[I]t doesn't bother me if I happen to agree with the mainstream media [in criticizing the Soviet Union]. Trotsky … was charged in the 1930s with agreeing with the fascists in his condemnation of the Soviet Union. And he pointed out that his critique was true; he wasn't going to abandon it if somebody else happened to say it for different reasons. (Noam Chomsky (Reference Chomsky1989), in response to being challenged for agreeing with the mainstream media in condemnation of the Soviet Union)
[I]t's okay to know it's Mussolini. Look, Mussolini was Mussolini. It's okay to – it's a very good quote, it's a very interesting quote, and I know it. I saw it. I saw what – and I know who said it. But what difference does it make whether it's Mussolini or somebody else? It's certainly a very interesting quote. (Donald Trump (Reference Trump2016), in response to being challenged for tweeting a quote sometimes attributed to Mussolini, “It is better to live one day as a lion than 100 years as a sheep”)
On one occasion he was being praised by some wicked men, and said, ‘I am sadly afraid that I must have done some wicked thing.’ (Diogenes Laertius (Reference Laertius and Yonge1853: 219), on the philosopher Antisthenes)
1. The problem
Like Noam Chomsky and Donald Trump, we are typically unmoved – except by irritation – when someone lambastes us for agreeing with a source that is morally or intellectually suspect by our own lights. But accusations of guilt-by-epistemic-association are asymmetrically popular: we tend to issue them in contexts of disagreement, yet think they bear little relevance for the truth of our own views. This is a mistake. We sometimes face what I call the problem of unwelcome epistemic company. This is the problem of encountering agreement about the content of your belief from an unwelcome source. The problem of unwelcome epistemic company is scarcely explored by philosophers, yet it is a problem of everyday life.
In what follows, I elaborate four iterations of the problem that unwelcome epistemic company poses. To this end, I introduce and discuss some simple cases. I then offer preliminary lessons of the problem, and I canvass possible responses, ranging from the extreme confidence of Chomsky and Trump to an epistemic prudishness that always seeks to reduce agreement with unwelcome sources.
2. Iterations
In circumscribing the problem of unwelcome epistemic company we must appreciate that not all unwelcome epistemic company poses a problem – even a prima facie one. Consider the following case:
Blue Sky: You believe that the sky is blue. Then you find out that Osama bin Laden also believed that the sky is blue.
This case does not even threaten to pose a problem, either epistemic or moral. Although bin Laden might be unwelcome epistemic company of some sort, his company poses none of the problems that I describe below. In particular, bin Laden's agreement about the color of the sky raises no epistemic or moral issues about your own belief that the sky is blue, or the role of your moral character in coming to that belief.
Here is a case, on the other hand, where you might face any of at least four different problems due to unwelcome epistemic company:
Refugees: You haven't read much on refugee crime rates in the United States, but you pay some attention to the news. Despite being generally in favor of increased acceptance of refugees, you find yourself with the belief that refugees, on average, commit more crimes than citizens. You then encounter a white nationalist who also believes that refugees commit more crimes than citizens.
Two of the problems that may be raised by Refugees are distinctly epistemic, and two are moral. The epistemic problems have to do with your belief's truth or normative status, and the moral problems have to do with its etiology or implications. I will address these in order.
2.1. Epistemic problems
In Refugees, agreement may provide some defeasible evidence that your belief is false. In other words, the problem of unwelcome epistemic company might take the following general form:
Falsity: The fact that S believes that p is evidence that ~p.
This is a problem having to do with the truth of your belief. But how bad is the problem? The amount of evidence that unwelcome epistemic company provides in cases like Refugees partly depends on just how unreliable one has reason to believe S really is. In Refugees, this is left unstated. If the white nationalist just gets the relevant matters wrong frequently relative to the average person, but still very rarely, then perhaps you need not worry too much, and what degrees of error bother you will depend partly on your tolerance for epistemic risk.Footnote 1
But suppose that your epistemic company does provide a great deal of evidence that your belief is false – say, because they get relevant matters wrong well over 50% of the time. (After all, even agreement from a source that gets relevant matters right 51% of the time is some evidence that the belief in question is true.Footnote 2) In that case, agreement provides evidence that your belief is false.
We are only very rarely able to assess the actual, or even approximate, degree of unreliability of any particular person (or even ourselves), whether in general or in specific domains. In light of this, the possibility of frequent error need only be salient in order that Falsity might apply to a case of unwelcome agreement. Even if you cannot assign an exact or even vague probability to someone's error, as long as there is a salient possibility that they are more unreliable than not within the relevant domain, their unwelcome agreement in that domain can warrant concern.
All that being said, even if your unwelcome company is extremely unreliable, the normative force of the evidence provided by Falsity will partly depend on your independent evidence for the proposition subject to unwelcome agreement. If you already know that you came to believe via a route that secures more warrant for you than the warrant that your company secures for the negation of your belief, then you may not face much of a problem at all. For example, suppose you calculate that 35+15=50, and a mathematically incompetent person calculates the same. Even if this person gets simple arithmetic questions wrong 95% of the time, the evidence provided by their agreement is decisively defeated by your prior, independent evidence for the proposition. Your total evidence makes it likely that this is just one of those unlikely cases where an extremely unreliable person got a calculation right. Although unwelcome agreement that 35+15=50 is perhaps some evidence that 35+15≠50, such evidence is decisively defeated by your overwhelming evidence to the contrary.Footnote 3 This is not a situation in which you do not face the problem of Falsity at all; rather, it is a situation in which the problem of Falsity is just not very severe.
The qualification about independent evidence is why it is important to the formulation of Refugees that “you haven't read much on refugee crime rates in the United States.” If you have read a great deal on refugee crime rates, then it is likely that you have independent evidence to defeat the evidence of unwelcome agreement. The need for this qualification in order for the problem to seem more serious suggests that the problem of unwelcome epistemic company, at least in the form of Falsity, is faced more commonly by laypersons than by experts. (By “experts” I mean to include both credentialed experts and autodidacts.) This is because laypersons are much less likely than experts to have independent lines of evidence that defeat the sort of evidence that unwelcome agreement provides.
So, if the white nationalist's beliefs about crimes committed by refugees are likely false, or there is at least a salient possibility that they are likely false, then you face the problem of Falsity. But there are three other distinct problems you might face. The following, like Falsity, is an epistemic problem:
Malfunction: The fact that S believes that p provides evidence that you acquired your belief p via an epistemically faulty process.
Perhaps white nationalists are likely to hold their beliefs about refugees on the basis of or otherwise due to some faulty epistemic process – even when their beliefs are true. If so, then the fact that the white nationalist shares your belief about refugees provides some evidence that your belief-forming process is in some way bad – even if your belief is true. In this form, unwelcome epistemic company may give you evidence about the reasoning or other process behind your belief. Whereas Falsity provides you with an alethic defeater, Malfunction provides you with a rationality defeater.
It is worth emphasizing that Malfunction is broader than just a claim about your reasoning processes. Perhaps you arrived at the belief in question intuitively, via perception, or otherwise on the basis of an immediate, direct judgment. For these kinds of beliefs, unwelcome epistemic company may pose the problem of calling into question the truth-conduciveness of the relevant cognitive faculties or processes. Alongside reasoning, intuitive reactions are one of many processes for acquiring beliefs, and so both sorts of processes are covered by Malfunction.
In the case of Refugees, the application of Malfunction seems especially plausible. It may be that white nationalists tend to make inferences about refugees as a class on the basis of sensationalist anecdotes that are prominent in the news. This is not a good form of reasoning, but it may be that you, too – despite your generally favorable attitude toward refugees – are engaging in it. This sort of epistemic similarity is just the kind of problem that unwelcome agreement can help us to discover.
As with Falsity, the evidence provided by Malfunction is defeasible. In addition to overwhelming independent evidence for the truth of your belief, you might also have overwhelming independent evidence for the differences in the rational bases between your belief and the white nationalist's. The importance of this qualification is further evidence that the problem of unwelcome epistemic company is worse for laypersons than for experts, because experts are more likely than laypersons (though certainly not guaranteed) to already have a thorough accounting of both their and their opponents’ rational bases for their beliefs.
2.2. Moral problems
Beyond the epistemic defeaters identified by Falsity and Malfunction, Refugees potentially raises two moral problems. The first is a problem for your own moral behavior or character:
Vice: The fact that S believes that p provides evidence that your belief that p is connected in some way to a moral vice.
Whether or not it is false or based on faulty cognitive processes, your belief about refugees might be the sort of view that attracts people with a certain kind of moral vice, e.g., the vice of racism as manifested by white nationalism. When you discover your agreement with the white nationalist, a distinctly moral possibility is made salient to you. Perhaps you have a vicious character trait that attracts people to the belief in question, or are at least carrying on in a manner characteristic of this trait, and this is part of how you came to hold your belief. In this way, your unwelcome epistemic company acts as a mirror of the sort of person you might be or might be becoming – a sort of person that you do not want to be.
Figuring out to what degree your unwelcome company raises the likelihood of Vice is tricky. After all, it could be that racist tendencies make it very likely that you will believe that refugees commit more crimes, but that believing that refugees commit more crimes does not make it at all likely that you have racist tendencies. Presumably, something structurally similar to this possibility drives Chomsky's thoughts in the quotation that opens this paper, wherein Chomsky appeals to the fact that the mainstream media has “different reasons” from his. But notice that someone motivated to figure out which of these probabilistic relationships obtains has in a sense already conceded that they face Vice as a problem. After all, figuring out where one stands here is precisely a way of responding to the problem of unwelcome epistemic company in the form of Vice; it is not a way of showing that there was no problem in the first place. Another way to put the point is that one is not called on, absent some special reason, to check whether one's view makes it likely that one is vicious in a particular way. Rather, this is something that one might be inclined to do precisely in the face of a particular kind of unwelcome epistemic company. This paper argues that this kind of double-checking can be rational.
Like the epistemic problems, the evidence that Vice provides is defeasible. Even in the formulation of Refugees, the fact that you are “generally in favor of increased acceptance of refugees” is some evidence against your being bigoted toward them, though it is hardly decisive. Here, being an expert rather than a layperson is not as strong a protection against the evidence that unwelcome agreement provides. Expertise in refugee issues does not necessarily include expertise in or even reflection on one's own moral attitude toward refugees, and expertise in general is just not the same thing as morally critical self-reflection.
Vice is not the only moral problem raised by Refugees. The case also raises the following moral problem, not about your own moral character, but about the implications of the content of your belief:
Implication: The fact that S believes that p provides evidence that you missed something important about the stakes of p that is relevant to the endorsement of p.
Perhaps you do not, either implicitly or explicitly, share the vicious character trait that motivates your unwelcome company to agree with you. Nevertheless, it may be that they agree with you because they see something about the matter of agreement that you do not, and it is precisely something that appeals only to people who are vicious in a particular way, or something that makes this particular matter more important than you at first realized. In the case of crimes committed by refugees, perhaps what really attracts white nationalists, though it does not function as a reason for their belief, is that high crime rates among refugees in turn helps to justify restricting acceptance of refugees, and this is (let us suppose) something that you have never considered. Whether or not this new information will ultimately change your mind, it is reasonable to react to the unwelcome epistemic company of the white nationalist partly by worrying that such information might exist, and double-checking accordingly. Given that you are someone who thinks that the acceptance of refugees is morally important, the evidence provided by Implication raises the stakes of the matter. If you end up maintaining your belief, it will be in spite of, not regardless of, this new evidence. What the white nationalist may give you is some good evidence that there is something that makes careful inquiry into the question especially important.Footnote 4
Refugees is a case where you very well may face any of at least four different versions of the problem of unwelcome epistemic company, two of which are epistemic and two of which are moral. But it is important to recognize that these versions of the problem are related to each other in complicated ways. For example, consider the following case, in which one might argue that only the moral problems arise:
Selectivity: You and Fred are in the same academic department in 1930s Germany. Fred is known to be antisemitic and is meticulous when it comes to finding faults in his Jewish colleagues’ work. When it comes to everyone else, however, Fred just doesn't pay that much attention. Despite not thinking of yourself as an antisemite, you find that your own view of your Jewish colleagues matches Fred's, and that you also don't pay that much attention otherwise.
Selectivity is a case where your belief is at least threatened by Vice and Implication, but it is not obvious whether it is threatened by Falsity or Malfunction. While I do not think that it is in fact common that someone is rationally and accurately attuned to all and only the faults of one's Jewish colleagues, less theoretically ideal cases that resemble this one do exist in the real world, and so it is a useful possibility to explore.
First, what is wrong with Fred, such that his agreement poses a problem? Fred's particular beliefs about his Jewish colleagues do not in themselves – ex hypothesi – immediately reveal any distinctly epistemic problems, but in their wider context or pattern, they reveal serious moral problems, and these problems underwrite the application of Vice and Implication to the case. Vice applies to encountering Fred's epistemic company, because Fred's record of accuracy and rationality about his Jewish colleagues is plausibly due to a moral vice: antisemitism. It takes a peculiar sort of selective attention to identify all and only the faults of one's Jewish colleagues. And it is morally objectionable to focus one's epistemic energies on all and only crimes committed by members of that group, even if one does so in an epistemically excellent manner. So when you find yourself agreeing with Fred, you acquire evidence both that your belief is epistemically appropriate (as far as truth and evaluation of evidence is concerned) and that this is due to something morally vicious.
Implication may also apply to Selectivity, because it is highly plausible that expending selective energy on identifying and pursuing faults only among Jewish colleagues in Nazi Germany serves to prolong and intensify systemic oppression of Jewish people. Perhaps this is something that (either implicitly or explicitly) appeals to Fred. As with Vice, this is wholly consistent with you being within your purely epistemic rights in believing what you believe. The upshot is that, because there are moral constraints on epistemic practice (in this case, how we direct our critical attention), and violating these constraints is compatible with obeying purely epistemic constraints, it seems possible to face the moral but not the epistemic problems of unwelcome epistemic company.
However, things are not so simple. In particular, there are reasons for thinking that Vice and Implication are reducible to Malfunction. I will consider these in order. Vice may reduce to Malfunction, because the presence of a moral vice (e.g., antisemitism) in the etiology of a belief seems to provide some (defeasible) grounds for debunking that belief via a rationality-defeater. This may be the case, but it need not be. Notice that the case of Selectivity itself actually avoids such a debunking – and, therefore, avoids a reduction from Vice to Malfunction. How does it do this? The moral vice of antisemitism does not play a rationalizing role in the actual content of the matter of unwelcome agreement. Rather, the moral vice is merely the putative cause of Fred's (otherwise excellent) epistemic skills being directed as they are. Compare a purely moral case of someone in a dominant racial group who helps all and only people who share their racial identity, on the basis of antipathy toward those who do not. There is one sense in which a person's actions (helping so-and-so) are morally right, but another sense (helping only so-and-so, for bad reasons) in which they are wrong. In such a case, the (narrowly) right actions, in the relevant context, provide evidence for a moral vice. Likewise, in cases that resemble Selectivity, there is nothing wrong with the beliefs per se, epistemically speaking. But, in the wider social as well as motivational context, these beliefs provide evidence of a moral vice. In finding that we agree with such a person, we may worry that we share this vice.
Notice further that, even in cases where we can extract the problem of Malfunction from the problem of Vice (say, because a moral vice does play a rationalizing role for the belief) this is not a full reduction. If it turns out that your belief that p is the result of faulty reasoning in precisely the sense that it is based on epistemically irrelevant racist attitudes, you still face two distinct problems. True, in this kind of case, the moral problem (having racist attitudes) itself poses an epistemic problem (having irrational beliefs), but that does not mean that having racist attitudes fully reduces to having irrational beliefs. Another way to put the point: having immoral attitudes is a problem, and using them irrationally is another.
What about Implication? It might seem that Implication reduces to Malfunction, because to the extent that you miss the implications of the content of your belief, your reasoning is impoverished. However, as with Vice, this merely shows that Implication can pose an epistemic problem in addition to a moral problem, as when a bad implication of a belief is also a rational ground for rejecting it. A controversial instance of this phenomenon involves what is sometimes called “pragmatic encroachment.” Some philosophers have recently argued that the epistemic standards for holding a particular belief might be raised in cases where the purely practical stakes of that belief are likewise raised.Footnote 5 If this view is right, then in cases where the evidence of Implication is evidence of high stakes, it might give you reason to think your belief is insufficiently justified. This is one way in which a moral problem of unwelcome epistemic company can in turn pose an epistemic problem.
But this need not be the situation. After investigating a matter of unwelcome agreement, you may find both that there was nothing wrong with your prior reasoning and that you were non-culpably ignorant of some unseemly implication of the content of your belief that attracts people with a particular moral vice but does not raise the stakes for that belief.
Perhaps less surprisingly, you might face the epistemic but not the moral problems of unwelcome epistemic company. Consider the following case:
Calculator : You are calculating a complicated tip at a restaurant. You are aware that your companion has a peculiar calculator programmed to frequently but not always make errors in basic arithmetic that are very common among human beings. The calculator agrees with you.Footnote 6
Calculator is a case where your belief is threatened by Falsity and Malfunction, but neither Vice nor Implication. Epistemically speaking, the issues you face are similar to those in Refugees, and their severity likewise depends on the frequency of error and your tolerance of epistemic risk. But (unless the calculator is really peculiar) there are no grounds for worrying about your moral character or morally relevant implications of your calculation.
3. Responses
How should one respond to the various problems of unwelcome epistemic company? I have already described both the cases and the interpretations of the cases in ways that reveal the flavor of my own answer to this question, but now I want to step back. I have imagined and commended various forms of doubting, revising, and rechecking one's beliefs or character when one encounters either a case best interpreted by Falsity, Malfunction, Vice, or Implication. Am I right to commend these responses?
The general problem of unwelcome epistemic company is a problem having to do with what epistemologists sometimes call “higher-order evidence.” Unfortunately for my purposes, the term is given a variety of disparate definitions across the literature. The term “higher-order evidence” may refer to any combination of evidence about evidence;Footnote 7 evidence about evidentiary relations;Footnote 8 evidence about cognitive processes,Footnote 9 or just indirect evidence about a proposition or belief. Unwelcome epistemic company may provide higher-order evidence in any of these senses. For example, cases of unwelcome epistemic company that pose the problem of Falsity seem to provide a kind of indirect evidence about the specific content of agreement, whereas cases that pose Malfunction seem to provide evidence about evidence, the evaluation of evidence, cognitive function, and much else.
Because the problems of unwelcome epistemic company have to do with a certain kind of higher-order evidence, appropriate responses will likely resemble responses to other problems having to do with higher-order evidence, especially the problem of peer disagreement. For example, faced with a case of unwelcome epistemic company, you will oppose any sort of revision of belief or double-checking if you endorse
Extreme Steadfastness : In response to unwelcome epistemic company in believing that p, you should maintain your credence that p and resist any introspective moral concern on the basis of the unwelcome agreement.Footnote 10
According to Extreme Steadfastness, unwelcome epistemic company poses no real problem at all. Of course, even those who reject Extreme Steadfastness as a general response to higher-order evidence may, like Chomsky and Trump, embrace it when confronted with unwelcome epistemic company. And there is a wide array of cases of unwelcome epistemic company that are, like Blue Sky , typically (and correctly) treated as irrelevant. Vegetarians are generally not bothered by their agreement with Hitler; critics of American interventionism in the Middle East are generally not bothered by their agreement with members of ISIS; Mennonite theists are generally not bothered by their agreement about the existence of God with the Westboro Baptist Church; and so on. Cases like these anecdotally support the popular thought that guilt-by-epistemic-association is no more a good foundation of criticism than its more general counterpart. But clearly not all cases are alike. Refugees bears on one's belief in a way that agreement with Hitler about vegetarianism does not.
What really does the work in the cases where we justifiably reject guilt-by-epistemic-association is not the mere logical compatibility to which Chomsky and Trump appeal, but rather that we judge that the unwelcome features of one's epistemic company simply do not bear in a relevant way (or to the relevant degree) on the belief's truth, normative status, etiology, or implications – as in Blue Sky, above. Sometimes there can be a dispute about whether the problem of unwelcome epistemic company arises. For example, in the case of critics of American interventionism and members of ISIS, those who make accusations of guilt-by-epistemic-association may believe that some single, objectionable underlying characteristic motivates both parties. But rather than simply state the banality that A member of ISIS believes that p is compatible with p, one should instead interrogate which kind of unwelcome epistemic company ISIS is supposed to represent, and whether it really poses a problem in this case. After all, no one would even think to suggest that Blue Sky is a problem, so something other than logical compatibility must be behind the complaint.
At the other end of the spectrum that is anchored on one side by Extreme Steadfastness one finds:
Extreme Dissociationism : In response to unwelcome epistemic company in believing that p, you should always reduce your credence in p by n degree and/or engage in introspective moral concern.Footnote 11
According to Extreme Dissociationism, one should take unwelcome epistemic company as a decisive reason to question one's beliefs or moral character, and to do so in a highly specified way. Extreme Dissociationism is perhaps not common in self-application, but popular arguments that appeal to guilt-by-epistemic-association seem to presuppose something like it. Many people point out to critics of Israel that there are antisemites who also criticize Israel (indeed, with something like the structure of Selectivity in mind); in 2016 many pointed out to supporters of Bernie Sanders that some opponents of Hillary Clinton were sexist; and so on. It seems that Extreme Dissociationism, although perhaps not something we commonly apply to ourselves, is a popular tool in the hands of those assigning guilt-by-epistemic-association.
Like its steadfast counterpart, Extreme Dissociationism is far too strong. Unwelcome epistemic company (and higher-order evidence generally) really does sometimes pose a problem. But sometimes it does not, even a little bit, as Blue Sky shows clearly. The crucial step, again, is to determine whether the features that make the epistemic company unwelcome bear on a belief's truth, normative status, etiology, or implications. If we were to avoid unwelcome epistemic company at all costs, then we would systematically guarantee getting matters wrong that our company gets right.
I commend the far more moderate – and flexible – principle of
Moderate Dissociationism: Unwelcome agreement that p always provides some defeasible reason to doubt, revise, or recheck your belief that p. This reason comes in degrees of strength, depending on the case.Footnote 12
According to Moderate Dissociationism, Extreme Steadfastness is flat-out incorrect, and Extreme Dissociationism has only a kernel of truth in the most extreme cases. Unwelcome epistemic company is neither always nor never a problem, but it very often makes salient the possibility of a problem. This possibility should not be dismissed without ruling out that one is not guilty by epistemic association in one or more of the four ways. Some cases might be easy (e.g., Blue Sky), but some cases are hard (e.g., Refugees).Footnote 13
Moderate Dissociationism is general enough to accommodate the facts that unwelcome epistemic company does not have to do only with the truth of one's belief; it need not only require abandonment of belief; it may raise problems that are both moral and epistemic in nature; and it comes in varying degrees of strength.
4. Disagreement
The problem of unwelcome epistemic company bears an unmistakable resemblance to the problem of peer disagreement. Very roughly, the problem of peer disagreement is the problem posed by encountering disagreement about p from someone who you reasonably judge to be as epistemically well-positioned as you with respect to p. In both the cases of peer disagreement and unwelcome agreement, your belief that p can be challenged by the mere fact of another person's or group's attitude toward p, in light of your evaluation of them as agents. In the case of disagreement, the fact that someone believes ~p, together with the fact that the same person is an epistemic peer, puts pressure on your belief that p.Footnote 14 In the problem of unwelcome epistemic company, the fact that someone believes that p, together with the fact that the same person is either epistemically or morally compromised, puts pressure on your belief that p. As noted above, the possible responses to the problem are at least partly parallel, because both problems involve issues having to do with higher-order evidence.
Another connection between unwelcome epistemic company and peer disagreement, worthy of investigation in its own right, is the role of moral considerations in each problem. It is tempting to think that the special relevance of moral characteristics is unique to the problem of unwelcome epistemic company, because there is arguably no analogue to either Vice or Implication in the case of disagreement. It may initially seem plausible that mere disagreement with an epistemic peer about p does not give one reason to suspect that there is some moral flaw in oneself, or that there are moral aspects of a view that one has missed in any systematic way.
But this is much too quick. Just as unwelcome epistemic company might involve agreement from someone you take to be morally inferior in some relevant respect, so epistemic peer disagreement might involve disagreement from someone who you take to be morally superior in a relevant respect. For example, you might disagree with an epistemic peer about the best policies for alleviating poverty, but acknowledge that your peer, unlike you, actually performs admirable work among the poor. This sort of case could very well pose an analogue to Vice in the case of disagreement. The analogue to Vice is the worry that you disagree with your epistemic peer due to some vicious trait that you have which your peer lacks – e.g., insufficient love for those who are poor. Of course, this analogy requires that moral peer-hood is not a proper part of epistemic peer-hood.
There is not and cannot possibly be, however, an analogue for Implication in the case of peer disagreement. If you are subject to Implication, then by necessity you do not stand in a relation of epistemic peer disagreement. After all, the analogue of Implication would say that you endorse your view only because you, but not your interlocutor, are missing something important about the matter of disagreement. But if that is the case, then you and your interlocutor are not epistemic peers.
The role of epistemic peer-hood itself provides a dissimilarity between the problems of disagreement and unwelcome epistemic company. Generally, an unstated assumption in the literature on disagreement is that disagreement with someone more expert than you poses an obvious problem for your belief, without the various puzzles attendant to the problem of peer disagreement. Supposing you are a layperson, we might say that encountering expert belief contrary to your own does nothing but intensify your epistemic problems.
With unwelcome agreement, the issue is trickier. Here I will say something brief, though I think that this issue is worthy of further exploration in its own right. Unwanted expert agreement arguably does not generally intensify the epistemic problems of Falsity or Malfunction, provided that expert beliefs are more likely than lay beliefs to be true and justified. Perhaps there are domains in which expert opinion is systematically epistemically corrupted, in which cases the condition wouldn't hold. But I do not think that this will be the majority of cases. That being said, unwanted expert agreement may often intensify the moral problems of Vice and Implication, provided that experts are more likely than laypeople to have beliefs that are well-integrated with their moral vices, or are more likely to be aware of the implications of their beliefs. I do not know how often these two contrasts between expert and lay opinion hold, but when they do, the relevance of expertise to the problem of unwelcome epistemic company will be asymmetrical in the way described.Footnote 15
Finally, the problem of unwelcome epistemic company, although I think it is an important and neglected problem, nevertheless does not seem quite as worrisome as the problem of peer disagreement. For example, it seems that the ordinary Christian theist should be more worried about disagreement with atheist, Islamic, and other peers, than about unwelcome agreement with religious extremists. Why might this be? Two related reasons suggest themselves. First, for any view that we think we correctly hold, there is some reason to think that it is prima facie likely that at least some people who are bad at reasoning, or bad in some other way, will happen to hold the view. There are not good grounds for expecting that no one will endorse the right views for the wrong reasons. We expect to find unwelcome epistemic company, and so it is unsurprising when we do. In other words, we are not very surprised by certain kinds of epistemic luck, where people stumble upon the truth in epistemically or morally degenerate ways. On the other hand, for any view we correctly hold, it is prima facie likely that people who are as good at reasoning and epistemically well-positioned as us will hold our view, and unlikely that they will not.Footnote 16 That is to say, we do not expect (to the same degree) to find peer disagreement, and so it is surprising when we do. We are at least somewhat surprised by certain kinds of competent failure.
Of course, we can still ask relevant questions: Why is competent failure more surprising than lucky success? I do not have a full answer to this question (which deserves an inquiry in its own right), but I suspect that the answer has to do with the way in which we expect reality to regulate our beliefs. An epistemic agent who is functioning well (or who is rational, or… pick whatever epistemic desideratum you like) should, to that extent, track reality. Properly functioning perception of a tree, for example, should yield the belief that there is a tree. It would be extremely surprising if it did not. Another way to put the point: the failure of a properly functioning perceiver to acquire the appropriate perceptual beliefs demands explanation. On the other hand, malfunctioning cognitive faculties are not necessarily tracking anything at all; or, in realistic cases, such faculties are only partially tracking reality. So, while it is not wholly unsurprising when a malfunctioning perceiver acquires a correct perceptual belief, it is certainly not very surprising. Except in sufficiently extreme circumstances (especially when a malfunctioning perceiver is more likely than not to get matters wrong), the lucky success of a malfunctioning perceiver does not demand explanation.
5. Conclusion
The problem of unwelcome epistemic company is, as far as I know, unexplored by philosophers, yet it is a problem of everyday life. This is unsurprising, because accusations of guilt-by-epistemic-association are asymmetrically popular: we tend to endorse them for people whose views we do not like and reject them in principle for views we do like. I think that the rejection is a mistake. There are relatively precise conditions in which different versions of the problem of epistemic company arise, and where different attendant responses are justified. By paying better attention to these conditions, we can achieve better knowledge about not only other people and the truth generally, but about our own beliefs and moral character.Footnote 17
