Tokyo Academic Review of Booksonline journal / powered by Yamanami Books / ISSN:2435-5712

April 30, 2022

Mark Schroeder, Reasons First

Oxford University Press,2021年

Reviewer: Ryo Chonabayashi

Tokyo Academic Review of Books, vol.47 (2022);

Introduction: The Idea of Reasons First

In his recent book, Reasons First, Mark Schroeder argues that the normativity discussed in epistemology is explained in terms of reasons. According to this thesis, normative phenomena in epistemology, such as justification, rationality, and knowledge, are all ultimately explained in terms of reasons. This thesis is entailed by the following stronger thesis. Not just epistemic normativity, but all kinds of normativity discussed in various areas, such as aesthetics, politics, legal matters and moral philosophy, are all explained in terms of reasons. This strong thesis certainly predicts that normativity in epistemology is explained in virtue of reasons. Schroeder’s project in this book is to test this strong thesis in epistemology (hereafter, I use the phrase “Reasons First” to refer to this strong thesis).

There is a further question of whether the fundamental normativity itself is explained by non-normative entities. Assuming reasons are such fundamental normativity, can reasons be explained in terms of some non-normative entities, such as desire? T.M. Scanlon, for instance, argues in his Being Realistic About Reasons, that reasons which is the fundamental normativity for him cannot be reduced to other non-normative entities, and he calls his view “reasons fundamentalism”. This question and the question of whether reasons are the fundamental normative entity within the domain of normativity are different (Broome 2018, p. 298), and Schroeder’s book engages with the latter.

Schroeder’s strategy to defend Reasons First is sophisticated. He explores traditional issues in epistemology, discussing each issue in detail, and argues that the best possible position in each issue is the one consistent with Reasons First. This overall conclusion motivates the strong claim that all normativity, beyond the normativity discussed in moral philosophy, but also the normativity discussed in epistemology, are all grounded in reasons.

The book consists of four parts that have twelve chapters in total. Of necessity being selective, below, I first expound the central claims of each part of the book and then raise one reservation about one of the central claims of the book. I will also raise one issue that is related to Schroeder’s discussion on the explanatory relevance of knowledge.

Before I start explaining the book’s contents, I shall briefly explain the kind of reasons Schroeder’s book focuses on and how he understands it. The kind of reasons Schroeder’s book focuses on are reasons that have the following three features. First, those reasons are act-oriented factors: “they are considerations that count for or against something we can do” (p. 29). Breaking a promise to my spouse makes what I said to her a lie, and this feature is a consideration that counts against this action. Second, those reasons can be competitors. The fact that breaking a promise to my spouse hurts her is a reason against such an act. But that act’s another feature, that not breaking that promise forces me to break another promise to my kids, is a consideration against the act. In this way, an act’s different features that play the role of reasons can compete. Finally, those reasons cad be acted on. I may break the promise to my spouse on the reason that not breaking this promise makes my kids disappointed. It is tempting to think that the kind of reasons with these three features are what philosophers often call “normative reasons” (cf., p. 225). These reasons introduced above are distinct from “motivating reasons”. Suppose I break a promise to my spouse. I can be asked why I do this. The reason I would provide in response to this question is a motivating reason: I might answer that not disappointing my kids is more important than keeping my promise to the spouse, not only for me but also for the spouse. Notice this claim of a motivating reason is correct only if the act in question is taken. But claims about reasons that have the three features discussed above can be true even if an act in question is not done. Even if I haven’t broken my promise to my spouse, we can still talk about the features of breaking the promise, its pros and cons. The claim, “I shouldn’t break the promise to my spouse because it hurts her”, may be true even if I haven’t broken the promise.

Schroeder also claims such reasons can be divided into two categories. The reasons in the first category are objective reasons. Objective reasons are just there regardless of whether we are aware of them. The reasons in the second category are subjective reasons. Subjective reasons are reasons we have, the reasons we are aware of. The fact that breaking a promise to my spouse hurts her is an objective reason against my breaking the promise. I might know this fact, and if I do, I also have a subjective reason against the act. Schroeder says false beliefs can act as subjective reasons (Chapter 4), so, in this case, my subjective reason can be understood as my belief that breaking the promise to my spouse hurts her. Schroeder says objective reasons “support correct courses of action – the ones that are a good idea to take” while subjective reasons support “rational courses of action – the ones that make sense from the agent’s subjective perspective” (p. 78). An important point to emphasise is that Schroeder seems to be thinking that the rationality of belief and action has a closer connection with relevant subjective reasons, not objective reasons. More strongly, Schroeder writes:

What makes an action the rational thing to do, if it depends on reasons at all, is determined by the balance of subjective reasons for and against performing that action. Similarly, what makes a belief the epistemically rational thing to believe is the balance of subjective epistemic reasons for and against that belief (p. 226).

With these preliminary understandings of the kind of reasons Schroeder discusses in the book, let us see the central claims of each part of the book below.

Part I

Part I introduces the project of the book. Schroeder says that in contemporary analytic moral philosophy, the idea of Reasons First is a default view. For instance, theorists tend to explain their views in terms of reasons. Schroeder quotes Derek Parfit’s presentation of egalitarianism as an example. Parfit explains egalitarianism as the view that “we always have a reason to prevent or reduce inequality, if we can” (p. 5, cf., Parfit 1997, p. 209)i. This tendency leads us to a further view that reasons play some fundamental roles in explaining various normative phenomena in moral philosophy. As already indicated, Schroeder suggests a strong thesis, Reasons First, according to which reasons are analytically and explanatorily prior to other normative entities, and in this sense, they are the most fundamental entity in the domain of normativity (pp. 4-5).

It is not the case that every moral philosopher accepts Reasons First. But, as Schroeder says, it is certainly a popular view in moral philosophy and Schroeder’s claim that the view has a “default” status (p. 6) has some plausibility. This situation in moral philosophy certainly supports Reasons First. But we have to remind ourselves of the strength of this thesis. Reasons First predicts that normative phenomena in all other areas can be also explained in terms of reasons. Thus, to defend Reasons First, we need to find more evidence for this thesis outside moral philosophy.

Schroeder’s test case in this book is epistemology. Epistemology is also concerned about normative issues. Is it rational/justifiableii to believe that we are not the victims of the Cartesian demon? How much evidence do we need to believe rationally? If we find other people reach a different conclusion based on the same evidence, should we be less confident in our own conclusion? These questions in epistemology are all normative questions. They ask how we should think, rather than how we think. Given the normative nature of epistemology, one would naturally expect that if Reasons First was true, normative phenomena discussed in epistemology could be explained in terms of reasons. Schroeder says this expectation faces some initial difficulties and explains why Reasons First has not gained the default status in epistemology.

The first explanation is based on the problem of unjustified belief. Recall the implication of Reasons First in epistemology. If Reasons First is true, these key entities in epistemology, knowledge, justification and rational belief are all explained in terms of reasons. To support this claim, we need to characterise the reasons that play such explanatory roles (p. 17). If the reasons that explain these epistemic items themselves need to be grounded by some other epistemic items, such as knowledge and justification, Reasons First fails. Reasons First says the reasons that ground epistemic normativity do not need to be grounded by any epistemic normative items. The problem of unjustified belief poses a challenge to this implication of Reasons First.

It seems plausible to assume that beliefs can act as reasons in epistemology. Schroeder tells us that in normal cases, what we believe affects what is rational for us to believe (p. 18). My belief that my spouse is trustworthy is a reason to believe what she says to me. Nevertheless, all of my beliefs cannot be reasons to form other beliefs. It also seems plausible to accept that one’s unjustified belief does not contribute to one’s gaining rational belief or knowledge. Suppose I believe my sons are happily playing in a park based on another belief that they are smiling. But this second belief is gained by my imagination. In this case, it is hardly the case that the first belief is rational. This shows that for a belief to be part of one's (normative) reasons to believe, the belief cannot be unjustified or irrational. Given this verdict, the analysis of subjective reasons for belief has to explain how beliefs that contribute to justification and knowledge differ from beliefs that cannot make such a contribution. Schroeder says, “[the] usual and most natural response to this” is to suggest some conditions a belief must satisfy to act as reasons. For instance, a belief that can be a reason for belief might have to be justified (Feldman 1988), or itself knowledge (Williamson 2007). But this move is in conflict with Reasons First. Reasons First says those epistemic notions, justification and knowledge, are analysed and explained in terms of reasons, not the other way round. If we have to appeal to either justification or knowledge to explain which beliefs act as reasons, we have to admit that the nature of reasons in epistemology needs to be explained by one of these epistemic notions (pp. 18-19). Note Schroeder thinks that this problem is conditional in the following way. The problem poses a challenge to Reasons First only if knowledge is also normative. If knowledge is not normative and it grounds subjective reasons as a non-normative entity, the problem loses its strength (p. 21).

The second explanation is based on the problem of sufficiency. Let us consider how reasons for action behave in cases where there are both equally good reasons for and against one action. Suppose one of my sons says we should go to a park and play there this afternoon while the other says we should stay at home and do some games inside. Suppose further that what they say (and various background considerations supporting their opinions) are equally good. In this situation, I have good reasons to go to a park while I also have good reasons against this action. In such a case, Schroeder says it is still rational to follow one of these suggestions (p. 19). On this suggestion, although one of my sons will get disappointed by this decision, it is still rational for me to go to a park this afternoon. Notice this reaction supports Reasons First in the realm of action. If it is still rational for me to do A even when there are both good reasons for and against A, reasons seem to be playing the fundamental role in the realm of action: since there are already reasons for A, the fundamental normativity in action, those reasons are enough to realise other normative phenomena such as the rationality of doing A in this case.

But, Schroeder observes, reasons in epistemology do not behave in this way. While beliefs in some cases seem to play the role of reasons in epistemology, all beliefs cannot be reasons for belief, as we observed above. So, perhaps, beliefs that are evidence plays the role of reasons in epistemology. It is natural to assume that my evidence that my spouse is at home (e.g., my spouse tells me so on the phone) is a reason to believe that my spouse is at home. Now, an interesting observation Schroeder provides is that some cases allude to the idea that a piece of evidence that plays the role of reasons in epistemology needs to be grounded by some other epistemic normativity, and this idea is in conflict with Reasons First. For instance, when the evidence (a reason in epistemology) between two competing hypotheses is tied, it is never rational to believe either of the hypotheses in question (p. 19). A rational reaction to this situation is being agnostic about these competing hypotheses. This case indicates that to explicate when evidence is sufficient for generating epistemic rationality, we might need to appeal to some independent feature, such as knowledge. Perhaps my evidence makes my belief rational since evidence has a close connection with knowledge. Assuming that evidence is only epistemic reasons, this conclusion conflicts with Reasons First: to explain the sufficiency of evidence and the mechanism of evidence’s function as reasons, we need to appeal to some non-reason normative notions in epistemology.

To defend Reasons First even in epistemology despite these two problems, Schroeder first suggests alternative positions and arguments within traditional epistemological issues. Then, he argues that his suggestions, in turn, provide good solutions to the two problems that are the obstacles for Reasons First. Through constructing those arguments within epistemology, Schroeder also tries to motivate what he calls the Core Hypothesis according to which “views about core topics of traditional epistemology have sometimes been distorted by lack of the perspective that we can gain by considering the commitments we incur in answering them in light of similar questions about normative inquiry outside of epistemology” (p. 9). As discussed in the next section, one such case that supports the Core Hypothesis is his discussion on subjective reasons. Schroeder observes that the dominant view of subjective reasons in epistemology is that they must be factive, but if there is a parallel between moral philosophy and epistemology on this issue, this dominant view needs more careful reflection. That is because it looks plausible to think that subjective reasons in moral philosophy do not have to be factive, and if there is a parallel between the two disciplines, subjective reasons in epistemology maybe not always factive. And, theories that endorse this non-factive view of epistemic subjective reasons might be attractive accounts.

Now, we can see the whole picture of Schroeder’s project. Some central issues in epistemology can be better dealt with if we are well informed about some findings in moral philosophy where the nature of normativity has been extensively discussed. Once we get well informed in that way, we find better options in epistemology. And, these possible options in epistemology further motivate the strong version of Reasons First despite the fact this thesis looks not plausible due to the problem of unjustified belief and the problem of sufficiency.

Part II

Now, let us see what the substantial suggestions Schroeder proposes in epistemology. The first issue he discusses is the issue concerning how perceptual experience provides evidence for the external world. Despite the fact that various sceptical scenarios are consistent with our evidence, how can perception still provide evidence for the facts about the external world? Schroeder exhibits and explores some representative views proposed in epistemology that explain this seeming gap between our perceptual experience and the external world.

Schroeder presents two groups of possible views on this issue. The division of the groups comes from their endorsing different conceptions of perceptual evidence. The first group of the views hold what Schroeder calls “the phenomenal conception of evidence”. According to this view, “the evidence provided by perceptual experiences includes only propositions about our own internally individuated, subjective psychological state” (p. 55). If we accept this view, we have to admit the gap between perceptual evidence and how the external world actually is. That is because, on this view, perceptual evidence is about ourselves, not directly about the external world. So, any view that endorses this conception of perceptual evidence needs to explain this gap, but this explanatory project is not easy.

Schroeder then introduces another group of views that rejects the phenomenal conception of evidence. These views endorse the idea that perceptual experience is world-implicating, and there is no gap between perceptual evidence and facts about the external world (p. 60). If this view is correct, sceptical scenarios such as the Cartesian demon scenario is not consistent with perceptual evidence against such a scenario. Schroeder’s overall assessment of these views is not very positive (though he thinks the idea that perceptual evidence is world-implicating is important). His dissatisfaction is a disjunctivist idea shared by these views, namely that the rationality of belief we gain in epistemically good cases and bad cases will be different. Consider the case that I see an object in front of me, and it looks like a red apple. Based on this perception, I believe that a red apple is in front of me. But suppose my perception is an illusion. Also assume that this illusory experience and the experience I would get if I were in front of an apple are indistinguishable. Such a case is an epistemically bad case while a normal case such as that I have the visual experience of a red apple in front of me and there really is a red apple in front of me is an epistemically good case. Schroeder says on the disjunctivist views, in a good case and a bad case, we have “different reasons” to believe that a red apple is in front of me. In the good case, the fact that there is a red apple in front of me is an objective reason that supports the belief, and this reason is not there in the bad case. In the bad case, the fact that my perceptual experience is an illusion is an objective reason that is against my belief, and this reason is also not there in the bad case. According to Schroeder’s reading of disjunctivism, there is no reason in the bad case that plays “a privileged role” in explaining the rationality of my belief, and we have to conclude that my belief in the bad case is not rational (p. 64). Schroeder thinks this is deeply a problematic implication of the disjunctivist views. Despite the fact that perceptual experiences we have in these cases are phenomenologically indistinguishable, and many other external elements are also the same (same environments, same ways of reasoning, etc.), why do we have to conclude that the rationality of belief in good cases and bad cases have to be different?

At this juncture, Schroeder raises the following question: do we have to choose one theory as the best one from these possible positions despite the fact all of them have some problems? Schroeder’s answer is “no”, and his answer supports the Core Hypothesis: inattention to the parallels between epistemology and moral philosophy limits available options in the former in a problematic way.

Schroeder argues that views that can avoid all of these problems become possible if we stick to the perception-world-implicating view while rejecting the view on evidence according to which evidence must be factive. Schroeder motivates the view called “the apparent factive view” according to which perceptual evidence is world-implicating, but it does not have to be true (when you see that P, you “come to the proposition” that it appears to you that P as evidence, even if it is not the case that P).

Through this rather complex discussion, how does Schroeder solve the problem of unjustified belief in a way that is consistent with Reasons First? To answer this question, let us focus on the disjunctivist theories of perception and consider the relation between these views and the problem of unjustified belief. Recall the problem of unjustified belief is a problem for the proponents of Reasons First in epistemology. The problem alludes to the idea that a belief can act as a subjective epistemic reason only if it is already either justified or knowledge. But if so, our beliefs’ role as subjective reasons in epistemology, including perceptual beliefs, cannot be explained only in terms of reasons: it has to appeal to some further epistemic notions such as justification or knowledge. Recall one of the disjunctivist theories, such as Williamson’s, endorses the view that our perceptual beliefs play an evidential role only if they themselves are knowledge. This view is a natural response to the problem of unjustified belief, and, importantly for our purposes, not consistent with Reasons First: the view says to explain the evidential role of beliefs, we need to appeal to knowledge (pp. 86-88). Now, we can predict that the apparent factive view endorsed by Schroeder may motivate a following different solution to the problem. Some beliefs that are not knowledge, including beliefs that are actually false, can still play the role of rationalisers. Consider the illusory red apple case again. In this case, it does appear to me that there is a red apple in front of me, though in fact there is no apple. My subjective reason that it appears to me that there is a red apple is not defeated by the objective defeater that there is no such apple. Thus, Schroeder would say, this subjective reason does rationalise my belief. Importantly, on this view, the rationality I gain in a normal case would be as good as the rationality I gain in a bad case, since the subjective reasons that work as rationalisers both in good cases and bad cases are the same. As we saw above, this verdict on the rationality we gain in good cases and bad cases is the one not consistent with disjunctivist theories while it is consistent with Schroeder’s suggestion. This is a promising proposal for Reasons First since this view seems to be capable of providing a mechanism of epistemic rationality that does not have to be grounded in terms of knowledge.

Part III

In Part III, Schroeder discusses the problem of sufficiency. The problem comes from various puzzles concerning the nature of evidence, assuming that evidence is only the reason in epistemology. The central one of such puzzles is what Gilbert Harman observes (Harman 2002), namely, the problem of near ties (p. 129). In the case of action, intuitively, acting A is rational if there are good reasons to do A, even when there are contrary reasons against acting A that are as good as the reasons that support acting A. On the other hand, our believing that P is not rational even if believing that P is supported by good evidence, if there is another evidence against this proposition that is as good as the supporting evidence.

Why does this pose a problem to Reasons First? This observation of near ties cases is problematic since the observation indicates that to explain when evidence will be sufficient, we need to appeal to something other than evidence. If evidence is only the epistemic reasons that matter for the rationality of belief, it must be the case that we are still rational to believe that A in the case above. But this is not what our intuition about epistemic near tie cases supports.

A natural conclusion we can draw from this observation is that the rationality of belief is very different from the rationality of action, the one suggested by Harman (2002, p. reference). But this is exactly what Schroeder hopes to reject: Harman’s conclusion alludes to the view that while rationality of action is grounded by reasons, rationality of belief is grounded by something else.

Schroeder’s proposed solution is simple and powerful. He rejects the idea that epistemic reasons are only evidence. And he argues that there are some non-evidential epistemic reasons, epistemic reasons in the sense that are relevant to the rationality of belief entailed by knowledge (p. 141). Schroeder’s substantial claim on this point is this: the costs of error and the availability of further evidence are always reasons against belief, while these considerations are not evidence against its content (p. 142). Note his proposal is not that non-evidential considerations can be epistemic reasons to believe something in general. Rather, his proposal is that such considerations can be epistemic reasons against believing something. He explains this point as follows: “evidence that P is the only kind of properly epistemic reason in favor of believing that P, evidence that ~P is not the only kind of properly epistemic reason against believing that P” (p. 144)

The costs of error in our belief are highlighted by the bank cases famously discussed in the literature on “practical encroachment” (p. 132). Consider the following two cases. In one case, you vaguely remember that the bank is open on Saturday morning, and you wonder whether it is the case that the bank is open on Saturday morning. In the second case, you have the same vague memory of Saturday bank, but you urgently have to put some money in your account before next Monday: if you fail to put the money in, the foreclosure starts. The costs of error in the second case are higher than the costs in the first case in this way. The proponents of practical encroachment would say that to justify your belief in the bank’s opening times, you ought to collect more evidence in the second case.

The significance of the availability of further evidence is illustrated by a pair of cases wherein one case gaining further evidence is easy while in the other case it is hard. Suppose we consider a historical question of whether a prominent general of the past has two daughters or only one daughter. Some documents support the two-daughters hypothesis while the others support the one-daughter hypothesis. Suppose further that the qualities of these contradicting documents are equally good from the perspective of the standard of rigorous historical inquiry. In this case, it is relatively difficult to gather further evidence to settle this issue since the issue is historical. But now consider the following case. I am at home and hearing some sounds from the next room. These sounds include both laughing voices and screaming. They support the following two hypotheses equally well: that my kids are playing together joyfully and that they are fighting with each other over the possession of their favourite toys. In this case, it is very easy to find further evidence that settles the issue. All I have to do is go to the living room and see what my energetic kids are doing (pp. 131-132).

It is tempting to think that these two non-evidential considerations sometimes work as reasons against belief. In the high-cost bank case, perhaps I shouldn’t believe that the bank is open on Saturday morning based on my vague memory. Well, it is really risky to have a false belief in this case since if I act on the basis of a false belief in this case, I will lose my home! In the living room case, perhaps I also shouldn’t believe that the children are playing peacefully on what I hear from there. Gathering further evidence is very easy, so what I should do is simply go to the living room and see what is happening.

In Chapter 7 and 8, Schroeder explains why those non-evidential reasons against belief are still relevant to epistemic rationality of belief (p. 145). Schroeder gives an analogous argument to make his case. We can make a distinction between the right kind of reasons and the wrong kind of reasons to, say, admire someone by thinking about the distinctive roles this mental state (admiration) plays (p. 157). Schroeder says, “[admiration] plausibly plays the role of giving us role-models to emulate. So, it makes sense that in order to play this role well, it will have to be sensitive to features that matter for whether someone’s life is worth emulating or aspiring to. Being offered money to admire someone is not something that admiration needs to be sensitive to, in order to do its job, because people who have the feature that other people offer third people money to admire them are no more worth emulating than anyone else” (pp. 157-158). In this way, we can make a distinction between right kind reasons to admire someone and wrong king reasons to do so. Right kind reasons and wrong kind reasons for belief can also be distinguished in the same way, according to Schroeder, and the distinction tells us that the availability of evidence and cost of errors against belief are the right kind of reasons for belief given the nature of belief (p. 159). After appealing to Stalnaker’s toy theory of belief according to which “the role of beliefs in our psychologies is to allow one to act in such a way as to fulfil one’s desires” (p. 159), Schroeder suggests a theory of belief, “the default reliance account”. On this view, “the role of belief is to give you something to rely on by default. On this view, beliefs are general-purpose, default assumptions about what is true”, “To believe that P is to have a perfectly general habit of accepting that P” (p. 168). With this theory of belief, we can now see how those non-evidential considerations can be categorised as right kind reasons against belief. To make good decisions based on our default assumptions, it is obviously relevant whether further evidence about these assumptions is easily available or not. If it is, we shouldn’t make our minds before further evidence comes. Also, the costs of error are very relevant. If the costs of accepting a general assumption are very high, such costs are good reasons against holding the assumption.

Part IV

The last part of the book deals with the issues concerning knowledge. Schroeder first discusses the distinction between right-properties and well-properties often discussed in moral philosophy. Consider Kant’s shopkeeper example. The shopkeeper treats his customers fairly, not from his sense of duty but from his inclination to keep his business running. Kant thinks this action does not have real moral worth (pp. 53-55, 4:397- 4:400). The shopkeeper's action becomes morally worthy if the keeper acts from the sense of duty. Here we see the distinction between actions that merely have right-properties and actions that have well-properties. Arguably, the shopkeeper’s action only has a right-property, but not well-property. An action done from the sense of duty has both the right-property and well-property.

There are several views on the relationship between right-properties and well-properties we can find in the discussions on moral worth. According to one of such accounts, right reasons accounts, you act with moral worth when you do the right action, and the reasons you act on are the ones that make the act right (p. 218). Suppose my spouse gets angry because I forget her birthday. In this situation, it is the right thing for me to apologise to her, since forgetting one’s spouse’s birthday is a terrible thing and this requires one to apologise to the spouse. And, a suitable way for me to apologise to her in this case is that I admit that it is wrong to forget her birthday and take it as my motivation for apologising to her. In this case, according to the right reasons accounts of moral worth, my apology with this motivation has moral worth: I apologise to my spouse in a good way. My apology would not have the same moral worth if my motivation for this apology is something else, such as that I want to quickly calm her down to do some other business. Obviously, this is an insincere apology, and does not have any real moral worth.

Schroeder thinks these discussions on moral worth, the discussions on the relationship between right-properties and well-properties in moral philosophy, are highly relevant to the issues concerning knowledge. Utilising the right reasons accounts of moral worth, Schroeder advocates what he calls “the Kantian Theory of Knowledge”. The central claim of this account is that knowledge is understood in terms of well-properties: knowing is believing well (pp. 221-220). Schroeder writes, “You know just in case among the reasons for which you believe are some that are both objectively and subjectively sufficient” (p. 228). On this account, first, to know that P, you have to have some subjective reasons that support that P. Those subjective reasons make your belief that P rational. At the same time, there must be matching relations between your subjective reasons and objective reasons that support that P. Schroeder says objective reasons in epistemology generate the standards of correctness of our beliefs (pp. 227-228), but such standards are determined by the balance of objective reasons, not just truth (p. 226). So, for instance, suppose I witness my sons taking another child’s toy in a park. Based on this, I believe that the culprits are my sons, and I should tell them off. My sons say to me that they don’t take the toy and other kids are real culprits. Also, nobody, except I, sees what happens. These may be reasons against my belief, but, in this case, my reason outweighs (cf., p. 234) what my sons say and the fact that nobody except me sees the event. In this case, I seem to have both objective and subjective reasons sufficiently, and there are matching relations between the subjective reasons I have and the objective reasons out there. I not only believe a correct thing, but also believe a correct thing in a good way.

Schroeder argues that this account of knowledge can provide better explanations of various phenomena a correct account of knowledge should have, such as knowledge’s explanatory power discussed briefly by Williamson (2000, pp. 62-64) and the issues concerning epistemic defeaters.

Discussions: (Non)Factivity of Subjective Reasons and Knowledge Explanations

I would like to raise two issues concerning the main parts of the book. The first issue is about the factivity of subjective reasons, and the second issue is about the explanatory role knowledge may play.

Concerning Part II, a crucial issue is whether we should understand subjective reasons in epistemology as factive or not. Schroeder finds this assumption is problematic, and his reason is that if we accept this assumption, we have to give implausible verdicts on the rationality of belief we gain in epistemically good cases and bad cases.

A couple of initial notes should be made on this point. First, the thesis that subjective reasons in epistemology are factive might somehow accommodate the idea that the rationality of belief in good cases and bad cases are the same, contra Schroeder’s discussion (cf., Ichikawa 2018). Second, some people might reject Schroeder’s preferred verdict on the rationality of belief in good cases and bad cases. Those people might say that external elements unknown to the subject does affect the rationality of the subject’s belief, and the rationality in good cases and bad cases can be different (e.g., Williamson forthcoming). These reactions in epistemology weaken the motivation of Schroeder’s suggestion, and it is interesting to further explore what the best account is that explains the relationship between epistemic rationality and epistemic reasons.

Let us now see the situation in moral philosophy. It might be the case that Schroeder’s inspiration for the non-factive thesis about epistemic reasons comes from his thought on subjective reasons in moral philosophy. Schroeder’s discussion on Williams’ classic gin and tonic (and gasoline) case reveals that he thinks a false belief can be one’s reason for action, and a false belief may rationalise one’s action (pp. 77-78). Suppose Bernie is given a glass by a bartender who says, “here is your gin and tonic”. Bernie believes that his glass is a gin and tonic. However, the glass actually contains gasoline, and Bernie is unaware of this. Schroeder analyses this case in the following way.

The fact that it is gasoline counts against taking a sip, but doesn’t make it more rational for him to not take a sip, because he is unaware of it. Similarly, the fact that he believes that it is a gin and tonic tends to make it more rational for him to take a sip, but it doesn’t help to make taking a sip a good idea, because it is not true […] Since actions can be correct without being rational and rational without being correct, we should therefore not be surprised if there can be subjective reasons without objective reasons, just like there can be objective reasons without subjective reasons. And Bernie’s case, I suggest, shows us that these relations simply cross-cut one another (p. 78).

Thus, according to Schroeder, Bernie’s false belief, a subjective reason in this case, makes it more rational for him to drink the liquid in the glass, while the objective reason against this action, namely the fact that the liquid contains gasoline, does not make Bernie’s not drinking the liquid more rational. I understand Schroeder’s project in Part II of the book is to extend this analysis to the case of epistemic rationality.

But we should note that some people would resist Schroeder’s thoughts on the relationship between (subjective) reasons and rationality in the gin and tonic case. Some would deny the claim that false beliefs play the role of reasons. Maria Alvarez even claims “[one] outcome of those debates [on reasons in the past decades] is the Factive Turn: a shift towards ‘factualism’, the view that reasons for acting or believing or wanting, or other attitudes – what are often called ‘normative’ or ‘justifying’ reasons – are facts”, and factualism is a “consensus” (Alvarez 2018a, p. 161). Importantly, those factualists deny that false beliefs or false propositions play the role of reasons. Why do these factualists think in this way? Alvarez says the following reasoning is one central argument for factualism. Reasons for actions are considerations that favour those actions. A consideration favours an action because it identifies some good-making features of the action. But if a consideration is not factive, it does not favour the action in question since it fails to identify good-making features of the action. Thus, a consideration that favours an action must be a fact (2018a, p. 163). So, on this view, the gin and tonic case should be understood differently from how Schroeder understands it. Bernie has a false belief that the liquid is a gin and tonic, and this false belief does not even play the role of a subjective reason. What Bernie has is merely an “apparent reason” (2018b, p. 3303), but there is no reason that really favours the action (namely, drinking the liquid).

Anti-factualists would rightly wonder whether on this factualist understanding of the gin and tonic case we have to conclude that Bernie’s action is not rational since the action is not supported by any reason, objective or subjective. Anti-factualists would say this conclusion is implausible. Bernie’s action seems to be rational: we shouldn’t criticise Bernie for his irrationality in this case. Alvarez responds that factualists can embrace the idea that Bernie’s action is rational though his false belief does not play the role of a “real reason”. Alvarez says factualists can embrace the idea that acting rationally requires acting for real or apparent reasons that make it reasonable to act as one does (2018b, p. 3308), and by appealing to this principle of rationality, factualists can argue that Bernie’s action is rational since his action is for an apparent reason that makes it reasonable to act in that way.

The discussion above indicates that there may be a different way to respond to Schroeder’s concern about the rationality of action and belief, a way that does not commit us to the denial of factualism both in moral philosophy and epistemology. One could argue that Schroeder’s concern can be accommodated, not by rejecting factivism both in moral philosophy and epistemology, but by constructing a theory of rationality that allows rational actions and beliefs without normative reasons. Or, one could argue that the right verdicts on the rationality of belief in good cases and bad cases are not what Schroeder thinks: the rationality of belief in these cases are different.

Given these suggestions both in epistemology and moral philosophy, it will be illuminating to consider further whether Schroeder’s own suggestion is the best response to the concern he raises. It seems to me Alvarez’s suggestion about rationality is problematic. Alvarez says an apparent reason, a thing that is not a reason, can influence the rationality of relevant action. But my intuition says if x can affect the rationality of my action, x should be a real subjective reason for me to appreciate: a thing that is not even a subjective reason would not influence the rationality of my action. But this observation might not necessarily lead us to the view Schroeder advocates. The observation may lead us to the view that the rationality of action and belief in good cases and bad cases are different. If this claim is accepted, plausible versions of factualism can easily come both in ethics and epistemology. The upshot is that we should make a reservation about Schroeder’s suggestion concerning the (non)factivity of subjective reasons, and further research on this issue both in moral philosophy and epistemology is needed.

Now, let us turn to the discussion on explanations in terms of knowledge. In Part IV, Schroeder considers the issue of the explanatory power of knowledge, originally suggested by Williamson. Williamson’s original case is this.

A burglar spends all night ransacking a house, risking discovery by staying so long. We ask what features of the situation when he entered the house led to that result. A reasonable answer is that he knew that there was a diamond in the house. To say just that he believed truly that there was a diamond in the house would be to give a worse explanation, one whose explanans and explanandum are less closely connected (Williamson 2000, p. 62).

Schroeder agrees with Williamson that knowledge plays an important explanatory role in this case, the role that cannot be played both by true beliefs and justified true beliefs that are not knowledge. Agreeing with Williamson on the explanatory role of knowledge, Schroeder further develops his version of knowledge explanations. Schroeder’s suggestion has two stages. First, he sharpens the explanandum that ought to be the focus of the knowledge explanation of the burglar case. Second, he employs his Kantian theory of knowledge and explains how the explanation in terms of knowledge is better than the explanation in terms of either the burglar’s belief or justified belief.

Concerning the first step, Schroeder writes:

What requires explanation, here, is not what the burglar does, but why he keeps doing it. So the explanatory power that we need from knowledge is an explanation of why someone who knows would continue to believe, even under circumstances like these (p. 232).

Concerning the second step, Schroeder first considers Williamson’s theory of knowledge. One part of Williamson’s theory says that knowledge exhibits safety: if one knows that p, there are no nearby cases in which the process that results in one’s knowledge results in a false conclusion. Schroeder argues that this understanding of knowledge alone cannot fully explicate the explanatory power of knowledge (p. 233). Schroeder then writes:

The Kantian Account […] says that the reasons for which an agent believes must be not only sufficient to make it rational to believe, but must, in addition, outweigh all of the objective reasons against belief— including any objective evidence that the burglar might learn while ransacking the house. So it is no wonder that the burglar spends all night searching. He does so because he still believes the diamond is there even after hours of not finding it; he still believes so because it is rational for him to still believe, and it is rational for him to believe because even before he started, the evidence on the basis of which he believed was good enough to outweigh the evidence that he has acquired in the meantime. His justification stands up to the facts. (p. 234)

I do not have anything to add to the discussion about knowledge explanations Williamson and Schroeder engage with. Magnus and Cohen (2003) argued that Williamson’s discussion fails to show the explanatory relevance of knowledge, and it may be the case that Schroeder’s further development of this issue has not yet met the objection Magnus and Cohen raised. After all, what Schroeder provides is a better explication of a possible explanatory link between the burglar’s knowledge and a certain way of his action. We haven’t been given an argument to the effect that Schroeder’s version of the knowledge explanation is better than some alternative explanations that do not appeal to knowledge. We need to see if Schroeder’s suggestion for the explanatory relevance of knowledge can be a good defence of Williamson’s original suggestion.

Schroeder’s discussion is interesting also for those who work on so-called “moral explanations” in metaethics. Some people argue that moral properties are explanatorily relevant to some empirical phenomena and the explanatory relevance of moral properties is good evidence for moral realism. Here are some examples of moral explanations.

[Moral Wrongness and Moral Judgement]

You see some people are about to ignite an innocent cat on the street and immediately judge that what they are doing is wrong. The best explanation of your moral judgement includes the fact what those people are trying to do is morally wrong and this is why you form this moral belief (Sturgeon 1984, p. 66, cf., Harman 1977).

[Injustice and Social Instability]

The injustice of slavery generated the instability of a society where slavery is employed. The social instability generated the subsequent anti-slavery movements and revolution (Sturgeon 1986, p. 122, Brink 1989, p. 187).

[Virtues/Vices and Actions]

The moral depravity of Hitler resulted in his order of the Holocaust and caused many people’s moral beliefs about him (Sturgeon 1984, p. 68).

[Rudeness and Embarrassment]

In a research seminar, someone rudely shouts, “this is rubbish!” and the other attendances get embarrassed and annoyed with the shout (Harman & Thomson 1996, pp. 81-83).

We should notice an analogous element between the discussions on moral explanations and discussions on knowledge explanations. Williamson suggested that explanations of some relevant phenomena in terms of knowledge are better than explanations in terms of belief or justified belief. The proponents of moral explanations similarly claim that moral properties, such as wrongness, injustice and rudeness, are explanatorily relevant to some phenomena and there will be explanatory losses if we do not have these moral explanations.

Now, for the proponents of moral explanations, how Schroeder argues for the explanatory power of knowledge is illuminating. Schroeder first clarifies what the explanandum needs to be, and then compares which theory of knowledge can provide a better explication of the explanatory relevance of knowledge. It seems we have not seen this way of arguing for the explanatory relevance of moral properties. But such an argumentative strategy might provide strong support for moral explanations and moral realism.

Consider the moral explanation in terms of injustice. A typical discussion on the plausibility of this explanation is of whether the explanation in terms of justice is in any way better than the rival non-moral explanation in terms of relevant facts about people’s psychology and social facts. A typical discussion often assumes that the explanandum in question is the fact that a particular society gets unstable, but this assumption may be sharpened in the way Schroeder does for knowledge explanations. The proponents of the injustice explanation might focus on some particular ways a society in question gets unstable, not instability in general. Furthermore, a typical discussion on the injustice explanation does not make it clear a background theory that accounts for the nature of injustice. Following Schroeder’s way of discussing knowledge explanations, the proponents of the injustice explanation might consider which theory of (in)justice provides the best case for the explanatory relevance of injustice. Does Rawlsian theory provide the best explication of how injustice is relevant to certain ways society gets unstable? Or, is an alternative theory capable of providing a more convincing explication? Through these discussions, the proponents of moral explanations could provide more strong cases, and based on these better cases, the proponents might have better responses to various objections raised in the literature.

Concluding Remarks

Overall, the book is full of interesting discussions and arguments and careful observations of various works recently done both in epistemology and metaethics, or meta-normative enquiry in general.

As the later reflective part of this review essay indicates, giving a closer look at both moral philosophy and epistemology will be illuminating, and we can expect that new interesting developments in both areas of philosophy will come by this approach. Schroeder’s book is certainly a very nice example of this approach.

Needless to say, the book is important reading for anyone with interests in moral philosophy and epistemology, particularly in the issues concerning the normativity and rationality discussed in these areas. Also, this book may be particularly interesting to those who work on the Knowledge-First projects initiated by Williamson, since many of Schroeder’s discussions can be seen as crucial challenges to some of Williamsonian theses.

There is no doubt this book will be a crucial discussion point both in epistemology and moral philosophy and the inter-dialogue between these core parts of philosophy.


I thank Nick Shackel, Masashi Kasaki, and Peter Shiu-Hwa Tsu for their helpful comments on an earlier draft of this article. All errors in this article are attributed to the author.

Further Resources

The Oxford Handbook of Reasons and Normativity (Daniel Star (ed.), 2018, Oxford: Oxford University Press) contains some important survey papers that give us a broad picture of the contemporary discussions on reasons. Schroeder’s own paper in this collection (“The Unity of Reasons”) is also a good start for considering the relationship between the reasons discussed in moral philosophy and the reasons discussed in epistemology.

There are some Japanese references available that engage with the issues discussed in this review. Shunsuke Sugimoto’s paper, 杉本俊介「行為の理由についての論争」, collected in 『メタ倫理学の最前線』(蝶名林亮【編】2019年、勁草書房), can be a good start, and a following exchange between Sugimoto and Abe is also illuminating(杉本俊介「行為の理由、生命倫理、内在主義と外在主義」、安倍里美「理由で倫理学をするということ」, collected in 『豊田工業大学ディスカッションペーパー』vol.20, 「特集『メタ倫理学の最前線』」). Although these papers may be a good introduction to the issues concerning reasons for action, they do not directly discuss the issues concerning the relationship between the reasons discussed in moral philosophy and reasons discussed in epistemology. Some Japanese references on virtue epistemology more directly discuss the relationship between moral philosophy and epistemology. One possible starter is Rie Iizuka’s piece on the moral virtues and the epistemic virtues (飯塚理恵「倫理的徳と認識的徳」, collected in『ワードマップ 心の哲学-新時代の心の科学をめぐる哲学の問い』信原幸弘【編】2017年、新曜社).


iThe page numbers without further reference are all from Schroeder’s book under review in this piece.

iiSchroeder uses the terms “rational” and “justified” interchangeably in the book though he is aware that externalists about justification might emphasises the important difference between these two notions (p. 17n).


  • Alvarez, M. 2018a. False Beliefs and the Reasons We Don’t Have. in Mitova, V. (ed.), pp. 161-176.
  • Alvarez, M. 2018b. Reasons for Action, Acting for Reasons, and Rationality. Synthese. vol. 195, 8, pp. 3293-3310.
  • Brink, D. 1989. Moral Realism and the Foundations of Ethics. Cambridge: Cambridge University Press.
  • Broome, J. 2018. Reason Fundamentalism and What Is Wrong With It., in Star, D. (ed.), Oxford Handbook of Reasons and Normativity. Oxford: Oxford University Press, pp. 297-318.
  • Feldman, R. 1988. Having Evidence. Reprinted in Conee and Feldman (2004). Evidentialism. Oxford: Oxford University Press, pp. 219-241.
  • Harman, G. 1977. The Nature of Morality. Oxford: Oxford University Press.
  • Harman, G. 2002. Practical Aspects of Theoretical Reasoning. in Mele and Rawling, (eds.), The Oxford Handbook to Rationality. Oxford: Oxford University Press, pp. 45-56.
  • Harman, G. & Thomson, J. 1996. Moral Relativism and Moral Objectivity. New Jersey: Wiley-Blackwell.
  • Ichikawa, J. J. 2018. Internalism, Factivity, and Sufficient Reason. in Mitova, V. (ed.), pp. 66-83.
  • Kant, I. 1785/1996. Groundwork of the Metaphysics of Morals. in Gregor, M. (ed,), Immanuel Kant: Practical Philosophy. Cambridge: Cambridge University Press.
  • Magnus, P.D. and Cohen, J. 2003. Williamson on Knowledge and Psychological Explanation. Philosophical Studies, 116 (1), pp. 37-52.
  • Mitova, V. (ed.), 2018. The Factive Turn in Epistemology. Cambridge, Cambridge University Press.
  • Parfit, D. 1997. Equality and Priority. Ratio, 10(3), pp. 202-221.
  • Scanlon. T.M. 2014. Being Realistic About Reasons. Oxford: Oxford University Press.
  • Sturgeon, N. 1984. Moral Explanations. in Copp & Zimmerman (eds.), Morality, Reason and Truth. Totowa, NJ: Rowman and Allanheld, pp. 49-78.
  • Williamson, T. 2000. Knowledge and Its Limits. Oxford: Oxford University Press.
  • Williamson, T. 2007. On Being Justified in One’s Head. In Timmons, Greco and Mele, (eds.), Rationality and the Good. Oxford: Oxford University Press, pp. 106-122.
  • Williamson, T. Forthcoming. Justifications, Excuses, and Sceptical Scenarios. in Dorsch, F. and Duntant, J.(eds.), The New Evil Demon. Oxford: Oxford University Press.

Publisher's Official Website

Oxford University Press

About the Author

Ryo Chonabayashi

Ryo Chonabayashi is an associate professor in philosophy at Soka University, Japan. His primary research interest lies in metaethics. His publication includes Can Ethics be Science? A Defence of Naturalistic Metaethical Theory (2016, Keiso Shobo [Japanese]), The Frontline of Metaethics (2019, Keiso Shobo [ed, Japanese]).