1 Introduction
It is possible to doubt whether there is such a subject as moral epistemology.
Simon Blackburn
1.1 A Working Hypothesis
As its title suggests, this book is an exploration of moral knowledge: its possibility, its sources, and its characteristic vulnerabilities. Although it is far from a comprehensive survey of questions that might be pursued under the heading of “moral epistemology,” it does address, and argue for answers to, a relatively wide range of such questions. These questions include the following: What are the strengths and weaknesses of the method of reflective equilibrium as an account of how we should make up our minds about moral questions? Is this, as many moral philosophers believe, the correct account of how moral inquiry should proceed? What would count as evidence for or against a fundamental moral conviction? Are perception and testimony potential sources of moral knowledge? What, if anything, would be wrong with simply outsourcing your views about moral questions to a moral expert, if you had reason to believe that the expert would be more reliable than you are about such questions? How fragile is our knowledge of morality, compared to other kinds of knowledge? Is it true, as Gilbert Ryle once claimed, that knowledge of the difference between right and wrong fundamentally differs from knowledge of other kinds in that it cannot be forgotten? To what extent are our moral views vulnerable to being “debunked” by empirical discoveries about why we hold them? What is the relationship between being able to justify a moral judgment and knowing that it is true? Should we invest more confidence in relatively abstract, general moral principles that strike us as true, or more confidence in our judgments about the rightness and wrongness of particular actions?
My hope is that the most significant contributions of this book will prove to be its specific proposals about the topics that I address, as opposed to its
lending support to any more general thesis about morality or our cognitive relationship to it. For the most part, these proposals have a substantial degree of independence from one another, so that, for example, even if more or less everything that I say about the topic of moral expertise is off the mark, the discussions that I offer of the method of reflective equilibrium or the way in which observation contributes to moral knowledge might still be substantially correct. Moreover, for most potential readers, not every topic that I address will be of equal interest. With that in mind, I have made some effort to make the discussions of the various topics relatively self-contained. When an argument does appeal to some claim that I have argued for earlier in the book, I mark this in a way that makes it as easy as possible for a reader to selectively follow up on the earlier discussion.1
That having been said, there is a larger theme that both unifies and informs my more specific proposals. I have a working hypothesis that is supported by the arguments offered here to the extent that they are successful. The working hypothesis is this: 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. Given the many ways in which we gain empirical knowledge of the world around us, this working hypothesis is much stronger than anything that I can hope to argue for (let alone establish) here. Nevertheless, I hope to compensate for this to at least some extent by focusing on potential pressure points—areas where the hypothesis seems least plausible, and where the chances for its falsification seem relatively high.
For example, much of what we know about the world depends on evidence that is provided by observation. If the working hypothesis adopted here is true, then it follows that our moral views are susceptible to empirical confirmation and disconfirmation. This implication is one that many philosophers take to be false. Against this, I will defend the claim that even relatively fundamental moral convictions are in principle susceptible to being confirmed or disconfirmed by non-moral observations.2 Similarly, it is clear
1 In addition, a short concluding chapter contains a summary of the main claims endorsed in the book, together with references to the specific sections where those claims are discussed. This too can potentially be used as guide for selective reading.
2 As will emerge below, I think that opinions to the contrary often rest on inadequate pictures of confirmation that we have reason to reject independently of anything having to do with morality.
on reflection that a great deal of our ordinary knowledge about the world depends on the testimony of others, including many cases in which we are not in a position to verify the relevant truths ourselves. But many have thought that full-fledged moral knowledge requires the knower to recognize or appreciate why the relevant moral claim is true. Against this, I will defend the possibility of arriving at full-fledged moral knowledge via testimony or (more broadly) deference to another person. I argue that what is legitimate in our unease about deferring to another person about morality does not derive from the impossibility of moral knowledge being transmitted or acquired in this way but rather from other sources.
Of course, the working hypothesis I endorse also entails that moral knowledge is susceptible to being undermined or lost in all of the ways in which non-moral knowledge is. Interestingly, some philosophers—Gilbert Ryle, Ronald Dworkin, and Thomas Nagel among them—have held that there are important respects in which our cognitive relationship to morality is more secure than our cognitive relationship to ordinary empirical knowledge.3 I explore several possibilities of this sort in Chapter 5, “Losing Moral Knowledge.” Although I defend the claim that moral knowledge is susceptible to being undermined in the same ways in which non-moral knowledge is, I also think that careful consideration of the proposed ways in which this might fail to be true both raises, and suggests answers to, a number of interesting and underexplored questions. These include questions about the extent to which some cases of moral corruption are best understood as cognitive processes (i.e. processes involving a loss of knowledge), as well as questions about the kinds of considerations that could in principle make it reasonable for us to lose confidence even in moral claims that strike us as obviously correct.
1.2 Some Methodological Preliminaries and Assumptions
A philosopher who announces that she intends to undertake an investigation of moral knowledge is apt to encounter skepticism about her project from a
Indeed, even those who defend the possibility of empirically confirming a moral claim sometimes assume an overly demanding picture of what this would involve. This has led to the view that the empirical confirmation or disconfirmation of a moral claim would require that the moral claim in question be in some sense reducible to non-moral, empirical claims. Against this, I will argue that even if a moral claim is irreducibly normative, it does not follow that non-moral observations cannot provide evidence for or against it.
3 For example, in the course of defending “normative realism” Nagel claims that “The truth here could not be radically inaccessible in the way that the truth about the physical world might be” (1986: 186).
number of different directions. Perhaps the most immediate and straightforward kind of skepticism that she confronts is skepticism about whether there is any such thing as moral knowledge at all. More generally, an inquiry into the ways in which we acquire moral knowledge might seem to presuppose answers to difficult metaethical questions that should be considered open, and therefore to beg the question against various metaethical views that are very much live theoretical options. Presumably, the existence of moral knowledge entails the existence of moral facts, no less than the existence of historical knowledge entails the existence of historical facts. And the claim that there are moral facts is denied, both by error theorists as well as by non-cognitivists who hold that the function of moral discourse is something other than the stating of facts. Even among philosophers who are prepared to countenance moral facts, the range of opinion as to the nature of such facts might seem too vast to allow for profitable discussion of moral knowledge as a relation that we sometimes stand in to these relata. In this respect, moral epistemology might seem premature—task that is better left until further progress has been made in moral semantics and moral metaphysics, areas which are sometimes treated as more fundamental.
While I feel the force of this line of thought, I do not believe that it warrants setting aside the kind of investigation that I undertake in this book. In part, this is because I am skeptical that moral metaphysics and moral semantics enjoy the kind of methodological priority to moral epistemology that this line of thought seems to assume. Consider the relationship between moral metaphysics and moral epistemology. While there is an obvious respect in which the question of whether there are any moral facts seems prior to the question of how we could come to know such facts, as a historical matter reflection on the second question has often influenced answers to the first. In particular, a prominent source of skepticism about whether there are any moral facts at all, or moral facts in anything like the relatively straightforward sense which common sense often seems to assume, has been dissatisfaction with existing accounts of how we might come to know such facts.4 (Although, of course,, there are other sources of skepticism about the existence of moral facts.) Significantly, all of the mechanisms under consideration in this book—including interlocution, perception, empirical confirmation and
4 Thus, one of the main motivations for Mackie’s (1977) error theory is his claim that, if we were aware of moral facts, “it would have to be by some special faculty of moral perception or intuition, utterly different from our ordinary ways of knowing everything else” (p. 38). Similarly, in the course of developing an argument against “evaluative realism,” Street (2006: 143) objects to the postulation of a “highly specialized, sophisticated capacity, one specifically attuned to the evaluative truths in question.”
disconfirmation, and reasoning toward a more coherent overall view—are generally acknowledged to be mechanisms that regularly deliver knowledge of non-moral facts. The only question is whether they also manage to deliver moral knowledge, or could deliver such knowledge if there were moral facts to be known. The more that it can be made plausible that these mechanisms are or would be suitable for delivering moral knowledge, the more one of the traditional reasons for doubting the existence of moral facts is weakened.5
Moreover, to the extent that the arguments that I offer here are successful, this would be relevant not only to the question of whether moral facts exist, but also to the question of what such facts would be like. For example, suppose that our moral beliefs are susceptible to empirical confirmation and disconfirmation regardless of whether they are reducible or equivalent to empirical claims. If this is the case, then at least one motivation for a reductionist conception of morality is undercut.6
Consider the following claims:
Rape is wrong.
Slavery in the antebellum American South was unjust.
One shouldn’t encourage children to smoke cigarettes.
It was wrong for the 9/11 hijackers to deliberately fly planes into the Twin Towers.
If one is in a position to help others in desperate need at little cost to oneself, then one should.
It is right to keep one’s promises.
It is good for parents to vaccinate their children against polio.
5 Byrne (2018) usefully distinguishes between “extravagant” and “economical” accounts of our knowledge of a given domain. An account is extravagant just in case it invokes a special purpose, dedicated faculty or capacity in accounting for our knowledge of the relevant subject matter, on the grounds that more general purpose capacities are inadequate to deliver such knowledge. In contrast, an account is economical just in case it holds that our knowledge of the subject matter is delivered by general purpose epistemic capacities, capacities that also deliver knowledge of distinct subject matters. For example, Chomsky’s (1998) account of our knowledge of language is extravagant as opposed to economical, because it entails that this knowledge requires a dedicated faculty, “a language organ.” In contrast, accounts of our knowledge of language that attempt to account for such knowledge in terms of the application of more general learning mechanisms count as economical. In terms of this dichotomy, the arguments of this book tend to support (to the extent they are successful) an economical as opposed to an extravagant account of our moral knowledge.
6 To be clear, and despite what might be suggested by the way in which the current study proceeds, I do not think that moral epistemology has any methodological priority to moral metaphysics, moral semantics, or moral psychology, either. My own view is that each of these sub-areas stands to gain from incorporating the genuine insights that ultimately emerge from each of the others, and that from the perspective of the division of cognitive labor it makes sense for different inquirers to set out from different starting points.
Plausibly, these claims are true, and known to be true by many people. Some philosophers hold that there are “Moorean truths,” philosophically incontestable fixed points that it is the job of the philosopher to respect, accommodate, and build upon in his or her theorizing, as opposed to either vindicate or refute.7 While I do not claim this status for the propositions listed above, they are among the closest approximations that the moral domain has to offer. Even if they are not beyond challenge, it seems reasonable to assume provisionally that they are true, in order to investigate how claims of the relevant kind might be known. This is a large project, to which the current study aims to make a modest contribution.8
Throughout the book, I invoke the concept of knowledge without offering an analysis (reductive or otherwise) of knowledge. In this respect, if not in others, this book is an exercise in “knowledge first” epistemology, in the tradition of Williamson (2000). Unlike in Williamson’s case, my proceeding in this way is not rooted in a conviction that there is no analysis of knowledge to be had, but rather in agnosticism about this question together with certainty that I myself do not know of such an analysis. I do, however, make the following two assumptions about knowledge.
First, regardless of what the standards for moral knowledge are—and even whether we have any moral knowledge at all—those standards are not material ly different from the standards for knowledge in general. In the absence of some compelling reason for thinking that things are otherwise, we should prefer a unified account of knowledge, one on which the standards that must be met in order to count as knowing a proposition do not vary from domain to domain. Notice that, even if we adopt this as a default assumption, it does not in any way beg the question in favor of the working hypothesis
7 For representative statements, see especially Armstrong (1999) and Lewis (2001: 418). For a systematic treatment, see Kelly (2005b).
8 Thomson (2013: 54–5) suggests that G.E. Moore (1925) should have included some moral propositions among his famous list of commonsense platitudes. Cuneo and Shafer-Landau (2014) is a recent attempt to vindicate the idea that there are some truths that have the status of methodological fixed points in the moral domain, although not all of the claims listed above are candidates for such truths, given the account that they offer. By contrast, McPherson (2009) argues that appeals to (alleged) Moorean facts face greater obstacles in the moral domain than elsewhere, while Dreier (2007: 241–2) takes issue with the tendency of “moral absolutist philosophers” to see themselves as defenders of common sense against the bizarre, on a par with traditional epistemologists who defend our ordinary knowledge claims about the external world against radical skepticism.
Although the methodological default assumption adopted here (i.e. that some moral claims like those enumerated above are known to be true) is a substantive one, it is also much weaker than those that are sometimes made. For example, Enoch (2018) claims that the default position in metaethics is non-naturalist realism (see especially pp. 30–3, “Obviously, the default position”), a view that he characterizes as entailing that moral facts are “abstract, outside of space-time, causally inert [and] utterly independent of human responses and attitudes” (p. 35).
that the sources of non-moral, empirical knowledge are also sources of moral knowledge. Generally speaking, even if the standards for knowing a proposition are the same across different domains, the sources of knowledge might vary greatly from domain to domain. For example, it is plausible that we can arrive at mathematical knowledge via a priori reasoning, although it is obvious that we cannot similarly arrive at geographical knowledge via such reasoning. Suppose then that a priori reasoning is a source of mathematical but not geographical knowledge. Even on that assumption, it does not follow that the standards or conditions one must meet in order to count as knowing a mathematical proposition differ from the standards or conditions that one must meet in order to count as knowing a geographical proposition. In both domains, knowledge might consist of (for example) sufficiently reliable true belief, or belief that is sufficiently safe (cf. Williamson 2000), or justified true belief that satisfies an additional “anti-Gettier” condition, or warranted true belief (in the sense of Plantinga 1993), or a belief that tracks the truth (cf. Nozick 1981), or . . . .
Notice that this is a point on which we should expect even the moral skeptic, who denies that we have any moral knowledge, to agree. In fact, it seems that the skeptic should insist that the standards for moral knowledge do not differ in any significant way from the standards for knowledge of other subject matters. For the claim that we lack moral knowledge is potentially much less interesting if the sense in which we fall short of moral knowledge involves our falling short of standards that differ from those required for knowledge of other subject matters.
The second assumption concerns the relationship between knowledge and (epistemically) justified belief. I will assume that, if one knows that p, then one is justified in believing that p. (Equivalently: if one knows that p, then it is not the case that one is not justified in believing p.) This assumption is widely accepted, including by many philosophers who differ significantly in their substantive accounts of knowledge and justification. However, it is a substantive assumption that is not completely uncontroversial.9 Of course, on the traditional view according to which knowledge can be (partially) analyzed in terms of justified true belief, it follows immediately that one is justified in believing anything that one knows. However, even if we attempt to understand knowledge in other terms, or as conceptually fundamental, the idea that
9 A notable dissenter is Robert Audi (2003: 235–9), who combines a strongly externalist account of knowledge with a strongly internalist account of justification. Audi’s discussion does not engage with the kind of consideration in favor of orthodoxy that I offer in the text above. For a more general critique of Audi’s account of justification, see Williamson, “On Being Justified in One’s Head” (2007).
one is justified in believing anything that one knows remains compelling. Consider the following assertion: “It’s true that you know that it rained this afternoon, but you’re not justified in believing that it rained this afternoon!” That assertion seems infelicitous, as does any assertion that replaces “it rained this afternoon” with an arbitrary proposition. The most straightforward explanation for this infelicity is the following: if one knows that p is true, then one is justified in believing that p is true. By contrast, on views on which unjustified beliefs can qualify as pieces of knowledge, we would expect that assertions of the relevant kind would be perfectly in order.
1.3 Towards a More Social Moral Epistemology
Another common thread that unifies many of the arguments and views presented in this book is that our access to moral knowledge has an important social dimension. This theme is most obvious in Chapter 3, “Moral Knowledge from Others.” There, I explore a number of philosophical issues raised by the possibility of arriving at moral views by relying on other people. I defend what I call the Moral Inheritance View, according to which a person whose earliest moral views are inherited from her social environment might very well have substantial moral knowledge even before she is in a position to begin critically reflecting upon or reasoning about those views. On this picture, a significant part of our earliest moral knowledge is due to our natural tendency to adopt beliefs that are held by those around us, or that are presupposed by common practices. More generally, other people are in principle potentially rich sources of moral knowledge. To the extent that we have reservations about the propriety of forming moral views by relying on others—as opposed to through the exercise of our own autonomous judgment—what is legitimate in those reservations does not derive from its being impossible to acquire moral knowledge in this way.
Although the theme is most obvious in Chapter 3, it plays a significant role in other chapters as well. For example, in Chapter 4, “Experience and Observation,” I argue that the phenomenon of moral testimony has important implications for the possibility of confirming moral views by non-moral observations (see especially section 4.3). I also argue that the fact that one is a member of a moral community, and thus is in a position to compare the moral opinions of others with one’s own, can contribute to moral knowledge, not only by affording evidence for or against those opinions, but in a more subtle way as well: by providing feedback that can serve to condition or
calibrate one’s capacity for judgment so that future exercises of that judgment are more likely to deliver knowledge (see especially section 4.4). In Chapter 5, lessons drawn from the social dimensions of moral knowledge play a role in my attempt to solve Gilbert Ryle’s puzzle about “forgetting the difference between right and wrong.”
The fact that our access to moral knowledge has an important social dimension is both bad news and good news when it comes to our ability to acquire and maintain such knowledge. When other people hold mistaken moral views, this can amount to misleading evidence, and the existence of such evidence can undermine moral knowledge, or make such knowledge more difficult to acquire or maintain than it otherwise would have been (see sections 4.3 and 5.6). In the moral domain, a particularly interesting and potentially important source of misleading evidence is the following: people who either are or seem to be morally upstanding people in other respects might engage in a practice or behavior that is in fact seriously morally wrong, in a social context in which its wrongness is not generally recognized. From the perspective of a member of the society who is currently uncertain that the practice is wrong, the fact that a significant number of people who are morally upstanding in other respects regularly engage in the practice might constitute strong (albeit misleading) evidence that the practice is not wrong. This mechanism for generating misleading evidence can pose formidable practical difficulties for would-be moral reformers, who seek to convince other members of their society that some common behavior or practice is seriously wrong (see section 4.3.5).
I emphasize the idea that our access to moral knowledge has a significant social dimension not only because I think that it is important, but also because I think that it has been at least somewhat underemphasized within moral epistemology. Although it is clear that the contents of our moral views frequently concern relations that obtain between different people, the processes by which we arrive at those views are often depicted as though they were solitary enterprises. In part, this tendency within moral epistemology might simply reflect (or be a special case of) a historical tendency to neglect the social aspects of knowledge within epistemology more generally.10
A significant portion of the history of Western epistemology since Descartes
10 In the words of Alvin Goldman, arguably the most prominent living epistemologist: Until recently, epistemology . . . was heavily individualistic in focus. The emphasis was on evaluating doxastic attitudes (beliefs and disbeliefs) of individuals in abstraction from their social environment. The result is a distorted picture of the human epistemic situation, which is largely shaped by social relationships and institutions. (2015: 1)
has been concerned with questions about how to respond to extremely radical forms of skepticism, for example, about whether and how someone can know anything beyond the present contents of her own mind. Unsurprisingly, such a focus tends to divert attention from epistemological questions that presuppose the existence of a rich social world of which we are members, questions that it is tempting to see as less basic or fundamental. In addition to this traditional concern with skepticism, many of the other topics that dominated the attention of mainstream epistemologists in the second half of the twentieth century—for example, the correct analysis of “S knows that p,” or the debates between foundationalists and coherentists and between internalists and externalists—similarly tended to encourage a focus on the individual knower in abstraction from the social environment that she occupies. In the final years of the twentieth century and the opening decades of the twenty-first, this situation has begun to change. We now see, in mainstream epistemology, sustained attention to topics such as the nature of testimony and the epistemology of “peer disagreement.”11 Like many others, I think that this is a positive development, and I hope that this book might serve to further encourage this tendency in the branch of epistemology concerned specifically with our cognitive relationship to morality.
I suspect that there is a special reason why the social aspects of knowledge have been underemphasized within moral epistemology in particular. For many moral philosophers, the method of reflective equilibrium is the correct account of how moral inquiry should ideally be conducted, and thus stands at the very center of moral epistemology. At least as it is often presented and understood, the method naturally encourages an individualistic picture of moral inquiry: in principle, an individual could flawlessly execute the method while sitting alone at her desk.12 Although the method of reflective equilibrium embodies genuine and important insights, I will argue that it does not have the kind of centrality for moral epistemology or moral inquiry that has been claimed for it. Exploring its virtues and its limitations is the central task of the next chapter.
11 For an overview of developments in this area, see Goldman and Blanchard (2016); for a representative collection of work, see the anthology edited by Goldman and Whitcomb (2011).
12 As we will see, there are at least some understandings of the method on which this is not the case; for discussion, see section 2.4.
2 Reflective Equilibrium, Its Virtues and Its Limits
2.1 Introduction
If asked for an account of their own methodology, or for an account of the way in which we can arrive at justified views about the subject matter with which they are professionally concerned, many moral philosophers would appeal to the method of reflective equilibrium. 1 Indeed, prominent moral philosophers sometimes suggest that when it comes to moral inquiry, the method of reflective equilibrium is, in effect, the only game in town. Thus, according to Michael Smith, it is among the “platitudes” about morality that properly conducted moral inquiry has “a certain characteristic coherentist form” of a kind that was given systematic articulation by John Rawls in his seminal discussion of the method (Smith 1994: 40–1). Similarly, Thomas Scanlon says:
it seems to me that this method, properly understood, is in fact the best way of making up one’s mind about moral matters and about many other subjects. Indeed, it is the only defensible method: apparent alternatives to it are illusory. (2002: 149)2
Kagan (1998: 16, 306) suggests that, in practice, something very much like the method of reflective equilibrium is accepted by anyone doing normative ethics, whether they realize it or not. Harman (1977: 79) argues that a fundamental division among moral philosophers is that between philosophical naturalists and practitioners of “autonomous ethics,” but claims that the method of reflective equilibrium is the shared method of both.
1 See, e.g., Boyd 1998; Daniels 1996, 2003; DePaul 1987, 1998, 2006; Dworkin 1996: 119; Ebertz 1993; Harman 1977, 2004; Huemer 2005: 117; McMahan 2000; Rawls 1971, 1993, 1999, 2001; Scanlon 2002, 2014; Smith 1994; Sturgeon 1986: 74, 2002: 184–5; Tersman 2018; and Wallace 1998: 113–14.
2 Compare DePaul (2006: 616) who argues that, when it comes to moral inquiry, “there is simply no reasonable alternative to reflective equilibrium.”
Moral Knowledge. Sarah McGrath, Oxford University Press (2019). © Sarah McGrath. DOI: 10.1093/oso/9780198805410.001.0001
The popularity of the method of reflective equilibrium should not obscure the diversity of ways in which the method is understood by those who profess their allegiance to it. Thus, some view the method as a kind of coherentist account of justification (Daniels 1996, 2003; Brink 1989; DePaul 1987; Smith 1994; Sayre-McCord 1996), while others view it as a species of foundationalism (Harman 2004; McMahan 2000). Some view the method as a rival to moral intuitionism (Daniels 1996: 26, 83) while others think that the two are perfectly compatible or even complementary (Huemer 2005:117; cf. Audi 1997: 49–58). Some think that the domains that can be fruitfully investigated by the method are limited to those that admit of investigation from the armchair, while others insist that the method of reflective equilibrium is the method of the empirical sciences as well (see especially Boyd 1988: 199–200, 207). Some think that the method descends from Aristotle (Nussbaum 1990:172–6), while others regard it as an invention of the twentieth century, and credit its discovery to Nelson Goodman (1953) and John Rawls (1971). Most view the method as a procedure for figuring out what one should believe, but its most famous and influential advocate, Rawls, often wrote of it as a procedure whose purpose is to unearth descriptive, psychological facts about our underlying “moral conceptions” (1971: 48–51; 1974: 288–91). Similarly, most think that the “considered judgments” on which the method operates are the considered judgments of the individual who is employing it, but others think that the only admissible considered judgments are those that are held in common by some larger group (see e.g. Smith 1994:40–1, following Rawls 1951).
There are, then, significant differences in how the method is understood by its proponents, some of which we will be concerned with below. Underneath these differences, however, lie common themes and commitments. At this point, it will be helpful to have a clear statement of the method that articulates some of these common themes and commitments. For this purpose, the recent account provided by Scanlon (2014: 76–7) is particularly useful, and worth quoting at length:
In broad outline, the method of reflective equilibrium can be described as follows. One begins by identifying a set of considered judgments, of any level of generality, about the subject in question. These are judgments that seem clearly to be correct and seem so under conditions that are conducive to making good judgments of the relevant kind about this subject matter. If the subject in question is morality, for example, they may be judgments about the rightness or wrongness of particular actions, general moral principles, or judgments about the kinds of considerations that are relevant to determining
the rightness of actions The method does not privilege judgments of any particular type—those about particular cases, for example—as having special justificatory standing.
The next step in the method is to formulate general principles that would “account for” these judgments. By this Rawls means principles such that, had one simply been trying to apply them, rather than trying directly to decide what is the case about the subject at hand, one would have been led to this same set of judgments. If, as is likely, this attempt to come up with such principles is not successful, one must decide how to respond to the divergence between these principles and considered judgments: whether to give up the judgments that the principles fail to account for, to modify the principles, in hopes of achieving a better fit, or to do some combination of these things. One is then to continue in this way, working back and forth between principles and judgments, until one reaches a set of principles and a set of judgments that “account for them.” This state is what Rawls calls reflective equilibrium.
This account of the method is broadly representative of the way in which the method has been understood by its proponents within moral philosophy.3 For example, here is a theme Scanlon emphasizes that would, I think, be generally accepted by proponents of the method of reflective equilibrium in the moral domain:
COHERENCE: In determining which moral views to hold, one should attempt to make one’s moral views more coherent and systematic.
In this context, the pursuit of greater coherence includes, but is not limited to, the elimination of conflicts among one’s moral views: the relevant notion of coherence is richer and more demanding than mere logical consistency. A set of moral views that consisted exclusively of views that did not overlap in content or in what they entailed would be perfectly consistent but would not on
3 See, for example, Daniels (1996: 48–9, 2011); DePaul (1998: 294–6 and 2006: 599–602); Harman (2004: 3); McMahan (2000: 100–1); Rawls (1971: 48–51, 1999: 289, 1980: 534, 2001: 29–31); SayreMcCord (1996: 141); Sturgeon (2002: 184–5; and Tersman 2018: 2. Compare Goodman 1953: 63–4, Lewis 1983: x–xi), and Scanlon (2003): 140–1.
More generally, when concreteness is called for in what follows, I will often focus on Scanlon’s (2014) discussion and defense of the method. In addition to the representativeness of his characterization, there is a second reason why this focus seems appropriate: Scanlon’s defense of the method is at least as sophisticated as any that has yet been offered. It is, for example, highly sensitive to, and resourceful in addressing, the kinds of concerns about the method that have been raised repeatedly since its popularization among moral philosophers as a result of Rawls’ (1971) influential presentation.
that account have a high degree of coherence. In order for a set of views to have a high degree of coherence, its members must positively support one another. Such support occurs, for example, when an independently plausible general moral principle “accounts for” or explains the truth of a moral judgment about a particular case that is itself independently plausible.
A second theme emphasized by Scanlon that would be generally accepted by proponents of the method is this:
NO PRIVILEGE: In revising one’s moral views in the course of pursuing greater coherence and systematicity, one should not privilege any view over any other simply in virtue of its level of generality.
Suppose, for example, that one realizes that a general moral principle that one has accepted up until now is in conflict with one’s judgment about the moral status of a token action. At this level of abstraction, there is simply no answer to the question of whether one should favor the judgment about the token action or the judgment about the general principle. The fact that one judgment concerns a general principle while the other concerns a token action is not itself a reason for resolving the conflict in one way rather than the other.
In the discussion that follows then, I will assume that accepting both COHERENCE and NO PRIVILEGE is a necessary condition for accepting the method of reflective equilibrium.4 However, we should not assume that accepting these two ideas suffices for accepting the method. For COHERENCE and NO PRIVILEGE might be accepted by a very wide range of philosophers, including some whose views about moral epistemology and moral inquiry differ radically from those held by Scanlon, Rawls, and other paradigmatic proponents of the method. For example, consider a hypothetical philosopher whom I will refer to as the Caricaturized Moral Intuitionist. The Caricaturized Moral Intuitionist holds that our ability to arrive at moral knowledge depends
4 Although I take this assumption to be well-motivated by relevant texts, it would not be universally accepted. In particular, in his “In Defense of Reflective Equilibrium” (2013), Walden contends that “It is a mistake to try to give a positive characterization of the method of reflective equilibrium” (p. 244). Rather, acceptance of the method is best understood as acceptance of a “negative thesis” (p. 255) to the effect that nothing substantive can be said in advance of an inquiry about what inputs, methods, and goals are appropriate for that inquiry. While I think that this methodological thesis is an interesting one, I don’t think it’s plausible to identify it with the method of reflective equilibrium. In addition to failing to accord with the characterizations of the method offered by its leading proponents (see the specific passages referenced in footnote 15 for textual substantial of this point), notice that on this usage “the method of reflective equilibrium” does not actually refer to a method—and the connections with reflection and equilibrium also seem to be lost. Terminology aside, I note here that none of the criticisms that I offer of the method in what follows, nor any of the virtues that I attribute to it, have any direct bearing on the negative thesis defended by Walden, at least as far as I can tell.
entirely on the operation of an occult, sui generis faculty of moral intuition, and no account of moral inquiry that neglects to mention the central role of this faculty could possibly be adequate. On this view, just as it is the job of the more familiar five senses to deliver information about the empirical world, so it is the job of the faculty of moral intuition to deliver information about the moral realm. The Caricaturized Moral Intuitionist holds that this faculty, like the more familiar five senses, is a fallible source of information, although it is generally reliable when used in hospitable circumstances and regularly provides us with knowledge of both general moral principles and particular moral claims.5
Notice that the Caricaturized Moral Intuitionist might very well accept both COHERENCE and NO PRIVILEGE. First, she might very well accept that it is important to seek coherence and systematicity among our moral views. Precisely because she regards the faculty of moral intuition as fallible and susceptible to error, she thinks that the attempt to make our moral views more coherent serves as an important check on its deliverances, and a way of weeding out mistakes that have resulted from previous misfirings. (Compare the way in which seeking coherence among the views of our surroundings that we arrive at via sense perception can play a role in weeding out mistaken perceptual judgments.) Moreover, the Caricaturized Moral Intuitionist might very well see the pursuit of greater coherence and systematicity as a way of positively extending the moral knowledge that is delivered by our faculty of moral intuition. She conceives of this process as analogous to the process by which scientific theorizing can extend the empirical knowledge that is provided to us by the more familiar five senses. Her view is not that reasoning aimed at the achievement of coherence cannot extend our moral knowledge; rather, it is that we would not be in a position to attain any moral knowledge at all if not for the operation of the faculty of moral intuition. Similarly, the Caricaturized Moral Intuitionist might happily accept the idea that no moral judgment is privileged over another simply in virtue of its level of generality. Perhaps her view is that while the faculty of moral intuition can deliver knowledge of both moral principles and more specific moral
5 Recall from Chapter 1 Mackie’s claim that if we were aware of moral facts, “it would have to be by some special faculty of moral perception or intuition, utterly different from our ordinary ways of knowing everything else” (1977: 38). As I am imagining her, the Caricaturized Moral Intuitionist agrees with Mackie that it is a necessary condition of our having any moral knowledge at all that we have some such special faculty; unlike Mackie, however, she holds that we are fortunate enough to have been endowed with such a thing. In the terminology of 1.1 (cf. fn. 4), the Caricaturized Moral Intuitionist subscribes to an “extravagant” moral epistemology, while it’s natural to understand the reflective equilibrium theorist as subscribing to an “economical” one.
claims, it is fallible with respect to both, and no more likely to be mistaken about one as opposed to the other. For this reason, she might hold that it would be a methodological mistake to privilege one moral judgment over another simply in virtue of its level of generality.
Thus, the Caricaturized Moral Intuitionist might very well accept both COHERENCE and NO PRIVILEGE. Nevertheless, she would adamantly deny that the method of reflective equilibrium is an adequate characterization of the epistemology of morals. When she reads accounts of moral inquiry offered by proponents of the method, these accounts strike her as at best radically incomplete. Indeed, she holds that we would have a great deal of moral knowledge even if we never engaged in anything like reflective equilibrium reasoning, and that on those occasions when such reasoning does deliver moral knowledge, this process is typically made possible by the fact that we have a substantial amount of moral knowledge that is provided by another source. By her lights, an account of how we acquire moral knowledge that neglects to mention the role of this other source is little better than an account of our empirical knowledge that leaves out the essential role played by sense perception.
The Caricaturized Moral Intuitionist offers us a badly inadequate moral epistemology, for we have no special faculty of moral intuition. Nevertheless, her attitude toward the method of reflective equilibrium is instructive. Like her, we should accept both COHERENCE and NO PRIVILEGE. But we should also agree with her that reflective equilibrium reasoning is less central to moral epistemology than it has sometimes been taken to be, and that this is true regardless of whether our concern is with the cognitive practices of ordinary moral agents or with those of more theoretically-minded philosophers. In particular, both ordinary moral agents and moral philosophers typically have a substantial amount of moral knowledge that does not derive from the method of reflective equilibrium. Moreover, on those occasions when the pursuit of coherence and systematicity does lead to an improvement in one’s moral views (either by leading one to jettison mistaken moral views or by extending one’s knowledge), this process is typically made possible by a background of substantial moral knowledge that is already in place; in the absence of moral knowledge from other sources, reflective equilibrium reasoning would be poorly suited to delivering such knowledge.
Two philosophically significant implications of this picture are the following. First, because standard accounts of the method of reflective equilibrium tend to neglect or at least obscure the role played by prior moral knowledge from other sources in the reasoning process, such accounts tend to under-characterize
good reasoning about morality. To the extent that such accounts are offered as accounts of how we should make up our minds about moral questions, they stand in need of amendment or at least supplementation, for the descriptions that they offer are consistent with paradigmatically bad reasoning as well as with paradigmatically good reasoning.
Secondly, to the extent that there is a genuine philosophical puzzle about “where moral knowledge comes from,” or how we could acquire such knowledge in the first place, “the method of reflective equilibrium” is not a good candidate for an answer to that puzzle. This is because the capacity of reflective equilibrium reasoning to deliver new moral knowledge typically depends on the reasoner’s already having substantial moral knowledge from other sources.
Perhaps some philosophers have understood the method of reflective equilibrium liberally enough so that anyone who accepts both COHERENCE and NO PRIVILEGE counts as accepting the method, including the Caricaturized Moral Intuitionist. While this strikes me as a suboptimal way of dividing up the terrain, I have no desire to argue over terminology. Especially if we allow for such an inclusive understanding of the method, we should distinguish between modest and ambitious interpretations of the method and its epistemological role. On a modest interpretation, anyone who endorses both COHERENCE and NO PRIVILEGE counts as endorsing the method. On the other hand, those who accept an ambitious interpretation of the method and its epistemological role tend to see what justification we ultimately have for our moral views as in some way arising out of, or dependent upon, reflective equilibrium reasoning. On an ambitious interpretation, the method of reflective equilibrium is a potential answer to an epistemological question about morality: how can we arrive at knowledge or justified beliefs about morality in the first place? Like the Caricaturized Moral Intuitionist, I accept the method of reflective equilibrium when it is interpreted modestly but reject it when it is understood ambitiously.
I will not engage in the exegetical exercise of attempting to determine which philosophers have accepted an ambitious interpretation of the method. But it seems pretty clear that at least some of its greatest proponents have tended to understand it in this way. Consider, for example, Nelson Goodman’s (1953: 61–3) seminal presentation of the method in the context of discussing the justification of logic and principles of inductive reasoning. There, Goodman explicitly champions the method as a preferable alternative to existing accounts of justification about the relevant subject matters, accounts that he takes to be unsatisfactory. Notably, Goodman’s discussion is frequently
cited with approval by moral philosophers who endorse the method as the correct account of justification in the moral domain.6 More recently, Scanlon (2014) claims that “the only way we have of establishing the truth of normative judgments is through direct, piecemeal application of the method of reflective equilibrium” (pp. 122–3, emphasis mine). Moreover, Scanlon explicitly presents the method of reflective equilibrium as an adequate answer to Mackie’s (1977) challenge of how we are able to achieve moral knowledge at all, given our apparent lack of any special faculty of moral perception or intuition (2014: 69–72). Both of these aspects of his discussion are at least strongly suggestive of what I have called an ambitious understanding of the method and its epistemological role.
According to the working hypothesis endorsed in Chapter 1, any source of ordinary empirical knowledge is also a potential source of moral knowledge. Given that one of the ways in which we acquire ordinary empirical knowledge is by attempting to make our beliefs about the world more coherent and systematic, the working hypothesis does not stand in any tension with the method when the method is interpreted modestly. However, when the method is interpreted ambitiously, the picture of moral epistemology that results does conflict with the picture suggested by the working hypothesis. For on an ambitious interpretation, the method is the source of our moral knowledge, or the primary source; and this idea conflicts with the alternative picture suggested by the working hypothesis, on which the sources of such knowledge are a diverse lot.
Let me conclude this section by summarizing the views about the method of reflective equilibrium for which I will be arguing in the rest of this chapter. On the one hand, I will argue that the method embodies genuine and important insights. These insights include the following:
(1) In moral inquiry, we never start from scratch. When we engage in moral inquiry, we typically come to the table already holding any number of substantive moral views, and there is no general requirement to set aside or bracket such views for purposes of the inquiry. On the contrary, exemplary moral reasoning typically makes heavy use of such convictions. It is not rationally required, or even desirable, to attempt to conduct moral inquiry from some more austere starting point (e.g. from some purely formal moral principle or principles, as suggested by Hare 1973, among others).
6 For an extended discussion of Goodman on the method, along with references, see Kelly and McGrath (2010): 329–34, 352.
(2) We can often improve our moral views by attempting to make them more coherent and systematic. Indeed, given that we have at least some substantial moral knowledge, the process of seeking greater coherence and systematicity among our moral views is likely to be a fruitful way of extending this knowledge.
(3) Contrary to what many philosophers have supposed, considerations having to do with the levels of generality of our moral judgments are not of any deep methodological significance.
(4) One consequence of (3) is that, when we pursue greater coherence and systematicity among our moral views, no moral judgment should be privileged over any other simply in virtue of its level of generality. For example, in cases in which an otherwise plausible general moral principle conflicts with an otherwise plausible specific claim about morality, there are no general norms which would favor one over the other simply in virtue of its level of generality.
(5) A further consequence of (3) is that there is no specific level of generality that is had in common by those propositions that constitute the proper starting point for moral inquiry. For example, and contrary to what some have supposed, it is not as though the proper starting point for moral inquiry consists exclusively of general moral principles as opposed to particular moral judgments, or vice versa.
Again, I take these to be insights of the method. But I will also argue that the method does not have the kind of epistemological significance that has sometimes been taken to have. More specifically, I will argue for the following claims:
(6) Neither the pursuit nor the achievement of reflective equilibrium is necessary for having substantial moral knowledge. An ordinary moral agent might have substantial moral knowledge even if she never engages in the pursuit of reflective equilibrium.
(7) Following the method is not sufficient for attaining moral knowledge, nor is it sufficient for attaining moral views that have any significant positive epistemic status. Following the method impeccably, and arriving at a state of reflective equilibrium in the canonical way, is perfectly consistent with having unreasonable moral views.
(8) Because of (6) and (7), the method of reflective equilibrium is not tenable when it is understood as an account of the circumstances in which our moral views are justified. Nor is the method of reflective equilibrium tenable when taken as a normative account of how we ought to make up our minds about moral questions. Given standard characterizations, the method is consistent with a great deal of good reasoning about morality, but it is equally
consistent with a great deal of bad reasoning about morality. Thus, on its standard characterizations, the method of reflective equilibrium undercharacterizes good moral reasoning.
(9) On its most defensible interpretations, the method of reflective equilibrium is characterized in explicitly normative terms: for example, the judgments which it treats as evidence are pieces of moral knowledge, or at least, things that we justifiably believe about morality prior to engaging in the method of reflective equilibrium. An upshot of this is that, on its most defensible interpretations, the method of reflective equilibrium takes for granted that we have some moral knowledge (or at least, justified moral beliefs) that we do not arrive at via application of the method itself. If we have any moral knowledge, we have some moral knowledge that is not due to engaging in reflective equilibrium reasoning.
(10) Because the ability of reflective equilibrium reasoning to deliver new moral knowledge generally depends on our already having substantial moral knowledge from other sources, the method is not a promising answer to the question of how we are able to acquire moral knowledge in the first place.
At the outset of this section, I noted the impressive popularity of the method among leading moral philosophers, including those who differ widely in their other commitments. Given this impressive popularity, it seems that someone who argues that the method does not have the kind of epistemological significance that has often been claimed for it inherits an additional dialectical burden: she needs to offer a plausible explanation of why the method has seemed to so many moral philosophers to be a compelling account of the kind of reasoning in which they routinely engage, a kind of reasoning that frequently seems to deliver genuine insights into the moral domain.7 It is hard to believe that so many moral philosophers could be radically deceived about what they have been doing all along. I attempt to discharge this burden in Section 2.7, “The Method and the Moral Philosopher.”
2.2 The Method Is Too Demanding
In my estimation, the most straightforward and compelling reason for thinking that engaging in the method of reflective equilibrium is not necessary for
7 As Sayre-McCord (1996: 142) says: “Recommending the method right off is the fact that it seems, in some sense, simply to work.”
acquiring moral knowledge is this: if there is any moral knowledge at all, the amount of it that is had by ordinary people is out of proportion to the frequency with which they engage in the activity of pursuing reflective equilibrium among their moral views, and the role that doing so plays in their lives.
Consider the claims that slavery is unjust and that rape is wrong. If there is anything that we might reasonably expect moral skeptics and non-skeptics to agree upon, perhaps it is this: these claims are not pieces of esoteric knowledge. If they are known by anyone, they are known by many. Suppose first that some variety of moral skepticism is true, so that no one knows that slavery is unjust (perhaps because slavery is unjust is simply not the right kind of thing to be the object of propositional knowledge). If no one knows that slavery is unjust, then a fortiori, no one knows that slavery is unjust on the basis of reflective equilibrium reasoning. Suppose on the other hand that moral skepticism is false, and that some people do know that slavery is unjust. Consider the description of reflective equilibrium reasoning offered by Scanlon, above. How plausible is it that all (or even most) of those who know that slavery is unjust know it on the basis of such reasoning? The claim that this kind of reasoning is responsible for the average person’s knowledge that slavery is unjust seems to involve an unrealistic, hyper-intellectualized picture of the basis for such knowledge. And the same point seems to hold with respect to canonical statements of the method due to Rawls and others.8
More generally, the kind of knowledge that many proponents of the method of reflective equilibrium take it to deliver seems like a poor candidate for knowledge that is available only to relatively reflective individuals. For example, Scanlon (2014) touts the method as the way in which we arrive at knowledge of normative truths about reasons, a category that he takes to include the most fundamental moral truths. In the opening pages of that work, he provides the following examples of the kinds of truths that he has in mind:
(1) For a person in control of a fast-moving automobile, the fact that the car will injure and perhaps kill a pedestrian if the wheel is not turned is a reason to turn the wheel.
(2) The fact that a person’s child has died is a reason for that person to feel sad.
8 Again, see the passages referenced in footnote 15 above. On the general point at issue here, compare Timmons (1999: 237) and Wedgwood (2007: 244). After noting that the method seems overly demanding when understood as an account that applies to ordinary moral agents, both Timmons and Wedgwood suggest that for that reason it is perhaps better interpreted as an account of theory acceptance that applies to moral theorists or philosophers. I consider this suggestion below.
(3) The fact that it would be enjoyable to listen to some very engaging music, moving one’s body gently in time with it, is a reason to do this, or to continue doing it. (2014: 2)
On the face of it, these examples seem like unlikely candidates for things that we know on the basis of any even moderately sophisticated reasoning. Notably, when Scanlon describes and defends the method in a later lecture, much of that lecture is devoted to a compelling description of the kind of holistic plausibility reasoning by which set theorists justify their beliefs in the less obvious axioms of set theory (including the axiom schema of replacement) which he also takes to be an application of the method (2014: 72–6). However, the kind of relatively subtle reasoning by which a set theorist might justify her belief in the axiom schema of replacement does not seem like a good model for the way in which an agent might come to justifiably believe that he should turn the wheel of his car when he realizes that there is a pedestrian directly in his path. If we look to mathematics for natural analogues to seemingly obvious and immediately grasped propositions about normative reasons like (1)–(3), better candidates for this role would seem to include the obvious and immediately grasped truths of elementary arithmetic such as 2+2=4. But of course, at least on the face of it, the method of reflective equilibrium does not seem to be a particularly plausible account of how we know that 2+2=4, either: it is no accident that our knowledge of such truths has historically inspired epistemological accounts that have appealed to notions like “self-evidence” (whatever the ultimate defects or merits of such accounts), as opposed to the more holistic picture of justification that is associated with the reflective equilibrium model.
In short, the view that whatever moral knowledge there is is knowledge that is arrived at via reflective equilibrium reasoning seems to predict that there is relatively little moral knowledge, and that what we have of it is disproportionately had by unusually reflective individuals. By my lights, that this is the actual situation is much less plausible than that either (i) there is relatively widespread moral knowledge, or (ii) there is no moral knowledge at all.
I will assume, then, that it is not a necessary condition for knowing that a moral claim is true either that that belief was arrived at or that it is sustained by reflective equilibrium reasoning. Even if you’ve never engaged in reflective equilibrium reasoning with respect to a moral belief that you hold, it does not follow that your belief falls short of knowledge.
Indeed, on my view, the plausibility of the claim that ordinary moral agents have at least some substantive moral knowledge (regardless of whether they’ve
engaged in reflective equilibrium reasoning or systematic moral theorizing) provides a compelling reason to believe one of the characteristic methodological claims of the reflective equilibrium theorist. Recall the idea that, in moral inquiry, we never start from scratch. According to this idea, whenever we engage in moral inquiry, we come to the table with moral convictions, and it is neither rationally required nor desirable for us to attempt to conduct the inquiry from some more austere starting point (e.g. one consisting exclusively of some purely formal principle or principles). One line of thought that supports this methodological claim runs as follows. Consider what we might call the knowledge platitude:
THE KNOWLEDGE PLATITUDE:
If you know something that is relevant to a question that you are trying to answer, then you should take that information into account in arriving at a view.
The knowledge platitude is highly plausible.9 Now, on the assumption that people typically have at least some substantive moral knowledge prior to engaging in reflective equilibrium reasoning or systematic moral inquiry, it follows from the knowledge platitude that they should take that knowledge into account if and when they do engage in such inquiry. (It’s not as though one loses whatever substantive moral knowledge one has simply by engaging in systematic moral inquiry.) Bracketing or setting aside substantive moral claims that one knows to be true is bad methodology, because it amounts to throwing away relevant information about the subject matter. Thus, the fact that a person might have substantive moral knowledge even without having engaged in systematic moral inquiry provides a good reason to believe that the reflective equilibrium theorist is correct in holding that the proper starting point for such inquiry need not be devoid of substantive moral claims.10
9 Notice that the knowledge platitude is not specifically concerned with moral knowledge, but with knowledge more generally. It seems clear enough that, if, for example, you were trying to make up your mind about which candidate is most likely to win some upcoming election, you should take into account any relevant information of which you have knowledge. On the face of it, there does not seem to be any reason to think that things are different when it comes to moral questions.
For more extended defenses of the knowledge platitude, see Kelly (2008) and especially Williamson (2000, chapter 9). The latter argues at length that the relevant norm remains compelling even if we abandon the so-called “KK thesis” and admit that we often know things without knowing that we know them.
10 Although I take the line of thought offered in this paragraph to amount to a compelling vindication of a major theme of proponents of the method, this particular way of vindicating the theme will not be available to anyone who interprets the method ambitiously as opposed to modestly. For the vindication depends on the idea that we have substantial moral knowledge prior to engaging in the method, an idea that will be rejected—mistakenly, if what has already been argued is correct—by
If what has been argued thus far is correct, engaging in reflective equilibrium reasoning is not necessary for moral knowledge. Similar considerations support the conclusion that engaging in reflective equilibrium reasoning is not necessary for having justified moral beliefs. In fact, there are two routes to this conclusion from the kinds of considerations canvassed in this section. In Chapter 1, I endorsed and briefly defended the common assumption that knowing that p entails justifiably believing p. If that assumption is correct, then, given that engaging in reflective equilibrium reasoning is not necessary for knowing some moral claim, it follows immediately that it is not necessary for justifiably believing that moral claim, either. However, even if we do not assume that knowing entails justifiably believing, the kinds of considerations canvassed in this section strongly suggest that one can have justified moral beliefs even in the absence of reflective equilibrium reasoning. Just as claims like rape is wrong or slavery is unjust seem like poor candidates for esoteric knowledge, so too they seem like poor candidates for being justifiably believed only by unusually reflective people: if they are justifiably believed by some, then they are justifiably believed by many.
Of course, even if it’s true that engaging in reflective equilibrium reasoning is not necessary for either knowing or justifiably believing moral claims, it might still be that such reasoning has an important epistemological role to play. For example, it might be that when one impeccably follows the method of reflective equilibrium, this is sufficient for the reasonableness of the moral views at which one arrives. Moreover, notice that even if a great deal of our moral knowledge is not attributable to the method of reflective equilibrium, this is perfectly consistent with the claim, made by Scanlon and others, that the method is the best way of making up one’s mind about moral questions. (Compare: it might be that the best procedures we have for arriving at knowledge of the empirical world are the procedures of the natural sciences. This is perfectly consistent with the fact that much of what one knows about the empirical world is not due to the application of the procedures of the natural sciences.) Finally, it might be that, even if the method of reflective equilibrium plays a limited role in the lives of ordinary moral agents, it is the correct normative account of how moral philosophers or theorists should conduct their inquiries. These are among the claims that I will argue against in what follows.
proponents of the method who understand it ambitiously. Thanks to an anonymous referee for making salient to me the need to clarify the dialectic on this point.
2.3 If the Method Is Too Weak, Then It Is Probably Not “The Best Method”
In the last section, I noted that, although there are good reasons to think that the reflective equilibrium framework does not supply a necessary condition for being justified in holding a moral view, those considerations are perfectly consistent with the claim that the framework does supply a sufficient condition for being justified.11 In the next section, I will argue against this sufficiency claim. But before doing that, I want to put on the table another way of thinking about the method of reflective equilibrium, on which it is not, at least in the first instance, an account of the conditions (either necessary or sufficient) under which our moral beliefs are justified. In particular, I want to look at Scanlon’s suggestion that the method as an account of “the best way of making up one’s mind about moral matters” and consider how that relates to the sufficiency question.
Let us start by considering a very straightforward sufficiency claim that a theorist might endorse:
(1) A person is justified in holding a moral belief if she holds that belief in a state of reflective equilibrium.
Now I suspect that few proponents of the method would find this principle attractive, because the idea that simply being in a certain state at a given time suffices for being fully justified in holding a moral belief fits poorly with the theme, frequently emphasized by proponents of the method, that the process of seeking or pursuing reflective equilibrium is as important (if not more so) than actually being in the state itself.12 If (1) is true, then someone who ended up in a state of reflective equilibrium in virtue of being hit in the head by a falling tree branch, or in virtue of having been hypnotized, would be fully justified in her moral views so long as she remained in that state. But this seems to be at odds with the idea that it is engaging in a certain procedure or way of reasoning that is crucial. Indeed, it seems at odds with the idea that the method of reflective equilibrium is a method. 13
11 Several paragraphs in both this section and the next are adapted from Kelly and McGrath (2010). I utilize this material with the permission of my co-author.
12 For representative statements of this thought, see Sayre-McCord (1996: 142) and Scanlon (2014: 79, 103).
13 Thus, the method of reflective equilibrium should be distinguished from any purely “current time slice” principle of epistemic justification. For example, the method should be distinguished from a principle according to which a person is justified in holding a moral belief a time t just in case that belief coheres well with the other beliefs that she holds at t, as well as from a principle according to
This line of thought might lead one to replace (1) with (2):
(2) A person is justified in holding a moral belief if (i) she holds it in a state of reflective equilibrium, and (ii) she reached that state by reasoning in the prescribed way.
where “the prescribed way” abbreviates the kind of procedure described by Scanlon in the passage quoted in Section 2.1, or something very similar. According to (2), holding a moral belief in a state of reflective equilibrium reached by reasoning in the prescribed way is a sufficient condition for being justified in holding that belief.
In the next section, I will argue that (2) is false. But for the moment, the issue is that of how (2) relates to the following claim:
(3) The method of reflective equilibrium is the best method for arriving at one’s moral views.
Suppose someone wanted to argue that (3) is false. To show that (3) is false, it would not be enough to show that a person who flawlessly follows the method could arrive at false moral views, or even that she could arrive at views that are radically mistaken. Because in general, it is not a condition of adequacy on a method of inquiry that it is guaranteed to deliver the truth, or even guaranteed not leave us much worse off with respect to the truth than we would have been had never applied it. Certainly, we do not hold our best scientific methods to the relevant standard. For example, contemporary cosmologists utilize relatively rigorous methods to arrive at theories about the history and structure of the physical universe on the basis of the evidence that is available to them. In a world in which the evidence that they have to go on is consistently misleading or unrepresentative, the impeccable application of their best methods to the data will not only fail to deliver the truth but will lead them further and further astray. But it would be a mistake to regard this fact as a good objection to their methods. We should not, I assume, hold a proposed method of moral inquiry to a higher standard than those to which we hold our best scientific methods. For this reason, even if there are circumstances in which employing the method of which a person is justified in holding a moral belief at time t just in case that belief is supported by the other mental states she is in at t, including both her beliefs and her moral intuitions or “seemings.” Notice that putative principles such as these purport to be purely synchronic principles of justification. Thus, they are completely silent on how we should go about arriving at or revising our moral views over time; indeed, they lack any diachronic import at all. By contrast, the accounts offered by Scanlon, Rawls, and the other representative proponents of the method listed in footnote 15 clearly do have significant diachronic import.
