Metaphysics, Epistemology, Utilitarianism, Intuitionism, and Egoism: A Response to Phillips on Sidgwick
PlanHaut de page
- 1 David Phillips, Sidgwickian Ethics (Oxford: Oxford University Press, 2011). All unattributed page (...)
1The shape of contemporary ethics owes a great deal to Henry Sidgwick, through his influence on Rawls, Parfit, and others. No one who reads David Phillips’s outstanding book can be left in the slightest doubt about Sidgwick’s continuing significance for both metaethics and normative ethics.1 Phillips’s scholarship and his substantive arguments are powerful and insightful, and I find them largely persuasive. So in these remarks I intend merely to raise a few questions about each of his four main chapters, several of which may well amount to little more than requests for clarification.
21. Phillips attributes to Sidgwick what he calls the ‘realist conceptual thesis’ (12-14):
It is part of our moral concepts that there is such a thing as moral truth and error.
- 2 Henry Sidgwick, The Methods of Ethics (London: Macmillan, 1874), pp. v-vi; 5-6. References to ME i (...)
3His evidence comes from the preface and the first chapter of the first edition of The Methods of Ethics.2 As Phillips himself says, these passages ‘treat ... realism [that is, the view that there is moral truth and error] as a presupposition of all moral enquiry’. But a presupposition does not have to be conceptual. So when Sidgwick says in the passage quoted by Phillips from the first chapter that the view that rightness depends on the agent’s beliefs about rightness implies that the ‘common notion of morality must be rejected’, he need not be taken as making a conceptual claim. What he is rejecting is not a particular understanding of the concept of morality, but a conception of it.
42. G.E. Moore is Phillips’s paradigm of a philosopher who claims to have a ‘master argument’ against any form of ethical naturalism. Sidgwick, in contrast, is said to offer only piecemeal, negative arguments against particular versions of naturalism (17). So ‘we should not expect to find in Sidgwick any diagnosis for [the] ... failure [of these particular analyses] beyond the general idea that they fail to capture something central to the fundamental concept expressed by “right”, “ought”, etc.’ (18).
5This ‘general idea’ is of course Sidgwick’s ‘fundamental distinction thesis’:
The fundamental notion represented by the word “ought” or “right”, which [ordinary moral or prudential] judgments contain ... [is] ... essentially different from all notions representing facts of physical or psychical experience. (ME 25; see 12)
6I fail to see, however, why this view does not itself amount to a ‘master argument’. Sidgwick’s claim is that any form of naturalism will be found to contradict the fundamental distinction thesis. This appears to be structurally analogous to Moore’s suggestion that any form of naturalism will close certain questions which we believe are open. So in the case of Bernard Williams, for example, we can predict that his failure will consist in unacceptably denying the fundamental distinction thesis (20-21); and in the case of Stephen, we find Sidgwick wheeling out that thesis using the very same terminology as in the key statement in ME (cited by Phillips at 27).
73. Phillips closes his chapter on metaethics with the suggestion that Sidgwick’s dualism might imply that he ought to have been, and perhaps sometimes was, an error theorist:
The central thought is that if we have a body of beliefs that we take to express positive truths about a certain subject matter, but these beliefs turn out to involve a fundamental contradiction, then we are not entitled to think that there are positive truths about this subject matter. (45)
8I am inclined to think that Sidgwick would have considered this position both too quick, and too pessimistic. But there is also an internal question. If he accepted the realist conceptual thesis and the realist substantive thesis (the view that there is moral truth and error), and on the basis of these would have denied non-cognitivism (28), then why would this not have led him to reject error theories for the same reasons?
9One further question. I wonder whether Phillips might garner further support for his interpretation here from the final paragraph inserted into the Methods after the first edition. There, in accepting the possibility that ethics may turn out to be analogous to science in our taking certain claims to be true merely because we have a strong disposition to believe them and they establish coherence, Sidgwick could be read as allowing for the ‘invention’ of ethical principles.
101. According to Phillips, Sidgwick fails to draw ‘fully or explicitly’ a distinction between weaker and stronger self-evidence – that is between a conception of self-evidence as less or more conclusive (59). In a footnote, he continues:
On the interpretation I favor, [Sidgwick] does in effect make the distinction in part. As I see it, Sidgwick has the idea of apparent self-evidence, the idea that the fact that a proposition seems compelling without reference to evidence beyond itself does not provide conclusive justification for believing that proposition, and that additional independent evidence in favor of believing it can be supplied by its passing the third and fourth tests [Phillips is here referring to the four tests for the highest degree of certainty at ME 338-42]. The additional idea he does not clearly have is the idea that some but not apparently conclusive evidence in favor of a proposition might be supplied by contemplating that proposition without reference to evidence beyond itself. (86 n9)
11I am inclined to think that Sidgwick is fully aware of the distinction between more and less conclusive conceptions of self-evidence, and in fact distances himself from the idea that any view held as self-evident can ever be said to be conclusive: ‘By cognition I always mean what some would call “apparent cognition”’ (ME 34 n2; see ME 211). So all we have is the idea of a proposition’s being maximally conclusive, in the sense of its having passed all four tests as fully as possible in the circumstances. But here I may be misunderstanding what Phillips had in mind.
122. Phillips states Sidgwick’s fourth test as: ‘The proposition must be universally accepted’ (60). This positive condition seems to me significantly stronger than the negative condition Sidgwick in fact gives at ME 442:
[I]f I find any of my judgments, intuitive or inferential, in direct conflict with a judgment of some other mind, there must be error somewhere: and if I have no more reason to suspect error in the other mind than in my own, reflective comparison between the two judgments necessarily reduces me temporarily to a state of neutrality.
13To pass the fourth test, a proposition I believe to be self-evident must not be one I know some epistemic peer of mine disagrees with. I do not even have to ask the question whether some epistemic peer might disagree with it, let alone check on whether everyone accepts the proposition (a test which no even slightly substantive proposition could ever meet).
143. Phillips raises an important and interesting puzzle about Sidgwick’s strategy in his chapter on the proof of utilitarianism (4.2), where he makes no reference back to the independent ‘criterial’ arguments of Book 3 and focuses entirely on what Phillips calls ‘bipartite’ arguments, which are ad homines and rely on the idea that the homines addressed have beliefs which are not self-evident but provide a route towards beliefs which are self-evident (65-78).
15Phillips’s own explanation of this is that Sidgwick himself believed that the criterial argument had failed, but thought that he could nevertheless, using bipartite arguments, show that utilitarianism is superior to dogmatic intuitionism. It seems to me that one could argue that Sidgwick should have taken this pessimistic view of the criteria argument, primarily because of the large amount of disagreement with utilitarianism by epistemic peers. But as far as I can see, there is no evidence that he in fact does, except for the ‘puzzle’ of 4.2 itself.
16I myself am inclined to accept something like Phillips’s second solution of the puzzle, according to which Sidgwick’s aim in 4.2 is to focus only on arguments for one method as against another (73). His objection to this suggestion is that Sidgwick believes that in that context criterial arguments can still be helpful. But this assumes that criterial arguments have not already been rejected. As I understand the context of 4.2, Sidgwick does indeed see the criterial argument as having failed – but only rhetorically, not as an argument in itself. He is considering what might be said to someone who is unpersuaded by the (perfectly good) arguments of book 3 to shift them from their current beliefs to a conclusion which, had they considered things properly in the first place, they might have arrived at using the criterial argument (see esp. ME 419).
174. Does Sidgwick give any credence to common-sense morality? Like David Brink and others, I am inclined to think he does. Consider in particular the passage from ME 373, cited by Phillips at 75:
[T]he truth of a philosopher's premises will always be tested by the acceptability of his conclusions: if in any important point he be found in flagrant conflict with common opinion, his method is likely to be declared invalid.
18Phillips disagrees (75-6), claiming that Sidgwick ‘clearly has ... a conception of criterial argument which gives no evidential role to common-sense morality’. This claim, however, is ambiguous. It could mean that criterial argument is in itself independent of common-sense morality; or that the claim that common-sense morality plays no such evidential role is part of criterial argument itself. Phillips I presume intends the stronger claim, and yet he provides no evidence for it. And he explains passages such as that on ME 373 as involving a confusion on Sidgwick’s part between philosophical and dogmatic intuitionism, so that his conception of the former there mistakenly includes a commitment to the latter.
19Given Sidgwick’s own formidable intellect, and the fact that the Methods was so carefully revised so many times, this strikes me as an implausible interpretation. More plausible is that his epistemology includes both elements of foundationalism and a commitment to some kind of Aristotelian dialectic, in which the results of philosophically intuitionist theorizing are checked against common-sense morality as a whole. (Recall Sidgwick’s claim in the preface to the 6th edition that he deliberately set out to ‘imitate’ Aristotle.)
205. There is a particularly interesting discussion of the implications of the four tests for Sidgwick’s epistemology at 81-4. Here Phillips argues that Sidgwick is not what he calls an ‘extreme foundationalist’ – that is, a foundationalist who would deny that a proposition’s passing the test could be part of our justification for accepting it. Phillips sees Sidgwick as a moderate foundationalist who does not deny this role to the tests, but who nevertheless is not a coherentist because the criteria are justified by appeal to what the opponent Sidgwick has in mind – the natural sceptic – already accepts.
21This seems to me an ingenious and suggestive interpretation. But I wonder whether Sidgwick might also be understood as an extreme foundationalist (as far as the tests are concerned) who sees the passing of the tests as itself part of what it is for a proposition to be genuinely as opposed to merely apparently self-evident. Then self-evidence itself would still be doing all the epistemic heavy-lifting.
221. Phillips claims that Sidgwick does not see hedonism as self-evident (97, 111 n7). His evidence is that Sidgwick treats it in 3.14 rather than in 3.13 (the chapter on philosophical intuitionism) and does not list it as an axiom. And in 3.14 he notes the disagreement about hedonism.
23I agree that Sidgwick’s not discussing hedonism in 3.13 requires explanation. I suspect that he wanted to dedicate a whole chapter to the good and hedonism, and it might be said to be telling that this chapter follows immediately on from 3.13 and that 3.13 ends with an explicit link to the ‘more indirect’ argument of 3.14. It is also true that Sidgwick is aware of the disagreement about hedonism, and tries to deal with it. But the same is of course true of utilitarianism and dogmatic intuitionism.
24We should expect hedonism to be self-evident for Sidgwick, since it is required for his version of utilitarianism, and he takes utilitarianism to be justified by intuition. And early in the book (ME 97-8) Sidgwick implies that the alternative to inductivism as a support for hedonism is intuitionism. Further, at ME 400, Sidgwick asks his reader to consider his own intuitive judgement, in the same way as he was asked to do ‘in considering the absolute and independent validity of common moral precepts’ (see also ME 406-7).
- 3 See esp. his The Elements of Morality, Including Polity, 2 vols. (London: John W. Parker, 1845).
- 4 See e.g. The Foundations of Ethics (Oxford: Clarendon Press, 1939), 5.
252. Phillips rightly sees Rossian deontology as a major competitor to Sidgwickian utilitarianism. But the claim that ‘Sidgwick, of course, never directly encountered Ross’s position’ (105), though obviously true, strikes me as potentially misleading. Whewell’s position is similar in various significant ways to Ross’s,3 and even more importantly so is that of Aristotle (Ross himself sees his own view as Aristotelian4).
26I think also that Phillips is correct in thinking that Ross’s account of promising is the best potential counter-example to Sidgwick’s utilitarianism. But Sidgwick also fails to recognize the important role played by the capacity for practical judgement in the ethics of Aristotle, Whewell, and Ross, and – I believe – in any plausible moral theory. Sidgwick held a scientistic view of ethics, according to which moral theories must aim to guide action in each particular case, without reliance on judgement. But this goal is a chimera. Any theory must itself be judged to be correct in the first place, and any plausible one, including utilitarianism, will require judgement, not explicitly guided by the theory, to be exercised in its application to particular cases. This seems to me another major failure in Sidgwick’s position which Phillips might have said more about.
- 5 The Methods of Ethics, 1st edn., cited at 116.
271. Phillips is inclined to commit Sidgwick to what one might call a ‘permissibility’ conception of the dualism of practical reason. He cites a passage from the first edition of the Methods which expresses the axiom of justice with reference to what is ‘right’, ‘reasonable’, ‘the dictate of reason’, and ‘my duty’5, and notes that ‘reasonable’ ‘in ordinary language tends to suggest permission’. But if Sidgwick were here mixing concepts of requirement with a concept of permission, he would of course be deeply confused. ‘Reasonable’ for him is, in such contexts, a technical term: the reasonable action is the one there is strongest ultimate reason to do. Nor does it make a difference that he uses ‘reasonable’ elsewhere in the ordinary sense (Phillips ibid.).
28My own view is that Sidgwick tends to use too many concepts to express his position, the passage here quoted by Phillips being an excellent example of that. That position could be stated purely in terms what we have ultimate reason to do, with no reference to permission, requirement, ‘ought’, duty, or whatever. So – pace Phillips 153 n17 – I would not want to attribute a ‘permissive’ interpretation to Sidgwick, nor even to advocate myself such a version of a ‘dual source’ view of practical reasons. For me to know what to do, all that is required is that I know what I have strongest ultimate reason to do. To ask whether I am permitted or required to act in that way is to ask an unnecessary and potentially confusing question.
292. I do find the version of permissivism Phillips advocates, and ascribes to Sidgwick (149-51), a lot more plausible than Sidgwick’s own pessimistic and ‘chaotic’ dualism, and I find it a little mysterious that Sidgwick never considers such a position. I suspect one reason may be that it relies on judgement about how ‘serious’ consequences are for the various individuals in the relevant outcomes of actions (147). Sidgwick would have seen this position as a version of ‘perceptional intuitionism’ and thus to be rejected out of hand. But of course one might attempt to make it somewhat more precise, by attaching weightings to one’s own interests as against those of others, and it is not obvious that the resulting view would be any more imprecise than the egoistic and universalistic versions of hedonism Sidgwick does discuss.
1 David Phillips, Sidgwickian Ethics (Oxford: Oxford University Press, 2011). All unattributed page references are to this book.
2 Henry Sidgwick, The Methods of Ethics (London: Macmillan, 1874), pp. v-vi; 5-6. References to ME in the text below are to the 7th edition (1907).
3 See esp. his The Elements of Morality, Including Polity, 2 vols. (London: John W. Parker, 1845).
4 See e.g. The Foundations of Ethics (Oxford: Clarendon Press, 1939), 5.
5 The Methods of Ethics, 1st edn., cited at 116.Haut de page
Pour citer cet article
Roger Crisp, « Metaphysics, Epistemology, Utilitarianism, Intuitionism, and Egoism: A Response to Phillips on Sidgwick », 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/671 ; DOI : 10.4000/etudes-benthamiennes.671Haut de page
Droits réservésHaut de page