Moral testimony has been an increasingly popular topic, as of late, with the debate centered around the asymmetry thesis: while testimony is a perfectly fine source for non-moral belief, there is something problematic about basing one's moral beliefs on it.Footnote 1 So-called “pessimists” about dependence on moral testimony defend some version of the asymmetry thesis.Footnote 2
My general aim in this paper is argue against a main strand of pessimism, what I call “epistemic” pessimism.Footnote 3 Epistemic pessimists’ argument for the asymmetry thesis is that moral testimony fails to transmit epistemic warrant as non-moral testimony does.Footnote 4 My specific aim in this paper is to reject epistemic pessimism by defending the following:
No Difference Thesis: there is no in principle difference between the transmission of epistemic warrant by moral and non-moral testimony.
My strategy is as follows: in Section 1 I clarify the epistemic pessimists’ challenge to moral testimony, and so the scope of this paper. Then, in Section 2, I give a direct argument in defense of the No Difference Thesis. In Section 3 I consider five attempts to undermine the No Difference Thesis, and argue that none of these attempts succeeds. So, in the absence of any other compelling criticisms, we are justified in rejecting epistemic pessimism and accepting the No Difference Thesis.
1. The scope of epistemic pessimism
Epistemic pessimists are not concerned to cast doubt on the possibility of moral knowledge, generally. Rather, they grant (at least for the sake of argument) that both non-moral and moral knowledge are possible. Instead, the epistemic pessimist argues that, while epistemic warrant for non-moral beliefs can be transmitted by testimony, epistemic warrant for moral beliefs cannot be so transmitted.
To assess the merits of epistemic pessimism, then, in this paper I will set aside worries about dependence on moral testimony that have been entertained in the literature but that are not principally about moral testimony. For example, one candidate explanation of the asymmetry thesis that has been widely entertained in the literature on moral testimony is that moral testimony is more problematic than non-moral testimony because there can be no moral knowledge.Footnote 5, Footnote 6 Alison Hills, for instance, explains this line of thinking as follows:
If there is no moral knowledge, then one standard reason for trusting testimony – that you can gain knowledge – would not apply to moral testimony. This would explain the difference between moral testimony and non-moral testimony too.Footnote 7
In a similar vein, but explicitly linking the denial of moral knowledge with non-cognitivist meta-ethical views, Sarah McGrath writes,
If some standard version of non-cognitivism is true, then our attitude toward pure moral deference can be explained: if (as Ayer would have it) my judging that eating meat is wrong is a matter of my expressing my own negative emotions toward eating meat, then it is unsurprising that we find moral deference problematic.Footnote 8
The basic idea seems to be that if moral judgments are not truth-apt, testifying about our moral judgments cannot transmit epistemic warrant because the very notion of our moral judgments being epistemically justified is a category mistake.
Another explanation of the asymmetry thesis that has been entertained in the literature roughly goes along the following lines: there can be no moral knowledge, and so no transmission of such knowledge by moral testimony, because moral disagreement precludes us from knowing the moral facts, if such there be.Footnote 9, Footnote 10 As Hills explains this view: “There is a lot of disagreement in ethics, and this should undermine our confidence in all potential testifiers.”Footnote 11
I have my doubts about the plausibility of non-cognitivism, and about whether non-cognitivist views have been aptly portrayed in the debate about dependence on moral testimony.Footnote 12 I also have my doubts about the formidability of the argument from disagreement as a challenge to moral knowledge. However, I will not take up these substantive meta-ethical issues here, because I take them to be orthogonal to the ongoing debate about dependence on moral testimony. These issues are not directly relevant to an examination of moral testimony because they are about moral knowledge in general, and not about testimonial moral knowledge in particular.
I acknowledge that in setting aside worries about the possibility of moral knowledge, I might lose the interest of those who, for metaphysical or epistemological reasons, do not accept the possibility of moral knowledge in the first place. Nevertheless, even those who do not accept the possibility of moral knowledge need not necessarily deny my central claims in this paper about dependence on moral testimony, for I am happy to allow that these claims have a conditional status, e.g., there are cases in which if a given speaker has moral knowledge, that knowledge can be transmitted via moral testimony.
2. The No Difference Thesis and the epistemology of testimony
One of the main topics that has been taken up in the epistemology of testimony is how we get epistemic warrant from depending on what others tell us. Two main theories of testimonial warrant have emerged: reductionism and non-reductionism.Footnote 13, Footnote 14 In this section my aim is to show that we have no particular initial reason to be suspicious of the transmission of warrant by moral testimony, for neither of these views of how and when testimony transmits warrant appeals to features of testimonial exchanges that are special to non-moral, rather than moral, matters. To be clear: it is not my purpose here to advocate for a particular epistemological view of testimony. Rather, my aim is to show that whichever view you espouse, there is no reason to be suspicious of moral testimony built into that view.
Reductionism about testimonial warrant is often traced back to Hume:
The reason, why we place any credit in witnesses and historians, is not derived from any connexion, which we perceive a priori, between testimony and reality, but because we are accustomed to find a conformity between them.Footnote 15
The basic idea is that in order to have epistemic entitlement to depend on testimony, hearers must have reasons for believing that the testimony is reliable that are independent of the testimony itself. These reasons might come from perception, memory, or inductive inference, for example.Footnote 16 In contrast, non-reductionism in the epistemology of testimony centers on the basic a priori claim that hearers have presumptive epistemic entitlements to depend on testimony.Footnote 17 Non-reductionists hold that testimony is a unique, fundamentally basic – non-reductive – source of epistemic warrant, and that “an assertion is creditworthy until shown otherwise.”Footnote 18 On their views, testimony is much like perception: just as perceivers can be immediately justified in believing what they see on the basis of their perceptions, so too can hearers be immediately justified in believing what they hear on the basis on others’ testimony.
Elizabeth Fricker puts the contrast between reductionism and non-reductionism as follows:
The solution [to the problem of justifying belief through testimony] can take either of two routes. It maybe shows that the required step – from ‘S asserted that p’ to ‘p’ – can be made as a piece of inference involving only familiar deductive and inductive principles, applied to empirically established premises. Alternatively, it may be argued that the step is legitimized as the exercise of a special presumptive epistemic right to trust, not dependent on evidence.Footnote 19
Reductionists take the former route, while non-reductionists take the latter; they maintain that hearers have a defeasible epistemic entitlement to depend on testimony even in the absence of non-testimonially based reasons for belief. These important differences notwithstanding, both reductionist and non-reductionist views share an important similarity. On both views, evidence of unreliability can override epistemic entitlements to depend on testimony, for such entitlements are defeasible. H has an epistemic right to depend on S’s say-so only so long as H does not have any undefeated defeaters. Reductionists and non-reductionists who want to defend against the pessimist attacks on dependence on moral testimony can readily allow, then, that just as hearers of non-moral testimony sometimes have undefeated defeaters, so too do hearers of moral testimony. Furthermore, since non-reductionists hold that hearers have default epistemic entitlements to depend on testimony, then for non-reductionists to be susceptible to epistemic pessimists’ views it must be shown that hearers’ default epistemic entitlements are defeased in wholesale fashion – that is, they must show that hearers of moral testimony have undefeated defeaters across the board.
Most importantly for present purposes, both reductionists and non-reductionists, in spelling out the details of their views more fully, do not make any claims about the content of testimony. So, on both reductionist and non-reductionist views, the transmission of testimonial warrant does not hinge on whether the content of testimony is moral or non-moral. As Lackey puts it,
There are no reasons, either from the literature specifically on testimony or from that concerning general epistemological issues, for regarding the subject matter of the testimony as relevant to its epistemology. If the speaker's testimony satisfies the central requirement – for example, it is reliable, virtuously produced, tracks the truth, and so on – and the hearer does not have any relevant defeaters, its content does not matter epistemologically.Footnote 20
On their face, then, if these views of testimonial warrant are successful in explaining how non-moral testimony confers warrant, then they are successful in explaining how moral testimony confers warrant. Without further reason to think that evidence of reliability is unavailable when it comes to moral testimony or that hearers of moral testimony have undefeated defeaters across the board, reductionism and non-reductionism support the idea that there is no general difference between the transmission of epistemic warrant by moral and non-moral testimony.
This provides a prima facie defense of the No Difference Thesis. To challenge this defense, those who endorse epistemic pessimism owe us an argument for the asymmetry thesis that shows that when it comes to moral matters, evidence of the reliability of testimony is unavailable or hearers invariably have undefeated defeaters. In the next section I examine five such arguments.
3. Objections to the No Difference Thesis
In this section I will consider five versions of epistemic pessimism that threaten to undermine the No Difference Thesis.
3.1. Hopkins's disagreement worry
Above I set aside appeals to disagreement that are intended to show the impossibility of moral knowledge generally. We might still wonder, however, if there is a special problem about dependence on moral testimony posed by moral disagreement beyond that which makes it hard, or even practically impossible, to get moral knowledge generally. That is one way we might read the following passage from Robert Hopkins:
Testimony requires reliable informants. But morality is a topic for which reliable informants are hard to find. There is simply too much disagreement on moral issues for one to be entitled to assume that any informant is reliable. This lack of consensus is in part due to the perversion of judgment through interest; and in part to the fact that even the disinterested do just disagree more in evaluative matters than in factual ones. Whatever its source, it is sufficiently prevalent to undermine one's confidence in others as moral informants. Hence one cannot legitimately rely on their testimony.Footnote 21
Here the worry is not that moral disagreement undermines all of moral knowledge, just moral knowledge by way of moral testimony. While disagreement inhibits our ability to reliably identify reliable moral testifiers, thereby undermining the transmission of epistemic warrant by moral testimony, it does necessarily undermine our confidence in our own moral judgments. Support for this line of thinking can be found in the literature on the epistemic significance of disagreement: some philosophers have argued that even when we know that other, well-informed, reasonable people disagree with us (our “epistemic peers,” as they are called in the literature), our confidence in the first-order reasons that support our judgments need not be swamped by our awareness of such disagreement.Footnote 22
I grant that this sort of worry poses a direct problem for the possibility of testimonial moral knowledge. Nevertheless I will argue that it is unsuccessful for three reasons: first, because it does not bottom out, so to speak, as a worry about moral disagreement; second, because many cases of dependence on moral testimony are not fundamentally subject to sustained moral disagreement; and third, because the worry threatens to collapse into a worry about the possibility of moral knowledge, generally, even though it does not bill itself in this way.
Non-moral disagreement is prevalent, and yet it is not taken to undermine the transmission of epistemic warrant by non-moral testimony. Consider Michael Huemer's list of some non-moral questions over which there is extensive disagreement: Who shot JFK? How should quantum mechanics be interpreted? What are the actual practices of other cultures? What are the economic effects of government social programs? What religion, if any, is correct? Are there paranormal phenomena? What causes illness?Footnote 23 Even in light of extensive disagreement over these questions, we still are comfortable saying that there are reliable testifiers in the fields of history, physics, anthropology, economics, religion, cosmology, and epidemiology. In these cases we do not take the mere fact of disagreement to undermine our ability to reliably pick out reliable testifiers, but rather we take it as a reminder to be careful in choosing on whose testimony we depend and to be critical in our dependence on the testimony of others. Epistemic pessimists owe us an explanation, then, of why the same strategies we use to navigate non-moral disagreement and carefully pick out non-moral testifiers will not work in cases of dependence on moral testimony. To simply point to pervasive moral disagreement in the absence of such an explanation is to give a wholly unsatisfactory explanation of the asymmetry thesis. For such an explanation “does not support an asymmetry between [moral] testimonial beliefs and all other testimonial beliefs. Rather, it supports an asymmetry between testimonial beliefs where there is a substantial amount of disagreement and those where there is not.”Footnote 24
As far as I know, Sarah McGrath offers the only explicit explanation given in the moral testimony literature of why we cannot use the same strategies to navigate moral disagreement and identify reliable moral testifiers that we use to navigate non-moral disagreement and identify reliable non-moral testifiers. She writes:
Certain scientific questions might be highly controversial among the population as a whole, but when a consensus or near consensus exists among those with the relevant expertise, one need remain in a state of agnosticism only for as long as it takes to discover the content of the consensus. It might be thought that there is a parallel defense of one's controversial moral beliefs … In general, identifying those with genuine expertise in some domain will be most straightforward when we have some kind of independent check, one not itself subject to significant controversy, by which we can tell who is (and who is not) getting things right… But significantly, we possess no … independent check for moral expertise.Footnote 25
The crux of McGrath's worry is not about moral disagreement full stop, but about the lack of an “independent check” that enables us to navigate such disagreement. Such a worry is clearly germane to a debate about testimonial moral knowledge, and so I attend to it below. But it is misleading to suggest that the worry is ultimately about moral disagreement, full stop; when adequately spelled out, it is ultimately a worry about our abilities to reliably pick out reliable testifiers when it comes to moral matters in the face of such disagreement.
Additionally, recall from above that optimists can grant that disagreement might sometimes render dependence on moral testimony problematic. They can concede this and still point to the many cases of dependence on moral testimony in which moral disagreement is not really at issue. Presumably anti-racist norms are not subject to deep disagreement, at least amongst the morally mature. If I am committed to anti-racist norms, but I am just really very bad at seeing when these norms apply and so depend on the testimony of my friends to pick out instances of racism, even blatant racism, the matters for which I depend are not subject to deep moral disagreement (again, at least among the morally mature). Furthermore, the fundamental reason that I depend on my friends’ moral testimony is not because I take them to be adept at navigating moral disagreement, but because I think they have more well-developed moral sensitivities than I do. That is, the worry about moral disagreement does not cover all cases in which dependence on moral testimony is at issue; in some cases it is principally a matter of thinking that some people have competence in some area that others lack.
This way of responding to the worry gives rise to a potential objection: maybe cases like the one above – in which I defer to friends to help me see how to live out my anti-racist values – are better understood as cases in which testifiers are, in effect, serving as deliberative proxies for those who depend on them.Footnote 26 Epistemic pessimists aren't particularly concerned about these sorts of proxy cases, so they're not effective tools for pressing back against pessimists’ worries. As pessimist McGrath writes,
Perhaps there are other cases in which moral deference is as straightforward as non-moral deference. Suppose that I know that you and I share the same moral sensibility, but that with respect to some particular question that is before us my judgment is impaired in a way that yours is not. Perhaps I am too angry or too drunk to think clearly, or I am too personally involved to see things aright. In such circumstances, I might view deference to your judgment as deference to my unimpaired self …Footnote 27
However, while I think there's a way to construe some of the cases I consider in this paper as proxy cases in the way that McGrath describes here, it strikes me as an implausible stretch to view all of them in this light. Consider Karen Jones's well-known example of Peter:
Peter: “Peter had a settled and serious commitment to the elimination of racism and sexism, but he was not very good at picking out instances of sexism and racism … Such blindness can sometimes indicate insincerity, but in Peter's case it did not … He could pick out egregious cases of sexism and racism, and could sometimes see that ‘sexist’ or ‘racist’ applied to more subtle instances when the reason for their application was explained to him, but he seemed bad at working out how to go on to apply the words to non-egregious new cases.” Peter lived in a co-op, and so when interviewing potential new members, Peter would often need to depend upon the testimony of the other members of the co-op to know whether or not the candidates were sexist or racist.Footnote 28
Peter is not too angry nor too drunk nor too involved to spot racism. Peter has a persistent moral blindspot. To see those on whom Peter depends as offering him guidance that we can regard as the very same guidance he would get from his “unimpaired self” is to stretch the notion of “unimpaired self” so far that we could, by parity of reasoning, view all cases of moral deference as proxy cases. After all, in some sense if we were all free of moral bias and ignorance our “unimpaired” selves would lack no moral information and so we'd be able to offer perfect moral guidance to ourselves.
More generally, we can have well-developed moral sensibilities, and even share some of those sensibilities with those on whom we depend, and still count as depending on them for new-to-us moral knowledge. In thinking about moral testimony, the examples I tend to focus on are those in which thicker sorts of relationships hold between hearers and testifiers, in part, because these are cases that, in my view, are less likely to be epistemically and morally problematic.Footnote 29 So one might worry that these are just the sorts of examples that are more liable to be proxy cases. But just because some more intimate relationship holds between a speaker and a hearer – e.g., they're fellow co-op members – does not at all show that the two share moral sensibilities to such an extent that they're aptly characterized as moral proxies for one another; indeed, I substantively disagree about a wide range of moral issues with some of the people on whom I'm most apt to depend upon for moral guidance – my husband, my best friend, my sister, etc. (although I'd like to think we do have shared moral sensibilities in the sense that we have a shared commitment to being moral, de dicto).
So many cases of genuine dependence on moral testimony are not fundamentally subject to sustained moral disagreement. Furthermore, the worry about moral disagreement does not just impugn dependence on moral testimony, but also paves the way for moral skepticism, more generally. As Hopkins argues, “[N]ot just unwillingness to believe what one is told, but reluctance to form moral beliefs by any means at all” follows shortly on the heels of a defense of this worry.Footnote 30 For if disagreement is such that it utterly undermines our trust in testifiers, it seems that it also threatens to undermine our own moral judgments. If we lack the resources to reliably adjudicate moral disagreements amongst others, and so lack the resources to reliably identify moral testifiers, then what resources do we have to adjudicate moral disagreements between others and ourselves and so maintain our confidence that we are right in the face of these disagreements? If the worry ultimately collapses into a worry about the possibility of moral knowledge, generally, then as I point out above, it is not directly relevant to the ongoing debate about dependence on moral testimony.
3.2. McGrath's calibration worry
McGrath offers the following argument for the asymmetry thesis:
When one observes that it is raining, one in effect has independent access, via perception, to facts in the target domain. In arriving at the view that a given weather-forecaster correctly predicted rain, one does not rely on techniques similar to those employed in arriving at that prediction. Because one is in a position to determine reliably which weather-forecasters got it right on particular occasions by some method other than those employed by the weather-forecasters themselves, one has a way of calibrating their techniques for accuracy … By contrast, there seems to be no analogous way to calibrate the accuracy or reliability of someone's moral judgment, because one lacks the relevant kind of independent access to the moral facts. If one attempted to rank others with respect to the accuracy of their moral judgment by checking how often they answered controversial moral questions correctly, it seems as though one could do so only by engaging in first-order moral reasoning and deliberation of one's own. It is thus unsurprising that clear and unequivocal evidence that someone possesses unusually reliable moral judgment is hard to come by.Footnote 31
McGrath's worry is that hearers lack “independent access” to the moral facts and so cannot reliably identify reliable moral testifiers. The reason that moral testimony fails to transmit epistemic warrant in the ways that non-moral testimony does is that when it comes to moral, but not non-moral, matters, we lack an “independent check, one not itself subject to significant controversy, by which we can tell who is (and who is not) getting things right.”Footnote 32 On McGrath's view, independent checks are necessary to ensure the reliability of testimony, but they are not available in the realm of morality.
What, more precisely, is an independent check? While McGrath does not give a concise definition, she does give a number of examples to illustrate the concept. In the passage above, she suggests that our direct sense perception provides an independent check on weather-forecasting methods: we can independently check who is a more reliable weather-forecaster by using our sense perception to determine whose predictions are more often accurate. She also argues that we can independently check the reliability of our long-distance vision by examining objects at a distance, and then examining those very same objects up close.Footnote 33 And, she says, we can independently check whether MIT provides good training for structural engineers by examining the stability of bridges built by structural engineers trained at MIT.Footnote 34 The basic idea at work in all of these examples seems to be that independent checks employ methods different than the methods used to arrive at the judgments being checked, and are themselves not subject to significant controversy. And so we can see why McGrath aptly labels her worry the calibration worry: to calibrate an instrument (the so-called “system under test”) we compare that instrument's deliverances to the deliverances of another instrument (the so-called “standard system”) that we know to be reliable. When it comes to morality, to get an independent check on the reliability of some judgment one would not engage in first-order moral deliberation, but would use some independent, non-controversial method to acquire evidence that the judgment was reliably formed.
It might seem, at first glance, that McGrath's calibration worry is particularly relevant for those who endorse a reductionist view of the epistemology of testimony. According to McGrath, the lack of an independent check impedes us from getting positive evidence of reliability, and reductionists, but not non-reductionists, require such evidence as a condition of testimonial epistemic warrant. But we can cash out McGrath's worry as a defeater that could serve to undermine moral testimony for those who hold a non-reductionist view. Let me explain.
Assuming McGrath is right, we are in general unable to tell apart reliable and unreliable moral testifiers because we do not have the means to independently check their testimony. This fact alone is not enough for the calibration worry to gain traction with non-reductionists. But consider what non-reductionist Coady writes of testimony: “We do not have to establish the many propositions which, if false, would invalidate our ready assent to what we are told, unless there is already some reason to believe that their truth is in jeopardy.”Footnote 35 So, on non-reductionist views, we do not have to establish the reliability of testifiers, unless we have reason to believe that their reliability “is in jeopardy.” When it comes to moral testimony, we have such reason: we know (or at least should know) that many people, including the people on whose moral testimony we might depend, are bad, even terribly bad, at making moral judgments. It looks like, then, if we are non-negligibly liable to depend on unreliable testifiers and we have no recourse for monitoring testifiers’ reliability, we are justified in believing that our beliefs based on moral testimony are unreliably formed – that is, we have a defeater for dependence on moral testimony. Consider an analogy with memory, widely considered a basic source of epistemic warrant: imagine that you somehow accumulate a wealth of evidence that you often grossly misremember, but your situation is such that you have no means for determining whether you have misremembered or the conditions under which you are most prone to misremembering. It is reasonable to think that in such a case there is a proposition which you are justified in believing to be true, i.e., that your memory often leads you astray, that indicates that your memory-based beliefs are unreliably formed. That is, you have a defeater for your memory-based beliefs. Returning to testimony, the issue for non-reductionists, assuming that McGrath is correct, is not that hearers’ lack positive evidence of testifiers’ reliability, but that hearers, in depending on moral testimony, have reason to believe that they are forming their beliefs in an unreliable fashion.Footnote 36
So McGrath's worry poses a problem for dependence on moral testimony regardless of which view of testimonial warrant one espouses. But let me suggest that McGrath's worry is, at least on one front, overblown. McGrath suggests that the challenge of identifying reliable moral testifiers is the challenge of finding
someone who consistently arrives at the correct moral answers to non-trivial moral questions … or at least, someone whose reliability with respect to such questions significantly exceeds that possessed by the average person, when the average person does not form his moral opinions by deferring.Footnote 37
But this is too high a bar. Individuals can be reliable about some moral matters but not others. Consider Karen Jones's Peter again. We can imagine that Peter, while very bad at picking out instances of sexism and racism, is wonderfully adept at navigating norms of rudeness, for example. As Jones puts it, there can be local relative epistemic advantages. However, McGrath seems to suggest that we must, in order to be justified in depending on moral testimony, have evidence that the testifier is global in her epistemic advantages.Footnote 38 Imagine if we applied this same standard in non-moral contexts; it would impugn what strike many of us as utterly uncontroversial cases of dependence on non-moral testimony. I know quite a bit about moral philosophy, but very little about the philosophy of time, for example. I know a lot about the rules of basketball, but almost nothing about the rules of soccer. And even though I know a lot about the rules of basketball, I know hardly anything about current NBA teams and players. That I do not know much about the philosophy of time, or soccer, or the NBA does not undermine my reliability as a testifier when it comes to moral philosophy and the rules of basketball, though. Similarly, that Peter is bad at picking out instances of sexism and racism does not undermine his reliability as a testifier when it comes to matters concerning norms of rudeness.
While my point above mitigates McGrath's worry, it does not resolve it; even once we lower the bar, so to speak, the calibration worry resurfaces. For example, even if we grant that Peter may be reliable when it comes to applying norms surrounding rudeness but unreliable when it comes to applying norms surrounding sexism and racism, there still remains a worry about how we could come to know that Peter is reliably getting it right with respect to rudeness. On McGrath's view, we would still need an independent check for the reliability of Peter's judgments concerning rudeness, and such a check is not available to us. In what follows, I will argue that we do have independent checks when it comes to morality. More importantly, I will argue that, even so, the requirement that we have independent checks in order to assess others’ reliability is, in general, a bad one.
Hearers often have the ability to assess the reliability of moral testifiers by some independent check, i.e., by means other than engaging in first-order moral deliberations regarding the content of the testimony. Consider what is for many of us an unfortunately all-too-familiar example: a coworker testifies to you about some moral question surrounding misogyny in the workplace and public life and concludes his testimony by informing you that these are the words he himself lives by. Now imagine that, in the past, you happen to have been at the receiving end of this person's consistent gendered microaggressions. In such a case, it's plausible to think that you have reason, both practical and epistemic, to doubt the reliability of his moral testimony in this case, even without engaging in first-order deliberations about its content.Footnote 39 Of course I am not suggesting that there is a necessary connection between the epistemic and the practical; one can make accurate moral judgments and yet still fail to live up to those judgments for a variety of reasons. It is possible, for example, that the testifier in this case has morally acute judgment but is just extremely weak-willed. But if somehow you were to find out that this is in fact the case, it would not mean that you did not initially have grounds for doubting his testimony, it would just mean that your grounds were undercut. This is because it seems to me far more plausible to think that by both perpetuating and benefitting from systems of injustice distorts one's epistemic understanding of those very systems, especially given what we know about epistemological costs of implicit bias.Footnote 40
So hearers can independently check the reliability of moral testimony. Nevertheless, I do not think that requiring “independent access” to the facts should be a requirement for calibrating the accuracy of moral or non-moral testimony, for it strikes me that general philosophical skepticism looms near such a requirement.Footnote 41 Consider a mundane non-moral example: imagine that you and I attended a party together last weekend. Earlier today, you recounted to me how nice it was to run in to one of our mutual acquaintances, Tom, at the gathering. At first, not immediately remembering seeing Tom at the party, I was confused by what you said. But after a few moments I correctly recalled the pleasant encounter with Tom. It strikes me that my remembering seeing Tom at the party does speak to the reliability of your memory (assuming I do not have a tendency to falsely remember things, etc.) – even though the “check” I employ relies on the same source, namely memory.
Sometimes we check the reliability of judgments using the very faculties or methods used to form those judgments. If we deny that this is an appropriate method of checking our judgments, and demand independent checks, then we threaten to undermine justification in many non-moral contexts as well as moral contexts. As Huemer points out, it is very difficult to see how one might check the reliability of introspection, memory, inductive reasoning, and reasoning in general, without relying on introspection, memory, inductive reasoning, and reasoning, respectively.Footnote 42 What's more, we have reason to think that even McGrath's own example of a supposedly paradigmatic independent check is not independent in the sense that she outlines. McGrath takes immediate sense perception to provide an independent check on methods of weather forecasting. She writes, “In arriving at the view that a given weather-forecaster correctly predicted rain, one does not rely on techniques similar to those employed in arriving at the prediction.”Footnote 43 However, weather-forecasting methods and models themselves depend, in large part, on historical data accumulated by sense perception.
3.3. LaBarge and Cholbi's credentials worry
So the requirement of an independent check is a bad one. With non-moral testimony we often use the same methods to check the reliability of the testimony as the testifiers themselves used. However, Scott LaBarge and Michael Cholbi argue that this is a strategy not readily available, if at all, to hearers of moral testimony. They defend the asymmetry thesis by arguing that when it comes to moral matters, as opposed to non-moral matters, hearers are not well-positioned to engage in first-order deliberations in order to assess the content of testimony. They have termed this worry the “credentials problem.” Unlike the calibration worry that emphasizes hearers’ lack of an “independent check,” the credentials problem centers on the idea that even if we abandon a requirement of an independent check hearers are still not able to distinguish reliable and unreliable moral testimony by way of first-order moral deliberations. LaBarge and Cholbi's idea, most basically, is that if hearers could reliably assess the content of moral testimony, then dependence on moral testimony would be otiose. But for those sincerely in a position to depend on moral testimony, they are in “no position to appraise the content” of the testimony because, for them, “the shape of a successful solution to our moral problems” is usually far from obvious.Footnote 44
You might think that a hearer, in order to be justified in depending on moral testimony, need not be able to assess the content of the testimony because she can have evidence of its credibility if she justifiably believes that the testifier is reliable. There are numerous and varied ways to distinguish reliable from unreliable moral testifiers that do not involve engaging in first-order moral reasoning about the particular proposition being testified to. As Sliwa points out,
Maybe you have seen the person make good moral judgments before and you know that they have thought about the issue at hand. Or maybe they have been recommended to you as a good advisor by someone whose judgment you trust. Maybe you have asked them some related moral questions and seen that they give reasonable answers.Footnote 45
So, even if a hearer is not a position to tell whether a testifier is right about the particular matter at hand because, to borrow Cholbi's words, she does not know what “the shape of a successful solution” will look like, the hearer might very well still have reason to depend on the testifier because she justifiably believes that the testifier is reliable. Such a line of thinking suggests an indirect solution to the credentials problem, in that it appeals to second-order considerations (about the reliability of the testifier) rather than first-order moral deliberations about the particular proposition to resolve the problem. But Cholbi argues that an indirect solution is not available to those in a position to sincerely depend on moral testimony because, just as they lack the sense of the “shape of a successful solution” to the moral issue at hand, so too do they lack an ability to distinguish reliable and unreliable moral testifiers.
For the sake of argument, let us grant Cholbi's (and McGrath's) point that we are unable to reliably identify reliable moral testifiers, in order that we may address in its strongest form his further claim that we lack the abilities to evaluate the content of moral testimony. Even granting that an indirect solution to the credentials problem is unavailable to hearers of moral testimony (because they lack the abilities to distinguish reliable and unreliable moral testifiers), I will argue that the idea that those in a position to sincerely depend on moral testimony are utterly unable to assess the content of that testimony is simply implausible.
When we depend on moral testimony, at least as mature moral agents, presumably we do so against a backdrop of an extensive network of moral beliefs and capabilities. So, when we depend on moral testimony we may very well have a well-developed sense of the “shape of a successful solution” to some moral problem. For example, the topic of three-person IVF has been getting quite a lot of press lately, on the heels of its approval in the UK's House of Commons and House of Lords. I have seen the headlines, but I have not had time to explore the issue in any depth, and so I have no idea what stance I, personally, would take on the issue. But because I have had the opportunity to think through some other, related reproductive issues more carefully, I have many justified beliefs that would help me to assess the content of moral testimony about three-person IVF. For instance, I believe that the issue is complex, that individuals’ rights to procreative liberty are at stake, and yet that there are other procreative options available to individuals interested in three-person IVF. Against my network of background beliefs I could, out of hand, dismiss all sorts of moral testimony about three-person IVF. I could, for example, dismiss the testimony of those who testify that three-person IVF is absolutely immoral because it involves “unnatural” interventions.
This is quite in contrast to Cholbi's depictions of what it looks like to depend on moral testimony. In arguing that hearers lack the ability to assess the content of moral testimony, Cholbi goes so far as to caution: “An unapologetic racist could offer moral [testimony] based on his racist paradigm, and the [testimony] could seem as coherent as that provided by another [testifier].”Footnote 46 But mature moral agents, even Peter – who, after all, “could pick out egregious cases of sexism and racism” – can readily dismiss the moral testimony of Cholbi's coherent racist, because even if they do not know the answers to complex moral issues surrounding race, they can know that racism is morally untenable.
One might worry at this point that, by bolstering the sense of the “shape of a successful solution,” I am only offering up a defense of a relatively weak form of dependence. I admit that the more one has such a sense, the weaker his dependence is. Still, having a relatively robust sense of the shape of a successful solution to a moral question or problem is compatible with a pretty high degree of dependence, even if it is less strong than the dependence would be if one had absolutely no idea what a successful solution to the matter at hand looked like.
To see this consider a simple non-moral example: Someone tells John that the Sears Tower has 108 stories. Prior to being told that, John had absolutely no beliefs about the particular number of floors in the Sears Tower, but based on the testimony John comes to believe that the Sears Tower does in fact have 108 floors. Had the testifier said that the Sears tower has 1000, or 20, or 300, or 50 floors, John would have immediately dismissed her testimony. So John has a rather robust sense of the shape of a successful solution to the question of how many floors there are in the Sears Tower; for example, he has a sense of what would count as clear error. Still, if somehow John found out that he had misheard the testifier – that the testifier in fact said that there are 118 floors – we can imagine that John would no longer believe that the Sears Tower has 108 floors, and instead he would come to believe, on the basis of the testimony, that it has 118 floors. That is, even though John has a well-developed sense of the shape of a successful solution to the question of how many floors there are in the Sears Tower, John's belief about the particular height of the Sears Tower still crucially hinges on the testimony in question.Footnote 47 Dependence on testimony – moral and non-moral – is almost never in a vacuum. We have a large set of background non-moral beliefs views that inform and limit the sort of non-moral testimony that we take seriously and depend upon. And the same is true of moral testimony. But even so, that does not preclude our dependence on testimony being so strong as to make all the difference between belief and disbelief.
Furthermore, I think Cholbi overlooks hearers’ abilities to assess the content of moral testimony, in part, because he fails to disambiguate the many varieties of dependence on moral testimony. He writes,
What the non-expert seeks from the moral expert is not technical or instrumental advice about how to pursue this or that moral end (though that may be part of the moral expert's expertise as well). The advice seeker does not want to know how to go about implementing some moral judgment, but instead wants to know what sorts of judgments to make. The goodness of a genuine expert's moral advice would thus not be exhausted by what the expert knows about how to realize particular ends or values. In the terms articulated earlier, the non-expert wants to know what shape the proper resolution ought to take and therefore seeks a determination of which reasons, and hence which ends or values, are decisive in particular complex moral situations.Footnote 48
But “how to go about implementing some moral judgment” was precisely what Peter was after when he turned to other co-op members for their moral testimony. The case explicitly states that Peter was committed to norms against racism and sexism, so we can imagine that Peter believed that the values expressed by those norms should be decisive in picking out new co-op members. The case also states that while Peter could pick out obvious violations of these norms, he struggled to pick out more subtle violations. And so, in turning to his fellow co-op members, Peter was in fact just looking for “instrumental advice about how to pursue this or that moral end.” And consider another case:
Wedding Testimony: Tom and Sara are planning a wedding and both of their families have offered to contribute money towards it. Sara's family, which is less wealthy than Tom's, offered a certain sum, which will cover less than half of the expenses. The couple is now wondering whether it would be permissible for them to ask Tom's family (which is wealthier) for a greater contribution. They decide to ask a friend. She tells them that it's permissible to ask the wealthier family for a greater contribution, and Tom and Sara, knowing that she's normally trustworthy and reliable, believe her.Footnote 49
We can imagine that Tom and Sara are sensitive to the main moral considerations in play (they can see that norms of fairness are at stake), but they are just unsure about how to weigh those considerations. Sometimes we, like Peter and like Tom and Sara, depend on testimony not because we are unsure about which values are at stake or which ends are worth pursuing, but precisely because we are unsure of how to live out existing moral commitments.
One might raise a different worry, at this point: are the preceding cases really cases of genuine moral testimony, or are they instead cases in which the testifier gives the hearer some non-moral information, too?Footnote 50 Pessimists aren't concerned about dependence on non-moral testimony, so if that's what's going on in the cases then I can't appeal to them to make headway against the pessimists’ concerns. I'm certainly happy to admit that ostensible cases of genuine moral testimony can sometimes, given the context, be better understood as cases of non-moral testimony. Eric Wiland illustrates this point with the following example:
At the Society of Act Utilitarians, one member tells another member ‘It would be wrong for you to order the expensive cocktail right now.’ On the face of it, this looks like moral testimony. But it isn't. Since it's common knowledge between the members that they are all act utilitarians, the testimony functions differently. The speaker aims to get the hearer to understand that ordering the expensive cocktail would not maximize happiness right now, itself a nonmoral claim … Moral testimony, then, is testimony that the hearer cannot accurately substitute a nonmoral claim for.Footnote 51
So even cases of what look to be genuine moral deference can, depending on how they're spelled out, really be cases of non-moral deference.
But is Jones's Peter case like Wiland's “Society of Act Utilitarians” case? I think not. Of course we're stuck to some extent in our philosophical theorizing with cases that are under-described. So you could fill out Peter's case in such a way that it is a case of non-moral deference, at least in part. For example, imagine that the reason the other co-op members don't want to accept a potential new member into their community is that because during the interviews the candidate member not-so-subtly addressed all of his remarks to only the white members of the interview committee – even when directly addressed by persons of color. Now, suppose that Peter didn't notice this at all. So while Peter is committed to anti-racist norms, he didn't see by his own lights how this commitment would be relevant to his vote for whether or not to accept the candidate into the co-op. In such a case, we can imagine that what Peter might have to learn from the other co-op members is that the candidate was racist because he consistently refused to acknowledge persons of color. This would be a case of impure moral testimony, because Peter is learning something non-moral, e.g., something about the candidate's behavior.Footnote 52
But imagine the case proceeding in a slightly different manner. Imagine that Peter, like the other co-op members, did notice that the candidate systematically refused to address the persons of color in the co-op, but even upon reflection he didn't see how such behavior was racist. Left to his own devices, Peter was just as likely to chalk the racist behavior up to the happenstance of which members the candidate “really clicked with.” In such a scenario, if Peter depends on the other co-op members’ testimony that “The candidate behaved in a racist manner,” I take it that we have a case of pure moral testimony. I am not alone in thinking that the Peter case could be interpreted both ways; indeed, pessimist Driver writes:
One might argue that this case is one of superior sensitivity to behavioral cues rather than a case of superior moral knowledge. However, if the knowledge that the other co-op members have is knowledge that ‘We ought not to let that person into the co-op’ that is moral knowledge since they are making a claim to know another claim that is a prescriptive claim.Footnote 53
Again, whether dependence on moral testimony is genuine concerns the extent to which such dependence involves reliance on the testifier for knowledge of the relevant non-moral facts. But this is entirely compatible with there being a range of possibilities of what sorts of moral information a speaker possesses and a hearer lacks even in cases of genuine moral deference. At one extreme, a hearer may depend upon a speaker's moral testimony that p and utterly lack an appreciation of the sorts of considerations that are truth-makers for p. Alternatively, a hearer may have a sense of what sorts of moral considerations support p, but in deliberating about whether p is the case she may turn to a speaker for guidance because she is unsure of how to weigh these considerations. Instead, it may be the case that a speaker and hearer share a commitment to the same moral norm, but the hearer depends on the speaker to figure out when the norm applies (I think this is probably the best way to understand the Peter case). As I see it, all of these types of dependence can count as cases of genuine moral deference, just so long as in each the hearer shares the same non-moral information with the speaker. So Cholbi's central concern, the so-called “credentials problem,” gets off the ground in part because he seems to overlook varieties of dependence on moral testimony.
Another means for gathering evidence about moral testifiers’ reliability using first-order moral deliberations, I have earlier suggested, is by looking retrospectively at testifiers’ track records. Cholbi, however, contends that we cannot do so; he writes,
Whether, for instance, an individual who claims to be an expert in investing money is an expert could be judged straightforwardly by the profitability of the investment plans she recommends. In contrast, there does not seem to be any straightforward basis on which one could, even retrospectively, appraise the advice of a … moral [testifier].Footnote 54
Contra Cholbi, one's reliability as a moral testifier can be post hoc verifiable. This can happen in various ways. For example, one reason that hearers might depend on moral testimony is that figuring out moral matters on one's own can be quite taxing. Perhaps in Wedding Testimony Tom and Sara turn to their friend because they are concerned about their own abilities to make a sound judgment about what to do given that they are both emotionally and physically exhausted from planning a large-scale wedding while holding down full-time careers, attending to their partnership, seeing friends, etc. In light of this, Tom and Sara rather unquestioningly depend on their friend's moral testimony. Now imagine that after the excitement of the wedding has come and gone, Sara has the opportunity to sit down with her and Tom's friend, who takes the time to explain why she thought it was permissible to ask Tom's family for a greater contribution. Sara, now much less stressed and much more clearheaded, carefully considers her friend's explanation and finds it thoroughly convincing. Sara's own deliberations on the matter speak to the reliability of her friend's judgment. Consider a non-moral analogy: imagine that you put a long proof up on the board and tell me that it is valid, but at the time I am unable to work it out on my own, so I just depend on your testimony. If I return later and work the proof through and determine on my own that it is indeed valid, then my own deliberations are evidence of the reliability of your judgment.
Alternatively, it might be that the very act of H depending on S’s moral testimony puts H in a position to check the accuracy of S’s testimony. The point, in brief, is that sometimes it is our dependence on moral testimony that puts us in a position to figure out what is right and why it is right.Footnote 55 Peter is unable to pick out instances of sexism and racism, but we can imagine that by depending on the moral testimony of other, more reliable, co-op members, he could over time develop his capacities for applying the norms, i.e., norms against sexism and racism, to which he is committed. After developing his capacities in this way, we can imagine that Peter would then be able to see, by his own lights, how the people on whose moral testimony he had depended had accurately identified cases of sexism and racism. The poet Ralph Hodgson captures the spirit of this point when he writes: “Some things have to be believed to be seen.”Footnote 56
3.4. Driver's transmission worry
Driver, like McGrath, LaBarge, and Cholbi, expresses concern about hearers’ abilities to reliably identify reliable moral testifiers. She writes,
Someone may possess the disinterest, or impartiality, etc., required to arrive at reliably true moral judgments, but lack the impartiality to deliver those judgments … To accept [one's moral testimony] I must not only have confidence in his judgment, but confidence in the impartial transmission on the judgment, and this may be harder to achieve in cases involving moral judgments.Footnote 57
According to Driver, worries about identifying reliable moral testifiers are compounded by worries about how their judgments get communicated. Even if one is confident that a testifier makes reliable moral judgments, one might lack confidence in that same testifier's ability to reliably convey those moral judgments via testimony. Consider Driver's striking illustration of this point:
Satan could well be an example of a being with superior moral knowledge, but it would be unwise to defer to Satan's judgment on what to do. I might be confident in his ability to know, but not confident in his accurate transmission of that knowledge, because I view him to be deceitful.Footnote 58
To give a more mundane illustration of the point: any dedicated teacher knows that there are all sorts of skills and strategies that one must build up in order to effectively and accurately communicate; it is not enough to simply know your material inside and out.
On Driver's view, then, there are two “levels” on which we need to assess the reliability of a moral testifier in order for our dependence on her testimony to be sound: first, we need to be confident that she can arrive at correct moral judgments; second, we need to be confident in her ability to provide us with those judgments by her testimony. In the preceding subsections I have already laid out numerous ways that we might acquire evidence of a testifier's reliability in making moral judgments, so here I will focus on what is unique to Driver's worry: her contention that it is particularly difficult to come by confidence in a moral testifier's ability to effectively and accurately transmit the content of her moral judgments.
Driver spells out her worry within a framework that assumes a reductionist view of testimonial epistemic warrant. Driver writes,
Edward Craig notes that being a good informant… ‘means more than just being right; in addition to that the good informant must possess some characteristic that make him recognizable as such and supports confidence in his information.’ To know on the basis of what the informant says is to be cognizant of this extra condition or characteristic.Footnote 59
According to Driver, the characteristics that make others’ recognizable as reliable informants – the markers of reliable moral testifiers – include experience, reasonableness, impartiality, and confirmation in judgment. In order to be justified in depending on moral testimony, hearers must be “cognizant” that testifiers satisfy some or all of these characteristics. But Driver's main point can be made without assuming reductionism in the epistemology of testimony. If hearers are in general unable to assess whether or not some moral testifier is a reliable transmitter of moral judgments, then, as with the preceding worries, hearers of moral testimony have a defeater.
But Driver's worry is ultimately untenable, as it does not support the asymmetry thesis. As I will argue, it draws attention to challenges in identifying reliable testifiers that crop up in moral and non-moral contexts.
Driver does not herself offer an explicit explanation as to why confidence in the reliability of the transmission of testimony is harder to achieve in moral, rather than non-moral, contexts. Consider Driver's own central illustration of how making accurate judgments and reliably testifying can come apart:
Perhaps parents are better knowers with respect to their children precisely because they lack impartiality with respect to their children. But note that in transmitting knowledge one may quite justifiably be skeptical of what a parent says about a child, precisely because of the lack of impartiality.Footnote 60
Most plausibly, the content of what parents say about their children is at times moral (as when one parent tells another that her child is exceedingly conscientious and kind-hearted) and at time non-moral (as when a parent tells another that her child is a precocious reader), and at times both. So even Driver's own example fails to motivate the asymmetry thesis. What is more, her example shows how bias infects the transmission of moral and non-moral testimony by providing an illustration of how transmission worries can crop up in non-moral contexts as well as moral contexts.
To be charitable, other parts of Driver's paper are suggestive of one possible explanation that she might have had in mind in claiming that confidence in reliable transmission might be, to quote her, “harder to achieve in cases involving moral judgments.” In those parts, Driver explores the ways in which moral experience can affect moral judgment. She notes,
Though the importance of experience in privileging moral judgment has been noted in the literature, it is surprising to note little discussion of the various ways in which experience can distort one's perceptions and lead to a loss of knowledge or failure to know.Footnote 61
Here, Driver suggests that moral experience is a double-edged sword. It might sometimes uniquely position one to gain moral knowledge and develop moral understanding,Footnote 62 but it also might sometimes block moral knowledge and understanding. As she notes, “The psychology literature, for example, is rife with examples of various fallacies people are prone to – hasty generalization, for example, when a single negative or positive experience takes on exaggerated significance.”Footnote 63 Driver's thought might be that when it comes to the transmission of moral judgments, moral experience can similarly prove a hindrance. Imagine, for example, someone whose own moral experiences have bolstered her abilities of moral discernment regarding gender and equity, and have provided grounds, in part, for her reliable moral judgments concerning sexism. We can imagine that those same moral experiences might, at least in some contexts, play a distorting role in the transmission of her moral judgments. Her own experiences of sexist discrimination, for example, might at times lead her to downplay the nature and extent of her concerns about sexism. Perhaps Driver's worry, more fully spelled out, is that when experience undergirds a testifier's moral judgment, it should make us wary of the reliability of the testifier's transmission of that judgment.
But this explanation, though offering a salutary warning about potential pitfalls in dependence on moral testimony, of course does not give a general reason to distrust moral testimony. For one, not all moral testimony is based on moral experience. For example, a friend of mine recently patronized a local restaurant that served drinks in glassware which featured images of the Washington, D.C. NFL mascot. She told the manager of the restaurant, “The Washington football team name and logo are racist and deeply derogatory.” My friend has professionally and personally thought very carefully about this and related issues; she is also a relatively affluent white woman and her testimony was not significantly based on her own personal moral experiences.
Or, sometimes moral testimony is based on personal moral experience, but not in such a way that it is likely to play a distorting role in the transmission of one's moral judgments. Consider the following example:
Rude Email: Anna shares the social norm against rudeness and she is trying hard to be polite, but she cannot always tell whether her tone of voice, her behavior or an email she is writing is rude. She worries about this because she doesn't want to be rude. Therefore, whenever she is uncertain, she relies on her friend's judgment.Footnote 64
We can imagine that her friend's moral sensitivity to issues surrounding the norm against rudeness have been built up out of personal moral experiences (times she has been treated well, times she has been treated rudely, and the like). But we can also imagine that these experiences have been relatively low stakes, so to speak; Anna's friend has never been dehumanized, or bullied, or profoundly alienated, she has just encountered a sampling of the mundane instances of rudeness (e.g., someone failing to hold the elevator door, being asked a personal question in public, being talked to curtly by a stranger, etc.) and civility that many of us encounter in our everyday lives. So we can imagine that her personal moral experiences have played a large part in Anna's friend adeptness at applying the norm against rudeness. But we can also imagine that these experiences – and, relatedly, even her commitment to the norm against rudeness – are not so central to her values and identity that they are apt to distort her moral testimony.
Finally, and most importantly, explaining how confidence in reliable transmission “is harder to achieve in cases involving moral judgments” by pointing to the distorting influences of moral experience is a strategy that can be used, mutatis mutandis, to undermine dependence on non-moral testimony. It is a strategy that overgeneralizes, and thus it is not a strategy fit to undergird the asymmetry thesis. Consider, for example, that this sort of explanation would make sense of why Driver thinks that we should be wary of parents’ testimony about their children: the very experiences that enable parents’ to know their children better than the rest of us might tempt them, when testifying to others, to exaggerate their children's positive qualities or achievements – both moral and non-moral.
3.5. Acting on the basis of dependence on moral testimony
Consider the following pair of well-known cases:
Low Stakes: Hannah and her wife Sarah are driving home on a Friday afternoon. They plan to stop at the bank on their way home to deposit their paychecks. It is not important that they do so, as they have no impending bills. Hannah says, “I know the bank will be open tomorrow, since I was there just two weeks ago on Saturday morning. We can deposit our paychecks tomorrow morning.”
High Stakes: Hannah and her wife Sarah are driving home on a Friday afternoon. They plan to stop at the bank on their way home to deposit their paychecks. Since they have an impending bill coming due, and very little in their account, it is very important that they deposit their paychecks by Saturday. Hannah says, “I know the bank will be open tomorrow, since I was there just two weeks ago on Saturday morning. We can deposit our paychecks tomorrow morning.”Footnote 65
Some might argue that while Hannah knows that the bank will be open in Low Stakes, in High Stakes she does not know that it will be open. The idea, put in general terms, is that our standards for knowledge and justification are sensitive to practical stakes. This idea, coupled with one plausible interpretation of internalism in ethics – roughly, that to sincerely make a moral judgment is, necessarily, to be motivated to act in accordance with that judgment at least to some extent – could be thought to pave the way for epistemic pessimism. One might attempt to defend the asymmetry thesis by claiming that dependence on moral testimony is different from dependence on non-moral testimony because moral judgments are intrinsically directed towards motivation and action, and hence testimony about such matters carries with it risk that is not present in non-moral testimony.
The first thing to note in response to this line of thinking is that not all moral testimony is directed to the hearer's action; on the basis of moral testimony one may form views upon which one will never be called to act. Consider the following example:
Unfair Boss: Imagine that after work one day you meet up with a friend and he tells you all about his terrible, horrible, no good, very bad day at work, and how his boss passed him up for a promotion, and gave the promotion to a co-worker. He testifies that his boss acted unfairly in doing so. You do not believe that your friend is better informed about potentially relevant non-moral facts (e.g., about his work-related abilities compared to those of his co-worker, or about his boss's criteria for making the decision). You also know that your friend is normally trustworthy and reliable.
If your friend's boss is a member of your weekly Bunco club, and you can reasonably expect to find yourself at intimate gatherings with him on a regular basis, then perhaps you ought to be less willing to readily accept your friend's moral assessment in this case because it carries with it the moral risk of treating your friend's boss poorly. But if you have never met your friend's boss and you reasonably never expect to meet him, it would be unlikely that depending on your friend's testimony, and so believing that his boss acted unfairly, would impact your moral agency moving forward.
That said, of course some moral testimony is directed at hearers’ actions. The concern may be that in these cases stakes are high. Furthermore, even if you are reasonably confident that some bit of moral testimony is not relevant to your moral agency, you could be mistaken. And the costs of your being mistaken are also high. So, it might be argued that it is appropriate to use different standards for believing moral testimony and for believing non-moral testimony.
But this goes too fast. Even if we grant that some moral testimony carries with it a high degree of moral risk, we can still recognize that some non-moral testimony is extremely high-stakes as well. Take, for instance, non-moral testimony concerning the amount of weight a balcony can support, whether the defendant in fact drove the getaway car, how to properly install a carseat, etc. On the other side, there can also be low-stakes moral testimony, e.g., moral testimony concerning the permissibility of telling some particular white lie, whether your co-worker's aunt is gracious, etc.
Since dependence on both moral and non-moral testimony can be both low- and high-stakes, there is no principled difference to be drawn between the two in terms of practical stakes. So it seems that we should endorse a sliding scale of confidence in testifiers in general based on the risks of believing.
But even that goes too fast, for it is not even clear that a worry about practical stakes is germane to a defense of epistemic pessimism. Here is a position that seems sensible: in high stakes cases, testimony can transmit epistemic warrant so as to give rise to knowledge, but nevertheless in such cases hearers are not entitled to act on the basis of that knowledge. This is like a judge not being able to act on the basis of his well-grounded, testimonially based beliefs, if the testimony constitutes evidence that would be inadmissible. That is, worries about the practical risks that can accompany dependence on testimony do not necessarily call into question epistemic warrant for one's beliefs, but rather whether one can justifiably act on those beliefs. In contrast, as I noted above, some epistemologists have argued that epistemic warrant does in fact depend, in part, on practical stakes.Footnote 66 They often appeal to cases like the bank cases to motivate their views. I am not committed either way on whether or not practical stakes partly determine epistemic entitlements. Instead, what I have shown so far – that there can be low stakes moral testimony, and high stakes non-moral testimony – is enough to demonstrate that there is no sense in which the stakes of testimony easily track the moral/non-moral distinction. So, even if epistemic warrant does depend on stakes, that fact should not be pressed into the service of a defense of the asymmetry thesis.
4. Conclusion
I have considered five attempts by epistemic pessimists to explain the asymmetry thesis. Each attempt, I argued, failed. Given the prima facie case I made in favor of the No Difference Thesis, we are left with no good reason to abandon the thesis, and good reason to endorse it. If there is something generally problematic about dependence on moral testimony, it is not that it is unable to transmit moral knowledge.Footnote 67
