We are pleased to present the latest installment of Ethics at PEA Soup, in which we host a discussion of one article from each issue of Ethics. The article selected from Volume 123, issue 1, is Katarzyna de Lazari-Radek and Peter Singer's "The Objectivity of Ethics and the Unity of Practical Reason." Ethics has kindly provided free access to the article here. We are also very grateful to Roger Crisp for providing the critical précis, which begins below the fold. Everyone is invited to join the discussion.
In this bold and interesting paper, Katarzyna de Lazari-Radek and Peter Singer (LRS) claim that a Sidgwickian response to certain attempts to use evolutionary theory to debunk moral judgements can itself resolve Sidgwick’s own ‘dualism of practical reason’. I am persuaded by the argument against debunking, but I believe the Sidgwickian response can do less than LRS suggest.
LRS begin by outlining Sidgwick’s dualism between, on the one hand, an egoistic axiom, according to which each of us ought to aim at her own good, and, on the other, an axiom of universal benevolence, which requires us to aim impartially at the good of all. They then explain Sidgwick’s arguments against the view that, once we understand the origin of our moral intuitions, we will see them as caused by factors outside our control and hence as unreliable. First, an intuition’s being ‘self-evident’ – that is, such that understanding it is sufficient to justify it – is quite consistent with its being caused. Second, we do not even have to show that the causes in question are likely to lead to true judgements, since this will lead us into a regress of justification. Finally, the causal judgements in question are within the domain of science, and this does not extend to propositions concerning what we ought to do.
Sidgwick distinguishes this form of general scepticism about moral intutions from more limited claims about particular ethical beliefs. Using this distinction, LRS first examine the general argument in more detail, and especially that form of it developed by Sharon Street. According to Street, moral realists, once they recognize that our evaluative attitudes have evolved, face an awkward dilemma. On the first horn, they accept that evolutionary forces have no tendency to select beings with objectively true evaluative attitudes, and so must draw the unpalatable conclusion that most of our evaluative judgements are unjustified. On the second, they claim that these forces were likely to select those able to grasp objective moral truths; but this claim goes against the most plausible scientific understanding of evolution, which sees it as heading in the direction of survival rather than truth.
Street suggests that, had we evolved to be more like, say, lions, we would have been readier to accept the killing of others’ offspring than we are. LRS note the echo here of Darwin’s suggestion that, were we like bees, we would think it a duty of a mother to kill her fertile daughter. Sidgwick responded that such arguments do not touch the abstract principle of utilitarianism, which allows for much variation in the rules of common morality. This, LRS plausibly claim, suggests that a modern Sidgwick, more informed that the real Sidgwick about the influence of evolution on morality, might readily impale himself on the first horn of Street’s dilemma, allowing that many of the rules of common-sense morality are not based on objective truth. (It is worth noting that reference to evolution might also enable the modern Sidgwick to avoid the somewhat implausible notion that the utility of common-sense morality suggests that human beings have been ‘unconscious utilitarians’.)
But, Street might object, if the principle of benevolence is objectively true, isn’t our arriving at it without any steer from evolution just a huge coincidence? LRS rightly point out that Sidgwick can offer a plausible explanation of how we understand such principles: we use our reason. And at this point he can embrace the second horn of Street’s dilemma. A rational capacity would advance success in reproduction, and it might do that most effectively in a general, ‘untargeted’ form, which would allow us to enquire into the foundations of mathematics or physics as well as to recognize self-evident moral truths.
LRS then turn from the general to the particular form of the sceptical argument. They cite Sidgwick’s claim that no theory of the origin of our ethical intuitions has been offered that might throw his own abstract principles into doubt, as arising from sources which were likely to make them false. LRS suggest that this is still the case as far as universal benevolence is concerned, since the kinds of judgement most consistent with reproductive success will recommend helping one’s own children rather than complete strangers. Since LRS are going to use Sidgwick’s arguments about evolution in an attempt to resolve his dualism, it is worth noting that egoism is in as strong a position as universal benevolence in this context to resist debunking evolutionary arguments. What we would expect to evolve would indeed be something like kin altruism, which is neither egoistic nor impartially benevolent.
It is true, of course, that some concern for self might be expected to arise through evolutionary development, and LRS later approvingly cite Folke Tersman’s attempt to use this point to debunk my own defence of a principle of self-interest, according to which each of us has a reason (not necessarily overriding) to advance her own good. So it might be claimed that egoism is just a development of that bias towards the self, a development tainted by its source in non-rational evolutionary processes. I myself am not persuaded by Tersman’s argument, and would want to appeal to some of the very Sidgwickian counters to debunking arguments which LRS state earlier in their paper. Egoism, or the principle of self-interest, are justified by appeal to their self-evidence, and the conclusions of reflection upon them, though that reflection must be fully informed by an impartial grasp of evolutionary development, need not be overturned by that grasp. If it is pointed out to me that the reason I think that 7 + 5 = 12 is that my hunter-gatherer ancestors needed to develop some system for sharing out food at the end of the day, my belief will be unshaken. But note also that if Tersman’s point has any force, it applies equally to the principle of universal benevolence. We would expect evolution to produce some concern for others, and universal benevolence can be seen as an extension of that concern in the impartial direction in just the way that egoism might be taken to be an extension of concern for oneself in the direction of partiality.
Since self-evidence can withstand reflection on the origins of beliefs, even kin altruism can resist debunking. LRS cite – without questioning it – Sidgwick’s somewhat remarkable claim that it is ‘certainly not’ [my italics; LRS paraphrase as ‘not at all’] ‘self-evident that we owe more to our own children than to others whose happiness equally depends on our exertions’. I have little doubt that, were people to reflect properly on this conception of extreme impartiality, the vast majority would reject it. But that of course is not the issue, as LRS point out: ‘This is not to say that the judgment that we have greater obligation to help our own children than to help strangers cannot be justified, but rather that if it is to be justified, it needs a form of justification that does not start from the idea that because we strongly feel that it is right it must be true’. This is hard to deny; but it is a point that applies as much to universal benevolence, and indeed egoism, as it does to kin altruism.
In support of the principle of universal benevolence, LRS claim that it results from ‘a process of careful reflection that leads us to take, as Sidgwick puts it, “the point of view of the universe”’. This idea, they suggest, has been converged on by leading thinkers in various traditions, including the Judaic, Christian, Confucian, Hindu and Buddhist traditions. Nor is there any plausible evolutionary debunking argument against it. There are, then, three elements to establishing that an intuition has the greatest degree of reliability: (1) reflection; (2) agreement among careful thinkers; (3) lack of any debunking argument.
By the point of view of the universe, LRS seem to mean something considerably less rigorous than Sidgwick’s own utilitarian conception of pure impartiality, seeing it in the Judaeo-Christian tradition, for example, as the Golden Rule. And if the egoist is permitted the same degree of latitude, she can claim that her view meets each of these three conditions as effectively as the principle of universal benevolence. The Golden Rule itself seems to imply that it is no less rational to have concern for oneself as for others, and self-love can plausibly be claimed to play an important role in the other three traditions mentioned by LRS, as well, of course, as the ancient western philosophical tradition, where if anything egoism rather than universal benevolence plays the more important role. And there has been no more careful thinker on these matters than Sidgwick himself! Further, as we have seen, the mere fact that some principle is partial is not enough to debunk it, even if – as with kin altruism, and not with egoism – it lines up with evolutionary expectations.
LRS’s conclusion, then, is that, because impartial universal benevolence withstands evolutionary debunking arguments, whereas partial principles such as egoism do not, Sidgwick’s dualism can be resolved in favour of benevolence. I have suggested that egoism, and indeed kin altruism, can withstand reflection as effectively as the principle of universal benevolence. I agree with LRS in rejecting appeal to reflective equilibrium in ethics (something I remember Joseph Raz’s pointedly describing as ‘unreflective equilibrium’). What is required is just the kind of rational, informed, impartial reflection on ultimate ethical principles advocated, and often (though not always) engaged in by Sidgwick. LRS are right too that Street’s evolutionary arguments are misdirected against moral realism. Such arguments might often be useful in debunking certain, unreflective, spontaneous moral responses, such as a visceral disgust at incest or homosexuality. But, at least as far as current evolutionary theory is concerned, they are largely irrelevant to first-order normative ethics (which philosophers do you know campaigning against incest or homosexuality?), as indeed is the neurological evidence based on fMRI scans used elsewhere by Singer, Greene, and others to support the principle of universal benevolence (put Frances Kamm or Judith Thomson in a scanner, and their brains will – I’m willing to bet – light up in the same way as Singer’s or Greene’s when they’re asked to state their fundamental ethical principles). There are no quick fixes here.
It is somewhat remarkable that, having concluded the Methods in a state of such internal incoherence, Sidgwick appears to have done little to try to resolve it. It was the same with disagreement with others: like LRS, when stating his own first-order view he largely ignored its implications and focused on agreement, though his own discussion of intuitionism demonstrates a clear awareness of the threat to self-evidence posed by disagreement with epistemic peers. What is needed now in normative ethics is a general facing up to the existence of such interpersonal disagreement, and a non-dogmatic and co-operative attempt to make progress towards greater convergence. Indeed this may be an area in which evolutionary theory (along with neuroscience, history, anthropology, psychology …) turns out to have real purchase.
- Roger Crisp