Replies to Crisp, Shaver and Skelton
PlanHaut de page
1It is a great privilege to have one’s work critiqued by such a distinguished trio of philosophers and Sidgwick scholars. I owe further debts to Anthony and Rob, who were the OUP referees for my book. As will have been quite evident from the preceding discussion, they would not want to be held responsible for the book’s detailed contents, on which they gave me much excellent commentary. But, in thanking them here, I do want to say in particular that it seems to me the published version is much better shaped than the first draft was, and that that improvement is the direct product of their sage advice.
2Between them, Roger, Anthony, and Rob raise more issues than I can possibly address in the space available. Roger asks really nice questions about all four of my substantive chapters, on metaethics, moral epistemology, utilitarianism versus dogmatic intuitionism, and utilitarianism versus egoism. Rob rejects my central conclusions about utilitarianism versus egoism. Anthony raises very important issues particularly in moral epistemology.
3I will have to be somewhat selective in my responses. But I will try to say at least something about all the Chapters on which my commentators remark, while devoting most of the space to the two Chapters to which Roger, Anthony, and Rob devote the most critical attention: the moral epistemology chapter and the utilitarianism versus egoism chapter.
- 1 Michael Smith, The Moral Problem (Oxford: Blackwell, 1994), esp. pp. 63-66.
4Begin with Roger’s first question, about whether Sidgwick is really committed to what I call (drawing on language of Michael Smith’s) the “realist conceptual thesis”.1 Roger says “a presupposition does not have to be conceptual…what [Sidgwick] is rejecting is not a particular understanding of the concept of morality, but a conception of it.”
5I suppose that the crucial passages for adjudicating this interpretive issue are those in which Sidgwick is rejecting a form of moral sense theory, the “suggestion” (as he puts it) that
- 2 Henry Sidgwick, The Methods of Ethics, 7th edition (Indianapolis: Hackett, 1981; original publicat (...)
The judgments or propositions which we commonly call moral…really affirm no more than the existence of a specific emotion in the mind of the person who utters them. (ME 26)2
6The question is whether Sidgwick takes this suggestion to express an alternative conception of the moral, or rather whether he takes the suggestion to fail to capture the concept of the moral at all. I am still inclined to read him in the latter way, inspired by passages like the following:
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. If I give up this conviction…I may no doubt still retain…a sentiment of repugnance to the opposite conduct: but this sentiment will no longer have the special quality of ‘moral sentiment’ strictly so called. (ME 27)
- 3 J. L. Mackie, Ethics (Harmondsworth: Penguin, 1977), p. 35.
7There is still the further question whether Sidgwick regards the claim that there is such a thing as moral truth and error as a presupposition or as “incorporated in the basic, conventional, meanings of moral terms” (to use Mackie’s language).3 But I don’t think Sidgwick regards the claim as merely part of something as optional as (one) conception of the moral.
- 4 G. E. Moore, Principia Ethica (Cambridge: Cambridge University Press, 1903), esp. Chapter 1.
- 5 Allan Gibbard, Wise Choices, Apt Feelings (Cambridge, MA: Harvard University Press, 1990), esp. p. (...)
8A second, very intriguing, question Roger raises is about whether Sidgwick’s argument for non-naturalism is really any different from Moore’s.4 Alternately put, it’s the question: what is the open question argument? In writing about this in my metaethics Chapter I adopted, probably too uncritically, a contrast Gibbard draws in Wise Choices, Apt Feelings between, on the one hand a “master argument” against all naturalistic definitions and, on the other hand, (piecemeal) “attacks on specific examples of naturalistic analysis”.5 Is there really any difference here?
9I still think that there is; but I do think the way I put it made the difference look too stark. The difference would be a matter partly of how much argumentative work is done by the initial intuition that the factual is one thing, the normative another, and partly a matter of just how the arguments against specific proposed naturalistic analyses are supposed to go. On the master argument view, the initial intuition would be close to decisive. By contrast, on the piecemeal arguments view, the initial intuition would establish a presumption, but would be much less close to decisive. And the argumentative resources used to supplement the initial intuition would also be different. On the master argument picture, (to do, necessarily roughly, something that takes a good deal of time and care to do precisely) for any proposed name of a natural property, “N” we have to just ask “Is A, which is N, good?” to see that N is not the same as goodness. By contrast, on the piecemeal argument picture, more work is required. We cannot simply appeal to a single, generic, all-purpose thought experiment. We must instead bring out specific, and different, problems with specific different naturalistic proposals. One example of such a specific objection to a specific naturalistic proposal would be the objection Sidgwick makes to moral sense theories a little earlier in the characteristically succinct and nuanced discussion I quoted from above: that at least their simple versions get it wrong about when people disagree:
It is absurd to say that a mere statement of my approbation of truth-speaking is properly given in the proposition ‘Truth ought to be spoken’; otherwise the fact of another man’s disapprobation might equally well be expressed by saying ‘Truth ought not to be spoken’; and thus we should have two coexistent facts stated in two mutually contradictory propositions. (ME 27)
- 6 Henry Sidgwick, “The Establishment of Ethical First Principles,” Mind Vol. 4, No. 13 (1879), pp. 1 (...)
10One central claim in my moral epistemology chapter is that there is a puzzle in Sidgwick’s moral epistemology. The categories necessary to frame the puzzle come from Sidgwick’s short 1879 paper, “The Establishment of Ethical First Principles”.6 He there distinguishes two general kinds of solution to the problem faced by the proponent of one putative ethical first principle who wants to give others “rational inducement” to accept this first principle. In the first solution:
I may begin by regarding some limited and qualified statement as self-evident, without seeing the truth of the simpler and wider proposition of which the former affirms a part; and yet, when I have been led to accept the latter, I may reasonably regard this as the real first principle, and not the former, of which the limitations and qualifications may then appear accidental and arbitrary. (EE1P 106)
11I call arguments of this first kind “bipartite ad hominem arguments,” or “bipartite arguments” for short.
12Sidgwick then goes on to characterize a second, “quite different”, way out of the moralist’s dilemma:
We may be able to establish some general criteria for distinguishing true first principles (whether ethical or non-ethical) from false ones; and may then construct a strictly logical deduction by which, applying their general criteria to the special case of ethics, we establish the true first principles of this latter subject. (EE1P 107)
13I call arguments of this second kind “criterial arguments.”
14With these concepts and labels available, we can articulate the puzzle. It is raised by the conjunction of three claims: (a) In Book III of the Methods, Sidgwick develops a criterial argument for utilitarianism; (b) In general Sidgwick thinks of criterial arguments as different from and more powerful than bipartite arguments; (c) In Book IV, in discussing the proof of utilitarianism, Sidgwick considers only bipartite arguments, not independent criterial arguments. The puzzle then is: why? Why does the independent criterial argument for utilitarianism disappear in Book IV?
15I go on to distinguish three possible solutions to the puzzle: (i) Sidgwick ignores the independent criterial argument for utilitarianism in Book IV because in Book IV, as against Book III, he recognizes that the independent criterial argument for utilitarianism doesn’t work; (ii) He ignores the independent criterial argument in Book IV because features of the rhetorical context of IV II, absent in Book III, make only bipartite arguments seem relevant – what are required in the context of Book IV are arguments addressed to specific opponents; (iii) He ignores the independent criterial argument in Book IV because he takes special features of common-sense morality to make the bipartite argument against the dogmatic intuitionist much more compelling than bipartite arguments are in general. I argue tentatively for solution (i).
16One thing I am concerned to do is to argue that there is a puzzle. To that extent, if my commentators agree that there is a puzzle, but defend a particular kind of solution to it, that is fine with me. But I do also argue tentatively in favor of a particular solution to the puzzle, namely solution (i). Roger argues instead in favor of solution (ii). He says “As I understand the context of IV II, Sidgwick does indeed see the criterial argument as having failed – but only rhetorically, not as an argument in itself.”
17I am less happy with this solution than Roger is. I am not persuaded that the rhetorical context of IV II makes criterial arguments irrelevant. Let me try to make, or to elaborate on, that point in two ways. First, consider again “The Establishment of Ethical First Principles”. I don’t see the context set there as different from the context set by Methods IV II. In that context Sidgwick treats independent criterial arguments as relevant and available. But in the context of Methods IV II he ignores them. Second, something like a criterial argument seems to be part of the complex bipartite argument characterized in IV II: after all, what is supposed to be on offer to the interlocutor is not merely a line of argument which “allows the validity, to a certain extent, of the maxims already accepted”; it is also an argument which “shows them to be not absolutely valid, but needing to be controlled and completed by some more comprehensive principle”. If a criterial argument is admissible as part of a bipartite argument in the context of IV II, how can an independent criterial argument not be admissible in that same context?
18Rob defends an interesting and different solution to the puzzle. Since it depends on his distinctive view about Sidgwick and the argument for utilitarianism, I will follow his lead and comment briefly on it after discussing that argument.
19A second epistemological matter on which Roger and Anthony both press me is the interpretation of the 4th of the criteria. Sidgwick does not (to my mind, anyway) give in the 7th edition of the Methods a short, illuminating formulation of this criterion (unlike the other three, for which the first sentence of his presentation in III XI provides such a formulation). Evidently most unwisely, I supply for him (inspired by EE1P) the following one sentence formulation “The proposition must be universally accepted”. Both Roger and Anthony argue that the criterion is really a negative one. Nothing nearly as strong as universal acceptance is required. Instead, what there must not be is rationally intractable disagreement with epistemic peers.
20I agree with them about the proper interpretation of the criterion, and regret my formulation of it. It is not that there is no textual support in the Methods for putting the criterion as I do. The first sentence of the passage introducing the criterion in the 1st edition of the Methods includes the following
Any defect in the universal acceptance of a proposition must pro tanto impair our confidence in its validity. (ME1, 320-321)
21But the phrasing is different by the 7th edition, and I agree that Roger and Anthony are right on balance to attribute the weaker version they articulate to Sidgwick. I am not yet persuaded, however, that anything I argue depends problematically on the stronger and mistaken interpretation. In arguing to the contrary, I take Anthony to be suggesting that meeting a negative condition cannot contribute to justification. So if the fourth criterion is read as negative rather than positive, the view that Sidgwick regards satisfaction of the conditions as contributing to justification is to that extent weakened.
22A third epistemological issue, on which both Anthony and Roger press me, is the conception of criterial argument, and whether criterial arguments involve granting common-sense morality probative status. Anthony in particular suggests that I end up with the puzzling view that “Sidgwick grants probative status to common-sense morality in the context of his criterial argument but not in the context of the bipartite argument.” I agree entirely that this would be puzzling, and it is certainly not what I want to say. Let me try to argue that I can avoid saying it.
23First, on (the conception I label) “criterial argument”: I take the general idea of criterial argument to have very broad scope. Consider again in this connection the passage from EE1P which introduces that idea:
We may be able to establish some general criteria for distinguishing true first principles (whether ethical or non-ethical) from false ones; and may then construct a strictly logical deduction by which, applying their general criteria to the special case of ethics, we establish the true first principles of this latter subject. (EE1P 107; italics added)
24In this passage (which, it is worth emphasizing, is from EE1P, not from the independent epistemological writings) Sidgwick clearly articulates a broad conception of criterial argument. The criteria invoked here give no special role to common-sense morality because they give no special role to morality at all. They can, after all, be applied to putatively self-evident non-moral as well as moral claims; and when they are applied to non-moral claims there is no reason to expect moral convictions to play any special evidential role. Criterial argument is a general strategy; its application to ethical first principles is a special case. Thus the general conception of criterial argument gives no special probative role to common-sense morality.
25Now, when criterial arguments are employed to try to establish the first principles of ethics, then appeal to the criteria – especially the fourth, discussed above – seems to give an epistemic role to common-sense convictions about ethics. For suppose epistemic peers have convictions about ethics which conflict with some ethical claim we initially take to be self-evident. Then, the 4th criterion, read as Anthony and Roger suggest, means that the self-evidence or justification of that claim is reduced or eliminated. But I persist in wanting to distinguish the set of epistemically important common-sense convictions about ethics in general from the specific subset of such convictions which are associated with one particular method – dogmatic intuitionism. Anthony urges that this distinction I want to make – between common-sense morality in the narrow and common-sense morality in the broad sense – has insufficient grounding in the text of the Methods. I persist in thinking, to the contrary, that it is adequately grounded in the very first chapter of the book, and that it is a distinction you really have to make once you note that Sidgwick thinks that each of his three methods is “implicit in our common moral reasoning” (ME 14). Putting together these two points– that criterial argument is a general strategy of argument whose application to ethical first principles is only one special case, and that, even when it is applied to ethics, the common-sense moral convictions that get a probative role are those of common-sense morality only in the broad sense – I still think I can reasonably deny that Sidgwick’s conception of criterial argument gives any special evidential role to common-sense morality in the narrow sense.
26Second, on bipartite argument: My view here is that the general conception of bipartite argument no more gives a special role to common-sense morality in the narrow sense than does the general conception of criterial argument. Of course a bipartite argument addressed to the dogmatic intuitionist (the proponent of common-sense morality in the narrow sense) will give a special role to the convictions of the common-sense moralist. But in just the same way a bipartite argument addressed to the egoist will give a special role to the convictions of the egoist.
- 7 David Brink, “Common Sense and First Principles in Sidgwick’s Methods,” Social Philosophy and Poli (...)
- 8 C. D. Broad, Five Types of Ethical Theory (London: Kegan Paul, 1930), p. 206.
27So I do not think that either the general conception of criterial argument or the general conception of bipartite argument gives any special epistemic role to common-sense morality in the narrow sense. As I see it, if there is a remaining problem here the problem is not that I am committed to a puzzling view about the different roles of common-sense morality in the narrow sense in criterial as against bipartite arguments. It is rather a problem that Anthony and I both face – that, as Roger urges, there are passages in the Methods which seem to give an epistemic role to common-sense morality in the narrow sense over and above anything either Anthony or I wants to allow. I claim that the idea that Sidgwick gives common-sense morality an evidential role is “largely mistaken,” and I have basically three tactics for (largely) explaining away these problematic passages: (a) Some such passages (including a number cited by Brink)7 show only that common-sense morality has an evidential role in the context of bipartite arguments addressed to the common-sense moralist. But that does not show that common-sense morality is independently evidential; (b) Remaining passages may reflect Sidgwick’s gradual and incomplete separation of the epistemic from the moral-theoretic components of intuitionism, about which C.D. Broad complained when he introduced the term “deontology” in its standard contemporary sense;8 (c) As above, we need to distinguish a narrow from a broad sense of “common-sense morality”. Common-sense morality in the narrow sense is just that part of ordinary moral thinking systematized as dogmatic intuitionism. Common-sense morality in the broad sense includes also those parts of ordinary moral thinking systematized as utilitarianism and as egoism. The fourth criterion gives an epistemic role to common-sense morality only in the broad sense.
28Roger is not convinced in particular by my second tactic, and suggests that, given Sidgwick’s formidable intellect and multiple careful revisions of the Methods, it is “more plausible that [Sidgwick’s] epistemology includes both elements of foundationalism and a commitment to some kind of Aristotelian dialectic”. By contrast Anthony (I take it) is committed to the claim that Sidgwick does not assign an evidential role to common-sense morality, but, as I noted above, is skeptical in particular of my third tactic of distinguishing narrow and broad senses of “common sense morality”.
29 To be really honest, I am not sure I am entirely convinced myself; I do think some commentators, including Brink, have exaggerated the textual evidence for the claim that Sidgwick takes common-sense morality to be independently evidential and that one has, for reasons including those my tactics bring out, to be quite careful with this textual evidence; and I think one could argue that Sidgwick ought not to assign an evidential role to common-sense morality. But I agree that it is tough to make a fully convincing textual case (a case which, again, I take it Anthony is committed to making) that Sidgwick does not assign any privileged evidential role to common-sense morality. In saying that Brink’s idea that Sidgwick assigns an evidential role to common-sense morality is only “largely” (as opposed to “wholly”) mistaken,” I take myself to have a bit more wiggle room than Anthony does. But, again, I take both of us to have plenty of work to do to be fully convincing on this issue.
- 9 Anthony Skelton, “Sidgwick’s Philosophical Intuitions,” Ethics and Politics Vol. 10, No. 2 (2008), (...)
30I want to offer just two observations about the issues covered in this Chapter. First, there is the question: how many self-evident principles does Sidgwick think there are? Is it just the (probably) four articulated in III XIII? Is there also some further self-evident principle connected to egoism articulated outside III XIII? And is Hedonism, defended by Sidgwick in III XIV, taken by him to be self-evident in the same way as the principles he articulates in III XIII? Rob thinks the only self-evident principles are those in III XIII. I think there is an additional apparently self-evident principle connected to egoism articulated outside III XIII. Roger urges here, and Anthony has argued elsewhere,9 that Sidgwick thinks hedonism too is self-evident. I (briefly) argued on the other side in the book, but for the record I am increasingly persuaded that Roger and Anthony may be right about this. And certainly if, as I do, one thinks that there is a self-evident principle connected to egoism articulated outside III XIII, one cannot go on to object to Roger and Anthony’s view about Hedonism on the grounds that all the self-evident principles to which Sidgwick is committed are articulated in III XIII.
31Second, I think Roger is right that Sidgwick fails to recognize the important role played by the capacity for practical judgment in any plausible moral theory. Ross’s view features both the key concept of prima facie duty and a key role for practical judgment in balancing competing prima facie duties in specific cases. I emphasized the former; but I agree that I could profitably have said considerably more about the latter.
32Rob wants to reject my overall verdict on the relative success of Sidgwick’s central normative claims, about utilitarianism versus dogmatic intuitionism and about utilitarianism versus egoism. I give two versions of the verdict. One is that Sidgwick’s treatment of the conflict between intuitionism and utilitarianism is less successful than he supposes, and his treatment of the conflict between utilitarianism and egoism is more successful. Rob, I take it, agrees with the first half of that version: he agrees that Sidgwick’s critique of dogmatic intuitionism is subject to an unfairness objection, and hence is less successful than Sidgwick thinks it is. The other version of my verdict focusses specifically on the arguments for utilitarianism and egoism respectively discussed in my final chapter. Here, I claim that the argument for utilitarianism is a failure while the argument for egoism is more successful. Rob disagrees with both halves of that version: he thinks there is more to the axiomatic argument for utilitarianism than I allow, and nothing important and correct in the argument for egoism. Let me focus on those two disagreements in turn.
33I claim that the axiomatic argument for utilitarianism is a failure. In summarizing my take on the argument, let me distinguish provisionally between my overall picture of the argument’s character and success, and my detailed verdict on just which claims in the argument are non-tautological self-evident principles:
34In my view the argument is supposed to invoke principles implied in the notion of “universal good”. This notion is supposed to yield two self-evident principles:
35(U) The good of any one individual is of no more importance from the point of view…of the Universe than the good of any other.
(R) As a rational being I am bound to aim at good generally…not merely at a particular part of it.
(U) and (R) together are supposed to entail
(B) Each one is morally bound to regard the good of any one individual as much as his own.
I take (B) to express a form of impartial consequentialism, such that by adding hedonism to it we get a form of utilitarianism.
36I take there to be two central problems with the argument, one of which Sidgwick acknowledges, the other of which he does not. The problem Sidgwick acknowledges is that the argument begins with the supposition that there is such a thing as universal goodness or goodness period: that there are (to use modern terminology) agent-neutral reasons. As Sidgwick famously says, egoists may deny that there are any such reasons, or, alternately put, that there is any such thing as universal goodness. The problem that Sidgwick does not acknowledge is that the existence of agent-neutral reasons does not entail that that is all the reasons there are. There may be both agent-neutral and agent-relative reasons.
37Rob dislikes, or anyway finds odd, the way I put the first of these problems. He says
Surely there is something I ought to desire, assuming myself to have an equal concern for all existence…It seems preferable to say that the egoist is uninterested in this thing – or, as Sidgwick says, refuses to take up the point of view of the universe.
38I still want to defend my way of putting it, and, for my part, I find something missing in Rob’s suggested alternative phrasing. What is missing is the authority of genuine oughts or reasons. If there is such a thing as genuine goodness, if there are agent-neutral reasons, the egoist cannot opt out of them because he is uninterested in them. They bind even those who are uninterested in them. That is why it is a crucial issue whether there is any such thing as genuine goodness, and an issue on which the clear-sighted egoist has to take a stand.
39My central claim is that Sidgwick cannot properly get from his starting point to his conclusion, in crucial part because it cannot be established conceptually that there are authoritative or genuine agent-neutral reasons. This central claim leaves it open just which, if either, of (U) and (R) are really non-tautological self-evident principles. Let me try to lay out as clearly as I can my view about the status of (U) and (R): does Sidgwick think they are tautological? Should he? Is he right?
40As I see it both (U) and (R) look problematically tautological. But, given the setup in III XIII, Sidgwick is committed to thinking that at least one of them is non-tautological. Given the choice, put this way, I argue it is (U) that Sidgwick should see as non-tautological. That is because, I claim, Sidgwick’s views about the concept of goodness make (R) come out tautological: because, for him, to think something good is simply to think that there is a reason to aim at it. But I think also that Sidgwick would still be wrong: that, properly conceived, both (U) and (R) are tautological. To argue for impartial consequentialism you need to show that there are genuine agent-neutral reasons, and that these are the only genuine reasons that there are. I think that in deriving (U), Sidgwick equivocates between a self-evident but tautological claim (roughly: if there is genuine goodness, each person has some reason to promote the good) and a non-tautological but also non-self-evident claim (roughly: if there is genuine goodness, the only reason each person has is to promote the good). So the axiomatic argument for utilitarianism is a failure. Of the two things that Sidgwick needs to establish (that there are genuine agent-neutral reasons, and that these are the only reasons there are), he recognizes in the end that he does not establish the first, and takes himself to establish the second only by sometimes illicitly construing (U) in a way that makes it not self-evident.
- 10 “Sidgwick’s Axioms and Consequentialism,” The Philosophical Review 123 (2014).
41Rob has his own very interesting take on the axiomatic argument for utilitarianism.10 I didn’t have the opportunity to consider that take in the book. If I had to try quickly to compare his take to mine, I would say that his key moves are (a) to make the conclusion, (B) come out as a weaker claim than I make it; (b) to do some very ingenious fancy footwork to make (U) and (R) come out non-tautological and distinct, and (c) to deny that the 1st edition material characterizing the argument as “only [evolving] the suppression of Egoism…implied in the mere form of the objective judgment that ‘an end is good’” (ME 1, 364) accurately characterizes the 7th edition version of the argument.
42Once Rob’s interpretation is available, should we go with his or mine? On balance I still go for my reading; but I see the attractions of his. It is a positive feature of his reading that it makes (U) and (R) come out non-tautological and distinct. It is less crucial to my reading, I think, just which of (U) and (R) is, or is thought by Sidgwick to be, non-tautological. You would get something very like my overall picture of the argument’s character and success if you said that Sidgwick ought to think it is (R) that is non-tautological, rather than (U); and even if (like Broad) you thought (R) really was non-tautological. Rob’s reading strikes me as more charitable, though I still think mine emerges more straightforwardly from the text. If (as I think and Rob does not) it is a key issue where and how Sidgwick takes himself to have established that there is such a thing as genuine goodness, it is an advantage of my reading that it gives prominence to this issue. On the other hand, as Rob points out in the last paragraph of his comments, it is an advantage of his reading that it helps provide an attractive solution to the puzzle in Sidgwick’s moral epistemology. Specifically, as Rob sees it, the arguments of Book III and Book IV do different and important work. The argument of Book III establishes the axioms; but more is needed to secure utilitarianism, and that more (he can then say) is done by the argument of Book IV.
43Anthony, I take it, wants also to press the idea that the argument of III XIII is supposed to do less work than I think it is supposed to do. But Anthony wants to emphasize in effect not only the distinction between other-things-equal and all-things-considered versions of the axioms, but also the distinction between first principles and methods. The argument of Book III, especially III XIII is concerned with the utilitarian first principle; new and different work in Book IV, especially IV II, is needed to vindicate the utilitarian method.
44This again is a quite intriguing idea, which of course Anthony can only sketch here. I have two initial reservations about it. First, it seems to me the language Sidgwick uses in IV II often suggests he is (still) there concerned with establishing first principles, not with establishing methods. Thus the chapter begins
In book ii, where we discussed the method of Egoistic Hedonism, we did not take occasion to examine any proof of its first principle. (ME 418)
45And the language of “principles” and “first principles” is prominent throughout the chapter (occurring again on every page of IV II, and not in sentences suggesting that the current concern is with establishing methods rather than principles). So, I suggest, it is hard to read IV II as concerned with establishing methods rather than with establishing principles.
46Second, I have some more general reservations about the very idea of establishing a method, as distinct from establishing a first principle. I think I know what it would be to establish the first principle of utilitarianism: it would be (roughly) to establish that the only valid ultimate reasons there are are agent-neutral reasons to maximize pleasure and minimize pain. To determine what would be involved in establishing a method, I take it, one would need to decide what the relation was between methods and first principles, and to decide whether methods could be established independently of first principles, or only via the establishment of associated first principles. The part of the Methods most relevant to these questions are, I take it, in I VI. The following two passages seem to me particularly important:
What then do we commonly regard as valid ultimate reasons for acting or for abstaining? This, as was said, is the starting-point for the discussions of the present treatise: which is not primarily concerned with proving or disproving the validity of any such reasons, but rather with the critical exposition of the different ‘methods’ – or rational procedures for determining right conduct in any particular case – which are logically connected with the different ultimate reasons widely accepted. (ME 78)
Not all the different views that are taken of the ultimate reason for doing what is concluded to be right lead to practically different methods of arriving at this conclusion. Indeed we find that almost any method may be connected with almost any ultimate reason by means of some – often plausible – assumption. Hence arises difficulty in the classification and comparison of ethical systems; since they often appear to have different affinities according as we consider Method or Ultimate Reason. (ME 83)
47These passages suggest to me first that, tricky as it is to establish first principles (or “valid ultimate reasons for action”), establishing methods would be a whole lot trickier. So I am reluctant to portray IV II (or Book IV in general) as concerned with the establishment of methods rather than with the establishment of principles. And, second, these passages incline me to suggest, at least tentatively, that the only things Sidgwick is ever concerned with “proving” or “establishing” are first principles or claims about valid ultimate reasons; while Sidgwick really does devote much space to methods, none of that space is concerned with “proving” or “establishing” methods at all.
48I think Rob and I disagree more fundamentally about the egoistic side of Sidgwick than we do about the utilitarian side. On the utilitarian side we largely agree about which arguments are successful; our disagreements are about which to attribute to Sidgwick. By contrast on the egoistic side we disagree pretty fundamentally about whether an argument for something related to egoism is any good; and that philosophical difference is one key driver of our interpretive disagreements.
49 Let me try informally to characterize what I take to be the key insight in that argument. The insight is that the separateness of persons means that each of us has special reason to care about good and bad things that happen to him or her, reasons that we do not have to care about good and bad things that happen to other people. The fact that I directly experience pains and pleasures I have gives me reason to care about my pains and pleasures that others do not have. The way of putting this key insight I draw from Sidgwick is this:
(2) If the distinction between any one individual and any other is real and fundamental, then I ought to be concerned with the quality of my existence…in a sense, fundamentally important, in which I ought not to be concerned with the quality of the existence of other individuals.
50There are two important qualifications here: (a) that I have special reason to care about good and bad things that happen to me does not mean I have no reason to care about good and bad things that happen to others. So the insight does not really support egoism; (b) Rob is right that the insight is most compelling when what we think about are hedonic goods and bads; it is at least less compelling for non-hedonic goods and bads.
- 11 Roger Crisp, Reasons and The Good (Oxford: Clarendon Press, 2006), pp. 131-145.
51Rob discusses one of Roger’s handy 2-doors cases, which I appeal to and run a variation or two on at the end of my utilitarianism and egoism chapter.11 But I think the way Rob treats the case illustrates how deeply we disagree about this. He writes as if we would only come to contemplate accepting (2) to explain specific intuitions about some 2 doors case (or cases). I think rather that (2) is itself intuitively compelling prior to any application of it to a specific 2-doors case.
- 12 W. D. Ross, The Foundations of Ethics (Oxford: Clarendon Press, 1939), pp. 8-9. Compare Derek Parf (...)
52There is one part of Rob’s discussion here I find puzzling. He objects that endorsing (2) as a normative premise “violates Sidgwick’s rule that normative conclusions cannot be deduced from wholly non-normative premises. For (2) says that a wholly non-normative claim entails a normative claim.” I think this cannot be right. Non-naturalists are typically committed to the possibility of true claims connecting the natural to the normative. As W.D. Ross observes, “[the non-naturalist Hedonist holds] that rightness is something indefinable, and merely [claims] that that which makes acts right is their tendency to promote pleasure. [He holds]…that a non-ethical characteristic, a psychological characteristic, is the ground of rightness but not its essence.”12 Whatever status the non-naturalist hedonist takes hedonism to have, (2) can have just as well. It cannot be that the rule that normative conclusions cannot be deduced from wholly non-normative premises rules out all premises connecting the non-normative with the normative. And I can see no good reason not to regard (2), like hedonism, as just one possible such premise.
53Roger, Anthony, and Rob ask many more good questions, and raise many more fascinating issues and troubling objections, than I have been able to address here. Let me just close by thanking them all once again for their comments, and thanking Anthony in particular for organizing this symposium, and the ISUS conference session from which it derives.
1 Michael Smith, The Moral Problem (Oxford: Blackwell, 1994), esp. pp. 63-66.
2 Henry Sidgwick, The Methods of Ethics, 7th edition (Indianapolis: Hackett, 1981; original publication: 1st edition 1874, 7th edition 1907). I will refer to it as “ME”; page references will be placed in the text.
3 J. L. Mackie, Ethics (Harmondsworth: Penguin, 1977), p. 35.
4 G. E. Moore, Principia Ethica (Cambridge: Cambridge University Press, 1903), esp. Chapter 1.
5 Allan Gibbard, Wise Choices, Apt Feelings (Cambridge, MA: Harvard University Press, 1990), esp. p.11.
6 Henry Sidgwick, “The Establishment of Ethical First Principles,” Mind Vol. 4, No. 13 (1879), pp. 106-111. I will refer to it as “EE1P”; page references will be placed in the text.
7 David Brink, “Common Sense and First Principles in Sidgwick’s Methods,” Social Philosophy and Policy Vol. 11, No. 1 (1994), pp. 179-201.
8 C. D. Broad, Five Types of Ethical Theory (London: Kegan Paul, 1930), p. 206.
9 Anthony Skelton, “Sidgwick’s Philosophical Intuitions,” Ethics and Politics Vol. 10, No. 2 (2008), pp. 185-209.
10 “Sidgwick’s Axioms and Consequentialism,” The Philosophical Review 123 (2014).
11 Roger Crisp, Reasons and The Good (Oxford: Clarendon Press, 2006), pp. 131-145.
12 W. D. Ross, The Foundations of Ethics (Oxford: Clarendon Press, 1939), pp. 8-9. Compare Derek Parfit, On What Matters Vol. 2 (Oxford: Oxford University Press, 2011), pp. 303-305; see also ME 79.Haut de page
Pour citer cet article
David Phillips, « Replies to Crisp, Shaver and Skelton », Revue d’études benthamiennes [En ligne], 12 | 2013, mis en ligne le 10 décembre 2013, consulté le 28 mars 2017. URL : http://etudes-benthamiennes.revues.org/676 ; DOI : 10.4000/etudes-benthamiennes.676Haut de page
Droits réservésHaut de page