1Nearly a century-and-a-half after the first edition of his philosophical masterpiece, The Methods of Ethics, Sidgwick’s star is still to rise again. In his own recent masterpiece, On What Matters, Derek Parfit (2011: xxxiii), echoing C.D. Broad (1930: 143), claims that the Methods is the best book on ethics ever written. A glance at contemporary literature or university syllabuses will show that, if Broad and Parfit’s view is anywhere near the truth, Sidgwick’s book is not receiving the attention it deserves.

2 Much of Parfit’s book concerns the nature of ethics and normativity. In this paper, I shall attempt to elucidate some of Sidgwick’s central views on these issues. These views seem to me often plausible, and always interesting and suggestive. I shall criticize some of them, but in the firm belief that one can learn a great deal through engagement with Sidgwick’s arguments. I shall begin by briefly considering his distinction between an ethics which seeks moral laws and one which inquires into the good, before moving on to Sidgwick’s conception of the practicality of ethics. Having examined the methological issues behind Sidgwick’s decision to restrict the scope of his argument to three main ‘methods’, I shall then move to Sidgwick’s views on the autonomy of ethics, the notion of ‘ought’, and moral motivation.

1. Jural and Teleological Views

3 Sidgwick suggests two ways in which ethics might be conceived (1.1.2.1-2/2-4). The first is the ‘jural’ view, according to which ethics seeks the ‘true Moral laws or rational precepts of Conduct’, while the second — which we might call the teleological view — suggests that ethics is an inquiry into the ‘True Good’ and the method for achieving it. The jural view is said to be more prominent in modern ethical thought, because such inquiry as there is into the ultimate good is directed at finding the actions that will lead to it. Sidgwick’s point here is hard to grasp, since he included within his definition of the teleological view inquiry into the means of attaining the good. But he goes on to make a further claim, that the position he calls intuitionism, according to which conduct is right when in accordance with certain principles, cannot easily be understood as teleological. Rather, it is concerned with explaining people’s moral duties independently of their own good or happiness. [1]

4 As Sidgwick himself realizes, we might assume that an intuitionist could adopt a teleological position by claiming that a person’s ultimate good in fact consists in her performing her duty. But Sidgwick says that modern intuitionists, and Christian thinkers in general, do not take such a line. This raises two important methodological questions, however. First, if a position reconciling intuitionism and teleology was held as part of common-sense in earlier times, why should it not be discussed now? Second, even if it has never been part of common-sense (as I suspect it has not), then why is that sufficient justification for our refusing to discuss it, if it has some independent plausibility?

2. The Practicality of Ethics

5 Philosophical ethics, Sidgwick tells us, like science aims to be ‘systematic and precise’ (1.1.1.2/1). [2] He later says that the assumption that moral rules should be precise ‘naturally belongs to the ordinary or jural view of Ethics as concerned with a moral code’ (3.2.3.1/228), and provides an argument for this view based on an apt analogy with law. If a law were vague, we would think it to that extent unreasonable: anyone subject to a legal obligation ought to be in a position to know what it is. Similarly, a moral philosophy which left it unclear on some occasion exactly what a person’s obligations were would, to that extent, have failed. [3]

6 A good deal of law is indeed highly precise. The UK Representation of the People Act 1969, for example, leaves no doubt as to when a person becomes eligible to vote in a parliamentary election: on their eighteenth birthday. But some law is less precise. Consider, for example, the definition of obscenity in the Obscene Publications Act 1959, still in force in the UK:

7

For the purposes of this Act an article shall be deemed to be obscene if its effect or (where the article comprises two or more distinct items) the effect of any one of its items is, if taken as a whole, such as to tend to deprave and corrupt persons who are likely, having regard to all relevant circumstances, to read, see or hear the matter contained or embodied in it.

8 If I have written some potentially obscene article, and am considering its publication in the UK, the law will not tell me whether or not it is safe for me to publish it. I have to rely on my judgement about the likely effects of its publication, and whether they might be described as depravation or corruption. Now it might be said that Sidgwick is right that, to this extent, the law is a failure, and legislators should seek further precision. But even if they did — perhaps by spelling out further what is meant by depravation and giving examples of what is and what is not to count as obscene — there will be an ineliminable role for judgement on the part of citizens. A law is not a failure if it is reasonably clear, and relies only to a reasonable degree on individual judgement. Exactly what counts as reasonable or not in ethics is a highly important question, but I believe that Sidgwick’s standard of reasonableness is almost certainly too high.

9 The scope of philosophical ethics, Sidgwick says, is the ‘methods’ of ethics, where a method is ‘any rational procedure by which we determine what individual human beings “ought” — or what it is “right” for them — to do, or to seek to realize by voluntary action’ (1.1.1.1/1). Such a ‘procedure’ need not be a process. Sidgwick allows that a version of intuitionism according to which we have immediate insight into the rightness of certain actions is proposing a method (1.1.2.3/4). Nor is Sidgwick to be understood as suggesting that only actions matter in ethics, and not, say, the feelings or the characters of agents. Indeed he elsewhere allows that the common-sense conception of virtue includes the emotions (3.2.2.2/222-3) and that ethics should construct ideals of character (3.14.1.3/393). But discussion of such topics is significant only in so far as it is related to the primary question of ethics — how we should act. Like Aristotle (1894: 1103b26-9), Sidgwick sees philosophical ethics as essentially practical.

10 Who are the ‘we’ who use these methods? We philosophers? I suggest not. Rather, Sidgwick is speaking of ‘we human beings’. At 1.6.3.4/85, for example, Sidgwick discusses the view that knowledge of common-sense moral rules was implanted in people by nature or God, these rules themselves being the best means to the general happiness. He says that, on this view, the method of utilitarianism is clearly rejected, even though the reason for any such rule is utilitarian. But it becomes clear that Sidgwick does not wish to restrict ethics to methods so understood. [4] Which ‘decision-procedures’ we should follow is indeed one important question; but another is the ‘criterion of rightness’, that is, what makes actions right, understood independently of decision-procedures (Bayles 1971). Indeed one way to understand Sidgwick’s later position is that he is, to some extent, recommending a position similar to that mentioned above, according to which there is a utilitarian justification for following the rules of common-sense morality, at least much of the time, and part of the task of philosophy is to work out which such justifications are plausible. At 1.6.3.1/83, Sidgwick notes that, by adopting some plausible assumption or other, one can connect almost any method with almost any ultimate principle. But he limits his own discussion to those methods ‘logically connected with the different ultimate reasons widely accepted’ (1.6.1.3/78). In other words, he is looking for certain widely accepted ultimate reasons for action (such as happiness), and the decision-making procedure that most naturally rests on that reason. For example, egoistic hedonism consists in the acceptance of the agent’s own happiness as ultimate reason or end, and its method will be the attempt by the agent to maximize her own happiness. But, since method and end can come apart — as in the case of Sidgwick’s own utilitarianism — one might wonder why Sidgwick appears to require that, before any ultimate reason can be given a place in his discussion, its related method must be present in common-sense morality. Imagine that common-sense morality made no room for utilitarian decision-making. The utilitarian principle could still capture what justifies the common-sense method or decision-procedure. [5]

11 Indeed, since ethics is an enquiry into what grounds or justifies our actions and any decision-procedure we adopt, we might wonder further why Sidgwick puts so much weight on methods, rather than ultimate principles. [6] It is those principles that will give us systematic ‘general knowledge of what ought to be’ (1.1.1.2/1), and a method turns out just to be the direct application of any principle in a decision about what to do. Ethics, as Sidgwick points out, is ‘sometimes considered as an investigation of the true… rational precepts of Conduct’ (1.1.2.1/2-3), and he himself implies that we are interested in the principles that determine which conduct is ultimately reasonable (1.1.3.5/5-6). Further, his emphasis on method over principle can lead to philosophical distortion. The focus on method may explain, for example, why Sidgwick is so ready to find utilitarianism within common-sense morality, merely because people sometimes decide what to do by trying to work out what would do the most good from the impartial point of view. [7] This is evidence merely of a principle of beneficence, alongside others, and that is a far cry from utilitarianism, according to which this is the only ethical principle. In other words, Sidgwick’s book should perhaps have been titled The Ultimate Principles of Ethics, those principles each being a different statement of our ultimate reasons for action.

3. Sidgwick’s Three Methods

12 According to Sidgwick, we continually inquire into what is ultimately reasonable because different and incompatible principles are present in common practical reasoning (1.1.3.5/6), and an answer given in terms of one of them will appear suspect from the point of view of the others. Which principles does Sidgwick have in mind?

13 The first set of principles are those that constitute the morality of common sense (1.1.4.3-4/7-8): the rules of prudence, justice, veracity, and so on. Sidgwick suggests that common sense ordinarily sees these rules as binding in any particular case independently of the consequences of the action in question or its alternatives. By this, Sidgwick should not mean that, according to common sense, its rules apply whatever the consequences. [8] So common sense may allow that in unusual cases an injustice may be permissible or even required to secure some great good. But in ordinary cases it will hold that, say, the keeping of a promise is required because of the nature of the action in question and not because of the bad consequences of breaking the promise or of the good consequences of the practice of promise-keeping in general. There is a difficulty, however, in that the position Sidgwick will contrast with intuitionism — utilitarianism — can be understood in such a way that all the moral weight is placed on the nature of an action rather than its consequences. This is because actions are open to various descriptions. Consider the action of maximizing happiness. A utilitarian will see such an action as absolutely required in every case (not just ordinary cases), regardless of the consequences of performing that action. So I am required by utilitarianism to maximize happiness even if, for example, one consequence of my doing so is that an innocent person is killed or tortured.

14 Sidgwick then notes (1.1.1.5-6/8) that many utilitarian thinkers see these common-sense rules as mere means to the general happiness of humanity or sentient beings, and contrasts this view with the principle of prudence which rests on the postulation of the individual’s happiness as an end. He goes on to point out that people often adopt ultimate ends independent of happiness, such as fame, and that this poses a potential problem for ethical theorizing if a different method has to be constructed for each end. It is not obvious, however, why this should be so, since the differences may be superficial. Compare the ‘method’ that involves the maximization of general happiness with that which requires the maximization of certain goods other than happiness. It is true that the actions resulting from the application of each method might be quite different; but the finding of the means to the different ends is a technical matter, and does not rest on any deep philosophical difference. But Sidgwick anyway goes on to suggest that common sense will not on reflection validate such ends as rational. Rather, fame will be seen as valuable only in so far as it is a source of happiness to the famous individual, a sign of excellence or the achievement of some benefit, or a spur to further achievement.

15 By choosing fame as his example of a common end, Sidgwick has loaded the dice against those proposing goods other than happiness or excellence as worthy of pursuit — knowledge, accomplishment, or friendship, for example, construed as good in themselves and not in so far as they constitute human excellence. Here (1.1.4.7/9) he merely asserts that, on the face of it, the only two goods with a strong and widely supported claim to being rational ultimate ends are happiness and excellence (i.e. the achievement, or partial achievement, of some ideal notion of human perfection). The reference to ‘strength’ here is important, and implies that there may be widely supported claims on behalf of other goods which Sidgwick sees as weak. This is an early example in the Methods of a potential clash between reflective common sense and Sidgwick’s own intuitions. [9]

16 Sidgwick does provide an argument for restricting his discussion of ultimate reasons to those based on the goods of happiness or excellence, and those based on duty as prescribed by unconditional rules (1.6.1.3/78). He suggests that this tripartite distinction mirrors what appear to be the most fundamental distinctions we use in understanding human life: that between the conscious being and the stream of experience, and the distinction within the stream of experience between action and feeling. Perfection is put forward as the ideal goal for a human being, considered diachronically, while duty concerns what should be done, and happiness is a matter of feeling. But on the face of it, even if we take some version of this trichotomy to be fundamental, it is hard to see why it must map onto the distinction between excellence, happiness, and duty in the way Sidgwick suggests. Happiness or duty could be said to be the ideal goal; perfection, as moral excellence, concerns both action and feeling. Why, anyway, should we see this distinction as fundamental? Why not that between, say, mind and body? And other goods could anyway be said equally plausibly to map onto the distinction. So beauty, for example, could be said to be primarily a quality of the conscious being as opposed to her actions or feelings.

17 Sidgwick’s failure seriously to consider goods other than happiness and perfection as ultimate ends lies behind another oddity of his position. He allows for happiness to provide the foundation for two different methods, egoism and utilitarianism, depending on whether it is the agent’s own happiness or general happiness which is to be sought. We might expect the same distinction between self-regarding and impartial perfectionist principles, but Sidgwick claims that past philosophers have always assumed, as far as ‘moral excellence’ is concerned, that promoting the virtue of others is never in tension with the agent’s own virtue. First, however, even if Sidgwick’s historical claim is correct, the question is whether common sense might accept the possibility of such a tension. It certainly seems that it should. All we have to imagine is some case in which, by refusing to act wrongly, I bring it about that two or more others act wrongly to the same degree. Second, his focus on moral excellence in particular enables him to ignore historical examples of sacrifice of the agent’s own non-moral excellence for that of others — such as, perhaps, the return of the philosophers in Plato’s Republic into the cave (2003: 519c8-520e3). Sidgwick does provide an argument for the restriction of excellence to moral excellence, which is that virtue is ‘commonly conceived’ as not only the most important component of excellence, but a component lexically prior in value to any other component. This, however, leaves open the possibility of a different conception, perhaps equally ‘common’. Consider a case in which I am in a position greatly to advance knowledge, accomplishment, and friendship, either in my own life or generally, but only through committing some minor moral misdemeanour such as telling some not terribly important lie. I suspect that many would permit, perhaps even require, the lie.

18 Sidgwick’s restrictive conception of excellence leads him to see perfectionism as merely a special type of intuitionism (the view discussed above, according to which common-sense moral principles are ultimately reasonable). But here we see another philosophical distortion arising out of his focus on method rather than principle. Even if the views are extensionally equivalent, they should not be identified. [10] According to intuitionism, the reason I should keep my promise, for example, is that promises should be kept. According to perfectionism, however, the ultimate reason here consists in the promotion of perfection.

19 The upshot of my discussion so far, then, is that Sidgwick should have considered, in addition to the three methods he does discuss, a further two: egoistic and general perfectionism. [11] Further, it is tempting to think that Sidgwick may be allowing his own intuitions to ‘filter’ the deliverances of common sense at an early stage in the argument. [12] A footnote to the sentence in which he says he will treat perfectionism as a form of intuitionism refers us to the argument in 3.14 against non-hedonistic conceptions of the good.

20 This concern raises further questions about Sidgwick’s own account of his aims in the Methods. In the preface to the first edition (viii), [13] Sidgwick claims that the protreptic goals of moral philosophers have distorted the development of ethics as an impartial science, and announces his intention to focus on the methods themselves and not their practical results:

21

I have wished to put aside temporarily the urgent need which we all feel of finding and adopting the true method of determining what we ought to do; and to consider simply what conclusions will be rationally reached if we start with certain ethical premises, and with what degree of certainty and precision.

22 Above, I claimed that Sidgwick’s focus on action suggests an Aristotelian regard for the practicality of ethics. In this preface, however, Sidgwick explicitly distances himself from Aristotle, stating that his aim is mere knowledge. Later, in the main text of the 7^th edition (1.1.5.2/13), he claims that his only ‘immediate aim’ [14] has been to clarify the three methods themselves and that he has therefore ‘refrained from expressly attempting any such complete and final solution of the chief ethical difficulties and controversies as would convert this exposition of various methods into the development of a harmonious system’ (see also 3.11.1.3/338). This is mysterious. [15] Sidgwick famously considered the Methods a failure (Hayward 1901: xix, n.), but not because it left the three methods unclear. The failure consisted in the dualism of the practical reason — the conflict between egoism and utilitarianism that made impossible the development of a ‘harmonious system’. [16]

23 Sidgwick’s methodology, then, is perhaps not quite as he suggests. The most plausible explanation for his discussing egoism, common-sense morality or intuitionism, and utilitarianism is that these are the three views he finds most plausible:

24

When I am asked, ‘Do you not consider it ultimately reasonable to seek pleasure and avoid pain for yourself?’ ‘Have you not a moral sense?’ ‘Do you not intuitively pronounce some actions to be right and others wrong?’ ‘Do you not acknowledge the general happiness to be a paramount end?’ ‘I answer ‘yes’ to all these questions. My difficulty begins when I have to choose between the different principles or inferences drawn from them.

25 Further evidence for this suggestion may be found in what Sidgwick says at 1.1.4.7/10 about the possibility of ends intermediate between the happiness of the agent and general happiness, such as the happiness of one’s family, nation, or race: ‘any such limitation seems arbitrary, and probably few would maintain it to be reasonable per se, except as the most practicable way of aiming at the general happiness, or of indirectly securing one’s own’. This is highly implausible, and it is an especially clear case of Sidgwick’s assuming too quickly that others will, on reflection, agree with him. Given a choice between utilitarianism, and a principle that allows or even requires the agent to give some priority to her own children, it is beyond doubt that the majority of those not already committed to utilitarianism will strongly prefer the latter.

26 As we shall see in the next section, Sidgwick’s views on the autonomy of ethics also led him to restrict the scope of philosophical ethics as he conceived it. In the end, however, the doubts I have raised, and will raise in the next section, about Sidgwick’s choice of three methods do not, I believe, undermine his project. Moral theories can plausibly be distinguished as broadly consequentialist on the one hand, and non-consequentialist or deontological on the other. [17] And the main opposition to moralism is, of course, philosophical egoism, which, though it is not currently much discussed in contemporary ethics, remains one of the most powerful and attractive normative theories. Indeed, Sidgwick could plausibly have started with these three methods, and gone on to explain how different accounts of the good will give rise to different forms of egoism and utilitarianism in particular. The arguments for his basic triad are problematic; but it stands independently of them.

4. The Autonomy of Ethics

27 As we have seen, Sidgwick is happy to see philosophical ethics as a science in so far as it aims at being systematic and exact (1.1.1.1-2/1-2). But calling it a science, he suggests, may be misleading, since the objects of ethics and of the ‘positive sciences’ are quite different; ethics is better described as a study. [18] Sidgwick notes that psychologists and sociologists who seek to explain human action individually or collectively often begin to speak of the actions or institutions they study as ‘good’, ‘bad’, ‘right’, or ‘wrong’, and, in doing so, they move out of their ‘positive’ discipline into ethics or politics. The attempt to explain human action, sentiments, and judgements is ‘essentially different’ from that to decide which of these is right or valid.

28 Ethics is ‘non-naturalistic’, in so far it sees what ought to be as an object of knowledge, [19] while it is usually held that a science must be an inquiry into some ‘department of actual existence’. Sidgwick is, then, both a cognitivist — in the sense that he accepts that there can be ethical knowledge, at least in principle — as well as a non-naturalist. Is he also committed to normative, ethical, or moral realism, if we understand such positions as involving a metaphysical commitment to moral or ethical facts? In fact, Sidgwick never speaks of such facts. [20] And his reference to the subject-matter of positive science as some aspect of ‘actual existence’ suggests that he was not inclined to think in those terms. [21] For him, it was enough that we can have knowledge of moral truth (1.3.3.3/34), and that philosophical ethics should be seen as the discipline through which we might acquire such knowledge. [22]

29 For this reason, Sidgwick appears to be in a position to sidestep the charge of metaphysical extravagance often made against ethical non-naturalism (see e.g. Mackie 1977: 38-42). Of course, in some weak sense he must be committed to ethical facts, if the postulation of such facts amounts to nothing more than the claim that ethics has an object which we can cognize and that this object is not reducible to non-ethical facts. But Sidgwick’s moral metaphysics is purely derivative from his epistemology. He does not begin with the claim that there is some ‘special realm’ of moral facts, like the Platonic forms, and then provide an epistemology in the light of that. In this respect, then, Sidgwick seems to show better philosophical judgement than G.E. Moore, who can fairly be said to have given non-naturalism (a position which he himself (1903: 17-21) credits to Sidgwick) [23] a bad name, by loading it down with a good deal of heavy — and from the point of view of normative ethics unnecessary — metaphysical baggage.

30 At 1.6.1.3/78-83, Sidgwick notes that his list of three ultimate reasons suggested for further attention — happiness, excellence, and duty — may appear to leave out some important further options for ultimate reasons: that something is God’s will; that it promotes ‘self-realization’; or that it is in conformity to nature. Sidgwick admits the immediate attraction of such views, since they might appear to provide a foundation for what ought to be in what actually exists. He then claims that, for this very reason, they are the concern not of ethics as he has defined it (which is of course as an autonomous discipline) but of philosophy more generally, which is concerned with the relations of all objects of knowledge.

31 This restriction of the scope of ethics is somewhat unsatisfactory. Sidgwick might be proposing his conception of an autonomous ethics for pragmatic reasons — as a research programme which, though it leaves out some significant options, will be for that reason more manageable. But this will leave his conclusions as merely temporary hypotheses, which, even if they appear highly plausible, will require testing against the ethical conclusions from ethics more broadly construed as part of philosophy. This seems especially regrettable given the pessimistic conclusion of the Methods: a proponent of one of the three alternatives might well be hopeful of greater philosophical success.

32 But Sidgwick’s restriction is best seen not as purely heuristic, since he continues:

33

The introduction of these notions into Ethics is liable to bring with it a fundamental confusion between ‘what is’ and ‘what ought to be’, destructive of all clearness in ethical reasoning: and if this confusion is avoided, the strictly ethical import of such notions, when made explicit, appears always to lead us to one or other of the methods previously distinguished.

34 Sidgwick proceeds to examine the ‘divine will’ view first, since here, he says, the connection between ‘what is’ and ‘what ought to be’ is clear. He raises the question of how we might ascertain God’s will, and suggests it must be through revelation, reason, or some combination of both. Revelation, he suggests, does not fall within the domain of ethics as he understands it. Given Sidgwick’s own philosophical intuitionism, this is not a strong response to the revelational position. He might rather have allowed for the view, and then impaled it on one or other horn of the Euthyphro dilemma. If the view is that God’s willing in itself gives us a reason to act, this appears arbitrary, and also fails to explain why on most theistic accounts God is said to will actions for which there often seem to be excellent reasons independent of his willing. But if God is seen merely as the best guide to what is independently right, then the view becomes more strictly ethical, and Sidgwick can run his argument that such views can be categorized within his triad. It is a form of this latter argument that he uses against the rationalist theist, claiming that her ethical position will be either one of his original three, or a version of the ‘self-realization’ or ‘nature’ views.

35 Sidgwick therefore moves to a discussion of the conformity-to-nature view (CTN). He claims first of all that human nature must be seen as designed, if the view is to be at all plausible. If nature were just a matter of aimless change, it becomes quite unclear how it could ground ends or moral rules (later, at 1.6.2.3/83, Sidgwick applies this claim to evolutionary theory). So let us consider the design view. First, to avoid the ethical conclusion that ‘anything goes’, we need to distinguish the notion of natural impulses from those that humans actually experience. But we cannot do this by claiming that the supremacy of reason is natural, since the starting assumption of the view is that reason requires conformity to nature. CTN-theorists have usually understood the natural to be either what is common or usual, or what is original or earlier as distinct from what develops later. But Sidgwick sees no ground for thinking that nature is opposed to the unusual or prefers what is earlier in time. Some widely admired impulses — such as enthusiastic philanthropy — are uncommon and emerged later than several other impulses often thought less valuable. Indeed, many ‘later’ developments, such as certain societal institutions, seem beneficial, and it would anyway be arbitrary to see them as beyond the ‘purposes’ of nature. Nor can any substantive ethical conclusions be drawn from premises about the physical make-up of humanity. Reflection upon the actual social relations within any human society will take us to the question of what such relations ought to be, and our original triad.

36 Many versions of CTN do fall to one or more of Sidgwick’s objections. But one strand of the view appears to remain standing. Sidgwick’s restriction of CTN to versions which postulate design in nature is unmotivated. Where human nature came from is irrelevant; what matters, according to CTN, is what it is. Many have accepted as plausible the Aristotelian view that the good of any being will depend on its nature, and that such a good is promoted through increasing the perfection of that nature. The question is what that nature is, and what emerges from reflection on the perfectionist tradition is that many thinkers have seen the acquisition of knowledge, social relationships, and so on as both characteristically human and aspects of the human good, independently of their relation to consciousness, pleasure, or moral excellence. Once again, we see an alternative to Sidgwickian hedonistic happiness and moralist perfection emerging as a candidate for ultimate end, an end which might quite plausibly be grounded in a form of perfectionism which avoids any ‘is’/‘ought’ objection through its reliance on a self-standing normative premise linking the good with human nature.

37 What about self-realization? Sidgwick postpones discussion of this until the following chapter (1.7.1.2-8/89-91), which concerns different interpretations of egoism (80 n.1). Sidgwick begins by noting the Hobbesian and Spinozan versions of egoism as a principle of self-realization. Sidgwick suggests that the intellectualism implicit in Spinoza’s conception would seem quite implausible to someone who saw the human good as consisting in action, the creation of artistic beauty, or obedience to reason or conscience. Egoism cannot be based on an appeal to those impulses which our self finds prominent, since any such impulses could meet this criterion. But it might be said that we should exercise our faculties and so on in ‘due proportion’. There are two ways in which this proportion might be understood. It might be understood to refer to whichever combination of tendencies we were born with; so we might be urged to ‘be ourselves’. But this can be plausible only as an instrumental recommendation. If adhering to our original nature made us very unhappy, no one would recommend doing so. Alternatively, the due proportion in question can be seen as merely another term for perfection.

38 Sidgwick’s conclusion is that the notion of ‘self-realization’ is too vague to be of service in philosophical ethics. This comes as something of a surprise, since he himself must have thought he had gone some way to clarifying the position in these pages of chapter 7. On the face of it, what Sidgwick should have concluded is that self-realization theories are really versions of moralistic or non-moralistic perfectionism; and he might then have referred the reader to his criticisms elsewhere of, respectively, common-sense morality and the claim that there are goods independent of pleasure.

5. Ought

39 Sidgwick seeks his answer about what we ought to do through examining the different methods and ultimate principles to be found in common-sense morality. So, to that extent, he is himself moving from sociological description to normative ethics, though his justification for the sociology is itself grounded in ethical epistemology. He is careful to distinguish the ultimate ought-judgements he is interested in from what Kant calls ‘hypothetical imperatives’ (1.1.4.1-3/6-7). These imperatives apply only to those who have already accepted the relevant end. If my art-teacher tells me I ought to hold my brush differently, she is saying this on the assumption that I want to improve my technique. If I’m in the class just to get out of the house, however, her ought-judgement no longer applies to me. Contrast that with the judgement that anyone ought to care about her own happiness. This is a categorical judgement; whether one wants to be happy or not is irrelevant.

40 If the ultimate ends I ought to pursue are independent of my wants or desires, it is tempting to follow Kant and see them as ‘prescribed by Reason’ and the motive to act through recognition of them as different from non-rational desires (1.3.1-8/23-8). If we do so, we shall be denying a Humean account of reason as motivationally inert. Sidgwick’s enquiry into this question is typically empiricist: he turns to introspection, and describes his own experience — which he plausibly believes to be widely shared — of what we mean by a conflict between reason and desire, mentioning as an example the bodily desire for indulgence judged to be imprudent. His denial of the Humean position is itself couched in terms of his earlier distinction between ethics and positive science:

41

I hold that this is not the case; that the ordinary moral or prudential judgments which, in the case of all or most minds, have some -- though often an inadequate -- influence on volition, cannot legitimately be interpreted as judgments respecting the present or future existence of human feelings or any facts of the sensible world; the fundamental notion represented by the word ‘ought’ or ‘right’ which such judgments contain expressly or by implication, being essentially different from all notions representing facts of physical or psychical experience. (1.3.1.4/25)

42 Sidgwick’s argument against Hume here is somewhat problematic. Hume’s position need not be understood as the view that moral judgements themselves must be understood as judgements about sensible facts. Rather, they may be seen as expressions of ‘calm passions’, which themselves are not based on any rational judgement. The Humean, then, might agree that the word ‘ought’ has a different sense from any set of mere assertions of facts, but deny Sidgwick’s suggestion that our moral phenomenology is best understood in rationalist terms as the cognition of normative truth. At this point, it has to be admitted, Sidgwick’s discussion appears lacking in contrast to the sophisticated conceptual and logical advances made in twentieth-century metaethics.

43 Nevertheless, Sidgwick’s discussion of these issues does anticipate several later developments. In further criticizing the view that ‘rightness’ is always to be understood hypothetically, as an attribute of means that ‘ought’ to be taken to some end, Sidgwick notes that we cannot understand Bentham, despite his own suggestions to the contrary, as claiming that the word ‘right’ means ‘conducive to the general happiness’. [24] For his ethical view would then be a mere tautology.

44 He does also come closer to confronting the Humean position more straightforwardly in his objections to the view that moral judgements such as ‘Truth ought to be spoken’ are equivalent to factual judgements such as ‘The idea of truthspeaking excites in me a feeling of approbation’. Sidgwick admits that such a feeling will ordinarily accompany a moral judgement, but finds the metaethical position itself to be ‘absurd’, because it cannot explain why we take ‘Truth ought to be spoken’ and ‘Truth ought not to be spoken’ as contradicting one another. The suggestion that, on reflection, all I will affirm when I make a judgement such as ‘the air is sweet’ is ‘the air pleases me’ may well be true in certain cases. But it is not so with the moral emotions: ‘The peculiar emotion of moral approbation is, in my experience, inseparably bound up with the conviction, implicit or explicit, that the conduct approved is ‘really’ right---i.e. that it cannot, without error, be disapproved by any other mind’ (1.3.1.7/27). Sidgwick’s argument against Humeanism, then, is ultimately an appeal to phenomenology of a kind which would undergird many defences of cognitivist positions in the century to follow. He goes on to provide an example of the distinction he has in mind between a rational ethical cognition and the feeling of moral approval. Consider a case of someone who habitually feels a sentiment in favour of veracity and then becomes convinced that, in certain circumstances, she ought to lie. The feeling of repugnance she may feel when lying seems quite different from that which led to her accepting the principle of veracity in the first place. [25] A similar example enables us to dispose of the suggestion that moral judgements concern not the individual’s own moral approbation but that of her society: consider the case in which someone comes to a moral view in opposition to that generally held in her society.

45 Sidgwick then turns to a metaethical position which did not survive into the twentieth century, according to which moral judgements are claims about the pain that the agent will suffer if she fails to do what the judgement requires (1.3.2.1-4/29-31). Sidgwick notes — quite correctly, I believe — that there is something plausible in this position, namely, the analogy on which it rests between law and morality. In the case of positive law, we might well claim that some law is not genuinely established in a society unless its violation results, or may result, in punishment. But in the case of morality, not only will we distinguish ‘P ought to ?’ from ‘P will be punished by public opinion if he fails to ?’, but we will sometimes make that former claim knowing that no such punishment will occur. In general, reflection will demonstrate the distinction between duties such as those of mere politeness, which exist purely as a result of their imposition by public opinion, and strictly ethical duties. And in the same way, those who hold a divine command view of ethics will see the claim ‘P ought to ?’ not as equivalent to ‘God will punish P for not-?-ing’ but as providing the grounds for God’s punishment.

46 In his own positive account (1.3.3.2/33), Sidgwick distinguishes between ‘ought’ in a narrow sense, in which ‘ought’ implies ‘can’, and a broader sense in which I might think, for example, that I ought to know what a wiser person would know, even though I am quite aware that I am unable to will myself into that position here and now. In this latter sense, ‘ought’ implies some ideal which I ought, in the narrow sense, to approximate as far as possible. But Sidgwick goes no further than this. Indeed he suggests that any reductive account of ‘ought’ will fail, since the notion underlying it and other terms such as ‘right’ is ‘too elementary to admit of any formal definition’ (1.3.3.1/32). [26] This is not to deny that we might be able to explain how that notion emerged. But its content is not itself to be understood in terms of that explanation: ‘I know of no justification for this transference of the conceptions of chemistry to psychology’. All we can do to elucidate ‘ought’ is to explain its relation to certain other notions. [27]

47 Let me end this section by returning to the distinction between categorical and hypothetical imperatives. In chapter 1, Sidgwick appears to wish to avoid confrontation with Humeans. At 1.1.4.6/8-9, he claims that his discussion of ethical methods does not in fact depend on the idea that there are rational cognitions of ultimate ends (see also 77 n.1). All that is required is ultimacy, whatever the source of the end. Here we see a recognition in Sidgwick of the independence of at least some central issues in metaethics from normative or first-order ethics. Humeans, then, are permitted categorical imperatives, as long as they are not characterized as rational.

48 Later, however, Sidgwick argues (against a Humean egoist) that, even if an ultimate end is not prescribed by reason, there is still a rational requirement to take the means to that end, which even a Humean must accept. [28] When your doctor says, ‘If you wish to be healthy, you ought to get up early’, she is not saying merely, ‘Getting up early is a necessary means to your being healthy’. Her statement implies the unreasonableness of refusing to take the means to some acknowledged end. Sidgwick’s thought here seems to be that the Humean must accept that there is a categorical requirement to take the means to one’s end. [29] And if there is this categorical requirement, then why should there not be categorical requirements to adopt certain ends in the first place?

49 Humeans, however, need not accept any rational requirement to take the means to one’s ends. [30] Just as there is nothing unreasonable in preferring the destruction of the world to the scratching of my finger, so there is nothing unreasonable about not taking the means to one’s ends. Indeed, it can plausibly be argued that this was Hume’s own view, since the means-end ‘unreasonableness’ he speaks of in the Treatise (2000: 2.3.3.6) is merely a mistaken judgement about causes and effects. As he puts it: ‘a passion must be accompany’d with some false judgment, in order to its being unreasonable; and even then ’tis not the passion, properly speaking, which is unreasonable, but the judgment.

6. Motivation

50 Sidgwick sees ethical cognition and motivation as closely linked. Near the beginning of the Methods (1.1.3.1-6/4-6), he raises the question why, if the aim of ethics is to tell us what ought to be and in particular what we ought to do, philosophers should concern themselves at all with epistemology. After all, physicists get on quite well without discussing the faculty of sense-perception.

51 Sidgwick’s first answer to this question appeals to the practical aim of ethics. [31] In the case of sense-perception, we cannot help believing what we see to be true, whereas we often end up doing what we know to be wrong on the basis of irrational desire. So it is important to understand the relation between moral judgement and the will.

52 He then goes on to note the frequency with which people ask: ‘Why should I do what I see to be right?’. One might see the question as empty, since in the end it can be answered only by reference to some further ethical principle about which the same question can be asked. But we also need to explain why the question is so persistent. One explanation is that, since we are moved to act by desires independent of moral judgement, those asking this question are seeking an answer that in itself will prompt an overriding desire to do the action. Sidgwick allows that this explanation is sometimes true. But he also claims that, often, when people ask this question they assume that they are already determined to do whatever argument shows them to be reasonable, even if it goes against their non-rational inclinations. The main explanation for people’s asking the question so often is the variety of mutually inconsistent principles adduced in answer to it.

53 Sidgwick believes that the nature of this moral determination or ‘moral sentiment’ does not matter in ethics (1.1.6.2/76-7). What does matter is its operating successfully, even if it is in fact, say, a desire for the pleasure of acting rightly or to avoid the pains consequent on acting wrongly. This suggests that what makes a moral sentiment rational is not so much its content, but its origin as a response to a rational judgement.

54 Later (1.3.3.4/34-5), Sidgwick strengthens his claim about the link between moral judgement and motivation. When I cognize or judge ‘I ought to ?’, he suggests, if I am using ‘ought’ in the narrower sense in which it implies ‘can’ and if I am rational, then this cognition will produce a motive to action — though it is not always overriding. [32] He goes on:

55

In fact, this possible conflict of motives seems to be connoted by the term ‘dictate’ or ‘imperative’, which describes the relation of Reason to mere inclinations or non-rational impulses by comparing it to the relation between the will of a superior and the wills of his subordinates. This conflict seems also to be implied in the terms ‘ought’, ‘duty, ‘moral obligation’, as used in ordinary moral discourse: and hence these terms cannot be applied to the actions of rational beings to whom we cannot attribute impulses conflicting with reason. We may, however, say of such beings that their actions are‘reasonable’, or (in an absolute sense) ‘right’.

56 Sidgwick appears to be assuming that a command can be made (a) only by a superior to an inferior, and (b) only when that inferior may be motivated to act in a way that conflicts with that command. This is questionable. Consider modus ponens. Many will wish to say that, if anyone (regardless of their ‘status’ as superior or inferior) believes both ‘If p, then q’ and ‘p’, they are required by reason to believe ‘q’. Whether they have any disposition or even ability to believe otherwise seems irrelevant. The laws of logic, then, apply to all beings, so that Sidgwick’s distinction between ‘ought’ and what is ‘ “reasonable” or (in an absolute sense) “right”’ is unnecessary. Even if God is superior to all other beings, knows what is right, and cannot help but do what is right, we may, if we wish, say that he ought to act in this way or that he has a duty to do so.

57 But this raises the question whether we might not avoid the notion of ‘ought’ altogether, and prefer Sidgwick’s broader notion of what is reasonable. In general, Sidgwick seems to retain more basic common-sense moral concepts than he needs, and as a result has to spend time elucidating and defining them, and plotting their mutual relations. If he were to expunge ‘ought’ and related notions from his moral vocabulary, he would avoid having to go into issues such as whether ‘ought’ is to be understood in terms of punishment. Sidgwick’s own understanding of ethics is stated clearly at 1.6.1.2, where he begins to summarize the results of the three preceding chapters: ‘The aim of Ethics is to systematise and free from error the apparent cognitions that most men have of the rightness or reasonableness of conduct, whether the conduct be considered as right in itself, or as the means to some end commonly conceived as ultimately reasonable’. Indeed even here the reference to ‘rightness’ can be eliminated, on the assumption that the ‘or’ is epexetic. As Schneewind puts it (1977: 228), for Sidgwick, ‘ [t] he point is to determine how we are to decide what it is ultimately reasonable to do, regardless of whether the ultimately reasonable is called “moral” or not’. [33] Parsimony requires any philosopher to ask her questions using as few concepts as possible, ceteris paribus, and indeed to introduce new concepts into her answers to those questions only as required. It may have turned out that Sidgwick could not state his own position without using the notion of ‘ought’. He might, for example, have ended up as a ‘dogmatic intuitionist’ with a list of moral duties. But the views he himself finds most plausible — egoism and utilitarianism — are best stated in terms of ultimate reasonableness, with no reference to morality at all. It is, of course, true that Sidgwick’s method will require him to mention (rather than use) the word ‘ought’ in explaining non-philosophical forms of intuitionism. But he could have pointed out the apparent flaws in that position without any serious enquiry into the logic of the word ‘ought’.

58 Let me now return to Sidgwick’s view of the relation between rational cognition and motivation. In one sense, Sidgwick might agree with Hume that reason is inert, in so far as it does not motivate even a rational agent on its own. [34] Rather, it gives rise to a sentiment which may then lead the agent to act. But in rational beings ‘as such’, recognition of a dictate of reason will produce such a motive. Sidgwick holds, then, some form of what has come to be known as motivational internalism, according to which there is a close link between ethical cognition and motivation. But it is, I suggest, a weak form of that view. In the summary at the beginning of 1.6, Sidgwick claims that apparent cognitions (‘apparent’, because they may be mistaken and hence not knowledge) are normally accompanied by moral sentiments. This suggests that he allows for cases in which a cognition — perhaps even a genuine one — does not give rise to any motivation. But Sidgwick will describe this as a case of irrationality, in so far as it is not an appropriate response by a rational being ‘as such’ to a rational demand. [35] This interpretation finds further support from Sidgwick’s claim at the end of 1.6.1.2/78 that, in a case in which I believe I ought to ? but lack any motive to do so, I cannot ordinarily see that lack of motive as a reason for not ?-ing.