In Epistemic Authority, I begin with an argument for the rationality and inescapability of epistemic self-trust, which is discussed in the paper by Elizabeth Fricker, argue next that consistent self-trust commits us to trust in others, and then argue that some of those others satisfy conditions for epistemic authority modeled on Joseph Raz's well-known principles of practical authority. One of the central theses that I adapt from Raz is the Pre-emption thesis, and this is the principal object of the critical part of Jäger's paper. I argue that epistemic authority applies in the domains of morality and religion, and that the account of epistemic authority I give shows that epistemic authority is compatible with intellectual autonomy for the same reason that political authority is compatible with autonomy in the tradition of political liberalism endorsed by Raz. That is, believing or acting on authority is a demand of conscientious self-governance. It is not only compatible with autonomy, but follows from it.

In describing the contours of epistemic authority, I propose analogs of all of Raz's theses of authority, but the most important ones are the Pre-emption thesis and the Normal Justification thesis. The Pre-emption thesis tells us what it means to act on authority. It says that when I am acting on authority, the authority's directive replaces my other reasons for and against doing the act, and is not simply added to them. An authority is a person who has the normative power to give me pre-emptive reasons.

As applied to belief, the pre-emption thesis says that I am believing on authority only if I take the authority's belief or testimony that p as replacing my other reasons for and against p. An epistemic authority is someone with the normative power to give me reasons of this kind.

The defense of someone's authority is given by Raz's Normal Justification thesis, which says that the normal way to defend the fact that one person has authority over another is to show that the alleged subject is more likely to act on reasons that apply to him if he acts on the authority's directives than if he tries to act on those reasons directly.

I have two epistemic analogs of the justification thesis. The first says that I should take someone as an epistemic authority when I conscientiously judge that believing what she believes or testifies is more likely to give me the truth than if I try to figure it out myself. The second says that I should take someone as an epistemic authority when I conscientiously judge that I am more likely to form a belief that will survive my conscientious self-reflection if I believe what the authority believes or testifies than if I try to figure it out myself.

Jäger offers three objections to the Pre-emption thesis as applied to belief. The first is the problem of competing authorities. The second is the switching problem. The third problem, and the one to which he devotes the most attention in that section of the paper, is the problem that pre-emption seems to require me to unhinge proper bases for my beliefs. Jäger accepts pre-emption in the case of disagreement between me and the authority, but he disputes my defense of pre-emption when the authority and I agree, arguing that we ought to use the idea of graded beliefs to handle those contexts.

The first problem is that of competing authorities. It would seem that two people could have the power to give me pre-emptive reasons because both could satisfy the first justification thesis – they could both be more likely to get the truth whether p than I am. It is true that the first justification thesis does not tell me which of two competing authorities I should accept as an authority for me. In many practical situations it does not matter which one I follow. Authority is intended to give me a more efficient and effective way to act on my own reasons. If I want a financial advisor, I pick one whom I know to be significantly better at financial decisions than I am, and who, as far as I can tell within practical limits, is no worse than others I could choose. I do not worry over the occasion when my advisor disagrees with my friend's advisor since that would miss the point of having an advisor. She is supposed to save me the trouble of concerning myself with all the reasons for and against each financial choice I can make.

But the epistemic cases are different because I want each and every belief to be true. I do not simply aim at a “good record” in believing, and so competing epistemic authorities is more problematic than competing financial advisors. I discuss this problem in Chapter 10 on reasonable disagreement, and argue that the second justification thesis can handle the problem. My determination of whether someone is an authority for me depends upon my judgment that taking a belief from this authority will survive my current and future conscientious reflection upon my total set of psychic states better than a belief I would come to on my own. Conscientious reflection includes reflection on the authoritative status of the competing authorities, my degree of trust in each of them, and my degree of trust in any beliefs I already have that might conflict with their respective judgments. Since I argue that trust is a reason (a first-personal reason), trust plays an important role in determining how I should handle cases of apparent conflict between myself and an authority, or between one authority and another. The cases of competing authorities show the importance of the underlying norms of trust that I outline in the book, norms that underwrite the justification of authority theses.

Jäger's second problem with pre-emption is the switching problem. Authorities sometimes change their minds. We see this a lot in medical research. One year researchers say drinking coffee is harmful. A couple of years later they are touting coffee's benefits. Sometimes they pronounce something against my previous belief, but then later change their minds and endorse my original position. Should we take a switcher as an authority? Again, my second form of the justification thesis is better than the first. Once I take a belief on authority and realize that the authority switches, I may tell myself that I was hasty in taking him as authoritative since the belief did not survive my conscientious reflection for very long. But it depends upon the case. In some situations, we should adopt the authority's belief even if we judge that he might switch later on. Life is short, and we often have to act in less than ideal epistemic circumstances. All we can do is to follow the best current scientific advice on whether to take calcium daily, how high a level of cholesterol is too high, and so on. In other situations, there is no rush to make a judgment. This is the case with most philosophical and religious issues, and some, but not all, moral issues. In these cases it is important to conscientiously judge whether a certain belief will survive changes in myself and in my evidence long into the future. I designed my second justification thesis partly because of cases like the switching problem and the fact that evidence can change, but also because of changes in the self. The authority can switch, but so can I. If I conscientiously judge that later the self upon which I reflect will likely change in ways that would conflict with the authority's current pronouncement, I have a reason not to accept the pronouncement. But notice that the first two objections do not raise problems with pre-emption, which is a thesis about what it means to believe on authority, but are problems with the theses giving the justifying conditions for taking someone as an authority.

Jäger's third objection is an objection to pre-emption, and it is an interesting one. He argues that pre-emption requires you to unhinge your previous adequate basing relations between your belief p and the reasons you had as soon as you find out that someone A is more likely to get the truth regarding p than you are, and A has the same belief. To answer this objection, let's look at the purpose served by believing on authority. I have proposed that believing on authority is a strategy in which I let an authority stand in for me when the authority is better able to do what I would do in some domain of belief if I had the requisite ability, the time, and the background knowledge. It is not a one-off policy. It is a strategy for intellectual economy and productive power. I am able to govern myself in my belief-forming activities more quickly and effectively if I believe on authority. I do not need to take the trouble to learn every field in which I want beliefs, nor do I need to take the time to find evidence and weigh all the relevant reasons for and against beliefs in every domain. But suppose I have already gone to whatever trouble it takes to get adequate reasons for believing p before I encounter an authority. In this case, the authority does not give me any advantage and I really do not need her. But if I find her anyway, what should I do? The justification thesis says I should take her as an authority, and the pre-emption thesis says that I should believe on her authority, “unhinging” my belief from the adequate reasons I already had. Is there anything odd about that?

I don't think so. Pre-empting does not require me to deny that my reasons are reasons or to deny that they are good reasons. I can even be aware that they are adequate reasons. I am not dismissing them or “disempowering” them. In fact, if the authority has these same reasons, as Jäger imagines in his same-belief-same reasons scenario, then I am even using them, but I am using them indirectly. I am using them as part of a general strategy to give me many more truths than I would have otherwise, and to do so more efficiently than if I depended upon myself alone. In this particular case I do not need the authority for efficiency or a greater number of true beliefs, but the rationality of the strategy makes it rational to use it without regard to whether it is needed in the particular case. So I may not be very good at investing, but suppose I have the same reasons as my trustworthy financial advisor to buy stock A. When my advisor says to buy stock A, I buy it for the same reason I buy the other stocks she says to buy and sell the other stocks she says to sell. The fact that I already have adequate reasons to buy stock A increases my confidence in myself, but I see no reason to think I should cease taking the authority as an authority in that particular case just because I do not actually need her authority on that occasion.

I think this example actually shows something else. It is doubtful that the financial advisor's reasons to buy stock A really are identical to mine, at least not if the mysterious element of good judgment counts as a reason (what I call a first person reason). Whatever she does in making her decisions, it is better than what I do, even in those cases in which what I do is good enough. If I am smart, I will take her advice pre-emptively.

Epistemic authority gives me the greatest help in reaching my epistemic goals when the same authority is used repeatedly, but most of Jäger's examples are one-time authorities, as in his case of identifying a red cardinal. Suppose I am not very good at identifying species of birds, but male cardinals are pretty distinctive, and even I can identify one up close. I seem to see a cardinal in my tree, but I ask an ornithologist. The ornithologist confirms that it is a cardinal. Shouldn't I base my belief on both my own reasons and the testimony of the ornithologist, asks Jäger? Hasn't the ornithologist given me reason to think my reasons should be active? I think not, although he has given me reason to think my reasons are good. The next time I seem to see a red cardinal, I will be more confident. As Jäger remarks, the ornithologist shares some of my perceptual reasons, but has other reasons besides. Unlike me, the ornithologist would not be fooled by a cardinal look-alike, if there is such a thing. The ornithologist identifies much more about the bird than I can, and in fact, the ornithologist with his trained eye sees more than I see. He has already included whatever in my experience indicates a cardinal, but he has included more. It would be an illusion to think that I have more reason to believe there is a cardinal in the tree if I base my belief on both the ornithologist's judgment and my own perceptual experience than if I base my belief on the ornithologist's judgment alone. But the veridicality of my perceptual experience is confirmed by the ornithologist, and that is an important thing for me to know.

This case is also an example in which Jäger expresses doubt that it is even possible to pre-empt. He asks, “If I were asked to do this, I would not know how to proceed.” He has already agreed that we pre-empt and ought to pre-empt when we disagree with the authority. But he has doubts about the possibility in the case in which I am continuing to gaze at the cardinal even after the ornithologist has rendered judgment.

I have no objection to the position that pre-emption is not possible in this situation. I do not assume that pre-empting is an act of will. In fact, I have no general position on the degree of control we have over pre-emptive reasons. Possibly there are cases like seeing a red cardinal in front of me in which I am not able to believe the ornithologist pre-emptively. It is even possible that we have little control over any of the reasons we have for doing what we do or believing what we believe. Nothing I have said requires a particular position on our control over our reasons. This leads to an interesting but distinct set of questions about the connection between norms of rationality and what we are able to do. Sometimes we cannot follow a norm of rationality. Other times we can do so, but not because we decided to do so. Can I decide to stop at a stop light because the law says so rather than because I can see a car coming that will hit me if I do not stop? Probably not, but my view of authority does not say otherwise. Plenty of people can and do stop at a red light because the law says so, and Raz's Normal Justification thesis shows why that is justified.

I think it is important to see that Raz's Normal Justification thesis is not intended to give necessary and sufficient conditions for someone's being an authority. The thesis is constructed to identify the “normal way” we show that someone is an authority over someone else. Likewise, my epistemic justification theses are not intended to give necessary and sufficient conditions for epistemic authority, but are meant to elucidate a general strategy in rationally governing myself. The bottom line in rationality, I believe, is to govern myself in a way that aims to make all my conscious states (beliefs, perceptions, memories, emotions) fit their objects while simultaneously making them fit each other. The ultimate test of whether I have succeeded in the first aim is that I have succeeded in the second. My judgment that someone is an authority is ultimately determined by conscientious reflection. My judgment that someone is an authority does not trump all my other judgments, much less all other parts of my self.

This brings me to the last section of Jäger's paper, on Socratic authority. I am grateful for his thoughts on the way authorities can help us gain understanding, a value that is more important than true belief. In Epistemic Authority I discuss understanding briefly in the chapter on moral authority, arguing that moral understanding is more important than moral true beliefs, and whereas belief on the testimony of an authority can give us true beliefs, it cannot give us understanding. However, it is significant that there are persons who can help give us understanding. In the chapter on authority in communities I discuss some examples of authorities who can help us govern ourselves in ways that we find ultimately fulfilling. Their authority combines the epistemic and the practical, but it is not limited to helping us reach ends we were able to identify in advance. A beginning philosophy student is justified in taking his philosophy professor as an authority on the basis of his conscientious judgment that the professor will help him reach goals he had in taking the course – typically, the goal of getting information about what the great philosophers thought. As he studies under the professor's guidance, his goals change. He begins to see what philosophical understanding might be like and it becomes a goal. But he probably will not begin to aim at understanding until he gets some understanding himself.

This situation is one of the reasons for the preferability of a more general justification of authority thesis. The fact that the student judges that the professor is more likely to know truths about philosophy than he is himself is a reason to take the professor as an authority, but it is more important that the professor is what Jäger calls a Socratic authority, someone who can help the student get understanding. What is the student's justification for taking the teacher as a Socratic authority? That question is easy to answer once the student is able to have enough understanding to know what understanding is and to form the goal of acquiring it. But I think the student can be justified in taking the professor as a Socratic authority even before he knows what understanding is. He can be justified by his conscientious judgment that if he follows the professor's directions, his future psychic states, including states he currently cannot even imagine, will survive his own conscientious reflection better than if he proceeds independently of the professor. This is an even more general justification thesis than the General Justification of Authority Thesis I offer in Chapter 7, and I thank Christoph Jäger for leading me to see this idea. I am also grateful for his very insightful reflections on understanding.

Elizabeth Fricker's generous and helpful paper on rationality and self- trust advances the discussion I give in Chapter 2, and raises important questions on the nature of trust, the defense of epistemic self-trust, and the value of harmony in rational self-governance.

First of all, I think Fricker is right that the core phenomenon of epistemic self-trust is one's ungrounded reliance on one's cognitive faculties to reliably yield true beliefs. Fricker agrees with me that there is a natural desire for truth under the assumption that that is not interpreted as the reflective desire to amass true beliefs, but is the recurring desire to have our questions answered. I propose that in the standard cases of trust, including pre-reflective trust, trust includes a belief, an affective component, and a behavioral component. Fricker disagrees with me about the belief component. She says that no meta-belief is necessary at the pre-reflective level, and I agree with that. I think there is a sense in which even the young child who continually asks questions has an implicit belief that the questions can be answered, but I agree that that does not rise to the level of a meta-belief, so I don't think that there is any important difference between us on this point.

Fricker agrees that in spite of the epistemic ungroundedness of epistemic self-trust, it is inescapable. She also agrees that it is rational, but she offers a very subtle and interesting discussion and critique of my arguments for the legitimacy of self-trust. In section 4 she gives three interpretations of my argument that either self-trust is a reason for belief or there are no reasons for belief. She calls one a transcendental argument, and the second an analytic argument for self-trust modeled on Strawson's argument for the rationality of induction. The third, and her preferred version, is a Kantian-style argument that there is no thinking outside the space of reasons, and self-trust is a presupposition of the space of reasons. I will not comment on each of these interpretations except to say that I am grateful for them and wish we had had a conversation about them while I was writing the book because Fricker's suggestions would have improved it. I hope to go back to that in future discussions with her.

Fricker gives the major part of her critical attention to an argument for the legitimacy of epistemic self-trust that she calls my Meaty Argument. First, she finds a thinned down version of the Meaty Argument plausible. She calls it the Pragmatic Argument (another good title). According to this argument, a kind of self-trust that includes belief, affect, and treatment is rational “because it puts an end to excessive self-reflection, and to the sense of unease one is prone to in light of one's awareness of the ungroundedness of one's inevitable ground-floor epistemic self-trust.” But Fricker objects to the general conception of human rationality that I use to underwrite the above argument.

One of the theses I propose that Fricker finds plausible is the thesis she calls NN (Normative Naturalism), according to which being rational is doing a better job of what we do naturally, but I elucidate what we do naturally in terms of a natural mechanism for resolving dissonance in the self that she finds problematic. My view is that it is rational to resolve dissonance in the self, and there are better and worse ways of doing so. We aim to make our states fit their objects – our beliefs true, our emotions fitting, our desires for the desirable, our perceptions veridical, our memories accurate. Ultimately, our only test that they do fit their objects is that they survive conscientious reflection upon the total set of our psychic states, which can go on into the indefinite future. What I call dissonance is the experience of conflict within parts of the self. We interpret that as a sign that something has gone wrong. Dissonance is a sign that one of my beliefs is false, or an emotion is unfitting for the circumstances, or I am misremembering something, and so on. Dissonance between a judgment and a decision to act is a sign that either the judgment or the act is mistaken – the familiar problem of akrasia. We have lots of ways to test the fittingness of a psychic state, but the problem of epistemic circularity can be generalized to a problem of the circularity of our total set of conscious states. We test one state by other psychic states now and into the indefinite future. That is why I say that ultimately, our only test that our states fit their objects is that they fit each other.

Fricker has a number of objections to the picture of Rationality as Dissonance Elimination (RDE) I propose. She remarks that when she has a desire and she believes it cannot be fulfilled, she does not experience dissonance in the same way in which she experiences dissonance when beliefs conflict. I think she is right about that and I think that stressing that point aids the view of the mind I am proposing. Second, she says she is not willing to say that it is always rational to resolve the dissonance between a belief and a desire. I agree with that also. And finally, she says that RDE does not tell us which of two conflicting states should be eliminated. Fricker is right about that as well. Another part of my psychological picture is needed for that issue.

Let's look at the conflict between belief and desire, the focus of Fricker's objection. I say very little about desire in the book, but in Chapter 2 I say that a desire fits its object as long as its object is desirable. So Fricker's desire to end torture is fitting as long as it is desirable to end torture. I take for granted that her desire is fitting. I also think her belief that torture will not be eliminated any time soon is true. So I think that both her desire and her belief fit their objects. I say in that chapter that we do not always need to resolve a conflict between a belief and a desire; we can live with the dissonance. But I also remark that harmony is better. Fricker replies that she does not have a reason to eliminate either the desire or the belief, and furthermore, she does not experience dissonance between them in a way that is phenomenologically similar to her experience of conflicting beliefs. I agree that she has no reason to give up either her desire that torture be eliminated, or her belief that it is not likely to be eliminated. Both states fit their objects. Furthermore, if Fricker is right that the experience of dissonance is different in belief/desire conflict situations than in belief/belief conflict situations, that explains why we feel no need to give up one of the states generating dissonance in the case of belief/desire dissonance. I think this is an important insight. It may mean that we interpret some kinds of dissonance as signs of a problem in ourselves, but we interpret other kinds of dissonance as a problem in the world. A desire can be unfitting, but that is when we desire what is undesirable. There is no reason to think it is unfitting just because it causes dissonance with the belief that it cannot be fulfilled. Of course, there are cases in which the belief that a desire is unfulfillable ought to lead me to give up the desire (e.g., if I acquire the desire to win the Nobel Prize in Physics), but I see now that I should not have made the remark that it is better to resolve dissonance between desire and belief (p. 31). I think this means that desires form a very complicated category, one that deserves more attention than I gave it in the book.

I think that one area in which Fricker and I disagree is in the treatment of emotions. Fricker says that she thinks of emotions as like desires, whereas I think of emotions as like beliefs in the respects I am discussing. We aim to make our emotions fitting just as we aim to make our beliefs true. If we admire someone who is not admirable or fear something that is not fearful, we have made a mistake of a certain kind. Our way of detecting mistakes is to reflect about our emotion together with our beliefs, experiences, and the emotions of other people we trust. I think that dissonance between an emotion and a belief needs to be resolved, unlike dissonance between desires and beliefs.

We cannot tell which of two conflicting states we should try to eliminate without predicting which one will survive current and future reflection on our psychic states. Sometimes it is obvious which one will survive, as when I believe that the meeting starts at 3:30, and then see the whole committee assembled in the room discussing item 3 on the agenda when I walk in at 3:30. But other times, both of the conflicting states have considerable depth in my psychic structure, and I conscientiously predict that both of them will survive considerable changes in my experiences and future beliefs. Moral and religious beliefs can generate conflicts of this kind. I have a belief that I conscientiously believe will withstand future changes in my experiences and other beliefs without generating dissonance, but I also trust someone else who has a contrary belief, and my trust in that person is based on aspects of myself that I also predict will withstand future changes in myself. The way I think a conscientiously reflective person should handle this problem is the topic of Chapter 10.

In conclusion, I would like to thank Christoph Jäger and Elizabeth Fricker for their very helpful and generous papers. I hope that my response has done justice to their careful work, and I look forward to pursuing the ideas in their papers in future discussions.