In this paper I assess the viability of a particularist explanation of moral knowledge. First, I consider two arguments by Sean McKeever and Michael Ridge that purport to show that a generalist, principle-based explanation of practical wisdom—understood as the ability to acquire moral knowledge in a wide range of situations—is superior to a particularist, non-principle-based account. I contend that both arguments are unsuccessful. Then, I propose a particularist-friendly explanation of knowledge of particular moral facts. I argue that when we are careful to keep separate the various explanatory tasks at hand we can see that a particularist-friendly explanation of the fact that (e.g.,) Jane knows that A is morally right might not be so difficult to come by. Moreover, I suggest that a particularist approach to explaining knowledge of particular moral facts may go some way towards discharging the challenge of moral scepticism.
Particularists claim that moral phenomena can be explained without relying on exceptionless moral principles. The debate over particularism in recent years has focused on whether the rightness and wrongness of actions can be so explained: generalists claim that in order to explain the rightness of some action, A, we must identify a feature, F, of A and a principle according to which any action that has feature F is right (or pro tanto right). Particularists, in contrast, claim that an explanation of the rightness of A can be perfectly adequate without mention of exceptionless universal generalizations – perhaps ceteris paribus generalizations or generic statements will do, or perhaps the fact that A is F can explain A’s rightness without mention of generalizations at all.1
However, the rightness (wrongness) of actions is not the only moral phenomenon in need of explanation. The fact (if it is a fact) that some people have moral knowledge must also be explained. If, for instance, an adequate explanation of Jane’s knowledge that action A is right must mention an exceptionless moral principle, then the particularist project of providing an account of morality without relying on exceptionless moral principles will be gravely jeopardized. And indeed, the chief argument in Sean McKeever and Michael Ridge’s [henceforth M&R] book-length critique of particularism is centred on the explanation of moral knowledge. According to M&R a generalist, principle-based explanation of practical wisdom—understood as the ability to obtain moral knowledge in a wide range of situations—is superior to a particularist, non-principle-based account. And consequently, they claim, we have reason to abandon the particularist project.2
In this paper I assess the viability of a particularist explanation of moral knowledge. First, I consider M&R’s arguments in favour of a generalist account of practical wisdom and I explain why I find them unconvincing (§2). In §3 I propose a particularist-friendly explanation of knowledge of particular moral facts. I argue that when we are careful to keep separate the various explanatory tasks at hand we can see that a particularist-friendly explanation of the fact that (e.g.,) Jane knows that act A is morally right might not be so difficult to come by and I draw some of the implications of the kind of explanations I propose for moral epistemology more generally.
2 M&R’s Arguments
A person of practical wisdom can obtain moral knowledge in a wide range of situations. How does she do this? According to M&R the best explanation of practical wisdom is that “practical wisdom involves the internalization of a finite and manageable set of non-hedged moral principles which together codify all of morality” (139). Their thought is that a person of practical wisdom has at her disposal a set of exceptionless moral principles from which, together with empirical information about the non-moral features of various situations, she can deduce the normative status of actions. This is a generalist model because exceptionless moral principles play an indispensible role in the explanation of the ability of the practically wise person to obtain moral knowledge. And this is the best explanation of practical wisdom, according to M&R, because unlike its particularist competitors it can accommodate two central features of practical wisdom: that practical wisdom extends to novel and unusual situations and that moral knowledge is not entirely a posterior. Consequently, inference to the best explanation leads to the conclusion that we have reason to endorse generalism and to abandon particularism.
To my mind, generalism does not have the explanatory advantages which M&R attribute to it. In order to see why, let us take a closer look at M&R’s arguments.
2.1 The Argument from Scope
M&R grant, for the sake of argument, Dancy’s holism in the theory of reasons. Holism is the view that a feature that is a reason in one case may be no reason at all, or an opposite reason, in another case. Holists distinguish between features that are reasons (favourers) and features that operate on reasons, such as enablers, disablers, intensifiers, and attenuators.3 On this model, whether a feature is a reason in a particular situation depends on the presence (or absence) of enablers and disablers. Moreover, the contribution that each reason makes to the overall normative status of a particular action is influenced by the presence (or absence) of intensifiers and attenuators. Let us call all these features—reasons, enablers, disablers, intensifiers and attenuators—the morally relevant features of a situation. A practically wise person must be sensitive to all morally relevant features of a situation because if she is not so sensitive her ability to obtain moral knowledge in a wide range of situations would be utterly mysterious. Moreover, if there are morally relevant features to which she is not sensitive her claim to know the moral fact in question may be undermined or defeated.
will invoke the idea that she already knows all of the kinds of considerations which can function as reasons, defeaters, enablers, intensifiers, and deminishers and indeed understands how they interact … The possibility of practical wisdom … is best explained by the hypothesis that the person of practical wisdom already knows and indeed can articulate all of the potential reasons, defeaters, etc. which might be in play. (142)
According to M&R a practically wise person has a “mental checklist” of all potential moral reasons. She runs through this checklist to determine whether any of the items on her list are present.4 Next, she must find out whether any of the candidate reasons that are present are disabled or enabled. So she runs through a complete list of potential disablers and enablers.5 Then she checks whether any item on her intensifiers/attenuators list is present. After running through all these lists she will be able to form an accurate judgment about which reasons are “active” in this specific situation and how “weighty” each reason is. Finally, she identifies an exceptionless principle that tells her how these reasons, given their various weights, balance out, and she uses this principle to deduce the normative status of the action in question. On M&R’s view, then, the fact that a person of practical wisdom knows the normative status of an action is explained by her ability to deduce it from an exceptionless principle. Hence, in order to explain Jane’s knowledge of the normative status of A we must appeal to an exceptionless moral principle.
Our account of the person on whom we can rely to make sound moral judgements is not very long. Such a person is someone who gets it right case by case. To be consistently successful, we need a broad range of sensitivities, so that no relevant feature escapes us, and we do not mistake its relevance either. But that is all there is to say on the matter. (1993:64)6
According to M&R the checklist model is superior to Dancy’s sensitivity model because it explains the scope of practical wisdom: practical wisdom, M&R maintain, should extend to novel and unusual situations. M&R invite us to imagine Wanda, a person of practical wisdom who visits an entirely alien culture. They claim that since Wanda is a person of practical wisdom, if she has access to all non-moral facts about this alien culture she should be able to determine the normative status of actions and practices there. M&R’s checklist model can explain this: since Wanda knows in advance all the morally relevant features and moral principles, her ability to know the normative status of actions depends exclusively on having access to non-moral facts.
Whether this proposed explanation of the scope of practical wisdom gives us a reason to favour generalism over particularism depends on three considerations: (a) the plausibility of the explanans—namely, the checklist model; (b) the plausibility of competing particularist explanations; and (c) our confidence in the truth of the explanandum: if practical wisdom doesn’t extend in the way M&R claim it does then the “explanation” of this alleged feature of practical wisdom will not help us to adjudicate between particularism and generalism.
Let us first assess the plausibility of the checklist model. The thesis under discussion seems to be a thesis about the cognitive process by which a person of practical wisdom obtains moral knowledge. If so, whether, in fact, a person of practical wisdom runs through mental checklists of morally relevant features and performs an explicit mental deduction is not a question that can be answered from the philosopher’s armchair.7 The only argument we have been given in support of this model is that it best explains the extended scope of practical wisdom. But those of us who are not practically wise cannot obtain moral knowledge by utilizing the same process that a practically wise person employs because we do not know in advance all the potential morally relevant features and exceptionless moral principles. So either we can obtain moral knowledge in some other way, in which case there is some other explanation of how moral knowledge is acquired, or we cannot obtain moral knowledge at all. If there is another way by which moral knowledge can be obtained, then we may wonder whether the checklist account is a better explanation of practical wisdom than the following hypothesis: The difference between us and practically wise people is not that we acquire moral knowledge by using different cognitive processes, but rather that practically wise people use the same processes we do but they simply use them better.
M&R may insist that only people of practical wisdom have direct (or basic) moral knowledge. But first, this seems to undermine the dialectical force of their argument, which is supposed to start from a premise that both particularists and generalists accept. Particularists are unlikely to accept the claim that only people who know all morally relevant properties and moral principles can have moral knowledge.8 Second, and more importantly, on the view under discussion moral knowledge is an “all or nothing” business. Either one knows all the morally relevant features and moral principles or one cannot obtain (direct) moral knowledge at all. According to the model under discussion, one must deduce one’s judgment from an exceptionless principle and the morally relevant features that are present. In order to obtain knowledge in this way one must correctly identify all the features that are present and the moral principle that applies to the situation at hand. If one uses the wrong principle it seems that one’s judgment would not qualify as knowledge. And even if one selects a correct principle, if one is systematically blind to certain morally relevant features then one’s identification of the correct principle might not be sufficiently sensitive to warrant an attribution of knowledge. This is because even if the principle in question had not been the correct principle to use in the situation due to the presence of additional morally relevant features to which the agent is blind, the agent would still have believed that this is the correct principle to use. Moreover, it is difficult to imagine a plausible account of moral education and moral development that accommodates this “all or nothing” aspect of knowledge of moral facts.9
Now let us compare M&R’s explanation to an alternative particularist story. According to one particularist model a person of practical wisdom obtains moral knowledge by exercising a skill of discernment. It may, therefore, be useful to consider the scope of other skills by way of comparison. For example, consider a skilled off-road motorcyclist. Let us suppose that she acquired her skill by riding in certain kinds of terrain. If she is a skilled rider, though, we should expect that she will have no difficulty riding in a slightly different terrain from that with which she is experienced. Moreover, we may expect that she will do better than unskilled bikers in unusual and unfamiliar terrain—perhaps we can expect her learning curve to be steeper in new environments. So to the extent that we have a handle on the notion of a skill, and to the extent that we can make sense of the acquisition of moral knowledge by the application of a skill, we should expect this skill to extend to new and even to somewhat unusual situations.
M&R insist that if Wanda is practically wise, she should be able to obtain moral knowledge in an alien environment. If this alien environment is sufficiently similar to ours then we should expect Wanda’s skill of discernment to carry over to this new environment. But what if the new environment is radically different? Presumably Wanda managed to acquire a skill that allows her to obtain moral knowledge in our normal environment. Perhaps we have reason to expect that given enough time and information Wanda would be able to acquire another skill—the skill required for the attainment of moral knowledge in this alien environment. Admittedly, Wanda will need more than just time and information to acquire this new skill—she will also need practice. But is it true that a comprehensive list of empirical facts is all one needs in order to obtain moral knowledge in an alien environment? Suppose that Wanda has access to a super-computer which can tell her the locations and properties of all the fundamental particles at any given time in this alien world. Are we supposed to expect her to be able to know the normative statuses of actions simply by consulting this computer? Is her ability, so understood, anything like what we might recognize as practical wisdom?10
Even one who shares M&R’s intuitions about knowledge in alien environments may admit that these intuitions can hardly be used as explanatory desiderata that any adequate theory of practical wisdom must accommodate. Consider, for instance, Wendy, whose moral judgments are extremely reliable in our normal environment and also in different, but similar, environments. Let us assume that in such environments Wendy’s judgments are as reliable as Wanda’s judgments and that she can justify her judgments as well as Wanda can. However, unlike Wanda, Wendy is unable to form any moral judgment in radically alien environments. It is far from obvious that we must conclude that Wendy is not practically wise. And consequently, it is not clear that a particularist account of practical wisdom that can explain Wendy’s, but not Wanda’s, abilities is not an account of practical wisdom.
So far, I’ve been treating M&R’s checklist model as a hypothesis about the psychological process by which a person of practical wisdom obtains moral knowledge. However, several features of M&R’s work suggest that we ought not to attribute this view to them. First, one might expect that if M&R wanted to establish an empirical thesis about actual psychological processes they would have provided some empirical evidence in support of their thesis. It would be quite remarkable if we could identify the psychological process by which moral knowledge is obtained simply by reflecting on the possibility of practical wisdom. Second, in the final chapter of their book M&R claim that their principled conception of morality—”generalism as a regulative ideal”—is compatible with both prototype theory and exemplar theory. These theories purport to explain how people classify particulars as falling under a certain concept. The classification process on both theories is very different from M&R’s checklist account. Indeed, the checklist model, understood as a psychological thesis about the process by which moral knowledge is obtained, seems incompatible with prototype and exemplar accounts. Perhaps, then, a more charitable reading of M&R is to interpret the checklist model as a “rational reconstruction” of the process by which a person of practical wisdom obtains moral knowledge rather than as the claim that a person of practical wisdom consciously uses the checklist model to reach moral judgments.11
While there is a great deal of disagreement over adequacy conditions for explanation, there is a fairly broad consensus that one necessary condition is that the explanans must be true. If so, the checklist model cannot explain moral knowledge unless it is true. Perhaps, then, the thought is that something like the checklist model operates at the sub-personal level. However, relocating the checklist process to the sub-personal level is incompatible with M&R’s insistence that the person of practical wisdom “already knows and indeed can articulate all of the potential reasons, defeaters, etc. which might be in play”, and it undermines the arguments for generalism we have been discussing. Moreover, it is not quite clear to me how to make sense of the checklist model at the sub-personal level. At best, it might be viewed as a highly metaphorical description of a sub-personal psychological process. Finally, it is no less surprising that we could discover the functioning of sub-personal processes merely by reflecting on the possibility of practical wisdom without doing any empirical research, than that we could discover the functioning of psychological processes at the personal level in this way.
It may be tempting to view the checklist model as one extreme on a spectrum of models of the attainment of moral knowledge, where the skill model occupies the other extreme. Viewed in this way we can imagine a range of “in between” views. Rather than interpreting M&R as advocating the checklist model per se, we can, perhaps, interpret them as recommending a model closer to the checklist-model-end, as it were.12 While read in this way M&R’s proposal becomes somewhat more palatable, it is not clear that anything weaker than the extreme checklist model can be used as an argument against particularism. To undermine particularism M&R must identify moral phenomena that cannot be explained without relying on exceptionless moral principles. Any model that is weaker than the extreme checklist model must allow for explanations without exceptionless principles and hence it will be compatible with particularism. So while moving away from the checklist-extreme increases the plausibility of M&R’s model, it renders the debate over a weakened checklist model tangential to the particularism-generalist debate.
One may certainly feel that the skill-based account of practical wisdom that we are considering is underdeveloped. Furthermore, one may have doubts as to whether such an account can be fleshed out and expanded into a fully fledged plausible theory about how moral knowledge is obtained. Yet M&R’s objection is not that the checklist model is well-developed while the particularist model is not. Instead they claim that there are moral phenomena that the checklist model explains and particularist models cannot explain. I hope to have shown that with respect to the phenomena considered in this section, a particularist explanation is no worse than a generalist checklist explanation. Indeed, given the implausibility of the checklist model and the lack of empirical evidence to support it, the particularist account, thin as it is, seems to offer a more promising approach to a satisfactory explanation of the phenomena we have been considering.
2.2 Moral Knowledge and the a Priori
How does a person of practical wisdom know a particular moral fact, e.g., that this action is right, or that in this case feature F is a reason to do action A? Presumably, a particular moral fact is a contingent fact—that a particular act is right depends on various contingent features of the action. For example, act A might be right because it brings about a lot of pleasure but that it does so is a contingent matter. Moreover, for holists, that the pleasure an action brings about is a right-making feature is also contingent. This suggests that knowledge of particular moral facts is a posteriori knowledge. But it may seem implausible to hold that knowledge that a particular act is right is entirely a posteriori because the rightness of an action, one might think, is not something that can be discovered by way of a purely empirical investigation.
Archparticularist Jonathan Dancy takes this challenge very seriously. In response he argues that although particular moral facts are, indeed, contingent they are, nevertheless, known a priori (2004, Ch. 8). M&R criticize Dancy’s position at length. It is an advantage of their favoured model, they argue, that it need not appeal to the mysterious category of the contingent a priori. On the view M&R favour we can have a priori knowledge of necessary exceptionless moral principles with purely descriptive antecedents. Knowledge of particular moral facts is gained by deduction from such a priori principles and a posteriori knowledge that their antecedents are (contingently) instantiated. For example, we know that act A is wrong because we know a posteriori that A (contingently) has feature F, and we know a priori that (necessarily) any act that has feature F is wrong. Thus, on the assumption that the person of practical wisdom internalizes a set of exceptionless moral principles we can explain the possibility of practical wisdom without attributing to her the curious ability to know a priori that a contingent fact obtains. So, the thought goes, exceptionless moral principles are essential for an adequate explanation of the possibility of moral knowledge and practical wisdom.13
As M&R see it, the only two alternatives to the model they favour are Dancy’s “contingent a priori” and a purely a posteriori account of moral knowledge. And since both these options are unattractive, we have reason to forgo particularism. I agree with M&R that both alternatives are not without difficulties; I find their critique of Dancy’s “contingent-a priori” compelling and there are familiar worries about a purely a posteriori account of moral knowledge. However, I believe that there are at least two other options open to particularists.
One option is that we can have a priori knowledge of exception-full or defeasible principles. Perhaps statements like “lying is wrong-making” and “pleasure is right-making”, understood as generic- or normic-statements,14 are both exception-full and knowable a priori. One might appeal to Ross-style self-evidence and argue that “normic statements” are self-evident. That keeping a promise is (normally) right-making is as plausible a candidate for self-evidence as Ross’ claim that keeping a promise is prima facie right.15 Alternatively, one could argue that normic statements seem right and make use of the principle of phenomenal conservatism.16 While normic statements cannot be used to deduce the normative status of particular actions, they may be a part of a proper explanation of particular judgments. In order to decide whether normic statements are knowable a priori we will need a detailed account of a priority which I am not prepared to give here. Yet by M&R’s standards we must not rule out the possibility that such statements are knowable a priori simply because they are not analytic.17
Another alternative is this. It is commonly thought that moral properties (globally) supervene on non-moral properties. Let us call this thesis the Supervenience Thesis or (ST). Moreover, it is plausible that if (ST) is knowable, it is knowable a priori.18 Particularists need not deny (ST), nor must they deny that (ST) is knowable a priori.19 (ST) guarantees that there are necessary connections between certain non-moral properties and moral properties. So particularists admit that there are true necessary statements of the form: (ST1) □∀x(Gx → Mx) where x ranges over actions, G is a non-moral property of being in a specific completely defined world-state and M is a moral property.20 Statements like this are certainly not analytic on any plausible construal of analyticity, but by M&R’s standards this doesn’t entail that they are not knowable a priori. So perhaps statements of this kind might serve as the a priori element, as we have been calling it, that M&R insist any account of moral knowledge must identify.
One might think that (ST1) is a moral principle and so that by identifying the a priori element of moral knowledge in the form of (ST1), I have, in effect, given up on the particularist project. But this would be a mistake. I presented the debate between particularists and generalists as a disagreement over whether exceptionless principles are essential for an adequate explanation of moral phenomena. Suppose we want to explain the rightness of some action, A. I take it that on any plausible theory of explanation the following statement would not qualify as an adequate explanation: action A is right because any action that instantiates the property of being-in-a-world-state-indistinguishable-from-this-one is morally right. Statements like (SP1) do not explain the rightness (wrongness) of actions and so these are not generalizations of the kind that generalists claim, and particularists deny, are essential to moral theorizing.
One may worry that to identify the a priori element of moral knowledge in the form of (ST1) doesn’t solve the problem with which we began. The challenge was to identify the a priori component of our knowledge of particular moral facts. But surely we cannot use generalizations like (ST1) to deduce the normative status of particular actions. So the problem with which we started remains unsolved.21 Now if the challenge is to indentify an a priori element that we, in fact, know and can use to deduce particular moral judgments that constitute moral knowledge, then I agree that the challenge has not been met. But we should keep in mind that M&R have not met this challenge either since they have not identified exceptionless moral principles that we know and can use in this way. So insofar as we must identify a known a priori element, particularists and generalists are on a par. Both particularists and generalists must appeal to features that are in principle knowable a priori. So, I conclude, whatever role the identification of an a priori element is supposed to fulfil in the explanation of knowledge of particular moral facts, if M&R’s principles can do it, then it’s not clear why (ST1) could not.
3 Explaining Knowledge of Particular Moral Facts
So far I have argued that M&R’s arguments—arguments that purport to identify phenomena that the checklist, principle-based model of moral knowledge explains better than a particularist skill-based model—are unconvincing. In this section I argue that particularist-friendly explanations of knowledge of particular moral facts are, in fact, not so difficult to come by—as long as we identify properly what it is that we are trying to explain.
The task of explaining knowledge of particular moral facts can be understood in several ways. One thing that we might wish to explain is the very possibility of moral knowledge. This task can be described as an attempt to answer the question whether, and if so how, knowledge of particular moral facts is possible. Alternatively, we may wish to explain particular instances of moral knowledge. This task can be described as an attempt to answer questions like this: How does Jane know that action A is morally right?
It may be useful to compare the explanation of knowledge of particular moral facts to the explanation of knowledge of other people’s mental states. Here, too, there are two possible explanatory tasks. One thing we might wish to explain is the very possibility of knowledge of other people’s mental states. This task can be described as an attempt to answer the question whether, and if so how, such knowledge is possible. Another separate goal might be to explain particular instances of knowledge of other people’s mental states. This task can be described as an attempt to answer questions like this: How does John know that Jim is distressed?
Explaining the very possibility of knowledge of other people’s mental states is notoriously difficult. However, coming up with possible explanations of particular instances of knowledge of other people’s mental states is simple enough. For example, consider the following answers to the question ‘How does John know that Jim is distressed?’:
- a) John knows that Jim received some disturbing news.
- b) John noticed that Jim is restless and out of focus.
- c) John is a very sensitive person.
- d) John was told so by Bill who is extremely reliable about such matters.
In “normal” contexts, I submit, answers like these are perfectly felicitous and they constitute adequate explanations of John’s knowledge of Jim’s distress.22 Indeed, if explanations like these are inadequate, we may be forced to conclude that John’s knowledge that Jim is distressed is inexplicable. The fact that we are typically not perplexed by claims like ‘John knows that Jim is distressed’ and our ability in ordinary contexts to offer acceptable explanations for such claims militate against the hypothesis that such claims are inexplicable.
Several aspects of these proposed explanations are worth emphasizing. First, none of them, nor their conjunction, guarantees that John knows that Jim is distressed. Moreover, there are no known exceptionless principles that we can add to these proposed explanations, or that we can reasonably expect John to know, that would guarantee that John knows that Jim is distressed. Second, if the mental supervenes on the physical, then there are some physical features that determine Jim’s distress. Presumably, if John knows that Jim is distressed he must, in some sense, be sensitive to these features. Nevertheless, John need not be able to articulate these features in order to know that Jim is distressed. In fact, it is difficult to imagine a (non-fanciful) context in which these features would partake in a proper explanation of John’s knowledge. Third, the explanandum we are after is that John knows that Jim is distressed, or schematically, that S knows that p (henceforth, ‘KSp’). An adequate explanation of p need not explain KSp and an adequate explanation of KSp need not explain p. For example, while in some contexts Jim’s brain-state may explain Jim’s distress, in non-fanciful contexts Jim’s brain state does not explain John’s knowledge of Jim’s distress. Likewise, while the proposed explanations (a)-(d) above may explain John’s knowledge they do not explain the fact that Jim is distressed. (a) is different from (b)-(d) because (a) entails that Jim received some disturbing news which, in turn, may explain Jim’s distress. This suggests that perhaps one explanation of KSp is that S knows something that explains p. But as the other proposed explanations indicate this is not the only way to explain KSp.
Returning to moral knowledge, explaining the very possibility of moral knowledge, like explaining the very possibility of knowledge of other minds, is a daunting task. In contrast, explanations of knowledge of particular moral facts are widespread and mundane, and it is quite easy to come up with potential answers to questions like ‘How does Jane know that act A is right?’:
- e) Jane knows that act A is a compassionate act.
- f) Jane examined the situation carefully, she understands what is at stake, she is sufficiently disinterested (or involved), she has been in a situation like this before (etc.)
- g) Jane is a very sensitive (or caring, or responsible) person.
- h) Jane was told so by Berta who is extremely reliable about such matters.
In “normal” contexts answers like these are perfectly felicitous and they constitute adequate explanations of Jane’s knowledge of the rightness of A. As in our previous example, if explanations like these are inadequate, we may have to conclude that the fact that Jane knows that A is morally right is inexplicable. That we are not typically perplexed by attributions of moral knowledge together with the availability of acceptable explanations suggests that the inexplicability hypothesis is no more palatable in the moral case than it is in the case of other minds.
In the moral case, like the mental case, the proposed explanans does not necessitate the explanandum. Although the moral supervenes on the non-moral, the set of properties on which the rightness of A supervenes need not be included in a proper explanation of Jane’s knowledge that A is morally right. Moreover, since we typically do not know what the subvenient base is, it is better to explain Jane’s knowledge of the rightness of A without attributing to her extraordinary knowledge of this kind. While (e)-(h) may explain the fact that Jane knows that A is right, they do not explain the rightness of A. Unlike (f)-(h), however, (e) entails that A is a compassionate act which, in turn, may explain the rightness of A. This suggests that perhaps one kind of explanation of Jane’s knowledge that A is morally right is that Jane knows something that explains the rightness of A. But as (f)-(e) indicate, this is not the only way to explain Jane’s knowledge.
There are, of course, important differences between the relation of the mental to the physical and the relation of the moral to the non-moral. For example, the supervenience of the moral on the non-moral is (arguably) knowable a priori, whereas the supervenience of the mental on the physical is (arguably) not knowable a priori.23 Also, (b) above—that John noticed that Jim is restless and out of focus—explains Jim’s distress because (arguably) Jim’s distress causes his restless behaviour. However, in the moral case it is not clear that the rightness of A causes anything. If moral properties are causally inert, then one cannot explain knowledge of such properties by knowledge of their causal consequences. These differences show that the analogy between the two cases discussed earlier can only go so far. But perhaps as far as it goes is far enough—some kinds of explanations seem to be applicable in both the moral case and the mental case.
Let us focus on the explanation of KSp that takes the form that S knows something that explains p—that is, explanations of the kind illustrated by (a) and (e) above. If, as I suggested, the fact that S knows something that explains p can (at least sometimes) explain KSp, then pace M&R, explaining knowledge of particular moral facts doesn’t pose a new challenge for particularism. If particularists succeed in demonstrating that we can explain the rightness and wrongness of actions without appealing to exceptionless moral principles, then what S knows when S knows something that explains p need not include knowledge of exceptionless moral principles. For example, if one can adequately explain the rightness of an act A—say an act of making a charitable donation to a worthy cause—by citing the fact that in doing A the agent of A was neither stingy nor extravagant, without the further commitment that all actions that are neither stingy nor extravagant are right (or even pro tanto right),24 then the fact that Jane knows that in doing A the agent of A was neither stingy nor extravagant may explain her knowledge that A is right without relying on or presupposing the availability of any exceptionless moral principles.
It may seem tempting to argue that the kinds of explanations I propose are felicitous but incomplete. The proposed explanans provide some pertinent information about the explanandum but they succeed as explanations only because they rely on listeners to fill in the gaps. For example, that Jane’s knowledge that act A is compassionate is a felicitous explanation of her knowledge that A is right, is a pragmatic phenomenon—in normal conversational contexts we can make use of information that remains unstated. In claiming that this explanation is adequate, however, we are helping ourselves to this unstated information. But this information, though unstated, is still a proper part of the explanation. And so the explanations, felicitous or not, are literally inadequate.25
It is certainly plausible that the explanations I proposed make use of information that remains unstated. Indeed, one could expand on these explanations in various ways. However, it seems implausible to suppose that these explanations are felicitous only because we can rely on the hearer to complete them by adding exceptionless principles that link the stated explanans to the desired explanandum – at present no such principles are known. One condition of adequacy for a theory of explanation is that some “explanations” are adequate explanations. We may, therefore, have reason to be sceptical of theories of explanation according to which felicitous “explanations” are irrevocably inadequate.26
But if, as I claim, explanations of moral knowledge are so readily available, why have philosophers been so puzzled by knowledge of particular moral facts? And why have explanations of the kinds I proposed been largely overlooked by philosophers?
One reason, I suspect, is that many ethicists still endorse, either explicitly or implicitly, a deductive model of explanation according to which in order to properly explain something one must identify features from which the explanandum can be deduced. The explanations offered above are not deductive—the explanans does not guarantee the explanandum—and so they may have been quickly dismissed because they have been thought to flagrantly fail a basic condition of adequacy for a proper explanation of the phenomena in question. However, a careful examination of the nature of explanation reveals that adequate explanations need not take the form of a deductive argument.27 Consequently, that the explanations offered above are non-deductive is not, in itself, a reason to discard them.
Another related reason why these explanations have not been considered concerns the type of generalizations on which they rely. When we seek to explain knowledge of particular moral facts we may expect that our explanation will tell us at once something informative about all instances of moral knowledge. This expectation parallels a similar expectation regarding explanations of the rightness of actions—namely, that an adequate explanation of the rightness of actions will tell us at once something informative about all right actions. Particularists, however, insist that this expectation is unwarranted. It may well be the case that the only true general claims we can make about all right actions will be in the form of uninformative trivialities like the following: right actions lie in the mean—they are neither excessive nor deficient—and they are performed at the right times, with reference to the right objects, towards the right people, with the right motive, and in the right way. Statements like this do not provide illuminating information about all right actions—they look more like tautologies or vacuous trivialities—and nothing of interest follows from them. However, there are other ways to understand the explanatory role of such statements. Instead of major premises in syllogisms from which the normative status of particular actions can be inferred, we can think of such statements as summaries of, or placeholders for, specific informative explanations. That is, such statements can be viewed as explanatory schemas that provide information about the structure of proper explanations and the kind of information that such explanations ought to include: to explain the rightness of a particular action we must identify a scale on which the action is neither excessive nor deficient and we must “discharge the hedges”—i.e., the right time, the right object, etc.28
Returning to explanations of knowledge of particular moral facts: perhaps the only true general claims we can make about all instances of moral knowledge will be trivialities like the following: all such cases are cases in which the agent gets it right. This, too, looks much like a tautology or a vacuous triviality and nothing of interest follows from generalizations of this kind. Thus, if such generalizations explain anything, they cannot do so by functioning as major premises in syllogisms. Their explanatory role is better viewed as explanatory schemas that tell us something about the kind of information that explanations of knowledge of particular moral facts ought to include. All cases of moral knowledge are cases in which the agent gets it right. Sometimes she gets it right because she knows that in doing A the agent of A was neither stingy nor extravagant. Sometimes she gets it right because she knows that in doing A the agent of A was neither fearful nor overconfident. And sometimes she gets it right because she knows that A is a compassionate action. These explanations are not derivable from the general statement about getting it right. Instead they give this general statement its substance and content. Indeed, general statements about getting it right will not be a part of a proper explanation of knowledge of the rightness and wrongness of particular action at all. So a proper explanation of Jane’s knowledge that A is right might be that Jane knows that A is neither excessive nor deficient and that A was performed at the right time, with reference to the right object, etc.,—when the generic features of this explanatory schema are replaced properly with Jane’s knowledge of specific features of the case at hand.
A third reason why the kinds of explanations I suggested have not been considered concerns the aim of giving such explanations. It may be tempting to think that one aim of the project of explaining moral knowledge is to help us figure out what we need to do in order to obtain moral knowledge. The explanation offered by M&R seems to achieve this objective – if we want to obtain moral knowledge we need to internalize a set of exceptionless moral principles from which, together with some empirical information, we will be able to deduce the rightness or wrongness of individual action. Since a complete explanation of Jane’s knowledge that action A is right will mention the principle from which Jane deduced the rightness of A, then once we have this explanation we will be in a position to obtain moral knowledge ourselves.29 That is, if in order to explain KSp we need to indentify conditions such that whenever they obtain then S knows that p, then any explanation can be used to predict that if these conditions obtain one will know that p. And so, one way to know that p, is to make sure that the relevant conditions obtain.
In contrast, the explanations I proposed cannot help us to figure out what we need to do in order to obtain knowledge of particular moral facts.30 Jane can know that A is a compassionate action or that A is neither stingy nor extravagant and still not know that A is right. For one thing Jane might not believe that A is right. What’s more, Jane might know these things about A and yet A might not be right. So it may seem as though the explanations offered here have no ‘potential predictive force’ and hence that they are inadequate.
This line of reasoning is, to my mind, unpersuasive. First, there is no reason to insist that an adequate explanation of knowledge of particular moral facts should be useful in or applicable for obtaining moral knowledge. It is constructive, here, to attend to a common distinction in normative ethics between explaining the rightness of actions and providing a practicable decision procedure.31 Utilitarians, for example, are quick to point out that even if one could show that one cannot use the principle of utility to determine what one ought to do, it may still be the correct explanation of the rightness/wrongness of actions.32 A similar distinction may be expedient in moral epistemology. It is one thing to explain moral knowledge and it is another thing to offer advice on how to obtain moral knowledge. Hence, that an explanation of moral knowledge doesn’t provide us with practicable advice on how to obtain moral knowledge is no objection to the proposed explanation qua explanation of moral knowledge.
Second, an examination of work on the nature of explanation shows that the “explanation-prediction symmetry thesis”, as we may call it, is not a constraint on explanation that one must accept. Indeed, on all but the classic D-N model of explanation this constraint is abandoned. As Scriven observed in 1959, “we have more data for explaining than we did for predicting” (469) because when we set out to explain p we are not asking whether p but rather we can use p as part of the data we have at our disposal.
A fourth reason why the kinds of explanations I proposed have not generally been considered is this: even after being presented with these explanations one might still be puzzled about moral knowledge. For example, one who thinks that moral properties are causally inert, non-natural properties, may still puzzle over how we could know whether, where, or when such properties are instantiated.
Here it is important to pay attention to what it is we have been trying to explain. Earlier we identified two explanatory tasks: explaining the possibility of moral knowledge and explaining knowledge of particular moral facts, and we said that we will focus on the latter task. With this in mind we can see that the resistance considered here to the explanations I proposed is based on a conflation of the two explanatory tasks – it is no objection to an explanation of p qua an explanation of p that it fails to explain something else. That there are features of moral knowledge that remain puzzling after an explanation of knowledge of a particular moral fact is given is to be expected.
However, one might think that the real challenge for moral epistemology is the task of explaining the possibility of moral knowledge – this is what philosophers have been interested in all along. I believe that the explanations I proposed may help us to make some headway towards addressing this challenge as well. If we have adequate explanations of instances of moral knowledge we may be able to explain the possibility of moral knowledge “bottom-up”. If we can properly explain Jane’s knowledge that A is right in many different situations, what aspect of the possibility of moral knowledge is left unexplained? Certainly we haven’t identified the mechanism by which moral knowledge is obtained and it may be quite interesting to try to learn more about this mechanism (or mechanisms). But if there is no puzzle about instances of moral knowledge, questions about the possibility of moral knowledge may seem less pressing and more manageable.
This leads us to a fifth and related reason why the kinds of explanations I proposed have not generally been considered: the “bottom-up” explanation of the possibility of moral knowledge seems to beg the question against the sceptic. In framing the issue in terms of explanations of knowledge of particular moral facts we helped ourselves to the explanandum—namely, that in those instances we are addressing the agent in question does have moral knowledge. When one attempts to explain something, one presupposes the truth of the explanandum at least for the duration of the explanation game. It is no surprise, then, that with this presupposition we are able to show that moral knowledge is possible because in effect we already assumed that it is possible.
It is certainly true that framing the issue in terms of explaining knowledge of particular moral facts we presuppose that such knowledge is possible and to that extent framing the debate in this way begs the question against the sceptic. However, one is unlikely to respond successfully to the sceptical challenge if one accepts the sceptic’s terms. Framing the debate in terms of explanation may help us to endorse a perspective from which the sceptic’s terms no longer seem attractive. The proposal is not meant to resolve the sceptical challenge but rather to dissolve it by offering a framework in which it is sensible to resist the sceptic’s terms.
So while we can understand why explanations of the kind I proposed have been overlooked, I conclude that none of the reasons considered here warrants dismissing such explanations as inadequate explanations of knowledge of particular moral fact. Moreover, taking such explanations at face value may help us to circumvent sceptical challenges to moral knowledge.
I hope to have shown that moral epistemology does not pose a special problem for the particularist research programme. I considered two arguments that purport to show that exceptionless moral principles are essential for adequate explanations of moral knowledge, and I claimed that both arguments are unpersuasive. In “normal” contexts, I argued, particularist-friendly explanations of moral knowledge of particular facts are not so difficult to come by. Moreover, I’ve suggested that if we have adequate explanations of particular instances of moral knowledge, we might also (indirectly) have an explanation of the possibility of moral knowledge. This “bottom-up” approach to moral epistemology is unlikely to appease the sceptic. Nevertheless it may help us dissolve the problem of moral scepticism by making it easier for us to resist the conditions of adequacy for explanations of the possibility of moral knowledge on which the sceptic insists.
Bales R.E., (1971) ‘Act-Utilitarianism: Account of Right-making Characteristics or Decision-Making Procedure?’, American Philosophical Quarterly 8:257–265.
Dreyfus H.L., and Dreyfus S.E. (1990) ‘What is Morality? A Phenomenological Account of the Development of Ethical Expertise’, in Rasmussen, D. (1990) Universalism vs. Communitarianism (The MIT Press) 237–264.
Scriven M., (1959) ‘Truisms as the ground for historical explanations’, in Gardiner P., (ed.) (1959) Theories of History (The Free Press) pp. 443–475.
Zangwill N., , (2006) ‘Moral Epistemology and the Because Constraint’, in Dreier J., (ed.)(2006) Contemporary Debates in Moral Theory (Blackwell Publishing).
* Earlier versions of this paper were presented at the Virtuous Agency Workshop at Uppsala University, the Centre for Ethics and Metaethics at Leeds, the Law and Philosophy Workshop at The Hebrew University of Jerusalem, the 4th RoME Congress at Boulder, and the Moral Philosophy Seminar at Oxford. I received many helpful questions, comments and suggestions from members of the audience on each of these occasions. I am particularly grateful to Daniel Elstein, David Enoch, Carrie Jenkins, Noa Leibowitz, Yair Levy, Sean McKeever, Sydney Penner, Michael Ridge, Pekka Väyrynen and Jeffrey Wisdom. Special thanks go to Tehila Sagy.
1 For this formulation of the particularism-generalism debate, see Leibowitz (2009a). The list of options here mentioned is illustrative rather than exhaustive. The specific details of what an adequate particularist explanation consists of may vary from one particularist theory to another. See Leibowitz (2011).
3 See Dancy (2004) Chapters 2 & 3.
4 One question that M&R do not address is how a person of practical wisdom knows that she is in a situation that calls for moral assessment in the first place. That is, how does she know when to run through her mental checklist of morally relevant features? It seems that even proponents of the checklist model must appeal to some skill or ability in their account of moral knowledge, otherwise they may have to resort to “checklists all the way down”.
5 If disablers and enablers can themselves be disabled or enabled, she also runs through a complete list of meta-disablers/enablers.
6 In his latest book (2004:142–3) Dancy expresses a similar position. He claims that moral knowledge is “more like knowledge-how than like knowledge-that” and that moral knowledge is obtained by the application of a skill of discernment.
7 There may be reasons to think that particular moral judgments are not acquired in this way. See, for example, Dworkin (1995) and Dreyfus and Dreyfus (1990) as well as various sources cited in M&R (2006) ch. 9 section 4.
8 In personal correspondence McKeever pointed out to me that he is happy to allow that moral knowledge is interpersonally transferable (e.g., by testimony) and so that there are other methods for the acquisition of moral knowledge other than the checklist model. The qualification “direct” or “basic” moral knowledge in the text is added in order to emphasize that the focus here is on moral knowledge which does not depend on there being other agents who possess moral knowledge. In any case, particularists are unlikely to accept the claim that unless one is practically wise one can obtain moral knowledge only indirectly.
9 The implausibility of the checklist model becomes apparent if we consider how we know that an act is dangerous (for example). It seems implausible to insist that in order to know that an act is dangerous one must know in advance all possible danger-relevant features (including enablers, disablers, intensifiers and attenuators) as well as the exceptionless principle from which one can deduce that an act is dangerous.
10 Schroeder (2009) raises similar concerns about M&R’s expectations about the scope of practical wisdom.
13 This model makes knowledge of principles prior to, and more fundamental than, knowledge of the normative status of particular actions. It is difficult to see how this model could accommodate the practice of revising principles in response to judgments about cases—a practice that M&R want to preserve (see, e.g., p. 158. But see Kagan (2001) for the view that judgments about cases do not generate data for moral theorizing.) Moreover, I suspect that if M&R’s argument succeeds it should be equally forceful against any form of pluralism. So its conclusion seems to be stronger than M&R recognize.
14 “Tigers have tails” is a paradigmatic example of a generic statement. “Tigers have tails” is true even though some tigers have no tails. The term “normic-statement” is introduced by Michael Scriven (1959) for statements that are not analytic and not refutable by a few counter instances. “The normic statement,” he writes, “says that everything falls into a certain category except those to which certain special conditions apply. And, although the normic statement itself does not explicitly list what count as exceptional conditions, it employs a vocabulary which reminds us of our knowledge of this, our trained judgment of exceptions” (1959:466).
15 Indeed, it is not so difficult to come up with scenarios in which it is anything but obvious that keeping a promise is right making. In contrast, it is much harder, and arguably impossible, to find examples that show that keeping a promise is not normally right making.
17 M&R take Moore’s Open Question Argument to show that moral predicates cannot be analyzed into purely descriptive language and so that moral principles are not analytic. See M&R’s Ch. 5 (and esp. section 5.2).
18 See, for example, Zangwill (2006).
19 Dancy explicitly states that he intends particularism to be compatible with (ST). See Dancy (2004:85–93). For a discussion of the compatibility of (ST) with particularism, see Leibowitz (2009a).
20 I am assuming for the sake of discussion that nihilism is false. This is common ground in the context of the particularism-generalism debate.
23 See, for example, Zangwill (2006).
24 I develop this version of a particularist-friendly explanation of the rightness of actions in Leibowitz (2012). I will say more about it below.
26 For a more detailed discussion of this issue as well as an argument for why a Hempel-style move to “explanation sketches” won’t do, see Leibowitz (2011).
28 The precise details of the particularist explanation of the rightness of actions may vary from one particularist theory to another. The explanation, or “explanation-sketch”, I suggest in the main text is based on the particularist theory I develop in Leibowitz (2012).
29 One explicit expression of this expectation can be found in Ladd (1952:499), where Ladd insists that proper explanations “should provide us with statements that have ‘potential predictive force’”.
30 At least not directly. They may be helpful in other ways. See Leibowitz (2012).
31 For a discussion of the distinction between explaining the rightness of actions and providing decision procedures see Leibowitz (2009b).
32 See, e.g., Bales (1971).
See M&R (2006) Chapters 6 & 7.
See Dancy (2004) Chapters 2 & 3.
Schroeder (2009) raises similar concerns about M&R’s expectations about the scope of practical wisdom.
See, e.g., Huemer (2007).
See, for example, Zangwill (2006).
See, for example, Zangwill (2006).
See Leibowitz (2011).
See, e.g., Bales (1971).