or you murder them gently)
And that together with ‘you murder someone’ yields no such inference.
But the wide scope reading seems unlikely on the face of it. Advice for what to do if a contingency arises is clearly not advice to make a disjunction true. Imagine a three horse race. I advise my bookmaker friend ‘if Galloper does not run, sell bets at evens on Trotter’. This is not: ‘sell bets at evens on (either Galloper running or Trotter winning). The evens bet that I advised might be a good one to sell, since Trotter is not really as good a horse as the third contender, Canter, so by selling the bet I suggested the bookmaker may expect to make money. But selling the bet on the disjunction may be a very bad idea, for instance if there is a much better than evens chance that Galloper will run. Nor is the advice to bring it about that either Galloper runs or you sell the evens bet, since you could follow this advice, but not the original, by arranging that Galloper runs. If you so clearly cannot export the advice in this case, it is very unlikely to be different if I arbitrarily choose to give it by using an auxiliary verb: ‘if Galloper does not run, you should/better/might want to/ought to/ sell bets at evens on Trotter’.
The proposed reformulation is inadequate in other important ways. Consider this conversation, in which Donald, Dick, Condi and George are four co-conspirators:
Donald: We are agreed, then, on a policy of imprisoning random Iraquis (IRI)
Dick: If we imprison them, we ought to humiliate them inhumanely (H)
Condi: No, if we imprison them, we ought to treat them with decency and
compassion (not-H). What do you think, George?
George: I agree with both Dick and Condi.
Surely Condi, Dick and Donald are right to gape. In the context of deliberation about what to do supposing that we are to imprison random Iraquis, George’s remark is completely at sea. It is simply not open to George to agree with both treating them humanely and treating them inhumanely.
This is obvious if we look again at the way we naturally formulate the conditionals. We might say: “if you are to murder someone, you ought to do it gently”. The activity is one of supposing that the end is given, and then recommending means, and this is a quite different activity from that of assessing the pair of <ends/means> together, which is all the proposed wide-scope recommendation of the disjunction shows us doing. In the deliberative context, IRI is being taken as given, just as the advice to sell the evens bet only becomes live if Galloper scratches. I think the best way of putting this is to say that the conditional has us consider the nearest normal world in which the end is to be achieved, and proposes a plan: a plan of what has to be done or is best to be done in that world, or to bring about that world.  The question of whether it was a good idea to achieve this end simply does not enter in, any more than when we say ‘if the giant slime is coming, flee for your lives!’ we express any attitude either to the probability or the desirability of the giant slime coming, or of any complex that has this as a component. The English variant closest to the Latin is perhaps more perspicuous here: when you murder someone, you ought to murder them gently.
With these points understood, it is clear that George’s contribution is in effect a contradiction, an endorsement of both of two incompatible plans and this would not be true, of course, if the conditional were simply a wide scope ‘ought’ governing a disjunction. Were that so, George’s remark would be perfectly intelligible as a way of saying or implying that we ought not to imprison random Iraquis. But I hope that most of us do not, simply because of that, agree with Dick in the above conversation. Whereas if the wide scope disjunctive account were correct, we might well do so. Similarly consider:
Donald: The Iraquis ought not to resent us being there.
Dick: If they do, we ought to beat them to death
Condi: I agree with Donald, but not with Dick.
If Dick’s remark were parsed as ‘it ought to be the case that either they do not resent us or they get beaten to death’, then it follows from what Donald said, and there would be no room for Condi’s position. But of course, there is.
In the context of deliberation, the conditional ‘if we murder someone, then we ought to do it gently’ is a perfectly acceptable recommendation of a plan for the nearest normal world in which we are in fact to murder someone. Much better do it gently! The ‘ought’ of planning detaches. The plan is conditional upon an antecedent being satisfied: it is only when or if we are to murder someone that we should follow the plan to do it gently. This is most obvious when the murder is unavoidable or irrevocable: the assassins we employed are on their way and beyond recall, but we can somehow get a painkiller to the victim before they arrive, so that is what we ought to do (Setiya, 2007). But mere supposition or postulation of the end takes us to the same deliberative context as actual irrevocability. The consequent is detachable even if the murder is not irrevocable or inevitable; it may be still under consideration, and the consequent only detached, in the way that I reminded us of at the outset, that any consequent is detached in a formal argument, potentially en route to a reductio or a modus tollens:
Dick: Are we to murder prisoners?
Colwyn: If we do so, we ought to do it gently
Donald: There is a major difficulty about that, since none of our soldiers know how.
Condi: Still, Colwyn is right, so perhaps after all we had better not murder prisoners.
Here the consequent is provisionally detached, an implication worked out (the plan requires resources we do not have), and turned into an objection to the proposal of murdering prisoners. The consequent is detached just as any proposition may be detached in the course of any inference, not necessarily as something to accepted in its own right, but under an assumption, provisionally en route to further inference, and then potentially to a backtracking on the original antecedent assumption. It is here that we must remember the remarks I made about detachment not being the same as acceptance.
All this is the context of deliberation. To repeat, in that context, the conditional ‘if we intend E we ought to do M’ signals the endorsement of a plant of action (M) in the normal world in which we are to perform or bring about E. Nothing is said about whether it was a good idea or morally acceptable, or inevitable or anything else, to have the intention itself. A consequence of deliberations in which the conditional works just as conditionals normally do (sustaining modus ponens, opening the way to modus tollens), may be to make it clearer than before that we had better abandon the intention itself. And in that context, ‘if we intend the end, then we ought to intend the means’ is clearly a good principle. ‘Taken to the closest normal world in which the end is to be achieved, plan on using the means’. Of course you should, and if Kant was right you must, on pain of forfeiting your claim genuinely to intend the end.
But this does not mean that if we switch to the different context, that of external judgement, that we need to see anything good either about having the intention, or about using whichever means the intention requires if it is to be fulfilled.
Although the language of reasons can be used carefully, so that the necessary distinctions are maintained, it makes it very easy to get all this wrong. Thus consider the question whether Iago’s villainous intention to destroy Othello ‘provides a reason’ or ‘provides a normative reason’ for him to manufacture lies about Desdemona. We naturally recoil from saying that it does: we do not want to hear ourselves recommending anything about Iago’s end, nor the means he adopts. On the other hand Iago does his planning impeccably; having turned his hand to the plough, he does not turn back, even if he is ploughing the wrong field. How are we to combine our out-and-out rejection of Iago’s intention and its handmaidens, with acknowledgement of his abilities as a planner?
Fortunately, we have ample ways of saying what needs to be said. There are two terrible things to say about Iago: he had villainous ends in view, and he chose villainous means to execute them. There is one, perhaps grudging, good thing to say about him: he is an able planner. When he contemplated and intended the closest normal world in which he is to effect Othello’s destruction, he planned efficiently and as it turns out successfully to bring it about. If we imagine instead an Iago who (at least apparently, if we remember Kant), intends Othello’s destruction, but does little or nothing effective to execute it, then things are reversed. There is one bad thing to say about him —he is not an effective or efficient planner—and two slightly better things can be said than are to be said in the Shakespearean scenario: first that he does not set up Desdemona, and second that his intention to destroy Othello seems relatively insecure or half-hearted. It is a mistake to try to shoehorn all these, and perhaps more distinctions, into the one verdict on whether Iago did or did not have a “normative reason” for his behaviour, or any part of it. The language simply will not bear the complexity of the distinction between the perspective of deliberation and that of external assessment, and it also encourages inattention to the crucial difference between description of Iago (give in terms of his reasons for doing one thing or another) and endorsement of one or another facet of the movements of his mind.
A conflation that assists in confusing this issue is that the conditional ‘if we intend the end we ought to intend the means’ can sound as if the antecedent locates a state of mind, and then it looks as if the issue is to be whether our having that state of mind provides some sort of reason for supposing that we ‘ought’, perhaps in some strong ethical sense, to intend the means. And that sounds in general outrageous: how can we bootstrap ourselves into having reasons or even obligations, so easily? But as I have already argued, in the context of deliberation the apparent reference to a state of mind is incidental. There is no inference from a state of mind to a plan, but only a supposition that something is to be done, to the conditional selection of a plan for doing it. And with this the appearance that means-ends rationality or means-ends normativity provide a problem for Humeans, a shining jewel that they cannot pick up, and hence that gives us every incentive to mine for others, disappears.
XI. The plasticity of reason
If we throw away attention to the particular nature of peoples’ flaws, preferring a blanket diagnosis of ‘unreasonable’ or ‘irrational’ whenever their minds move in ways we think inferior, we not only lose important textures and distinctions, but we also lose most chances of engagement and imrovement. For ‘unreasonable’ and still more ‘irrational’ not only function as general terms for denigrating the movement of peoples’ minds. They usually have further, sinister connotations that the defect is irredeemable, that it is not sensitive to discursive pressure, that it licenses us to treat the subject as a patient or in other ways as beyond the human pale, or out of the game. Let us return to errant Sally. We can say, of course, that Sally is irrational or unreasonable—her mind is guided in bad ways. What we cannot do is invest the term with more interest than it gains from gesturing at the more specific and insightful descriptions of the particular flaws that infect Sally’s character. But if we are to improve Sally, it is her particular flaws that need particular attention. We might want to cherish Sally a little more, be careful how we praise Molly when Sally is present, be more careful of providing opportunities for envy and jealousy, and so forth. In harsher climates, we might have wanted to frighten or bribe her. Whatever rationalists, intuitionists, realists, Kantians, or Platonists may say, these are the only tools anyone has. We may win in the end. Sally may not be irredeemable after all. For one implication of all this is that reason is every bit as pliable as sentiment.
Dept of Philosophy
University of Cambridge/University of North Carolina, Chapel Hill.
Bateson, M, Nettle, D & Roberts, G. 2006. “Cues of being watched enhance cooperation in a real-world setting”, Biology Letters, 2. 412—414.
Bennett, Jonathan, 1974. “The Conscience of Huckleberry Finn”, Philosophy 49, No 188
—1984. A Study of Spinoza’s Ethics, Indianapolis: Hackett, 1984, p. 324—8.
Blackburn, Simon. 2005. ‘Success Semantics’ in Lillehammer, Hallvard & Mellor, D. H. Ramsey’s Legacy. Oxford: Oxford University Press.
Bonevac, D. 1998. ‘Against Conditional Obligation’, Nous, 32:1
Bratman, M. 1987. Intention, Plans, and Practical Reason, Cambridge, Mass: Harvard University Press.
Broome, John. 2004. “Reasons.” In Reason and Value: Themes from the Moral Philosophy of Joseph Raz, edited by Wallace et al., 28-55. Oxford: Clarendon Press.
Cowley, Christopher. 2005. “A New Defence of Williams’s Reasons-Internalism.” Philosophical Investigations, 28:4, 346-368.
Dancy, J. 2000. Practical Reality. New York: Oxford University Press.
Fitzpatrick, W. 2004. “Reasons, Value, and Particular Agents: Normative Relevance without
Motivational Internalism.” Mind, 113: 285-318.
Foot, P. R. 1978. “Reasons for Action and Desires.” In Practical Reasoning, edited by Joseph Raz, 178-184. Oxford: Oxford University Press.
Grice, G. R. 1978. “Motive and Reason.” In Practical Reasoning, edited by Joseph Raz, 168-177. Oxford: Oxford University Press.
Holton, R. 1999. “Intention and Weakness of Will” Journal of Philosophy, 96 (1999) pp. 241-62
Kant, I. 1785. Royal prussian Academy Edition, IV, 417. Also The Groundword of the Metaphysic of Morals, ed. Paton. New York: Harper Row.
Korsgaard, C. 1986. “Skepticism about Practical Reason.” Journal of Philosophy, 83: 5-25.
———. 1996. The Sources of Normativity. Cambridge: Cambridge University Press.
Lewis, D. 1996 “Elusive Knowledge”, Australasian Journal of Philosophy 74.
Nagel, T. 1970. The Possibility of Altruism. Princeton: Princeton University Press; reprinted 1978.
Parfit, D. 1997. “Reason and Motivation.” Proceedings of the Aristotelian Society, Supplementary Volume 71: 99-130.
Plato 2008. Republic, trs. Waterfield. Oxford: Oxford University Press.
Quinn, W. 1995. “Putting Rationality in its Place” in Virtues and Reasons: Philippa Foot and Moral Theory, eds. Hursthouse, Lawrence and Quinn, Oxford: Oxford University Press.
Raz, Joseph. 1975. Practical Reason and Norms. London: Hutchinson & Co. Ltd.
———. 1978. (ed.) Practical Reasoning. Oxford: Oxford University Press.
———. 2003. The Practice of Value. Oxford: Clarendon Press.
Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge, Mass.: Belknap Press.
Setiya, K. 2004. “Against Internalism.” Noěs, 38: 266-98.
—2007. “Cognitivism about Instrumental Reason,” Ethics (2007): 649-673.
Shafer-Landau, R. 2003. Moral Realism: A Defence. New York: Oxford University Press.
Smith, M. 1994. The Moral Problem. Oxford: Blackwell.
Spinoza (2001) Ethics ed. G. R. Parkinson. Oxford: Oxford University Press.
Suikkanen, Jussi. 2004. “Reasons and Value — In Defence of the Buck-Passing Account.” Ethical Theory and Moral Practice, 7:513-535.
Thomason, R. 1981. ‘Deontic Logic and the Role of Freedom in Moral Deliberation’ in R. Hilpinen, ed., New Studies in Deontic Logic, Dordrecht: Reidel.
Wallace, R. J. 1999. “Three Conceptions of Rational Agency.” Ethical Theory and Moral Practice, 2: 217—42.
———. 2002. “Scanlon’s Contractualism.” Ethics, 112(3): 429-470.
Wallace, R. Jay et al. 2004. Reason and Value: Themes from the Moral Philosophy of Joseph Raz. Oxford: Clarendon Press.
Williams, B. 1979. “Internal and External Reasons.” Reprinted in Moral Luck. Cambridge:
Cambridge University Press, 1981: 101-13.
— 1995. ‘Internal Reasons and the Obscurity of Blame’ in Making Sense of Humanity, Cambridge: Cambridge University Press
Wong, D. 2006. “Moral Reasons: Internal and External.” Philosophy and Phenomenological Research, 72: 536—58.
 I should say that I have misgivings about the term, and usually find myself writing it sous rasure. I believe Fodor has said that ‘cows go “moo”, but philosophers go ‘norm’”, and I agree.
 The exceptions I have in mind are Huckleberry Finn style cases. See Bennett 1974.
 I do not in this paper highlight any differences between ‘rational’ and ‘reasonable’. Pruned of theoretical accretions I think they come to little more than, as Edward Craig once put it to me: being reasonable just means being reasonably rational.
 In my 1998, p.254, I call this the ‘leading, characteristic mistake of a whole generation of theorists wanting to go beyond Hume’, and ten years later I can add around a third of a new generation.
 I gratefully adopt this useful expression from Michael Bratman.
 The separation of contexts of deliberation from contexts of judgement was recommended in Thomason 1981.
 I talk of the closest normal world, following Bonevac 1998. The view is that conditionals take us to the closest normal world in which their antecedent is satisfied. This is not necessarily the closest world, since the actual world may be abnormal. A consequence is that conditional logic is not monotonic. Conditionals in ordinary discourse do not accept strengthening: ‘if you turn the ignition the car will start’ may be true, while ‘if you turn the ignition after taking out the battery, the car will start’ is not.
 A question addressed in these terms by Michael Bratman…