The sources of moral knowledge

Does moral knowledge require a special or distinctive epistemology? Or is it enough just to develop a general account of knowledge and justified belief, which can then just be applied to our moral beliefs? Among those theorists who have searched for a special epistemology for moral beliefs, one common idea is that moral knowledge has a distinctive fundamental source. In particular, according to many of these theorists, there are certain special mental states—often labelled “moral intuitions”—that serve as the primary source of our moral knowledge. As most of its proponents accept, this view comes with some deep and difficult questions about the nature of these moral intuitions. Broadly speaking, there are two main kinds of answers to these questions that philosophers have explored. First, some philosophers have proposed that these moral intuitions are fundamentally akin to a priori intuitions, of the sort that are often supposed to be central to reasoning in mathematics.1 Secondly, some other philosophers have proposed that these moral intuitions have a deep connection of some sort with our moral emotions.2 In principle, it may be possible to reconcile these two answers, by arguing that our moral intuitions have a deep connection with our moral emotions, but that nonetheless at least some of these intuitions are in an important sense a priori—in the sense of being available in principle to all thinkers who possess the relevant concepts, regardless of the particular experiences that those thinkers have had.3 By contrast, those who hold that it is enough to apply a general account of knowledge and justified belief to the special case of moral beliefs often appeal to the idea of “reflective equilibrium.” This idea was originally proposed by John Rawls (1972: 46–53) as part of his account of the method of moral theory. Strictly speaking, for Rawls, reflective equilibrium is an ideal, which he never claimed to actually have reached; so, for Rawls, the appropriate method of moral theory is the pursuit of reflective equilibrium. For our purposes, however, the important point is that several philosophers have thought that the pursuit of reflective equilibrium is simply the appropriate method for rational intellectual inquiry in general; for this reason, these philosophers have concluded that no special or distinctive epistemology for moral beliefs is required.4 Moral knowledge can be acquired in any of the ways in which we acquire ordinary empirical knowledge, and our efforts to acquire and preserve such knowledge are subject to frustration in all of the same ways that our efforts to acquire and preserve ordinary empirical knowledge are. Any source of ordinary empirical knowledge is also a potential source of moral knowledge, and any threat to our ordinary empirical knowledge (or our ability to acquire such knowledge in the first place) is also a threat in the moral domain. This makes it sound as if she favours the “generalist” side in this debate—that is, the view that moral epistemology need only take the form of applying a general account of empirical knowledge and justified belief to the special case of moral beliefs. Admittedly, McGrath clearly does not endorse the most familiar version of this generalist view, according to which the idea of “reflective equilibrium” provides a complete general account of knowledge and justified belief in a wide variety of domains, both moral and non-moral. This is made clear in Chapter 2, where she argues—entirely convincingly, as it seems to me—that the idea of reflective equilibrium cannot by itself provide a general account of the sources of moral knowledge. Instead, she argues, the most defensible version of the method of pursuing reflective equilibrium actually presupposes that there are other sources of moral knowledge as well (39). In the following three chapters, she offers further arguments in support of this hypothesis. Thus, in Chapter 3, she argues that we can learn moral truths from the testimony of others—and more generally by deferring to the beliefs of others; in Chapter 4, she argues that our perceptual experiences and observations play an extensive role in giving us access to moral knowledge—while allowing that it may also be true that some of our moral knowledge is purely a priori; and in Chapter 5, she argues that moral knowledge can be lost, both through being forgotten and through being defeated or undermined, in essentially all the same ways as non-moral knowledge. As I shall now explain, there are two different ways of interpreting McGrath's working hypothesis. At least with respect to many kinds of knowledge, it seems possible to distinguish between the primary and the secondary sources of a kind of knowledge. For example, consider the following view of our knowledge of colour. According to this view, the primary source of our knowledge of colour is visual experience. When we know an object's colour in this primary way, we know what colour the object has by seeing that it has that colour. However, this view can allow that there are also many secondary sources of our knowledge of colour. For example, once one has made sufficiently many observations of the colours of objects, one can reason inductively to general beliefs about the colours that objects of different kinds typically have; for example, by observing sufficiently many lemons, one can rationally come to hold the general belief that normally lemons are yellow. One can then draw inferences from these general beliefs; for example, even if one has not seen a particular lemon, one can infer from the fact that it is a lemon that it is probably yellow. One can also use such general beliefs about colours to infer a proposition about something's colour by an inference to the best explanation. For example, if one knows that the driver of a car stopped the car on reaching some traffic lights, one may infer that at the time the traffic lights were probably red. One may also learn something's colour by relying on someone else's testimony—or in general by deferring to someone else's belief about its colour. In general, however, these secondary sources of our knowledge of the colours of objects have a kind of asymmetric dependence on the primary source: they all presuppose and rely on beliefs about colour that are not acquired from these secondary sources—whereas knowledge that derives from the primary source does not in the same way rely on knowledge that is acquired from any other source. Thus, there are two ways of interpreting McGrath's working hypothesis. On one interpretation, it implies that the primary sources of moral knowledge include all the sources of ordinary empirical knowledge; this interpretation effectively rejects the idea that the distinction between primary and secondary sources has any non-trivial application to moral knowledge. On a second weaker interpretation, it does not imply this. Instead, it implies only that moral knowledge can be acquired from all these sources in the same way that knowledge of colour can be acquired from all these sources; this interpretation is quite compatible with the idea that the primary source of moral knowledge is a special and distinctive source, which it is the task of moral epistemology to give an account of. It seems clear on reflection that McGrath's arguments only support the weaker interpretation of this working hypothesis. For example, in Chapter 3 (70–73), she argues that it can be rational to defer to others in our moral beliefs. But she never argues against the thesis that all knowledge that we acquire through this kind of deference is essentially second-hand—in the sense that all knowledge acquired through deference can be traced back (perhaps through a chain of beliefs each of which was acquired through deferring to earlier beliefs in the chain) to some beliefs that were acquired in some other way than through such deference.5 Similarly, in Chapter 4, she argues that experience and observation can contribute to moral knowledge in any of the ways in which they contribute to our ordinary, non-moral knowledge of the world. Specifically, she identifies four ways in which experience and observation “contribute” to moral knowledge (109). Three of these four ways do not involve experience's serving as a source of moral knowledge, but rather “contributing” in some other way: specifically, experience can serve as a factor that enables one to have such moral knowledge (for example, by allowing one to possess the relevant concepts), or triggers one's acquisition of such knowledge (for example, by directing one's attention towards the relevant questions) or conditions one's skill at acquiring such knowledge (for example, by providing a kind of feedback on one's earlier moral beliefs, enabling one to refine or improve this skill). Admittedly, she also argues in Chapter 4 that one's experience and observation can sometimes “directly” confirm or disconfirm one of one's moral beliefs (115). But all her examples of this rely on one's prior credences in moral propositions. Specifically, an observation can directly confirm or disconfirm a moral proposition whenever one rationally has a prior conditional credence in the proposition, conditionally on the supposition of the content of the observation, which differs from one's prior unconditional credence in that moral proposition (121). It is clear that, in this way, one might know by taste or smell that an object is green: for example, one might taste a bit of food in complete darkness and learn by taste or smell that it is spinach, which would confirm the proposition that it is green. This is clearly compatible with the thesis that visual experience—rather than gustatory or olfactory experience—is the primary source of our knowledge of colour. Similarly, the possibility of sensory experiences' confirming or disconfirming moral beliefs in this way does not show that such observations are among the primary sources of moral knowledge. Finally, the arguments of Chapter 5 are about the ways in which we can lose moral knowledge—for example, through its being forgotten or defeated. Even if moral knowledge can be lost in all these ways, that obviously does not imply that moral knowledge does not have a distinctive primary source. Thus, McGrath's arguments seem only to support the weaker version of her working hypothesis. In fact, there are reasons for thinking that moral knowledge does have a distinctive primary source. McGrath herself, in her discussion of “moral expertise”, appeals to the idea that “in the moral domain, we typically lack an independent basis for attributing genuine expertise that is comparable to the independent bases that we frequently possess in many other domains” (102). But why is this? If moral knowledge can come from a great many different sources, why couldn't we just check each of these sources against all the others—and thereby come to see that a particular person's beliefs count as an unusually reliable source? It is much easier to explain why one typically lacks “an independent basis for attributing expertise” if all of one's moral knowledge must ultimately be supported, directly or indirectly, by one's own moral intuitions. (Since we all know of psychopaths, and vicious people like hardened criminals, who have wildly unreliable moral intuitions, we cannot simply trust someone else's moral intuitions without first using our own intuitions to assure ourselves that their intuitions are reliable.) A similar point arises about McGrath's argument in Chapter 5 that the proposition that you could express by saying, “I have an unreliable sense of right and wrong,” is typically a “blind spot proposition” for you (166). If moral knowledge can come from all sorts of primary sources, then it should be easy to find that one's beliefs about right and wrong are unreliable—if one finds all sorts of conflicts and discrepancies between these primary sources. Again, it is much easier to explain why it is so hard for one to realize that one's beliefs about right and wrong are unreliable if all these beliefs must ultimately be supported by one's own moral intuitions, which rarely are in direct conflict with each other; for this reason, one's intuitions will rarely be capable of revealing their own unreliability. Initially Ted believes that participating in a romantic relationship with someone of the same sex is seriously morally wrong. He strongly believes that Fred, with whom he has worked closely on a professional basis for many years, is an ordinary decent person who would not repeatedly engage in behavior that is seriously morally wrong. Ted then discovers that Fred has been in a same-sex relationship for many years. In response to this discovery, Ted becomes less confident of his belief that it is seriously morally wrong to take part in a same-sex relationship. McGrath treats this as a straightforward case of an empirical observation's disconfirming Ted's earlier belief. But treating the case in this way is somewhat lacking in psychological realism. Those of us who remember the 1980's are quite familiar with people like Ted among our personal acquaintances. What actually seems to have happened with people like Ted is something like this. In their earlier outlook, the thought of homosexual relationships led them to focus their attention on certain kinds of sex acts, which they found it somewhat disturbing to imagine. On some theories of what moral intuitions are, this feeling of discomfort in imagining these sex acts itself counted as a (misleading) moral intuition; on other theories, this feeling is not strictly a moral intuition, but it is intelligible why people like Ted might have treated it like a moral intuition. At all events, this feeling is what explained their earlier belief. Later, realizing that some of their friends were involved in homosexual relationships led them to a radical shift in attention: instead of focusing on sex acts, the focus of their attention shifted to the mutual love, loyalty, and support that these relationships involve—giving rise to a completely different intuition about these relationships, which in turn supported a change in belief. Since this change in these people's moral beliefs about homosexuality centrally involves this shift in attention, it is in many ways more akin to a “Gestalt shift” than a straightforward case of empirical disconfirmation. There is a similar lack of psychological realism, as it seems to me, in McGrath's “Moral Inheritance View” (60–66). According to this view, as children we absorb a rich body of moral beliefs from our social environment; at least so long as our social environment is a sufficiently reliable source of moral truths, these beliefs can count as a body of genuine moral knowledge. In articulating this view, she compares children's acquisition of moral beliefs to their acquisition of beliefs about geography. In the case of geography, however, it is clear what the primary source of geographical information is: specifically, it comes from the direct observations and memories of travellers (such as explorers and surveyors). By contrast, her description of the Moral Inheritance View does not include any account of the primary source of moral beliefs. In this way, her presentation of this view makes a child's acquisition of moral knowledge sounds much like the way in which religious doctrines are often inculcated in children—where it is often left obscure where the information in these religious doctrines originally comes from. However, this description of children's acquisition of moral beliefs leaves out a crucial element that distinguishes moral beliefs from such religious doctrines. In addition to acquiring moral beliefs, as children we also acquire a moral sense—a faculty of moral intuitions or moral sentiments—that supports these moral beliefs and can guide us towards new moral beliefs that we could not simply have inferred by any generally rational inference pattern from our pre-existing moral beliefs. These reflections seem to support the conclusion that, if an account of the sources of moral knowledge is to be psychologically realistic and at the same time avoid being seriously incomplete, it must include some account of moral intuitions. There is in fact only one section of this book where McGrath comes close to discussing moral intuitions. This is in her discussion of “armchair moral knowledge” (140–50). Here she proposes, in effect, that whenever a moral intuition serves as a source of moral knowledge for a subject, the relevant moral proposition “intellectually seems true to her because it is true” (148). While this proposal is preferable to the alternatives that she considers, it raises many questions. What is an “intellectual seeming”? (Is it just an inclination to believe the relevant proposition?) How exactly can such a seeming be explained by the truth of the moral proposition that is its content? What psychological mechanism is in play in the cases that can be explained in this way? Without answers to these questions, the appeal to “intuition” or “intellectual seemings” is liable to seem, as J. L. Mackie (1977: 38) famously complained, like a “lame answer” to the question of how we know moral truths. In fact, however, there is extensive evidence from empirical moral psychology for the view that emotions play a pervasive role in moral thought.6 Since emotions are closely related to desires and other conative attitudes, the idea that moral intuitions have a close connection to moral emotions seems particularly plausible given that moral thought has seemed to so many philosophers over the centuries to have an essentially motivating or action-guiding role. For these reasons, it seems to me that the most promising approach to explaining the primary source of moral knowledge is to develop an account of moral intuitions on which they have a close connection to moral emotions. Incidentally, an approach along these lines could enthusiastically endorse McGrath's emphasis on how deeply social our moral knowledge is (8–10). On an approach of this kind, the social dimension of moral thought is tied up with the essentially social and interpersonal character of moral emotions—which seem to be closely connected to what P. F. Strawson (1962) described as “reactive attitudes.” At all events, however, a complete moral epistemology requires some account of moral intuitions—and McGrath's book does not study this central question in much depth or detail. This point, of course, is not in itself any criticism of McGrath's arguments themselves. On the weaker of the two readings that I distinguished above, her working hypothesis seems to me both true and important. It is important for two reasons. First (as she notes), a large number of distinguished philosophers have defended views that are inconsistent with it: for example, many philosophers have denied that we can ever learn moral truths through testimony; many have denied that empirical evidence can ever confirm a moral proposition; and many have denied that empirical evidence can ever defeat or undermine a moral belief that we hold. As she convincingly shows in this book, these denials are mistaken. Secondly, McGrath's hypothesis is important for another reason. If our moral judgments were radically different from our beliefs about ordinary factual matters with respect to the conditions under which they are rational or justified, that would cast serious doubt on cognitivism—the view that moral judgments are genuine beliefs or cognitive states. By contrast, if McGrath is right that the epistemology of moral beliefs is much more similar to other beliefs than is often supposed, that casts doubt on non-cognitivism—the view that moral judgments are mental states of a fundamentally different kind from genuinely cognitive beliefs. After all, if moral judgments walk and quack like ordinary beliefs, surely they just are beliefs? In my opinion, all cognitivists should welcome McGrath's arguments with open arms; and everyone who is interested in the fundamental question of whether or not moral cognitivism is correct should ponder her arguments seriously.