Epistemic Self-Trust and Doxastic Disagreements

Abstract

The recent literature on the epistemology of disagreement focuses on the rational response question: how are you rationally required to respond to a doxastic disagreement with someone, especially with someone you take to be your epistemic peer? A doxastic disagreement with someone also confronts you with a slightly different question. This question, call it the epistemic trust question, is: how much should you trust our own epistemic faculties relative to the epistemic faculties of others? Answering the epistemic trust question is important for the epistemology of disagreement because it sheds light on the rational response question. My main aim in this paper is to argue—against recent attempts to show otherwise—that epistemic self-trust does not provide a reason for remaining steadfast in doxastic disagreements with others.

1 Introduction

How much weight should you give to the beliefs of others, especially to the beliefs of those who disagree with you? It may often be tempting to ignore the beliefs of others and remain steadfast in response to disagreements. But would such an attitude be justified? Or would it be more rational for you to take the beliefs of others into account, including the beliefs of those who disagree with you?

The recent literature on the epistemology of disagreement engages with this important issue in our everyday epistemic lives, albeit with a somewhat narrower focus. The key question in this literature tends to be: how are you rationally required to respond to a doxastic disagreement with someone you take to be your epistemic peer, i.e. with someone who has access to the same evidence and who you expect to be as reliable as yourself in judging that evidence? Conformists argue that you are rationally required to reduce confidence in your original belief and perhaps even to suspend belief altogether.¹ Non-conformists tend to focus on epistemic reasons for maintaining confidence in your original belief, even if they admit that some adjustment is sometimes rationally required.²

A doxastic disagreement with someone, especially with someone you take to be your epistemic peer, also confronts you with a slightly different question. This question, call it the epistemic trust question, is: how much should you trust our own epistemic faculties relative to the epistemic faculties of others? While the epistemic trust question has not received quite as much attention in the debate on the epistemology of disagreement as the rational response question, answering the former is important for the epistemology of disagreement because it sheds light on the latter.

To illustrate, suppose it turns out that, at least prima facie, you should not trust your own epistemic faculties more than those of others, and perhaps even that you should trust the epistemic faculties of others as much as you trust your own. Call this type of answers to the epistemic trust question the Symmetry View. If the Symmetry View is correct, it is less likely that ignoring the fact of a doxastic disagreement with someone else and remaining steadfast will be the rational response. The Symmetry View may thus lend support to Conformist intuitions.

Vice versa, suppose it turns out that, again at least prima facie, you should trust your own epistemic faculties more than those of others. Call this answer to the epistemic trust question the Asymmetry View. The Asymmetry View can lend support to Non-conformist intuitions. Indeed, a number of recent contributions to the literature on the epistemology of disagreements have made use of some version of the Asymmetry View to explain why it is rational to retain confidence in your original belief when you face a doxastic disagreement with someone else.

They have argued either that the view implies that it is less rational for you to consider others as peers than is generally assumed (Schafer 2015) or, when you are facing a doxastic disagreement with a peer, that it explains why there are less pressures to conciliate than Conformists claim (Wedgwood 2007, 2010; Enoch 2010; Pasnau 2015).

My aim in this paper is to argue against the Asymmetry View and thus to undermine the support for steadfastness that it provides. My argument will focus on trust in one’s own epistemic faculties and the question of what justifies such epistemic self-trust. Epistemic self-trust is the three-place relation that obtains between a person, her epistemic faculties—faculties related to rational inquiry—and truth. It is the positive attitude that we take to relying on our epistemic faculties to get truth when forming and evaluating our beliefs. I will defend an understanding of epistemic self-trust that involves a leap of epistemic faith (Foley 2001: 173) against one that generates its own epistemic justification. I will also show that while epistemic self-trust is often justified on non-epistemic grounds, this justification only obtains as long as there are no epistemic reasons to believe that we may have trusted our epistemic faculties too much. Based on this understanding of epistemic self-trust, I will argue that doxastic disagreements with others, especially with epistemic peers, are one important source of epistemic reasons that should lead us to reconsider the epistemic trust we place in ourselves and, therefore, the beliefs we have formed by relying on our epistemic faculties.

2 The Epistemic Asymmetry View and Steadfastness

I will consider three main lines of argument in support of steadfast intuitions that build on the Asymmetry View. The first two, to be introduced in this section, rely on an epistemic version of the Asymmetry View. The Epistemic Asymmetry View holds that it is generally epistemically rational to trust one’s own epistemic faculties more than those of others. Stated more precisely, the Epistemic Asymmetry View holds the following:

Epistemic Asymmetry View: absent specific epistemic reasons to believe that others are more reliable than you, it is epistemically rational to trust your own epistemic faculties more than those of others.

When there are specific epistemic reasons to trust the epistemic faculties of others more than your own, you should respond to those reasons and lower confidence in your original beliefs or even defer to the opinions of others. The scope of the Epistemic Asymmetry View is restricted to cases where there are no reasons to believe that others are more reliable than you. In those circumstances, it claims, it is epistemically rational to trust your own epistemic faculties more than those of others.

The Epistemic Asymmetry View can lend support to Conformist intuitions in two ways. A first argumentative strategy in support of steadfastness deploys the Epistemic Asymmetry View to argue that when you face a doxastic disagreement with an epistemic peer, it is epistemically rational for you to give more weight to your own belief. Ralph Wedgwood (2007, 2010) has developed an argument of this sort. His argument focuses on the case of moral disagreements, but it does not hinge on the distinctiveness of the moral case. He writes (2010: 244):

It seems plausible that there is a sort of rational asymmetry between one’s own moral intuitions and the intuitions of other people: it is rational to have a special sort of ‘fundamental trust’ in one’s own intuitions, but it is not even possible to have the same sort of ‘trust’ in the intuitions of others.

On Wedgwood’s view epistemic self-trust is epistemically basic in a way that trusting the epistemic faculties of others is not. While you rely directly on your own epistemic faculties in forming beliefs, you cannot form a belief by relying directly on the epistemic faculties of others. There is thus an asymmetry between epistemic self-trust and trust in the epistemic faculties of others that makes it epistemically rational to trust your own epistemic faculties more than those of others. In a doxastic disagreement with a peer, this asymmetry supports giving more weight to your own belief.

David Enoch (2010), similarly, takes the basicness of epistemic self-trust to justify giving more weight to your own beliefs in disagreements with an epistemic peer, but his argumentative strategy is slightly different from Wedgwood’s. The starting-point of Enoch’s argument is the claim that your own perspective is inelminable from an assessment of how reliable others are. As he puts it: you cannot rationally treat yourself merely as one “truthometer” among many.³ This asymmetry in the relation between you and your own epistemic faculties and those of others has implications for how you should respond to a disagreement with an epistemic peer. Enoch argues, against the Conformist, that the disagreement itself is a reason to downgrade the reliability of someone you thought you had reason to take to be your peer; it is not a reason to reduce confidence in your own belief. And this is so, not because you are entitled to give extra weight to your belief because it is yours, but because trust in your own epistemic faculties entitles you to take your belief as true and to take the disagreement itself as evidence that you should trust the other person less.⁴

Building on this last thought, the Epistemic Asymmetry View can also be deployed in a second type of argument in defence of steadfast intuitions. It can be used to argue that the normal case is one where we should regard others as less reliable than ourselves and, therefore, not as our epistemic peers. The point here is not so much to present a challenge to the Conformist view of how you are rationally required to respond when facing a doxastic disagreement with an epistemic peer. Instead, the point is to support Non-conformist intuitions by arguing that genuine peer disagreements, situations in which we may have epistemic reason to conciliate with each other, are much more rare than is commonly thought.

Karl Schafer (2015) pursues this strategy.⁵ He starts by noting that we may have all sort of epistemic reasons to treat others as either more or less reliable than ourselves. When we have reasons to believe that the other is more reliable, for example, then epistemic rationality requires that we defer to her views. Disregarding such special epistemic reasons to regard others as either more or less reliable than ourselves, however, there is the question of how reliable we should consider ourselves compared to others—I have called this the epistemic trust question. Schafer argues that, absent special reasons to believe otherwise, we should consider ourselves more reliable than others.

Schafer’s argument also starts from the basicness of epistemic self-trust. We rely on our epistemic faculties in forming our beliefs and we rely on them even when we consider the reliability of others. Schafer interprets the basicness of epistemic self-trust to imply an entitlement: we are entitled to rely on our epistemic faculties in forming our beliefs and we make use of this entitlement when we consider the reliability of others.⁶ Schafer argues that this entitlement implies that, absent special reasons for a belief about how reliable others are, how much others agree with us becomes the main evidence for how reliable they are: if they agree with us, this is evidence that we should trust their epistemic faculties as much as our own and if they disagree with us, this is evidence that we should trust their epistemic faculties less. This also implies, however, that unless we have special epistemic reasons to think that others are more reliable than us, we should normally regard others as less reliable than ourselves and give less weight to their beliefs than to our own. That is the Epistemic Asymmetry View and in Schafer’s argument, as in Enoch’s, this view is a consequence of the evidential situation that you find yourself in when assessing the reliability of others that is created by the basicness, or ineliminability, of epistemic self-trust. But whereas Enoch deploys the Epistemic Asymmetry View to make a point about disagreements with epistemic peers, Schafer argues that this asymmetry makes it epistemically rational not to think of others as our epistemic peers unless we have special reasons to do so.⁷

3 Against the Epistemic Asymmetry View

There is a problem with both ways of deploying the Epistemic Asymmetry View to defend steadfast intuitions, however. The problem arises from the Epistemic Asymmetry View itself, or, to be more precise, from the way in which the fact that epistemic self-trust is basic is thought to support the Epistemic Asymmetry View and steadfast intuitions.

I take it to be uncontroversial that epistemic self-trust is basic in a way that trust in the epistemic faculties of others is not. Without trust in our faculties, we would not hold many of the beliefs that we do hold. And while we might consider the opinion of others when we form our beliefs, in doing so, we will still rely on our own epistemic faculties. Trust in our own epistemic faculties is thus basic for the formation and evaluation of many of our beliefs in a way that our epistemic trust in others is not. This shows that there is an asymmetry in how epistemic self-trust and trust in the epistemic faculties of others are grounded—trust in our own epistemic faculties is basic, while trust in the epistemic faculties of others is derivative.

We should not conflate this asymmetry in grounding with asymmetry in rational trust in our own epistemic faculties relative to those of others, however, and not assume that the basicness of epistemic self-trust directly supports the Epistemic Asymmetry View, therefore. The grounding claim is, in the first instance, a descriptive claim—it is a claim about how epistemic self-trust is involved in how we form many of our beliefs and in how we come to trust the epistemic faculties of others. The Epistemic Asymmetry View, by contrast, makes a normative claim—a claim about the epistemic rationality of trusting our own epistemic faculties more than those of others. Even if we accept that we rely on our epistemic faculties all the time and that the way in which we rely on our epistemic faculties is more basic than the way in which we rely on the epistemic faculties of others, we need not accept the claim that it is epistemically rational to trust our own epistemic faculties more than those of others without further argument.⁸

It might be thought that the indirect argument in defense of the Epistemic Asymmetry View that both Enoch and Schafer offer is of the right kind. They argue that the asymmetry in grounding has implications for the evidence that you have for how reliable others are. While the basicness of epistemic self-trust gives rise to an entitlement to treat your epistemic faculties as reliable, how much you trust the epistemic faculties of others depends on the evidence you have for their reliability. Absent specific epistemic reasons to believe that others are either more or less reliable than you, how much others agree or disagree with you is the only evidence you have for their reliability. If they agree with you, this is evidence that they might be equally reliable as you. And if they disagree with you, this is evidence that they are less reliable than you are. This sort of argument would, if successful, offer support for the Epistemic Asymmetry View—for the claim that, absent specific epistemic reasons to believe that others are more reliable than you are, it is epistemically rational to trust your own epistemic faculties more than those of others, especially of those who disagree with you.⁹

But the argument is not successful. The problem concerns the normative significance of the basicness of epistemic self-trust. Earlier, I have argued that we should not conflate the descriptive grounding claim and the normative claim that is at the core of the Epistemic Asymmetry View. It is true, however, that the basicness of epistemic self-trust supports certain normative claims. I do not want to deny that the basicness of epistemic self-trust is of normative significance. But the normative significance of the basicness of epistemic self-trust does not support the Epistemic Asymmetry View.

The normative significance of the basicness of self-trust is a case of ‘ought implies can’. The requirements of epistemic rationality cannot imply that the way we generally form our beliefs is epistemically impermissible. We rely on our epistemic faculties in forming most of our beliefs, and we rely on them even without a positive epistemic reason to believe that they are reliable. Restricting the permissible reliance on our epistemic faculties to cases where we have reason to believe that they are reliable would make most of our belief formation impossible and this undermines the plausibility of such a requirement. We should thus not expect the rationality of epistemic self-trust to depend on the presence of epistemic reasons that show that our epistemic faculties are reliable. Vice versa, epistemic self-trust must generally be epistemically rational even in the absence of positive epistemic reasons to believe that our epistemic faculties are reliable.¹⁰

It is thus true that the basicness of epistemic self-trust has some normative implications. But the ‘ought implies can’ argument only shows that we are entitled to trust our epistemic faculties even in the absence of epistemic reasons to believe that our faculties are reliable. It does not show that there is an epistemic reason to believe that our epistemic faculties are reliable and that suspending belief about the reliability of your epistemic faculties would be epistemically irrational. The entitlement to trust your epistemic faculties—grounded in their basicness—thus does not entail an entitlement to believe that your epistemic faculties are reliable.¹¹

Given that the entitlement to rely on our epistemic faculties does not entail an entitlement to believe that our epistemic faculties are reliable, the rest of the argument in support of the Epistemic Asymmetry View, and thus of steadfast intuitions, is also in trouble. If the basicness of epistemic self-trust does not imply an epistemic reason to believe that our epistemic faculties are reliable, it also cannot support the comparative claim that is at the heart of the Epistemic Asymmetry View—that we are entitled to trust our own epistemic faculties more than those of others, at least as long as there are no specific epistemic reasons for believing that others are more reliable than us.

What would support the comparative claim is the following thought. If we are entitled to believe that our epistemic faculties are reliable, then the reliability of our beliefs gives us reason to treat our beliefs as true (or likely to be true) and to measure the reliability of others on the basis of how much they agree with us. It is reasoning along those lines that justifies discounting the beliefs of those who disagree with us. If the entitlement to rely on our epistemic faculties does not entail an entitlement to believe that our faculties are reliable, by contrast, then it is not true that the question of how much we should trust the epistemic faculties of others reduces to how much they agree or disagree with us. Instead, we also need to consider the possibility that a disagreement is the result of us trusting our own epistemic faculties too much.

The point can be illustrated with the help of Schafer’s list example. Schafer draws an analogy between the problem of how much you should trust the epistemic faculties of others and the following thought experiment (Schafer 2015: 31f):

Suppose I have a list of 100 yes/no questions and answers to each. And suppose that I know that each of these answers is 95% likely to be correct. Now suppose I am given a second list with the same 100 questions and answers to each, where these answers agree with the answers in the original list in the first 95 instances and disagree with them in the remaining 5.

Schafer argues that given no prior information about the reliability of the second list, the only evidence you have for assessing the reliability of the second list is how much the answers on the second list agree with the answers on the first list. And while each question on which the two lists agree is evidence that the second list is as reliable as the first (95%), disagreement is evidence that the second list is less reliable. Given that there is less than full agreement, Schafer argues, you should thus conclude that the second list is less reliable than the first.

This argument fails to engage with the possibility that the second list is 100% reliable. And given that the reliability of the first list is only 95%, we should at least consider whether the disagreement between the two lists is best explained as a result of the limited reliability of the first list, not as a result of the greater error-proneness of the second list. To see the point more clearly, consider the case where you take the reliability of the first list to be only 80%. As this implies that you should expect the first list to yield the wrong answer for 20% of the questions, when you notice a disagreement between the two lists, you should at least consider the possibility that the second list is more reliable than the first.¹²

My point is that your agreement-based evidence that the second list is less reliable than the first list might be misleading and the potential for it to be misleading is greater the less reliable the first list is. Limiting yourself to agreement-based evidence only makes sense if you have good reasons to take the first list’s answers to the questions to be true or likely to be true. If the possibility that the first list’s answers may be wrong is taken into account, limiting yourself to agreement-based evidence is, at best, incomplete.

In the epistemic self-trust case, what you should believe about the reliability of your own epistemic faculties is, as I have argued, an open question. And without an entitlement to believe that our epistemic faculties are reliable, this sort of defense of the Epistemic Asymmetry View collapses and so do arguments in support of steadfast intuitions that build on this view.¹³ I will further flesh out this point in the last section. But before getting to that, it will be useful to consider an alternative way in which the Asymmetry View might be deployed in support of Non-conformist intuitions first. This alternative way rests on a non-epistemic version of the Asymmetry View.

4 The Non-Epistemic Asymmetry View and Steadfastness

The alternative argument aims to show that although it may be epistemically rational to conciliate in a disagreement with an epistemic peer, there are other, non-epistemic reasons for trusting your epistemic faculties more than those of others and for remaining steadfast in a disagreement with a peer. The non-epistemic version of the Asymmetry View on which this argument relies may be characterized as follows:

Non-Epistemic Asymmetry View: there are non-epistemic reasons that make it rational to trust your own epistemic faculties more than those of others.

Robert Pasnau (2015) pursues this strategy in support of Non-conformist intuitions. Like Schafer, he does not question the Conformist’s claim that a disagreement with a peer gives you an epistemic reason to adjust your belief. Pasnau accepts that there are no epistemic reasons to trust your own epistemic faculties more in these circumstances. Instead of arguing against Conformism directly, Pasnau argues that trust in one’s epistemic faculties may legitimately pull one in the opposing direction: while there are epistemic reasons to conciliate, there are other, non-epistemic, reasons grounded in epistemic self-trust that support remaining steadfast.

Pasnau understands epistemic self-trust as a doxastic value. Epistemic self-trust, as already discussed, plays an essential role in our belief formation—we come to hold most of our beliefs in virtue of trusting our epistemic faculties and there is value in that. This is not to say, however, that trust in your epistemic faculties gives you an epistemic reason for the beliefs you have formed in virtue of that trust. Instead, we should recognize that honoring the value of epistemic self-trust gives you a different type of reason for sticking to your guns.

Understanding epistemic self-trust as a doxastic value has implications for how you should respond to a disagreement with an epistemic peer. Pasnau distinguishes between a strong and a weak hypothesis about the significance of the doxastic value of epistemic self-trust in this regard. The strong hypothesis is that self-trust requires giving more weight to one’s own belief than to the belief of one’s peer. The weak hypothesis is that self-trust merely licenses giving more weight to one’s own belief. Pasnau aims to defend the latter (2015: 2323):

[T]he doxastic value of self-trust has sufficient weight to make it the case that it would not be wrong, at least in some cases of peer disagreement, to give significantly less weight to the views of one’s epistemic peer.

On Pasnau’s view, while there may be epistemic reasons to give considerable weight to the belief of the peer you disagree with, there is a trust-related reason to give more weight to your own belief in cases of peer disagreements. In the context of epistemic peer disagreements, Pasnau argues that the normative significance of self-trust is such that, while not making it epistemically rational to give more weight to one’s own belief relative to the belief of your peer, all things considered it may “not be wrong” to do so.

I agree with Pasnau that it is promising to understand epistemic self-trust as a value that sits on the outside of the framework of epistemic rationality. One advantage of this way of thinking about epistemic self-trust is that it avoids the problems that I have identified with the Epistemic Asymmetry View. It gives us a better handle on why accepting that epistemic self-trust is basic and common does not imply that it gives us an epistemic reason to trust our epistemic faculties more than those of others.

But the alternative view also gives rise to some questions. What does it mean to say that one is licensed to give more weight to one’s own belief than to that of one’s peer? Pasnau makes it quite clear that he does not just mean that there is an explanation, based on self-trust, for why one might be inclined to do so. He means something normative. But as he denies that there is an epistemic reason for trusting one’s epistemic faculties more than those of others, what sort of reason would that be? The sense in which it might not be wrong to place more weight on one’s own belief requires further clarification. In addition, there is the question of whether understanding epistemic self-trust as a doxastic value can provide resources for resisting pressures to conciliate in doxastic disagreements with a peer.

5 The Practical View of Epistemic Self-Trust and the Rationality of Self-Doubt

My aim in this final section of the paper is to both clarify the nature of the reasons that can justify epistemic self-trust and to argue—against Pasnau—that a doxastic disagreement with a peer puts pressure on originally justified epistemic self-trust. I shall develop and defend an account of epistemic self-trust that builds on Pasnau’s insight that epistemic self-trust may be justified on non-epistemic grounds. As we will see, however, this account, which I take to be independently plausible, does not support Pasnau’s non-epistemic version of the Asymmetry View.

I call the account of epistemic self-trust that I will put forward in this section a practical view. On this view, there are practical reasons that can justify epistemic self-trust in the absence of epistemic reasons for believing that our epistemic faculties are reliable. Because the practical view does not involve a belief in the reliability of our epistemic faculties, it contrasts with what we may call the epistemic view of epistemic self-trust. According to the latter, epistemic self-trust involves a belief that our epistemic faculties are reliable, or at least more reliable than those of others.

I will argue that the practical view implies that the justification for epistemic self-trust is undermined when there are epistemic reasons to believe that our epistemic faculties have not been reliable. This account of epistemic self-trust and its justification will allow me to show that epistemic self-trust does not license us, on non-epistemic grounds, to remain steadfast in a doxastic disagreement with an epistemic peer and to explain why epistemic self-trust tends to come under pressure in such disagreements.

In developing the practical view of epistemic self-trust, it will be helpful to consider trust in general, first. I take trust in general to be a three-place relation—as an attitude that we take towards someone, perhaps oneself, in some respect. When we trust a person in some respect, we rely on her to do what we have trusted her to do and take a positive stance towards that reliance.¹⁴ On an increasingly common view, one that I adopt here, this positive stance need not involve a belief that the person will do what we have trusted her to do (Holton 1994; Faulkner 2007, 2011; McGeer 2008; Marušić 2015). There is a difference between relying on someone in the sense that one expects—predicts—them to do something and relying on someone in the trust-sense. While the former involves a belief, the latter type of reliance need not involve a belief. I can trust you to phi, say to remember our agreement, without believing that you will phi. Trust, on this understanding, is thus a practical stance. As Holton (1994: 67) describes this understanding of trust:

Trusting someone does not involve relying on them and having some belief about them: a belief, perhaps, that they are trustworthy. What it involves is relying on them to do something, and investing that reliance with a certain attitude. This is to take a practical stance.

Some draw a sharp distinction between trusting someone to phi and believing that someone will phi. Berislav Marušić (2015: 180), for example, argues that we misrepresent the phenomenon of trusting someone to phi if we interpret it as entailing a belief that the person will phi. On this view, when we have a belief that a person will phi, trust is not involved. This may be too strong a claim, but I cannot pursue this here.¹⁵ My argument will only rely on the weaker claim that trust does not entail, or typically involve, a belief, allowing that sometimes when we rely on someone in the trust-sense we also hold a belief that they are reliable.

Epistemic self-trust, similarly, is a practical stance, or so I want to claim. It need not involve a belief that our epistemic faculties are reliable. Instead, it is better seen as the (epistemically) “ungrounded reliance on our epistemic faculties” (Fricker 2015) and the positive attitude that we take towards this reliance. The positive attitude that we take towards this reliance plays a role not only in doxastic contexts—when we form and evaluate our beliefs—but also in practical contexts—when we act on the basis of those beliefs.¹⁶

I argued earlier that the basicness of epistemic self-trust entails an entitlement to take our epistemic faculties as reliable, even in the absence of positive epistemic reasons for belief that they are reliable. The practical view that I am developing here helps us see more clearly why the entitlement to rely on our epistemic faculties should not be equated with an entitlement to believe that our epistemic faculties are reliable. Instead, we are entitled to rely on our epistemic faculties even if doing so involves an epistemic leap of faith (Foley 2001: 173).

At first, the idea of epistemic self-trust as a practical stance may seem puzzling. If epistemic self-trust is a positive attitude that we take towards the reliability of our epistemic faculties to get truth, then, surely, it must be possible to epistemically evaluate this attitude? That is correct, but it does not imply that epistemic self-trust is an epistemic attitude, not a practical stance, or that epistemic self-trust necessarily, or even typically, involves a belief that our faculties are reliable. We may rely on our epistemic faculties in the trust sense without having given much thought to their reliability. And we often trust our epistemic faculties even though we are aware, or could become aware, that we do not have sufficient epistemic reasons to believe that our epistemic faculties are reliable.¹⁷

Understanding epistemic self-trust as a practical stance does not preclude the epistemic evaluation of epistemic self-trust, however. That is because epistemic self-trust should be interpreted as an epistemically focused attitude (Cassam in press). Like other epistemically focused attitudes—think of intellectual diligence, for example—epistemic self-trust is an attitude that we take towards epistemic goals such as knowledge and epistemic procedures such as rational inquiry. And insofar as trusting our epistemic faculties to get us the truth amounts to taking a positive stance towards epistemic goals such as knowledge and epistemic procedures such as rational inquiry, it is possible, and indeed appropriate, to ask whether our epistemic self-trust is conducive to reaching knowledge and engaging in rational inquiry or not. As an epistemically focused attitude, epistemic self-trust is thus subject to epistemic evaluation.

This understanding of epistemic self-trust as a practical stance that is subject to epistemic evaluation has the following implications for the justification of epistemic self-trust.¹⁸ First, epistemic reasons limit justified epistemic self-trust: epistemic self-trust is not justified when there are epistemic reasons to believe that our reliance on our epistemic faculties is epistemically unwarranted. As already argued above, justified epistemic self-trust must be possible in the absence of epistemic reasons to believe that our epistemic faculties are reliable, but this does not entail that such trust is justified on epistemic grounds. Epistemic self-trust does not generate its own epistemic justification and justified epistemic self-trust is subject to epistemic constraints.

The point is important, not least, because empirical evidence shows that we tend to trust our epistemic faculties more than we should, i.e. even in cases where there are epistemic reasons that suggest that our epistemic faculties are not reliable.¹⁹ Our tendency to trust our epistemic faculties, while doxastically and practically valuable in the absence of epistemic reasons that show that our epistemic faculties are not reliable, can thus come into conflict with the demands of epistemic rationality. When we become aware of evidence that suggests that we may have trusted our epistemic faculties too much, we should reconsider the belief in question—we should reduce our confidence in this belief or perhaps even suspend belief. In these cases, epistemic self-doubt becomes epistemically rational. So, while it is epistemically permissible to trust our epistemic faculties in the absence of good evidence that they are reliable, we cannot permissibly rely on our epistemic faculties “against the evidence”.²⁰

When we have no epistemic reasons to believe that our epistemic faculties are unreliable, however, then trust in our epistemic faculties may be justified on non-epistemic grounds and this is the second dimension of justified epistemic self-trust. I want to claim that both doxastic and practical considerations are relevant for the justification of epistemic self-trust in those circumstances. We rely on our epistemic faculties not just when we form and evaluate beliefs for the sake of figuring out what to believe, but also when we try to figure out how to act. The doxastic value of epistemic self-trust is thus not limited to its role in role in forming and evaluating beliefs. The doxastic value of epistemic self-trust also, and importantly, relates to its role in rational action. At least if the cognitive model of action is true, rational action is not possible without a connection between our actions and our beliefs. Forming beliefs about our practical context, evaluating those beliefs, and settling our beliefs in light of possible conflicts are all important for being able to act in rational fashion. Epistemic self-trust is one factor among others that enables us to be practically rational and to respond to practical reasons.

To be sure, practical considerations cannot justify epistemic self-trust in circumstances where epistemic self-trust is epistemically unwarranted. Since epistemic self-trust is the reliance on our epistemic faculties to get truth, practical considerations cannot trump epistemic reasons we have for believing that our epistemic faculties have malfunctioned. If there were allowed to trump epistemic considerations, the very essence of epistemic self-trust would be lost. Practical reasons can thus only provide a justification for epistemically permissible self-trust. If there are no epistemic reasons that render epistemic self-trust impermissible, then practical reasons can justify our epistemically ungrounded reliance on our epistemic faculties. Trust in our epistemic faculties is warranted when it is an appropriate practical response to the fact that there are insufficient reasons to believe that our faculties are reliable, but not when we have epistemic reasons to believe that our faculties are unreliable.²¹

This practical view of epistemic self-trust puts us in a position to evaluate Pasnau’s claim that epistemic self-trust is a doxastic value that licenses non-conciliation in a disagreement with an epistemic peer. First, with regard to the doxastic value of epistemic self-trust, Pasnau is right to claim that our trust in our epistemic faculties gives rise to a whole range of beliefs and that there is value in that. I have argued that this value is primarily practical, as trusting our epistemic faculties even in the absence of reasons for believing that our faculties are reliable can help us be good practical agents.

But I have also argued that such epistemic self-trust is only conditionally justified. When there is no evidence that we have trusted our epistemic faculties too much, epistemic self-trust has doxastic value—both in relation to belief formation as such and in relation to rational action based on such beliefs. When we have evidence that shows that we have trusted ourselves too much, however, then epistemic rationality requires that we reduce confidence in the belief in question or perhaps suspend belief altogether. Practical considerations can supplement epistemic considerations in the justification of epistemic self-trust, but the former cannot trump the latter without undermining the aim of getting truth that is constitutive for epistemic self-trust.

Pasnau’s claim about the doxastic value of epistemic self-trust can thus only partially be supported: it is true that epistemic self-trust can license us to hold beliefs in the absence of epistemic reasons to believe that our epistemic faculties are reliable, but it is not true that it “would not be wrong” to trust our epistemic faculties against the evidence. The practical view of epistemic self-trust, and the two-step approach to the justification of epistemic self-trust that it supports, explain why non-epistemic considerations cannot encroach on epistemic considerations. It is true that the rationality of epistemic self-trust does not just depend on epistemic considerations; doxastic and practical considerations matter as well. But if the doxastic value of epistemic self-trust were allowed to silence epistemic considerations, the essence of epistemic self-trust would be lost.

In addition, and importantly given the main topic of this paper, the practical view of epistemic self-trust also allows us to see why the doxastic value of epistemic self-trust does not license you to remain steadfast in a doxastic disagreement with a peer. An epistemic peer is someone who has access to the same evidence as you do and who you expect to be equally reliable. You do not have to know or have good epistemic reasons to believe that the other is equally reliable as you. You are in a doxastic disagreement with an epistemic peer if you lack good epistemic reason to think that you are more reliable than the person who disagrees with you—my argument against the Epistemic Asymmetry View above suggests that we should not take this to be a rare occurrence.

While there are good non-epistemic reasons to trust your epistemic faculties, these reasons do not justify trusting your epistemic faculties when there are epistemic reasons that suggest that your epistemic faculties may not be reliable—that you have trusted your epistemic faculties too much. A doxastic disagreement with an epistemic peer, however, is precisely the sort of situation that can give you an epistemic reason to reconsider your original belief.²² Such a disagreement is thus the kind of case in which your originally justified epistemic self-trust can come under pressure.

Let me explain this further. As we saw, trusting your epistemic faculties involves a leap of epistemic faith. All sorts of input from the world can, and should, lead you to question the extent to which you trust your epistemic faculties and the beliefs that you formed on that basis. For example, you may have trusted your epistemic faculties in forming a belief about the comparative performance of several applicants for a job. You then remember reading about the effects of implicit bias in hiring. This gives you reason to pause and to ask yourself whether you should trust the process through which you formed your beliefs about the merits of the different candidates and, on that basis, whether you have reason to revise your original belief.

A doxastic disagreement with a peer, similarly, is one such input from the world that raises the question of whether you have leapt too far. In the hiring example, suppose that you communicate your final ranking of the candidates to the other panelists. You then learn that in the ranking of another panelist—one that you generally take to be your epistemic peer in these situations—the candidate that you have ranked last comes out first. This, similarly, gives you reason to pause and to ask yourself whether you should trust the process through which you formed your beliefs about the merits of the different candidates. You had assumed, trusting your epistemic faculties, that your ranking was the one that was warranted by the available evidence about the candidates. You now realize that someone you consider to be equally reliable in evaluating the evidence has ranked one of the candidates very differently. Whose epistemic faculties have malfunctioned? Could it have been yours? Both scenarios confront you with the possibility that you might have trusted your epistemic faculties too much in forming your original beliefs.

Admittedly, not all doxastic disagreements with a peer raise the question of whether you have leapt too far with the same urgency. Suppose your disagreement with your peer concerns partial belief and the evidence supports both credences. In a case like this, the doxastic disagreement will exert less pressure on your original belief than in a case where the disagreement concerns full belief (you believe p and your peer believes not-p) and the evidence supports only one belief. So, not all doxastic disagreements with a peer will necessarily have the effect of undermining your justification for epistemic self-trust. But there will be at least some cases in which the fact that you and your epistemic peer disagree gives you reason to think that (at least) one of you has leapt too far. In those cases, the fact that an epistemic peer disagrees with you gives you an epistemic reason to reconsider the trust you placed in yourself in forming this belief and to adjust your original belief. Pace Pasnau (2015), epistemic self-trust does not license you to remain steadfast in those cases; it is what comes under pressure in at least some disagreements with an epistemic peer.

6 Concluding Remarks

In this paper, I have explored the significance of epistemic self-trust for the epistemology of disagreement. My main aim has been to argue against attempts to defend steadfast intuitions on the basis of an asymmetry in how much you are entitled to trust your own epistemic faculties compared to those of others. Focusing, first, on epistemic versions of the Asymmetry View, I have shown that while it is true that epistemic self-trust is basic in a way that trust in the epistemic faculties of others is not, this does not have the asymmetric normative implications that would be needed to defend steadfast intuitions. I then discussed Pasnau’s non-epistemic version of the Asymmetry View and argued that although Pasnau is right to draw our attention to the doxastic value of epistemic self-trust, he is wrong to claim that non-epistemic considerations can legitimately encroach on epistemic rationality and, in this way, support steadfast intuitions in the case of a doxastic disagreement with an epistemic peer. Such disagreements tend to give you epistemic reasons to doubt the reliability of your epistemic faculties.