I. INTRODUCTION
In ‘The Scalar Approach to Utilitarianism’ Alastair Norcross argues that scalar consequentialism is the most plausible form of consequentialism and argues, therefore, that it is the form of consequentialism that consequentialists ought to adopt. He writes:
Utilitarianism should not be seen as giving an account of right action, in the sense of an action demanded by morality, but only as giving an account of what states of affairs are good and which actions are better than which other possible alternatives and by how much. The fundamental moral fact about an action is how good it is relative to other available alternatives. Once a range of options has been evaluated in terms of goodness, all the morally relevant facts about those options have been discovered. There is no further fact of the form ‘x is right’, ‘x is to-be-done’, or ‘x is demanded by morality.’Footnote 1
I will argue, however, that Norcross's arguments are seriously flawed. In section II, I will consider Norcross's argument that there are no facts of the form ‘x is right’ or ‘x is to be done’, and will argue that his arguments are invalid and confused. In section III, I will consider Norcross's claim that ‘abolishing the notion of “ought” will not seriously undermine the action-guiding nature of morality’Footnote 2 and that scalar consequentialism is sufficiently action-guiding. In relation to the arguments considered in section II, I argue that Norcross's arguments are simply confused and that he doesn't seem to recognize just how radical scalar consequentialism is. In relation to the arguments considered in section III, I will show that Norcross's arguments can only justify the much weaker claim that scalar consequentialism is action-guiding to some extent. This is not sufficient to make scalar consequentialism plausible.
II. NORCROSS'S ARGUMENTS FOR SCALAR CONSEQUENTIALISM
Norcross has a number of arguments in favour of scalar consequentialism. I will consider each in turn.
A. Rightness is derived from a gradable property
Norcross's first argument is based on the simple fact that rightness is derived from goodness, and goodness is a scalar property. And this, he suggests, is reason to think that rightness too must be gradable. The idea seems to be that if x is derived from y, then x ought to have the same properties as y. Or perhaps Norcross would not commit himself to this more general claim, but this certainly seems to be his thinking regarding the property of being gradable.
However, Norcross does not have an argument for this position. In fact, the argument is presented in the form of a rhetorical question. First, he asserts that, according to consequentialism, ‘the property of an act that makes it right or wrong – how much good it produces relative to available alternatives – is naturally thought of as a matter of degree’.Footnote 3 He then asks, ‘Why, then, is rightness and wrongness not a matter of degree?’Footnote 4
Norcross might insist that this was not meant to be an argument in and of itself. After all, he does follow this question by saying, ‘I will argue that, from the point of view of a consequentialist, actions should be evaluated purely in terms that admit of degrees.’Footnote 5 It is clear, therefore, that he doesn't take it to be a self-evident position that can be defended merely by appealing to the fact that rightness is derived from goodness.
Nevertheless, it does seem clear that, when Norcross asks the question, he does expect people to struggle to give an adequate answer. At the very least, he seems to think that the burden of proof is on the maximizer and the satisficer to give an answer, and if an answer cannot be given then this is indeed a reason to favour the scalar approach.
However, this is not a heavy burden, and an adequate answer can be provided. It is not at all unusual to derive an absolute property from a gradable one. One can provide a huge number of counterexamples. Consider the property of being the winner of a long jump competition. The rules of the competition state that whoever jumps the furthestFootnote 6 is the winner. Clearly, the length of jump is something that comes in degrees. One jump can be longer than another. But we do not conclude that the property of being the winner comes in degrees. We do not say that the person who jumped the furthest is more the winner than the others. The person who jumped the furthest and only the person who jumped the furthest is the winner. Winning is all-or-nothing.Footnote 7
Likewise, consider ‘best’. If Pele is a better footballer than Maradonna, then Maradonna clearly cannot be the best footballer. Something is or is not the best: it is not a matter of degree. But clearly the property of being best is derived from a gradable property – goodness. Furthermore, this is also true of the property of being good enough. Now we can say that Pele is better than Maradonna, but still insist that Maradonna is good enough. (Indeed, we can say that he is much better than merely good enough.) But nevertheless, the property of being good enough is still absolute, or in Norcross's term: all-or-nothing. We do not say that Pele is more good-enough than Maradonna. Rather, we say that Pele is better than Maradonna, but both are good enough – unlike Tony Adams who falls short of the mark.
Thus, if we say that an act is right (or permissible) if and only if it is good enough (or if and only if the consequences of the act are good enough), then it will follow that rightness is derived from goodness, but also that rightness is an all-or-nothing affair. And there is nothing at all mysterious about this.
Furthermore, if we work on the assumption that the derived property should share all the characteristics of the property it is derived from (which seems to be the assumption behind Norcross's challenge), there is a real worry that we will never be able to derive a new property at all. Rather, we will simply duplicate the original property and give it a new name. I conclude, therefore, that, far from being odd that rightness should have some properties that goodness does not, it is a necessary part of the process if we want to derive a new property rather than merely duplicating (and renaming) the original.
B. Norcross's rejection of cost-sensitive consequentialism
Considering the possibility of a form of act-consequentialism that makes demands, but is not as demanding as maximizing consequentialism, Norcross writes:
The position of the boundary between right and wrong may be affected by such factors as how much self-sacrifice is required of the agent by the various options, and how much utility or disutility they will produce. For example, it may be perfectly permissible for the doctor to stay at home, even though the best option would have been to go and help with the epidemic.Footnote 8 On the other hand, if all the doctor could do and needed to do to save the villagers were to send a box of tablets or a textbook on diseases, then he would be required to do all he could to save them.Footnote 9
Norcross, however, goes on to argue that we should reject this cost-sensitive form of consequentialism on the grounds that we cannot justify the claim that there is a cut-off somewhere on the scale such that any act above that cut-off is permissible, and any act below it is impermissible. Norcross has two arguments for this conclusion:
i. The no significant difference argument
In this argument, Norcross tries to show that we cannot find a cut-off between right and wrong. Norcross considers the obligation of beneficence. He supposes that Jones is obliged to give 10 per cent of his income to charity. This would suggest that there is a significant difference between giving 9 per cent and 10 per cent, but a less significant difference between giving 8 per cent and 9 per cent. After all, on this account, the first is a choice between right and wrong, while the latter is a choice between two acts, both of which are wrong. Thus, Norcross suggests that the first choice should be much more significant than the latter.
Norcross, however, argues that no consequentialist should accept this. From a consequentialist point of view the difference between 8 per cent and 9 per cent is just the same as the difference between 9 per cent and 10 per cent.
To demonstrate why this is so, Norcross then considers two people, Smith and Jones. Jones is trying to choose between giving 11 per cent or 12 per cent, and Smith is torn between 9 per cent and 10 per cent. Smith and Jones earn the same amount, and thus an extra 1 per cent from Smith will be an extra $5,000, and an extra 1 per cent from Jones will also be $5,000. If we are concerned with the good that can be produced as a result, $5,000 from one person is as good as $5,000 from another. Therefore, a consequentialist shouldn't be any keener to persuade Smith to give 10 per cent than he should be to persuade Jones to give 12 per cent.
In saying this, however, Norcross is mixing two issues. The first is the issue of permissibility, and the second is the issue of what a consequentialist ought to prefer of other agents. There is nothing inconsistent in saying (a) that (if we are concerned only with the good produced as a result) we should be indifferent to the question of whether the extra $5,000 comes from Smith or from Jones, and (b) that if Jones moved from giving 11 per cent to giving 12 per cent this would be a move from acting impermissibly to acting permissibly.
Norcross is committed to the claim that: ‘the sole determinant of the appropriateness of blame or punishment is how much utility will be produced by doing so’.Footnote 10 He should also be committed to: ‘The sole determinant of the appropriateness of persuading someone to give an extra 1 per cent of his income to charity is how much utility will be produced by doing so.’ But as soon as this is stated explicitly, it should be clear that this argument doesn't tell us anything about permissibility. For the consequentialist, whether or not the person has already given enough, or more than enough, is simply irrelevant. The consequentialist doesn't care whether or not Smith and Jones act permissibly. He is only concerned with how much utility will be produced.
Norcross argues that the consequentialist has no more reason to persuade Smith than he does to persuade Jones. And he concludes from this that the difference between 9 per cent and 10 per cent cannot be any more significant than the difference between 11 per cent and 12 per cent. And he concludes from this that we cannot conclude that the difference between 9 per cent and 10 per cent is the cut-off between right and wrong, while the difference between 11 per cent and 12 per cent is a less significant difference – the difference between one supererogatory act and another slightly better supererogatory act.
But this argument is flawed. All it shows is that, actually, the consequentialist doesn't care whether Smith and Jones act permissibly or not. He only cares about getting the extra $5,000 from each of them.
Imagine a slightly different situation. Suppose Smith is torn between giving 0 per cent and giving 1 per cent, and Jones is torn between giving 80 per cent and giving 81 per cent. Furthermore, Smith is leaning towards 0 per cent and Jones towards 81 per cent, so it would be easier to persuade Jones than Smith. Let us also suppose that the consequentialist has limited time in which he can try to persuade Smith and Jones to part with their money. Given the time constraint, he is unlikely to be able to persuade both. Where should he concentrate his efforts? When we ask this question, we can even stipulate that Smith and Jones are – as a matter of fact – both required to give 10 per cent, because this stipulation won't make any difference.
According to the consequentialist's reasoning, he ought to start by trying to persuade Jones. But what does this tell us about the cut-off between the impermissible and the permissible? Nothing. As I have shown, this conclusion is quite consistent with the fact that Jones acts permissibly while Smith does not. But this fact simply doesn't enter into the consequentialist's reasoning (at least as Norcross represents it). Thus, this just highlights the fact that the consequentialist does not care about fairness or the distribution of burdens, and does not care about permissibility. This tells us something about the consequentialist's attitude to permissibility, but it does not tell us anything about permissibility itself.
Thus, in one sense, Norcross's argument fails: the conclusion does not follow from his premise. In another sense, however, there is some kind of success, because it simply illustrates that, in relation to our treatment of others at least, the distinction between the permissible and the impermissible seems to be irrelevant from a consequentialist point of view. Whether I act permissibly or impermissibly will have no bearing on how the consequentialist treats me. This prompts the question: why are consequentialists so worried about this particular issue?
Thus, even if we reject Norcross's claim that scalar consequentialism is the most plausible form of consequentialism, it could be argued that scalar consequentialism is at least the most honest form of consequentialism, as it doesn't pretend to care about permissibility.
ii. The arbitrary cut-off argument
Norcross also writes: ‘A related reason to reject an all-or-nothing line between right and wrong is that the choice of any point on the scale of possible options as a threshold will be arbitrary.’Footnote 11 Essentially, Norcross's concern is that any cut-off will merely be the laying down of ‘arbitrary conventions’.Footnote 12 But it is not clear why Norcross should think that it would be the mere laying down of conventions, particularly given that he recognizes that it is plausible to think that the cut-off should depend on the level of self-sacrifice required.Footnote 13 If the location of the cut-off depends on a particular criterion, it is not arbitrary.
The only claim that we need to commit ourselves to in order to reject scalar consequentialism in favour of a cost-sensitive form of consequentialism is the claim that the self-sacrifice demanded of an agent can be compared with the benefits to others, such that we can judge whether or not the benefits are great enough (compared to the level of sacrifice required) to require an agent to meet the demands.Footnote 14
This may be clearer if we express it in terms of reasons. The cost-sensitive consequentialist merely claims (1) that moral reasons are comparable with non-moral reasons, and (2) moral reasons do sometimes, but not always, outweigh or defeat the non-moral reasons. In relation to the example of the doctor who could go to the aid of a town that has been hit by an epidemic, putting himself at great risk, we simply say that the doctor has a moral reason to go to the aid of the people in the town, but he also has a prudential reason to stay at home. The question then is: is the moral reason great enough to outweigh or defeat the prudential reason? If so, the doctor will be morally required to help; if not, he won't be required to help.
In short, if moral reasons and non-moral reasons are comparable, such that we can weigh these reasons against each other in order to reach an all-things-considered conclusion, then it simply won't be true that we are merely laying down arbitrary conventions. Thus, it seems that, if we want to defend scalar consequentialism, we have to argue for the incomparability of reasons, to demonstrate that moral reasons can never outweigh or defeat prudential reasons, and therefore can never give us a conclusive reason.Footnote 15
C. Epistemological concerns
Of course, it may be vague, and clearly there are epistemological worries: how do we weigh reasons? Are we reliant on judgement? What if your judgement is different from mine? I concede that these are real concerns, but they do not lead to the conclusion that we are merely laying down arbitrary conventions. Furthermore, we cannot appeal to these epistemological concerns in order to defend scalar consequentialism. To do so would be to say: I don't know what morality demands of me, so I will conclude that it demands nothing. This move is simply invalid. It's like someone trying to guess the number of sweets in a jar at a fun fair and thinking:
I really have no idea. Any number I give would be a complete guess.
And then, in a moment of inspiration, he thinks:
Yes, that's it. I have no idea. Any number would be a guess. Therefore, I conclude that there must be none. Yes, that's it. The jar is empty.
Excitedly, he grabs the man at the stall.
I've got it. I know the answer. It's none. There aren't any sweets. The jar is empty. Do I win the prize?
Clearly, this man's reasoning is flawed, and he walks away empty-handed.
A slightly less absurd variation on this reasoning is to suggest that, because we don't know where the cut-off is, we should act as if there is no cut-off. Or, because we don't know what morality demands of us, we should act as if it doesn't demand anything at all. It might be claimed that this kind of neutrality on the matter is the appropriate response to our inability to locate the cut-off. But, of course, to act as if morality demands nothing is not a neutral response. Ultimately, given that we have to act one way or another, it is not clear that there is a neutral response available. Regardless of the epistemological concerns, we have to make some sort of judgement of what we think morality demands of us. And, if this is the case, even if we don't know – even if we can't know – what morality demands of us, we should at least try to give a plausible answer. But to judge that morality doesn't demand anything at all is possibly the least plausible conclusion available.Footnote 16
D. Epistemological and ontological scalar consequentialism
I should stress that I am not suggesting that Norcross argues in this manner. He does not. As I have said, his argument is based on the claim that any cut-off would be based on an arbitrary convention. I was merely responding to a possible response to my objection to Norcross.
Slote, however, does seem to appeal to something like this reasoning in his account of scalar consequentialism.Footnote 17 However, Slote does also consider a stronger version of scalar consequentialism, which doesn't rely solely on epistemological concerns but rather claims that ‘scalar act-consequentialism might represent the sole objective truth about the morality of actions’.Footnote 18 I call the former epistemological scalar consequentialism and the latter ontological scalar consequentialism. Norcross is arguing for ontological scalar consequentialism: the claim that there are no moral facts other than comparative facts about one act being better than another. There are no further facts of the form ‘x is right’ or ‘x is required’ or ‘x is impermissible’.
Epistemological scalar consequentialism, on the other hand, accepts that there are acts that are right and others that are wrong, but insists that we can never know which are which, and then concludes that we should limit ourselves to making comparative judgements of better and worse, and not even try to work out which acts are permissible and which are impermissible. For the reasons given in the section above, we should not accept this conclusion even if we accept the premises. Therefore, in this article, I am primarily concerned with ontological scalar consequentialism.
III. SCALAR CONSEQUENTIALISM AS A GUIDE TO ACTION
In section II, I argued that Norcross is committed to the claim that consequentialists do not (or should not, if they are to be consistent) care about whether or not others act permissibly. When it comes to punishing people, or other expressions of praise and blame, it should not matter whether they have acted permissibly or not. Nevertheless, consequentialists might care about the permissibility of their own actions. In particular, consequentialists may be concerned that, if we do not have a cut-off between the permissible and the impermissible, then morality will not guide their actions.
In relation to this problem, Norcross has two lines of argument. In the first, he tries to undermine the importance of the concepts of right and wrong in our moral thinking. In the second, he argues that, contrary to first appearances, scalar consequentialism can in fact function as an adequate guide to action.
A. Right and wrong
Norcross writes that: ‘It is clear that the notions of right and wrong play a central role in the moral thinking of many. It will be instructive to see why this is so.’Footnote 19 Norcross identifies two main reasons. The first is ‘a long list of examples which present a choice between options which differ greatly in goodness’Footnote 20 and the second is ‘the imperatival model of morality’.Footnote 21
i. Examples in which the options differ greatly in goodness
When Norcross talks about a long list of examples in which the available options differ greatly in goodness, I assume that Norcross is talking about people in general and not of philosophers in particular. If he was discussing the ‘list of examples’ that could be found in moral philosophy, this claim would surely be false. Moral philosophers tend to be interested in the difficult cases, and therefore have a diet of examples in which the options do not differ greatly in terms of goodness. The two options may differ greatly in their nature, but in terms of the goodness of the two acts, presumably they do not differ greatly, and it is for this reason that we do not know what to do. Should I accept a job in a laboratory that pursues research into biological and chemical warfare, or should I reject the offer, although I know that I will struggle to get another job and that this will cause my wife more stress and will be to the detriment of my children?Footnote 22
Thus, I suggest Norcross must be talking about the examples that we grow up with as children. A child hits another child, and he is told, ‘Don't do that! It's wrong.’ Norcross considers the example of a choice between helping an old lady across the road, and mugging her.
If these are the only two options presented, it is easy to classify helping the old lady as the ‘right’ thing to do, and mugging her as ‘wrong’. Even when there are other bad options, such as kidnapping her or killing her and eating her, the gap between the best of these and helping her across the road is so great that there is no question as to what to do.Footnote 23
As a scalar consequentialist, however, it is not clear how Norcross can justify this claim. For a maximizing consequentialist there is no question as to what to do, and (presumably) for a satisficing consequentialist or a cost-sensitive consequentialist there is no question as to what to do. The satisficer and the cost-sensitive consequentialists may have difficulties in locating the cut-off between right and wrong, but there is little or no doubt that it will rule out the various bad options considered in this example. For the scalar consequentialist, however, there is still a question of what to do. Scalar consequentialism does not rule out any options – not even the very worst actions, such as killing and eating old ladies. Norcross just doesn't seem to have recognized quite how radical a position scalar consequentialism is.
At this point, one might wonder if Norcross means something else by scalar consequentialism, but I simply don't know what else he could mean. Furthermore, Norcross himself states that, on his view, there are no facts about what we ought to do:
Once a range of options has been evaluated in terms of goodness, all the morally relevant facts about those options have been discovered. There is no further fact of the form ‘x is right’, ‘x is to-be-done’, or ‘x is demanded by morality’.Footnote 24
So, for Norcross, killing and eating old ladies must remain an option.
ii. The imperatival model of morality
According to the imperatival model of morality, morality issues (decisive) demands. If a person obeys the demand, he does right, and if he disobeys, he does wrong. Norcross associates this model of morality with divine command ethics and some forms of deontology. But surely most forms of consequentialism also fit this model: the maximizing consequentialist, for example, demands that you do the best act possible (or produce as much good as possible).
Norcross then claims that one of the motivations for accepting this model is its simplicity. I am not convinced that this is right. And, even if it is one of the motivations, it surely isn't the main motivation. The motivation is quite simply that we have a very strong intuition that there are some acts that ought not to be done – such as killing and eating old ladies.
Norcross then concludes that the main justification for the imperatival model of morality is merely pragmatic. But this argument is invalid. Norcross states that simplicity is ‘Part of the motivation’Footnote 25 and ‘one of the motivations’.Footnote 26 To move from these premises to a conclusion about the ‘main justification’Footnote 27 is simply invalid. Not only does he move from ‘part of’ and ‘one of’ to ‘the main’, he also moves from ‘motivation’ to ‘justification’.
He then continues:
After all, if we don't have the justification that the commands issue from a deity, it is always legitimate to ask what grounds them. That certain states of affairs are good or bad, and therefore should or should not be brought about, seems like a far more plausible candidate to be a fundamental moral fact than that someone should act in a certain way.Footnote 28
But the claim that ‘certain states of affairs are good or bad, and therefore should or should not be brought about’ is simply not consistent with scalar consequentialism. Scalar consequentialism does not tell us that certain states of affairs should or should not be brought about – it can only tell us that it would be better or worse (than other alternatives) to bring them about. Again, Norcross seems to have forgotten just how radical scalar consequentialism is. In fact, all Norcross has done here is given a consequentialist answer to the question: what grounds the commands? According to the consequentialist, considerations of the relative goodness of states of affairs ground the commands. And, therefore, maximizing, satisficing and cost-sensitive forms of consequentialism also fit the imperatival model of morality. As a result, Norcross seems to have undermined his own objection. He asks, what could ground these commands? And then he tells us: consequentialist considerations.
B. Scalar consequentialism and reasonsFootnote 29
In addition, though, Norcross also argues that it is simply not true that scalar consequentialism cannot guide our action. Norcross offers the following characterization of the argument against scalar consequentialism:
1. If a theory doesn't guide our action, it is no good.
2. If a theory doesn't tell us what we ought to do, it doesn't guide our action.
3. Scalar utilitarianism does not tell us what we ought to do.
Therefore, scalar utilitarianism is no good.Footnote 30
Norcross then claims that the first premise is ambiguous, and claims that if ‘guide our action’ just means ‘tell us what we ought to do’, then the first premise will be question-begging. Thus, he interprets ‘guide our action’ to mean ‘provide us with reasons for acting’. Having clarified the argument in this way, Norcross argues that premise 2 is false. That is, he argues that a scalar theory can provide us with reasons for acting, and therefore can guide our actions, even though it does not tell us what we ought to do.
In fact, it should be noted that Norcross does not merely claim that satisficing consequentialism can offer guidance to some extent. He also claims that: ‘Abolishing the notion of “ought” will not seriously undermine the action-guiding nature of morality.’Footnote 31 This, of course, is a stronger claim.
In order to defend this position, Norcross in fact offers two arguments, the first working on the assumption that internalism is true, and the second on the assumption that externalism is true. However, Norcross is not as clear (or as consistent) as he could be in his characterizations of internalism and externalism. Partly for this reason, and partly because I think it will be more illuminating to do so, I will not focus solely on Norcross's arguments but will, instead, consider scalar consequentialism in terms of reasons generally. But in doing so, I will address Norcross's arguments.
So what should the scalar consequentialist say about reasons? Norcross writes: ‘If the fact that an act is wrong gives us reason to avoid it, then the fact that it involves the production of a bad state of affairs, by itself, gives us reason to avoid it.’Footnote 32 The argument seems to be simply that facts about wrongness and facts about badness can both provide us with reasons. If Norcross's intention is simply to reject the claim that scalar consequentialism cannot be action-guiding at all, then this claim may be plausible. If it is not clear why I say this, consider Mulgan's Magic Game example:
The Magic Game. Achilles is locked in a room, with a single door. In front of him is a computer screen, with a number on it (call it n), and a numerical keypad. Achilles knows that n is the number of people who are living below the poverty line. He also knows that, as soon as he enters a number into the computer, that number of people will be raised above the poverty line (at no cost to Achilles) and the door will open. There is no other way of opening the door. Because of the mechanics of the machine, any door-opening number takes as much time and effort to enter (negligible) as any other.Footnote 33
Scalar consequentialism doesn't tell us that pressing n is right, or that doing anything else is wrong, or impermissible, or ought not to be done. However, it does tell us that pressing n is the best act available. Even without saying anything about rightness or permissibility, doesn't it follow from this fact alone that Achilles has a reason to press n? Wouldn't we think that Achilles was confused if he said, ‘I know that pressing n is the best thing to do, but nevertheless there is no reason to press n.’
The objection against scalar consequentialism, however, should not be that it is not action-guiding at all. As the example above suggests, that objection doesn't seem convincing. Any theory that ranks options according to their goodness provides at least some guidance. The objection should be that scalar consequentialism is not sufficiently action-guiding. My claim now is that Norcross's argument is not adequate as a response to this objection. It is clear, however, that Norcross did intend his argument to be a reply to this objection, and not only the first. As I have already stated, Norcross claims that, ‘Abolishing the notion of “ought” will not seriously undermine the action-guiding nature of morality.’Footnote 34 Again, Mulgan's Magic Game sheds light on the issue.
Norcross's approach clearly suggests the following response to the Magic Game: maximizing consequentialism will claim that, because pressing n is the only permissible option, Achilles has a reason to press n. Equally though, scalar consequentialism states that, because pressing n is the best option, Achilles has a reason to press n. Thus, Norcross's claim seems to be that both theories give the same answer, and therefore maximizing consequentialism is no better off in terms of action-guidingness than satisficing consequentialism. But this is simply not true. Both accounts tell Achilles that he has a reason to press n, but the maximizer says more than this. The maximizer states that Achilles has a conclusive reason to press n. As a result, he tells Achilles that he ought to press n, and that it is impermissible to do anything else. Clearly, this offers significantly more guidance than a theory that merely tells you that you have a reason to press n, but says no more.
How should the scalar consequentialist respond to this observation? Let us consider, in more detail, what a scalar consequentialist might want to say about Achilles and the Magic Game.
Option 1
He might claim that because pressing n is the best option available, we can say more than just ‘Achilles has a reason to press n’. We can also say that Achilles has more reason to press n than to do anything else. He has most reason to press n. But now it looks like we should also say that Achilles ought to press n. Perhaps the scalar consequentialist will resist this final step, and simply deny that we ought to do what we have most reason to do, but it is not clear that this move would be plausible, or even coherent.
Option 2
Alternatively, the scalar consequentialist might accept the claim that Achilles ought to press n, but insist, nevertheless, that it is not wrong or impermissible to fail to do what you ought to do. On this account, moral oughts would be similar in nature to prudential oughts. For example, we might say that one ought to brush one's teeth, without concluding that it would be impermissible not to. Furthermore, this needn't depend on the claim that prudential oughts are typically hypothetical oughts. Admittedly, many people will want to say that the reason it is not impermissible to fail to brush your teeth is because it is merely a hypothetical ought, which depends on your having a desire to have healthy teeth, or to avoid toothache. But this needn't be the case. One could also claim that it is not merely a hypothetical ought. You really ought to care about your teeth – whether you do or not – and therefore you ought to brush your teeth. Nevertheless, one could still claim that it was not impermissible to fail to brush your teeth. It is not impermissible simply because it is not a moral ought.
However, when applied to moral oughts, it is not clear that this is a plausible option. Typically, we do think that it is impermissible to fail to do what morality states that you ought to do – and this is one of the things that distinguishes moral oughts from prudential oughts.
Furthermore, this simply isn't an option that is available for Norcross, because he explicitly states that he takes ‘wrong’ and ‘ought not to be done’ to be interchangeable.Footnote 35 Likewise, he also identifies scalar consequentialism with the abolishing of the notion of ‘ought’ (as well as ‘right’ and ‘wrong’) from moral discourse.Footnote 36
Option 3
Finally, the remaining option is for the scalar consequentialist to concede that Achilles ought to press n (because this is what he has most reason to do), and even concede that it would be wrong not to. But now it looks like we have simply rejected scalar consequentialism. After all, scalar consequentialism is defined by its claim that morality does not issue demands: it does not tell us what we ought to do, and it rejects the concepts ‘right’ and ‘wrong’.
C. Scalar consequentialism is not sufficiently action-guiding
Norcross is right to claim that judgements about the comparative goodness of acts (or states of affairs) can give us reasons, but it doesn't follow from this that scalar consequentialism gives us all the guidance we need. In particular, it has two major flaws.
First, it seems to be unable to give a coherent account of what we ought to say in response to Mulgan's Magic Game example. Here, we have a strong moral reason to press n, and no competing reasons. As such, it seems that this gives a conclusive reason to press n. This is what we ought to do. But this conclusion isn't available to the scalar consequentialist.
Second, the scalar consequentialism lacks an ability to compare different types of reasons, in order to reach conclusions about which reasons are conclusive. For example, it cannot compare moral reasons with prudential reasons to conclude that, on this occasion, the moral reason is conclusive. It cannot tell me that the moral consideration in favour of saving a person's life is stronger than, and outweighs, the prudential consideration in favour of sitting back and doing nothing, in order to avoid breaking a nail. In this case, merely saying that ‘saving the person's life would be good’ is not sufficiently action-guiding.
IV. CONCLUSION
In this article, I made a distinction between epistemological and ontological scalar consequentialism. I argued that the former can be dismissed very easily. In relation to the latter, I argued that Norcross has simply failed to demonstrate that there are no facts of the form ‘x is right’ or ‘x is to be done’. He is simply mistaken when he suggests that there is a problem with deriving absolutes like right and wrong from gradable properties such as goodness, and he cannot justify his claim that the choice of a threshold will always be arbitrary. Also, his ‘no significant difference’ argument only shows that the consequentialist doesn't care about permissibility, not that we cannot distinguish permissible acts from the impermissible.
In section III, I argued that Norcross was right to claim that a scalar theory can be action-guiding but that he was mistaken to think that it could be sufficiently action-guiding, or that ‘Abolishing the notion of “ought” will not seriously undermine the action-guiding nature of morality.’Footnote 37
Scalar consequentialism is not the most plausible form of consequentialism. If there are any reasons to favour it over other forms of consequentialism, Norcross hasn't identified them. Furthermore, on closer inspection, the theory doesn't look plausible at all, and has major problems explaining why strong reasons don't lead to oughts even in cases where there are no competing reasons (as in Mulgan's Magic Game example).
If anything can be said in favour of scalar consequentialism, it is only that it is the most honest form of consequentialism, because it doesn't pretend to care about permissibility.Footnote 38
