Introduction
We get around in the world in large part by relying on the testimony of others. We rely on other pedestrians for our beliefs about directions in cities we are unfamiliar with, we rely on doctors for our beliefs about what is healthy, and we rely on meteorologists for our beliefs about what the weather will be. But we do not usually rely on others for our moral beliefs—even if they are experts and we are not. For example, consider the following case:
Advice: Katie reads a moral advice column where people send in their quandaries for a moral expert to ponder. The expert considers their questions and returns a one-sentence response instructing them what to believe. Katie asks the expert if it is morally obligatory to give money to charity. The expert responds: “You are morally required to give half of your money to charity.” Katie reads the response and comes to believe that she is morally required to give half of her money to charity solely on the basis of the expert’s testimony.
Many philosophers think there is something off-putting or troubling about Katie forming her moral belief solely on the basis of another person’s testimony.
Moreover, the fact that her deference is off-putting is somewhat puzzling. Why? There are two reasons. First, there is nothing off-putting about giving similar deference to experts in other domains (e.g., accountants, doctors, religious leaders, etc.). Second, it is generally agreed that moral deference is epistemically powerful in that it can transmit both knowledge-that and knowledge-why.Footnote 1 Call this the “Puzzle of Moral Deference.”
Most explanations of this puzzle rely on an epistemic and a moral claim. The epistemic claim normally amounts to saying that moral deference, as a method of forming moral beliefs, is epistemically deficient. For example, it does not give one access to any morally relevant properties of actions, i.e., the properties that bear on the moral status of an action (e.g., Enoch [Reference Enoch2014], Fletcher [Reference Fletcher and Shafer-Landau2016], and Lord [Reference Lord and Shafer-Landau2018]), or it cannot (or does not) yield moral understanding-why (Nickel [Reference Nickel2001], Hopkins [Reference Hopkins2007], Hills [Reference Hills2009 and Reference Hills2010], McGrath [Reference McGrath2011], and Callahan [Reference Callahan2018]).Footnote 2 The moral claim is that because of the epistemic deficiency, moral deference is also in tension with gaining certain moral achievements, e.g., performing actions with moral worth or fittingly having the full range of affective and conative reactions to morally valenced situations. This is because having access to morally relevant proprieties or having moral understanding-why is necessary (or at least very important) for these moral achievements. Thus, for these authors, what explains the off-puttingness of moral deference is the fact that it is in tension with gaining certain moral achievements.Footnote 3 Call this explanation “The Achievement Explanation.”
I agree with proponents of the Achievement Explanation that there is something off-putting about moral deference.Footnote 4 However, I argue that the Achievement Explanation should be rejected. First, I argue that the Achievement Explanation proves too much, i.e., it overgeneralizes. It entails that using many—if not most—of the methods people have for gaining or maintaining moral beliefs should be off-putting (e.g., inference to the best explanation, reductio ad absurdum, and intuition). But there is normally nothing off-putting or troubling about using these methods to form or maintain moral beliefs.Footnote 5 The Achievement Explanation entails this undesirable conclusion because the aforementioned methods of moral belief formation seem to have the same epistemic deficiencies as moral deference and thus are also in tension with gaining the relevant moral achievements. Second, I argue that there is a way in which moral deference is sometimes actually superior in the sense that it puts one in a better position to gain the important moral achievements in question. Thus, proponents of the Achievement Explanation should find moral deference less off-putting and less problematic than using other belief-formation methods. But the opposite is true. Thus, proponents of the Achievement Explanation are left having to explain why moral deference is so troubling, but all these other methods of moral belief formation are not—given that moral deference is sometimes epistemically superior by their own lights. Call this The “New Puzzle of Moral Deference.”Footnote 6
The plan of the paper is as follows. In section 1, I clarify what cases of moral deference philosophers are interested in and what positions they take on them. In section 2, I briefly summarize some of the main explanations of the Puzzle of Moral Deference and show how they are versions of the Achievement Explanation. In section 3, I give detailed examples of several methods for forming moral beliefs that have the same epistemic deficiencies that moral deference is purported to have (and thus also conflict with gaining certain moral achievements), but which are not off-putting. In section 4, I explain why a certain kind of moral deference is epistemically superior to the methods mentioned in section 3 and thus better for gaining the relevant moral achievements. In section 5, I briefly conclude.
1. Moral deference: A brief primer
1.1 The phenomenon
Consider the following case that is popular in the literature:
Vegetarianism: Katie and her friends want to know if eating meat is morally permissible. They each examine the same body of evidence and come to know all the nonmoral facts relevant to eating meat contained in that body of evidence. After looking at this evidence, Katie is unsure whether it is morally permissible to eat meat and suspends her judgment. All her friends, however, are convinced that eating meat is morally wrong. They jointly tell Katie that eating meat is morally wrong and solely on the basis of this testimony, Katie forms the belief that eating meat is wrong.Footnote 7
Vegetarianism is meant to be a paradigmatic case of troubling moral deference. However, there is not clear agreement about exactly which features of cases are necessary for generating the Puzzle of Moral Deference.Footnote 8
The following are the features that everyone seems to agree are required for generating the puzzle. First, the deferrer is a mature moral agent. This feature is important, but often not sufficiently spelled out. The deferrer is much like you and me with respect to their moral knowledge, moral understanding, and moral sensibilities. That is, they have a fair amount of moral knowledge and understanding. This is one reason that there is no puzzle of moral deference concerning children or immature moral agents. Relatedly, mature moral agents must be capable of gaining moral understanding of why the moral claim under consideration is true. Both Hopkins (Reference Hopkins2007) and Hills (Reference Hills2009) make it explicit that if one cannot gain moral understanding about why the claim under consideration is true, then it shouldn’t be off-putting or impermissible for one to defer.Footnote 9 In fact, Hills (Reference Hills2009) argues that if you cannot acquire moral understanding of why p, but you can come to acquire knowledge that p via deference, then you should defer. This is because knowing that p will help you do the right thing.Footnote 10
Mature moral agents are also reliable recipients of testimony. This is because being a reliable recipient of testimony is necessary for being able to get testimonial knowledge. Footnote 11 For example, if a deferrer would believe just any moral proposition a testifier told her, it’s unclear that she would gain moral knowledge when the testifier was reliable and asserting a true moral claim. Hopkins (Reference Hopkins2007) and Hills (Reference Hills2009) both explicitly claim that mature moral agents won’t believe just any moral claim or any moral testifier.Footnote 12 Part of being a reliable recipient of moral testimony is having a grasp of the kinds of considerations that bear on the moral status of an action and perhaps also a grasp of how much they tend to bear on the moral status of an action, e.g., causing lots of pain significantly counts against doing an action, but causing mild irritation does not. If the deferrer didn’t know what kinds of considerations were normally morally relevant, then she would not be able to reliably receive moral testimony.
Second, the testifier is an epistemic superior of the deferrer (e.g., the testifier is an expert or a group of the deferrer’s epistemic peers).Footnote 13 Third, the deference is purely moral, i.e., the deferrer only defers about thin moral content. In other words, the deference is only about things like the moral status of actions: right/wrong, permissible/impermissible, good/bad, obligatory/optional, etc. Fourth, the deferrer forms an outright or full moral belief on the basis of the moral expert’s (or her epistemic peers’) testimony.
Fifth, the outright moral belief is only based on that moral testimony.Footnote 14 This last condition needs some clarification. It doesn’t say that no other considerations that bear on the truth of the content of the belief play an explanatory role in why the deferrer forms the moral belief. For one, the deferrer will have some evidence that the testifier is a reliable moral testifier or that she shares some of the same core moral values as the deferrer.Footnote 15
Moreover, the deferrer, as I imagine her, filters the content of the moral testimony through her own moral beliefs and sensibilities. By this, I mean that if the content of the moral testimony conflicted with the deferrer’s strongly held antecedent moral beliefs or with her strong moral intuitions about the case, she’d be less inclined to accept the testimony. Thus, as noted above, it’s not as if she would believe just any moral claim. In fact, many cases of moral deference will involve the deferrer knowing what the morally relevant considerations are and even grasping their individual weight but not being able to figure out exactly how they weigh up or interact with each other, or deferring to someone else about the total weight of reasons.Footnote 16
Importantly, these considerations are not the deferrer’s bases for belief. Rather, they’re better thought of as enabling conditions. Had the deferrer not filtered the testimony though her moral beliefs and sensibility, she would be at risk for psychological or normative defeaters for the content of the testimony.Footnote 17 Moreover, we need not think of the deferrers as epistemic pushovers who give no weight to their moral beliefs or sensibilities. In fact, Hills (Reference Hills2009) is explicit that if a person thinks that their moral beliefs or sensibilities are “too far off the mark,” it is fine for them to defer.Footnote 18
The reason for focusing on this narrow set of propositions is to eliminate possible confounding factors. First, the testifier is restricted to an epistemic superior to the deferrer because otherwise the oddness of deference could be explained by the fact that the deferrer is deferring to an epistemic peer or inferior.Footnote 19 Second, the deferrer’s belief is solely based on the testifier’s say-so to ensure that our assessment of her belief is not based on any other grounds of her belief. Third, the deference only concerns purely moral propositions because impure moral propositions mention grounding claims, e.g., “David’s action was wrong because it constituted failing to maximize happiness.” The problem with providing this additional content is that the deferrer might use this content as another basis of her belief and thus not completely base her belief on the testifier’s word. Or, she might infer from (1) the claim that the action constituted failing to maximize happiness and (2) her background knowledge that failing to maximize happiness makes actions wrong to (3) that David’s action is wrong. In either case, her belief is not completely based on the testifier’s say-so. Fourth, the testifier’s belief concerns thin moral content because thick moral content entails descriptive content that the deferrer could use to infer the moral status of the acts, events, persons, etc. under consideration. Finally, the deference concerns outright belief as opposed to credences because the former settles a person’s view and make further inquiry unjustified in many cases.Footnote 20
1.2 The central questions
Much of the literature on moral deference has focused on the normative question: Does one have a strong reason (of some sort) to not defer on moral questions? Pessimists say yes, while optimists say no.Footnote 21 Second, there is the explanatory question: What explains our aversion to moral deference? Herein, I will focus on the explanatory question. That is, I will focus on to what extent the accounts I consider can explain our aversion to moral deference. However, what I argue bears directly on the normative question as well. Pessimists usually think we have a strong reason to not defer on moral matters because deference is somehow incompatible with obtaining certain moral achievements, e.g., performing actions with moral worth. However, the reason they think that moral deference is incompatible with these moral achievements is that moral deference is incompatible with gaining moral understanding. It follows from their arguments that we have a general obligation to not use any belief-formation method that is incompatible with gaining moral understanding. Thus, pessimists end up having to claim that we have a general obligation to not use the belief-formation methods I discuss (e.g., abduction, reductio ad absurdum, intuition, etc.), however, this seems patently false. So, my arguments against the Achievement Explanation are also arguments against certain pessimistic answers to the normative question.Footnote 22
2. How moral deference is deficient
Proponents of the Achievement Explanation tend to offer one of two explanations why moral deference is incompatible with gaining certain moral achievements: (1) it does not give us access to the morally relevant properties of actions, and (2) it does not (or cannot) yield moral understanding-why. However, these two camps are related. Why? Because coming to understand why some moral claim is true seems to require having access to at least some of the morally relevant properties involved.
2.1 No access to morally relevant properties
Enoch (Reference Enoch2014) is the most explicit about explaining the Puzzle of Moral Deference by virtue of the fact that it does not give one access to morally relevant properties. He claims that deference only provides “opaque” evidence, by which he means evidence that does not involve any access to morally relevant properties. He writes, “I think it is not premature, then, to conclude that the problem with expertise-based moral knowledge is an instance of the more general problem of opaquely based moral knowledge” (237). What is so important about forming beliefs on “transparent” evidence? Enoch writes:
[W]hat is fishy about moral deference … is that forming a moral judgment by deference and then acting on it is much less of a moral achievement than forming the true judgment without deference, because it does not constitute the appropriate response to the morally relevant features of the case. Moral deference indicates failure to achieve something we are expected to achieve [my emphasis].
Because one does not form the moral belief on the basis of morally relevant properties one does not respond to the “morally relevant features of the case,” and this leads to one’s acting on that belief to constitute “much less of a moral achievement.”
Fletcher’s (Reference Fletcher and Shafer-Landau2016) explanation of the Puzzle of Moral Deference is also a version of the Achievement Explanation. To see why, consider the following principle:
Sentiments Deference Denial* (SDD*): Moral sentiments are at least difficult to [rationally] form on the basis of pure, direct, testimony.Footnote 23
Fletcher defends SDD* by noting certain features of sentiments that make them hard to rationally acquire solely through testimony. Chief among these features is that sentiments are responses to particular properties of actions or events. I take this to mean that sentiments are responses to particular tokens of morally relevant properties.Footnote 24 Rationally having a range of moral sentiments seems like an important moral achievement.
Lord (Reference Lord and Shafer-Landau2018) thinks that what is amiss with moral deference is that it does not put one in a position to gain “appreciative knowledge.” This kind of knowledge allows one to “fittingly have the full range of affective and conative reactions” to a particular case. The reason that deference does not put one in a position to gain this kind of knowledge is that it does not put one in a position to be acquainted with the particular morally relevant properties of a given action.
Both appreciative knowledge and acquaintance are important, according to Lord, for numerous reasons. First, appreciative knowledge can put one in a position to “fittingly have the full range of affective and conative reactions” to a particular case. For example, imagine that one witnesses a murder. In response to seeing this horrific event, one is in a position to fittingly have the aforementioned reactions. Moreover, knowing that the murder was wrong gives one a reason to be repulsed by it and a reason to desire that the murder never occurred. Second, being acquainted with certain moral properties via experience (e.g., witnessing a murder) or imagination can put one in a position to recognize these properties (or similar ones) in other circumstances. Third, such acquaintance can also put one in a position to grasp the weight of certain properties (e.g., the weight of the reason one has to not kill an innocent person).
2.2 No moral understanding-why
Some philosophers think that moral deference is off-putting and deficient because it cannot yield moral understanding-why and this understanding is necessary for certain moral achievements. What is involved in understanding why p where p is a moral proposition? Understanding why p involves more than just knowing that p and knowing that q is why p. One also needs to base one’s belief that p on q in the right way, where this often means that one must “see” or “grasp” that q is why p. For example, imagine that a mathematician tells you that some claim is correct and that some set of premises makes up a proof of it. In order to understand why that mathematical claim is correct, one must see that the premises entail the conclusion.Footnote 25
Hills (Reference Hills2009, Reference Hills2010, Reference Hills2016) defends the most comprehensive account of moral understanding-why and so I will focus on her account. Hills’s account can be split into two parts: objective and subjective conditions. Objective conditions for understanding why p are conditions outside the agent’s control that must obtain in order for her to understand why p. There is only one objective condition for Hills—understanding why p is factive. In other words, in order to be able to understand why p, p must be true. Subjective conditions for understanding why p, on the other hand, are those that the agent can meet herself. There are three subjective conditions for understanding why p: if one understands why p, then: (1) one believes that p, (2) one has a belief (or beliefs) as to why p, i.e., one has a view about the explanation or reason why p, and (3) one grasps a reason why p, i.e., one grasps or “sees” the relationship between p and whatever explains why p.
Importantly, in order to be able to understand why some moral proposition is true, one needs to know at least one morally relevant property of that action. So, if one cannot get access to any morally relevant properties via moral deference, then one cannot gain understanding via deference.
Here is some of what the following authors have to say about moral deference:
Hills (Reference Hills2009) writes:
But as with many abilities [of which, moral understanding is one], you will not always or even typically acquire these abilities through testimony. Instead, you have to practice.Footnote 26
Hopkins (Reference Hopkins2007) writes:
[H]aving the right to a moral belief requires one to grasp [read: understand] the moral grounds for it.Footnote 27
McGrath (Reference McGrath2011) writes:
Similarly, the realist might say, an agent who is dependent on pure moral deference for her knowledge that an action has certain right-making features is typically not in a position to perform the action on the basis of those reasons, in the way that someone who genuinely grasps the connection between the two [i.e., has understanding] is.Footnote 28
Thus, moral deference is deficient because even if it gives one access to morally relevant properties it cannot yield understanding.Footnote 29
According to pessimists like Hills, moral understanding is important because it is necessary (or at least very important) for certain moral achievements—reliably doing the right thing on one’s own, being able to justify oneself to others, being (fully) virtuous, and being able to perform actions with (full) moral worth. Thus, when deferrers fail to gain moral understanding, this negatively affects their ability to be fully virtuous or perform an action for a right-making reason, and so on.
3. There is nothing wrong with opaque methods of forming moral beliefs
In this section, I argue that there are many methods of forming moral beliefs that have the same epistemic deficiencies that moral deference is purported to have and thus are also in tension with gaining the same moral achievements. But, I argue, using these methods is not off-putting. The upshot of this argument is that proponents of the Achievement Explanation must explain why is it that moral deference is off-putting, but all these other methods, which have the same deficiencies, are not intuitively problematic.
3.1 Inferential belief-formation methods
3.1.a Abduction (inference to the best explanation)
In abduction (i.e., inference to the best explanation), one tries to figure out what conclusion one's evidence best supports. Here’s a case involving a moral proposition. Emily is curious about the following moral principle:
Doing/Allowing: It is morally worse to cause harm than to allow harm (all else being equal).
Emily has the intuition that Doing/Allowing is correct but doesn’t want to base her belief on her intuition. She notices that the principle has a lot of explanatory power. For example, it can explain our intuitions in a wide array of cases, e.g., why it’s intuitively permissible, but not required, to redirect the trolley to the side track; or why passive euthanasia is intuitively permissible, but active euthanasia is not; or why we do not think that people who fail to donate to child-saving charities are not monsters but child-killers are. Moreover, Emily reads all the arguments against Doing/Allowing and correctly finds them to be unsuccessful. Importantly, the reasons these arguments fail have nothing to do with why the Doing/Allowing distinction is true. The arguments are invalid or rely on cases that are not actually cases of the Doing/Allowing distinction. Emily reasons that the best explanation of these data is that the Doing/Allowing distinction is true. Solely on the basis of her abductive reasoning, she forms the belief that it’s morally worse to cause harm than to allow harm (all else being equal).
Recall that just because she bases her belief in the Doing/Allowing distinction only on her inductive reasoning, it doesn’t follow that she can’t say anything about why the principle is true. She might suspect that it is grounded in ownership of one’s body.Footnote 30 Recall, she has the intuition that it’s true. Again, the deferrers we are interested in are mature moral agents. However, because she does not know whether Doing/Allowing is true, she is not in a position to know for sure why it is true. Moreover, recall that she will be filtering her belief about Doing/Allowing through her own moral beliefs and sensibilities. She would not believe something like, “It’s always permissible to murder children for fun,” even if there was a strong inductive argument for it. So, she is a reliable moral believer.
Now notice that the kind of reasoning Emily uses does not include any reference to morally relevant properties, nor does it give us any access to morally relevant properties. That is, coming to believe Doing/Allowing is true provides one with no insight into why this is true. And, without any morally relevant properties, one cannot come to understand why the claim is true or be in a position to gain the relevant moral achievements.Footnote 31 Yet, there does not seem to be anything off-putting about forming a moral belief solely on the basis of this reasoning. But, if the Achievement Explanation is true, then we should expect forming moral beliefs on the basis of such abductive reasoning to be as off-putting as moral deference.
3.1.b Belief conflict
Another common source of our moral attitudes is whether a claim is consistent with our most deeply held moral convictions. Imagine that Heidi is a hedonist and thus thinks that the only consideration that could tell against some action is that the action causes pain. Imagine further that Heidi’s hedonism is completely justified—even though hedonism is a false view. Hedonism is a core part of Heidi’s identity as well, and thus she would not easily discard it. Now imagine that Heidi realizes that she also believes (though she is not sure why) that it is wrong to steal something from someone even when no one will ever know about the theft. Given her justified commitment to hedonism, however, she should not think that such actions are impermissible. Once she sees the conflict, she comes to disbelieve that such cases of theft are wrong solely on the basis that it conflicts with her more strongly held conviction about hedonism.
Of course, when Heidi reasons to the conclusion that it is not wrong to steal things if one can “get away with it,” she is not put into contact with any of the morally relevant properties. Thus, her reasoning has not told her why the claim is true (or false). She believes that the action is permissible because it does not cause anyone any pain. However, what the proponents of the Achievement Explanation are concerned with is having access to the morally relevant properties that make it the case that the action has the moral status it does. Because “victimless” theft is wrong, there are no properties that make it the case that it is permissible. Nevertheless, there does not seem to be anything off-putting about disbelieving some particular moral claim because it conflicts with one’s core values—especially when one is epistemically justified in holding those values.
3.1.c Higher-order evidence conflict
One good reason to change a moral belief is if one receives higher-order evidence about the reliability of one’s formation of that belief. For example, imagine that Phoebe formed the belief that sending her children to an expensive private school is morally permissible on the basis of a strong intuition. Imagine further that Phoebe finds out that she was secretly given an antireliability pill that caused her to have the strong intuition on which she based her belief. She comes to find out that her judgment was perfectly antireliable and thus that she should believe the opposite of what her intuition supported. Phoebe then comes to believe that it is not morally permissible for her to send her children to an expensive private school.
Now notice that the kind of reasoning Phoebe uses does not include any reference to morally relevant properties, nor does it give us any access to morally relevant properties. All she comes to know is that she should believe the exact opposite of what her intuition tells her to believe and thus she is given no information about why the opposite belief is true. And, without any morally relevant properties, one cannot come to understand why the claim is true or be in a position to gain the relevant moral achievements. Yet, there does not seem to be anything off-putting about her forming a moral belief solely on the basis of this reasoning. In fact, it seems like she is epistemically required to form this new belief. But, if the Achievement Explanation is true, then we should expect forming moral beliefs on the basis of such reasoning to be as off-putting as moral deference.
3.1.d Moral consistency
Jan Baalsrud, a Norwegian resistance fighter during WWII, was hiding from the Nazis with a family in a small village in Norway. The family had to decide whether to continue to hide Jan and put themselves at risk or turn him over to the Nazis. The mother of the family sympathized with Jan, but thought they were morally required to turn him over because of the risk it posed to multiple people—perhaps even the whole village. However, her son Marius disagreed. He asked his mother, “What if I were a resistance fighter being hidden by a family in Oslo? What ought that family to do with me?” Marius’s mother realized at that moment that there was no moral difference between Jan’s current position and her son’s potential position.Footnote 32 Then, solely on the basis of recognizing that there is no moral difference between Jan’s actual situation and her son’s hypothetical situation, she comes to believe that her family should keep hiding Jan.
Marius’s mother’s moral revelation did not put her in contact with any new morally relevant properties concerning Jan’s situation. What Marius’s question brought out to his mother was that she was not properly responding to the reasons she had (or that were in her ken) to continue to hide Jan. Moreover, it may have been the case that she did not gain any moral understanding because she did not grasp why it would have been morally required for the hypothetical family in Oslo to hide her son. She might have justified her judgment by claiming that Marius is a good boy, or her son, or that she would be heartbroken if someone turned him over to the Nazis.
Perhaps I’m underestimating how much Marius’s mother learns from her son’s question. Perhaps in most cases like this, she would understand why the hypothetical family ought to hide Marius if he were on the run, and once she sees that there is no difference between the hypothetical case with her son and Jan’s actual case, her antecedent understanding extends (or is transmitted) to Jan’s case.Footnote 33 The proponent of the Achievement Explanation might argue that this is the difference between moral deference and moral consistency reasoning: moral deference does not involve extending antecedent moral understanding.
While I agree that Marius’s mother does extend her moral understanding in this case, I think that moral deference also allows people to extend their moral understanding in much the same way. Imagine a slight variation of this case where Marius’s mother has not yet thought about whether the case of Jan is morally equivalent to a hypothetical case involving her son. She has moral understanding of why the hypothetical Oslo family would be morally required to hide her son if he were on the run: because, let’s say, they would have a duty to protect resistance fighters. She grasps the relationship between having a duty to protect resistance fighters and being required to protect any particular resistance fighter—as long as there are no defeaters. She also grasps the following hypothetical: Her family would be morally required to hide Jan if his case were morally equivalent to her son’s hypothetical case. It seems like the only thing preventing her from understanding why her family is required to hide Jan is her not believing that Jan’s cases and her son’s hypothetical case are morally equivalent. Now, imagine that Marius tells her that there is no moral difference between Jan’s situation and his hypothetical situation and she defers to him. It seems like once she comes to believe that there is no moral difference between the two cases, she comes to understand why her family is required to hide Jan in much the same way that she did when she reflected on Marius’s question.Footnote 34
A natural reply from the proponent of the Achievement Explanation is that there is an important asymmetry between how Marius’s mother acquires moral understanding in the two cases. In the original case, Marius’s mother comes to grasp or see that there is no difference between her son’s case and Jan’s case and, on the basis on this grasp, she forms the belief that her family is morally required to hide Jan and finally ends up understanding why they are required to do so. However, in the amended version, she comes to believe that there is no difference between her son’s case and Jan’s case, then she comes to believe that they are required to hide Jan because they have a duty to protect resistance fighters. And since she already grasps the relationship between having a duty to protect a resistance fighter and being required to protect any resistance fighter, she also grasps why her family ought to hide Jan.
The path from deference to moral understanding seems a bit longer than the path from seeing that there is no moral difference between the cases to moral understanding. Perhaps the difference is that some nontestimonial sources of moral belief extend antecedent moral understanding while testimony converts conditional moral understanding into unconditional moral understanding. However, the path from belief to moral understanding cannot matter to proponents of the Achievement Explanation because what they care about is not how one acquired moral understanding but that one acquired moral understanding. After all, they value moral understanding because it is necessary (or very important) for the aforementioned achievements and how one acquired moral understanding does not affect one’s ability to gain these achievements.
Important for my purposes is that this kind of conditional moral understanding is common in mature moral agents and forming certain moral beliefs can directly lead to the acquiring of unconditional moral understanding. One way to gain unconditional moral understanding is the way that Marius’s mother did, i.e., by coming to believe that two actions are morally equivalent when one understands why one of those actions is right or wrong. Another way is to understand the correct moral view and come to find out that some action is right or wrong. For example, let us imagine that some form of Kantianism is correct and that an agent understands the basics of this view. If she were to come to believe that action A is wrong, she would simultaneously come to understand why it is wrong, i.e., because it constitutes treating a person as a mere means. Finally, a person could come to acquire unconditional moral understanding in the following kind of case. First, she had a basic understanding of moral theories, but does not believe any of them to be correct. Second, she believes that some action A is right or wrong but does not know why. And, third, she defers to somebody about which moral theory is true. Because she has antecedent moral understanding of the moral theories, she can extend this understanding to any action she comes to believe is right or wrong.
3.1.e Reductio ad absurdum
One of the strongest and most common ways of arguing for a moral claim is via reductio ad absurdum, i.e., by showing that its negation leads to either a contradiction or to an absurd conclusion. In addition, forming moral beliefs via reductio seems common. Consider the following:
1. Suppose for reductio that no animals have any rights.
2. If no animals have any rights, then it is always permissible to torture puppies for fun.
3. But it is obviously false that it is always permissible to torture puppies for fun.
4. So, it is false that no animals have any rights.
Notice that this argument gives us good reason to think that at least some animals (e.g., puppies) have some rights (e.g., not to be tortured).
Again, the reasoner will likely have some idea of why some animals have some rights, e.g., they feel pain, they have desires, they are loved by humans, etc. Nonetheless, she may not think she has enough evidence to form the belief that some animals have some rights. At the very least, she will believe that if some animals have some rights, then it must be because one of the following: they feel pain, they have desires, humans love them, etc. Then, once she considers the above reductio, she is convinced that some animals have some rights, because otherwise, something absurd would be true.
Now notice that this form of argument justifies belief in certain propositions without providing any indication of the morally relevant properties. For example, we are left clueless as to why some animals have some rights. Thus, this is an opaque method of moral belief formation and as such does not help one gain the relevant moral achievements. But notice that there is nothing off-putting about forming moral beliefs solely on the basis of a reductio nor do we intuitively have a strong reason to avoid using this method. In fact, if there were something fishy about using this belief-formation method, then much of applied ethics would be in trouble. After all, lots of applied ethics involves arguing against a view between it has an absurd ethical conclusion.
Perhaps one will object that in the above reductio, the reasoner started out with some idea of why some animals have some rights, e.g., they feel pain, they have desires, humans love them, etc. Thus, once they come to believe that some animals have some rights, they are also propositionally justified in believing that some animals have some rights because they feel pain, or they have desires, or humans love them, etc. However, the same propositional justification can come from deference. Imagine that the reasoner believes that if some animals have some rights, then it must be because they feel pain, have desires, or are loved by humans. Imagine further that she defers to a moral expert that some animals have some rights. As soon as she defers, she is just as propositionally justified in believing that some animals have some rights as she would have been had she used the reductio.
Another reason that using reductio might be intuitively better than trusting moral testimony is that the former gives one access to a reason why one’s moral belief is true. For example, the person who comes to believe that some animals have some rights will know that the negation of this claim leads to an absurd conclusion, i.e., that it’s always permissible to torture puppies for fun. However, reductios don’t give one access to any morally relevant properties, i.e., the properties that make it the case that an action is right or wrong. Notice that the reason provided by the above reductio would not be the kind of reason that one could act on to give one’s action moral worth.
Thus far, I have only argued that using opaque inferential belief-formation methods is not always off-putting, but perhaps the reason that trusting moral testimony is off-putting is because trusting testimony is an opaque and noninferential belief-formation method. I will now argue that using an opaque and noninferential belief-formation method is also not always off-putting.
3.2 Noninferential belief-formation methods
3.2.a Intuition
According to one account of what intuitions are, they are intellectual seeming states.Footnote 35 They present the world as being a certain way and have a phenomenology similar to that of perceptual seemings. Forming beliefs on the basis of intellectual seemings is just like forming beliefs on the basis of perceptual seemings. A house seems to me to be green and on the basis of this seeming I come to believe that the house is green. Likewise, pain seems to me to be bad and on the basis of this seeming I come to believe that pain is bad.Footnote 36
Does coming to form a belief on the basis of moral intuitions give one access to morally relevant properties or yield understanding of why something is the case? I think not. For example, I might have a strong intuition that pain is bad and form the corresponding belief without also getting access to the properties that make pain bad or understanding of why pain is bad. This will certainly be true of fundamental moral truths because there is not any further reason why they are true and so one cannot acquire access to the morally relevant properties or gain understanding of why they are true.
More important, even if one was given access to morally relevant properties by intuitions, this access would be given in virtue of having the intuition, not forming a belief on the basis of it. I could have a strong intuition (e.g., that killing is morally worse than lying) and fail to form a belief on the basis of it. But, what is off-putting about moral deference is forming a belief on the basis of what another person says. There is nothing off-putting about receiving testimony just like there is nothing off-putting about “receiving” an intuition. Likewise, there is nothing off-putting about asking for moral advice just like there is nothing off-putting about consulting one’s intuitions.
3.3 Combining methods
Finally, one can use all of these belief-formation methods in conjunction and still not have access to the morally relevant properties of an action and so not be in a position to gain the relevant moral achievements. But there certainly does not seem to be anything deficient or off-putting about forming a belief that p on the basis of some combination of an abductive argument, a conflict with a deeply held belief, conflicting higher-order evidence, moral consistency, a reductio ad absurdum, and an intuition. But, again, if the Achievement Explanation were true, we should expect that we would find forming a moral belief on this basis to be off-putting. Moreover, if we agree with pessimists, we should think that we have a strong reason to not form moral beliefs on the basis of this combination of methods.
4. Why moral deference is superior
4.1 Moral testimony vs. mere deference
The literature on moral deference tends to elide the difference between two distinct phenomena: “moral testimony” and what I will call “mere deference.” Here is the intuitive difference between the two. When a person’s deference is based on testimony, she forms her belief that p on the basis of being told that p by a testifier directly. The testimony is aimed at or intended for her. The testifier knowingly puts herself in a position where the audience might rely on her testimony. When a person’s deference is mere deference, she forms her belief that p on the basis of another person’s assertion (or belief) and that assertion is not aimed at or intended for her, nor is it public. Examples of this include forming a belief on the basis of eavesdropping or reading another person’s diary.Footnote 37
The distinction between moral testimony and mere deference is important for my project because my claim that moral deference is epistemically superior to the aforementioned nontroubling methods of moral belief formation is meant to only apply to moral testimony. This is because all of the epistemic advantages that I argue for derive from the interpersonal nature of testimony. In particular, I follow several authors in holding that when one asserts a proposition to another person, one takes on certain normative commitments.Footnote 38 For example, one takes on the commitment to defend the content of one’s assertion in some range of cases or one is committed to the defensibility of that content.Footnote 39 In this way, I think asserting (to another person) is importantly analogous to promising.
However, I do not think that one takes on these commitments in simply believing things or expressing them privately, e.g., in a diary. Like Watson (Reference Watson2004), I think it is only when one makes these beliefs public (via assertion) that one takes on this commitment. In cases of mere deference, the source of the belief does not take on any such commitments because the source of the belief doesn’t make their beliefs public. Thus, the epistemic advantages that I will argue accompany belief based on moral testimony do not accompany beliefs based on mere deference.Footnote 40
4.2 Epistemic superiority
Before expanding on why I think moral testimony is sometimes epistemically superior to the aforementioned belief-formation methods, I need to say what I mean by “epistemically superior.” I’m only concerned with what I will call a person’s “epistemic position.” One’s epistemic position is determined by one’s opportunities or abilities regarding something of epistemic value. Some people’s opportunities and abilities regarding things of epistemic value are superior to others. The better the epistemic position, the easier it is to gain: (a) more valuable doxastic states (e.g., true beliefs), or (b) higher-quality doxastic states (e.g., knowledge or understanding), or (c) more or higher-quality evidence. For example, a mechanic is in a better epistemic position vis-à-vis understanding why the car stopped running than someone who knows nothing about cars. Why? Because the mechanic already has related understanding. In particular, the mechanic has an intellectual ability that the novice does not. She has much more knowledge about cars than the novice and she so can much more easily come to gain knowledge about why the car will not start. So, in virtue of her understanding, she is in a better epistemic position than the novice.
The epistemic position that is relevant to defenders of the Achievement Explanation is one’s ability to access the morally relevant properties of actions. When one defers about the moral status of an action, one does not seem to come to know what the morally relevant properties of the action are. Nonetheless, I argue that those who base their moral beliefs on moral testimony are in a superior epistemic position relative to users of the other belief-formation methods.
4.3 Why moral deference is epistemically superior
In what ways is moral testimony epistemically superior to the other methods of belief formation discussed herein? First, there seems to be a certain kind of responsibility that a testifier takes on when she asserts a proposition to an audience and that audience takes their word for it. Footnote 41 In particular, it seems like the testifier takes on the responsibility of meeting third-party challenges to the audience’s belief. That is, when the audience’s testimony-based belief is challenged and the audience cannot answer the challenge, she can pass the buck to the testifier.Footnote 42 For example, imagine Katie believes that human-organ markets are morally bad on the basis of a moral expert’s testimony. Chad challenges Katie to defend her belief. She tells Chad that a reliable moral expert told her, and she provides reasons for thinking that that expert is reliable. But, given that it is important to Chad to get the answer right, he presses Katie for more reasons. At this point, there seems to be a further response available to Katie: she can pass the buck to the moral expert. That is, she can direct Chad to ask the moral expert for a further defense of her belief.
But this extra move is not available in many cases involving forming moral beliefs using the other methods I have discussed. For example, if one believes that killing is worse than letting die on the basis of an intuition, one cannot defend one’s belief by directing a challenger to one’s intuition.
Moreover, being able to pass the buck puts one in a better epistemic position vis-à-vis maintaining one’s moral belief. What I have in mind is that most people are concerned with having moral beliefs that are resilient, i.e., beliefs that can be rationally maintained even after being challenged by others. When one’s moral belief is challenged and one cannot offer a satisfactory reply, one might be rationally forced to reduce one’s credence or give up one’s belief. But, if one can pass the buck to someone else, there is less of a burden to reduce one’s credence or give up one’s belief.Footnote 43
Second, testimony is a dynamic epistemic source in a way the other belief-formation methods discussed are not. A testifier has a reason to update when she knows that her audience (actually or likely) still holds a view that she now thinks is clearly wrong and that her audience holds this view on the basis of her testimony. Here, one has a strong reason to tell one’s audience that there are reasons to no longer hold that view. Of course, practical and moral reasons will importantly restrict when one has an all-things-considered requirement to update one’s audience. But, when it is relatively effortless to do so, one will often have a strong moral reason to do so.Footnote 44
Here is a more formal expression of what I have in mind:
Reason to Update: When B believes some moral proposition, p, only because B deferred to A’s testimony and A knows this, A has a pro tanto moral reason to tell B that (maybe) not-p if: (1) A no longer thinks her assertion was warranted because it failed to meet the norm of assertion in the first place, or (2) A no longer thinks what she asserted is true because she now has compelling counterevidence, or (3) A knows that B is about to engage in an action with high moral stakes and is doing so in large part because B trusts A’s assertion.
This norm is related to having reasons to update those who are currently relying on you for practically relevant information. For example, imagine that you see Susan on Sunday afternoon and she asks you when John’s party is. You tell her it is Tuesday night. However, on Monday, you find out John’s party is actually Wednesday night. It seems like you ought to tell Susan that John’s party is actually on Wednesday night—at least if it is relatively effortless to do so. I take moral beliefs to be a specific example of this broader phenomenon because they so often have practical upshot. Moreover, because moral beliefs often have the ability to affect actions with high stakes, one will often have a strong—though defeasible—moral reason to update one’s audience.
Relying on someone’s moral testimony puts one in a better epistemic position regarding moral belief maintenance. If a testifier has a strong reason to inform you of a false (or otherwise improper) belief that you have, and they know they have this reason, then one’s testimony-based beliefs have a kind of back up. One is more likely to know if one’s belief is false or unjustified or if there are defeaters than one would be if one relied on mere deference. In other words, one’s belief is more sensitive, i.e., there are fewer close possible worlds in which p is false and one believes that p. Footnote 45 This is the advantage of sharing beliefs with other people where those other people know you share that belief and they were responsible for you having that belief.
Importantly, abduction, reductio ad absurdum, intuition, etc. do not have strong reasons to inform you of a false or improper belief. For example, if you form a belief on the basis of abduction and you missed an important defeater, abduction will never tell you that you missed it. The same is true for the other belief-formation methods.
Third, it seems that if a testifier asserts that p to an audience and that audience comes to believe that p because she takes the testifier’s word for it, then the testifier is required to answer the audience if she asks the testifier why p is true—at least in cases of moral assertion. For example, if a moral expert tells Katie that human-organ markets are bad and she defers to the expert on this matter, she seems to have a moral entitlement to be told by that moral expert why organ markets are bad. That is, the moral expert is obligated to answer Katie if she asks why organ markets are bad.Footnote 46 This strikes me as especially true when the audience encounters evidence against the proposition that the testifier asserted. For example, if Katie thinks of a possible counterexample to the claim that human-organ markets are bad, she is especially entitled to an explanation as to why the moral expert thinks organ markets are bad. Being able to tell one’s audience why one’s moral assertion is true is motivated by the thought that moral understanding is the goal of moral inquiry and moral understanding requires knowing the grounds of the proposition one understands. Thus, moral testimony seems to put the audience in a good epistemic position vis-à-vis the morally relevant properties of actions.
However, one does not have a right to be answered by abduction, reductio ad absurdum, intuition, and so on. They do not talk back in this way. In particular, your epistemic position vis-à-vis access to morally relevant properties is not made better by virtue of forming a belief on the basis of one of these other methods.
Thus, moral testimony improves one’s epistemic position vis-à-vis both morally relevant properties of actions and moral understanding. This is because those who rely on moral testimony have a moral entitlement to be told by the testifier why the testifier believes as they do. In virtue of being put in a better epistemic position vis-à-vis morally relevant properties and understanding, moral testimony puts one in a better position vis-à-vis all the moral achievements mentioned earlier. So, proponents of the Achievement Explanation must admit that moral testimony is superior to the other belief-formation methods mentioned above by their own lights.
5. Conclusion
I have argued that we have reason to reject the Achievement Explanation because it has implausible implications. First, it entails that using many of the methods people have for gaining or maintaining moral beliefs should be off-putting. But using these methods is not off-putting. Second, I argued that moral testimony is epistemically superior to the other methods I mentioned and puts one in a better position vis-à-vis gaining the relevant moral achievements. For proponents of the Achievement Explanation, moral testimony should actually be less off-putting than these other belief formations. But it is not. Thus, proponents of the Achievement Explanation face the New Puzzle of Moral Deference: they must explain why moral deference is so troubling, but all the other methods of moral belief formation I discussed are not—especially given that moral deference can sometimes put one in a better position vis-à-vis gaining the relevant moral achievements. This means that even if the Achievement Explanation is a satisfactory answer to the original puzzle of moral deference, accepting it comes at the cost of having to answer a more perplexing puzzle.
Acknowledgments.
For helpful discussions and suggestions, I would like to thank Robert Audi, Zach Barnett, Eilidh Beaton, Justin Bernstein, Justin Clarke-Doane, Jamie Fritz, Camil Golub, Elinor Mason, Sarah McGrath, Caleb Perl, Richard Rowland, Russ Shafer-Landau, Vishnu Sridharan, Michael Vazquez, Silvan Wittwer, two anonymous reviewers, and audiences at the 2017 Austin Graduate Ethics and Normativity Talks (AGENT) at UT Austin, “Knowing What to Do: A Workshop on Ethical Epistemology” at the Ohio State University, the 2018 Northwestern/Notre Dame Graduate Epistemology Conference at Northwestern University, the Australian Catholic University Conference on Moral Epistemology, “The Fourth Foundations of Normativity” workshop at the University of Edinburgh, and “The Future of Normativity” at the University of Kent. Special thanks to Tony Coady, Steven Gubka, Ross Jensen, Errol Lord, Lisa Miracchi, and Daniel J. Singer for written comments on an earlier draft of this paper.
Max Lewis is currently a PhD candidate at the University of Pennsylvania. He works primarily on metaethics, epistemology, philosophy of language, and their intersection.
