Epistemic Worth

Abstract

Actions can have, or lack, moral worth. When a person’s action is morally worthy, she not only acts rightly, but does so in a way that reflects well on her and in such a way that she is creditable for doing what is right. In this paper, I introduce an analogue of moral worth that applies to belief, which I call epistemic worth. When a person’s belief is epistemically worthy, she not only believes rightly, but does so in a way that reflects well on her and in such a way that she is creditable for believing what is right. While the notion of epistemic worth is independently interesting, the main aim is to show that appealing to it provides a response to arguments against the view that truth is the fundamental norm for belief and, thereby, to arguments for the view that knowledge is the fundamental norm for belief. The direction of travel does not only run from ethics to epistemology. In closing, I tentatively suggest that some of the points to emerge when developing the account of epistemic worth might prompt revisions to the account of moral worth.

1. Introduction

Actions can have, or lack, moral worth. When a person’s action is morally worthy, she not only acts rightly, but does so in a way that reflects well on her and in such a way that she is creditable for doing what is right. In this paper, I develop and defend an analogue of the notion of moral worth that applies to belief, which I call epistemic worth. When a person’s belief is epistemically worthy, she not only believes rightly, but does so in a way that reflects well on her and in such a way that she is creditable for believing what is right. The notion of epistemic worth, I suggest, is of interest to both epistemologists and ethicists.

To the epistemologists: the account of epistemic worth bears on debates about the norms governing belief. Consider:

According to Truth, it is right for Miyuki to believe that Francis Ford Coppola directed The Conversation, since that is true, and it is not right for her to believe that he directed Jaws, since that is not true.

A number of philosophers, myself included (Whiting 2010; 2012; 2013a; 2013b), defend the idea that a norm of truth of this sort governs belief.[1] Several go further, myself included, and claim that Truth is fundamental in the sense that other norms to which believing is subject are derived from or explained by it. On this view, as one might put it, truth is the fundamental right-maker for belief. Clayton Littlejohn presents proponents of this view (which he dubs ‘truth-first’) with a challenge:

To illustrate, suppose that Miyuki believes that Coppola directed The Conversation but for no reason whatsoever—due only to some dogmatic conviction, say. Or suppose that she believes it on the basis that, if Coppola directed The Conversation, then he directed Rumble Fish, and he directed Rumble Fish. In these cases, Miyuki is criticisable. Epistemically speaking, all is not well. How can a proponent of Truth explain this? After all, Miyuki’s belief that Coppola directed The Conversation is true. So, it satisfies Truth.

Littlejohn thinks that proponents of Truth are unable to meet this challenge. In place of Truth, and following Timothy Williamson (2000), he suggests that knowledge is the fundamental norm of belief:[3]

Plausibly, if Miyuki believes that Coppola directed The Conversation for no reason, or for bad reasons, she lacks knowledge. So, a proponent of Knowledge can easily explain why Miyuki is open to criticism: it is not right for her to believe what she does. Alternatively: she violates the norm governing belief.

One response to Littlejohn’s challenge is to deny that epistemic assessment has what he calls an inward-looking focus. That seems to me hopeless. In any event, it is not the strategy I pursue. While I take the account of epistemic worth in broad outline to be one that is available to all parties, my aim is to show how it can be filled in in such a way that, by appeal to it, a proponent of Truth can explain the inward-looking focus of epistemic assessment. I go on to suggest that in the same way, and given some additional assumptions, a proponent of Truth can account for other data that—again, following Williamson (2000)—many take to support the view that Knowledge is the fundamental norm for belief. The discussion serves, not only to defend Truth, but also to outline an appealing view of epistemic assessment, one that promises to dovetail with an independently plausible view of ethical assessment.

To the ethicists: it is interesting to consider whether and how the notion of moral worth might generalize to other domains, not least because we might learn some lessons about moral worth by reflecting on its epistemic counterpart. While the primary focus in what follows is the development of the notion of epistemic worth and its application to the debate concerning the norms of belief, in closing I will suggest (in an exploratory spirit) that thinking about what makes for epistemically worthy belief might prompt revisions to the account of what makes for morally worthy action.

2. Preliminaries

Before proceeding to the main discussion, I will address some preliminary matters.

First, I do not deny that knowledge is a norm for belief. Indeed, and to anticipate, I suspect that knowledge is a measure of epistemic worth. More generally, all parties can and should agree that there are various standards for assessing beliefs—truth and knowledge among them. The issue at hand is one of relative priority; specifically, it concerns whether Truth is explanatorily prior to whatever norm knowledge figures in. I suggest that it is.[4]

Second, according to both Truth and Knowledge, it is only right for a person to believe a proposition when it is true. Some disagree (e.g., Feldman 1988). I will not attempt here to defend the point on which proponents of Truth and Knowledge agree, which is a task for another occasion.

Third, I try to show how, by appeal to Truth, one might derive some other more or less demanding norm to which belief is subject. I am not the first to do so.[5] However, the other attempts are not directed toward the distinctive challenge Littlejohn presents; that is, they are not trying to account for epistemic assessment’s focus on the reasons for which a person believes. The aim instead is to explain the sense in which it might be acceptable from an epistemic point of view to believe what is false, or to arrive at a norm that might serve as a guide in belief-formation.[6] Where the explanatory aims are similar, the views advanced are significantly different from the one I develop here. Rather than criticising alternatives, I will focus on motivating a positive proposal and exploring its significance.

Finally, I take no stand here on the relationship between (what I call) epistemic worth and epistemic justification. There are ongoing debates concerning what it is for a belief to be justified, debates which can appear intractable and which I am keen to sidestep. Framing the discussion in terms of worthiness serves to reveal instructive parallels between the epistemic and ethical domains which might otherwise be invisible. Should it turn out that a belief is worthy if, or only if, it is justified, that would be an interesting discovery; it is not something I assume from the outset or something I defend in what follows.[7]

3. Moral Worth

A familiar thought in ethics is that a person can do the morally right thing and yet her action might lack moral worth. Alternatively: a person can act rightly and yet not deserve credit for doing so. Alternatively again: a person might do what is right but not in a way which reflects well on her.

Kant’s shopkeeper famously illustrates this (1785/1997: 4:397). The shopkeeper does not overcharge his customers. He is right not to do so—it is fair. However, the shopkeeper refrains from overcharging, not for reasons of fairness, but for reasons of profit. The shopkeeper knows that, by keeping prices low, customers will return, and for that reason does not overcharge. In this case, while the shopkeeper does what is morally right, his action lacks moral worth; he does not deserve credit for doing the right thing. In acting on the basis that he does, the shopkeeper manifests a concern for profit, not for fairness.[8]

A common diagnosis of why the shopkeeper’s action lacks moral worth is that, given the reasons for which he acts, it is an accident that he acts rightly (cf. Herman 1981). In acting only for reasons of profit, the shopkeeper could easily fail to act in a way which is morally right. It is a matter of luck that, in acting on the basis that he does, the shopkeeper does what is right. As Kant puts it, if an action is performed on non-moral grounds, its conformity to the ‘moral law’ is ‘only very contingent and precarious’ (1785/1997: 4:390).

This line of thought suggests that, for an action to be morally worthy, there must be a non-accidental connection between the reasons for which a person acts and her doing the right thing. In turn, this points to the following principle:[9]

Some comments. First, an action might be right in one respect—say, that it is fair—but not right in another respect—say, that it is harmful. By the same token, an action might be right in some respect but not right overall. In a similar fashion, a person’s action might be morally worthy in some respect—say, that it manifests concern for fairness—but not in all respects—say, that it manifests indifference to harm. By the same token, a person’s action might be morally worthy in some respect but not overall. Moral Worth₁ concerns when a person’s action is morally worthy in some respect. It is an interesting question how to arrive at an account of overall moral worth, but not one I need to address here.[10]

Second, being morally worthy is here a matter of being creditable for doing the right thing. No doubt there are cases in which a person is praiseworthy in some way despite doing the wrong thing. Suppose that Macie promises to meet a friend at 7pm. She receives a text from her friend telling her that she will be at the cinema at 7pm. Macie goes to the cinema but her friend is not there—she was unavoidably delayed. In this case, due to blameless ignorance, Macie breaks her promise. There is nonetheless reason to think well of Macie—she manifests a concern for fidelity. She might deserve credit for her efforts—what she does not deserve credit for is doing the right thing, since she did the wrong thing. Hence, her action lacks (what I and others call) moral worth.[11]

Third, while rightness is here a deontic notion, worthiness is an evaluative notion. Being morally worthy is a good thing. For present purposes, I do not need to explain why. That said, a plausible thought is that, if a person acts in a morally worthy fashion, she is appropriately oriented in relation to goods such as honesty or fidelity, and in general a proper orientation to what is good is itself good (cf. Hurka 2001).

Finally, I take no stand here on the first-order, substantive issue of what sorts of considerations make it right to act. In places, I will make some remarks along these lines but for illustrative purposes only.[12]

A virtue of Moral Worth₁ is that it delivers the desired verdict in cases like Kant’s shopkeeper. His action is not morally worthy because the reasons for which he refrains from overcharging, reasons of profit, are not reasons which, individually or collectively, make it right not to overcharge.

Consider, however, a revised version of that case.[13] The shopkeeper refrains from overcharging for the reason that it is fair. The shopkeeper acts on that basis because he thinks that, if he does what is fair, customers will return. That is, the shopkeeper is only concerned with fairness insofar as it is conducive to profit. In this version of the case, the shopkeeper’s reason for acting is a moral reason, and yet his action lacks moral worth. As before, it is an accident of sorts that the shopkeeper does what is right. Were profit to point in a different direction to fairness, the shopkeeper would act wrongly.

This suggests a further revision:[14]

Talk of acting on a right-maker as such is shorthand for a person’s acting for a reason because or in virtue of the fact that it makes so acting right. I will say more later about what this might involve. For now, it is enough to have an intuitive grip on that notion of the sort reflection on cases like the above affords.

Moral Worth₂ gets the right result in both the original shopkeeper example and the revised version. Consider, however, the following.[15] Ananya reads in the newspaper that there is a famine in a certain war-torn country. On the basis of the newspaper report, Ananya donates money to charity, which is the right thing for her to do. In this case, it seems that Ananya’s action might be morally worthy, at least in some respect—she might manifest concern for suffering. However, Ananya’s reason for donating, namely, that the newspaper reports that there is a famine, is not among the reasons why it is right for her to do so. What makes it right for her to donate are facts such as that, by doing so, she relieves suffering. The newspaper report does not stand in an explanatory relation to the rightness of her act but an evidential one.[16]

In response, one might suggest that Ananya’s reason for donating is that there is a famine, rather than that the newspaper reports that there is a famine. In that case, Ananya’s reason for acting might be among the reasons why it is right to do so. In that case, in turn, there is no counterexample to Moral Worth₂.

However, I can just stipulate that Ananya’s reason is that the newspaper reports that there is a famine. In that case, her action might still possess moral worth. However, I do not want to get too hung up on the details of the case; inevitably, it is underdescribed and there are various ways of fleshing it out. The example serves to illustrate an independently plausible point, namely, that a person can deserve credit for doing the right thing by acting in light of or on the basis of evidence which reveals or indicates respects in which acting is right, but which does not explain why it is right in those or any other ways.

This suggests a revision to the account of moral worth:

Assuming, in line with most conceptions of evidence (cf. Hawthorne & Magidor 2018: 137), that a proposition is evidence for itself, this formulation allows that a person’s action might possess moral worth if she acts on the basis of a right-making consideration, say, that it keeps a promise. That consideration is (trivially) evidence of a right-making consideration.[17]

A virtue of Moral Worth₃ is that it gets the right result in Ananya’s case. The newspaper report is evidence that, by donating, she will relieve suffering. That donating will relieve suffering is why it is morally right in a respect for her to do so. So, Moral Worth₃ allows that Ananya’s donation has moral worth in some respect.

According to Moral Worth₃, morally worthy actions are a response, not to evidence that acting is right in some respect, but to evidence of some respect in which acting is right, say, that it keeps a promise. Arguably, a person motivated by evidence of the moral status of acting, rather than by evidence of features of the situation in virtue of which it has that status, is criticisable, rather than praiseworthy. She displays what Michael Smith calls ‘moral fetishism’ (1994: 75; cf. Markovits 2010: 204).[18]

Moral Worth₃ is a step in the right direction. I consider later whether it needs further revision. For now, it serves my immediate purpose, which is to introduce an analogous notion of epistemic worth.

4. Epistemic Worth

As the previous section reminds us, ethical assessment has, to adapt the language Littlejohn uses, an ‘outward-looking’ dimension, which focuses on the relationship between what a person does and the facts, and an ‘inward-looking’ dimension, which focuses on the relationship between what a person does and her reasons for doing it.[19] In this section, I suggest that one can find the same structure in the epistemic domain.

It is straightforward to revise Moral Worth₃ to arrive at an epistemic counterpart:

The notion of epistemic worth is an unfamiliar one. It is uncommon to say such things as, ‘Miyuki’s belief has worth’. Recall, however, that to say that an action is morally worthy is just to say that a person deserves credit for doing what is morally right. So, to say that a belief is epistemically worthy is, here, just to say that a person deserves credit for believing what is epistemically right. This idea is natural enough. It would not be odd to say such things as, ‘Miyuki deserves credit for believing what she does’.[21]

Like the accounts of moral worth, Epistemic Worth₁ is a meta-normative proposal and, as such, it is neutral with respect to first-order normative debates in epistemology. To appreciate this, consider a toy theory according to which one way in which it is right to believe a proposition is that it coheres with popular opinion. Suppose that Boris has evidence that some proposition coheres with popular opinion, say, that the UK sends the EU £350m each week. If Boris believes that the UK sends the EU £350m each week on the basis of that evidence, as evidence of coherence with popular opinion, it follows from Epistemic Worth₁ that he believes in an epistemically worthy fashion. Of course, this is an implausible verdict, but that is because the toy theory is implausible. In what follows, I will consider more serious contenders.

As mentioned above, I take no stand here on the substantive issue of what the right-makers for action are. But, as a proponent of Truth, I do take a stand on the corresponding issue of what the right-makers for belief are. As discussed above, an action might be right in one respect but not right in another or overall. The same, I suggest, does not hold for belief. There is only one respect in which it is right to believe a proposition, namely, that it is true, and only one respect in which it is not right to believe a proposition, namely, that it is not true. Moreover, a proposition cannot be true in one respect but not true in another. So, when it is right in some respect to believe a proposition, it is right overall to do so, and vice versa.[22] This is the thought, I suggest, which Truth captures.

Since there is only one respect in which it is right to believe a proposition, there is only one respect in which so believing is epistemically worthy, namely, that it manifests concern for the truth or, more carefully, for whether the proposition is true or false.

In view of this, I can restate the above principle as follows:[23]

This proposal raises a question. What does the evidence on the basis of which a person believes have to be like for her believing to have epistemic worth? A proponent of Epistemic Worth₂ might dismiss this query. In general, it is not a condition on an analysis being informative that it be accompanied by an analysis of the analysans. The point is well taken. But, as I will show, plugging different conceptions of evidence into Epistemic Worth₂ delivers very different results. How plausible the account is, then, turns on what conception of evidence is operative.

A tempting suggestion is that, for a belief to possess epistemic worth, the evidence on which it is held must make it highly probable that the proposition is true. To see that this is false, consider a case of a sort familiar from recent epistemology. Lily holds a ticket in a fair lottery. The winning ticket has been drawn but she has yet to learn the result. As Lily knows, it is highly probable that her ticket lost, say, 0.999. On this basis, Lily believes that her ticket lost. As it happens, this is true. So, given Truth, it is right for Lily to believe that her ticket lost. Nonetheless, Lily does not deserve credit for believing what it is right for her to believe, that is, for believing the truth. Hence, her belief lacks epistemic worth.

Again, the case is underdescribed. But a common and independently plausible suggestion is that, if a subject believes that her ticket lost only on the basis of statistical evidence or evidence which makes it likely that her ticket lost, it is an accident or matter of luck when she believes the truth (cf. Greco 2004; Pritchard 2005). In the above case, Lily could easily have formed a false belief that her ticket lost on the very same evidence. If accidentality precludes moral worth, accidentality precludes epistemic worth.

To put the same thought differently, the Kantian insight is that, for an action to possess moral worth, there must be a modally robust connection between the basis on which a person acts and her doing the right thing. Epistemic worth is nothing but the epistemic counterpart to moral worth. So, for a belief to possess epistemic worth, there must be a modally robust connection between the basis on which a person believes and her believing the right thing. In lottery cases, the connection between the statistical evidence and the truth of the belief is not modally robust. So, the belief lacks epistemic worth.

This suggests that believing the truth on the basis of evidence that only makes the truth of the proposition highly probable is insufficient for epistemic worth. But what is the alternative?[24] Again, what is required is a modally robust connection between the evidence for a proposition and its truth. Fortunately, a connection of this sort is the subject of considerable discussion in the recent literature. Consider the notion of safe evidence.[25] Evidence for a proposition is safe just in case, in the nearby possible worlds in which that evidence obtains, that proposition is true.[26] If evidence for a proposition is safe, then, given the evidence, that proposition could not easily be false.

Importantly, evidence which makes a proposition highly probable need not be safe. Lottery cases illustrate this. While Lily’s ticket lost in most nearby worlds, there are nearby worlds in which it won.

On the assumption that evidence consist of truths, the notion of safe evidence is factive in the sense that, if there is safe evidence for a proposition, that proposition is true (cf. Martin Smith 2016: 106). To see this, suppose that E is safe evidence for P. In that case, P is true in nearby worlds in which E obtains. Since E is true, E obtains in this world. This world is trivially among the nearby worlds. So, if E is safe evidence for P, P is true in this world.

Given this, I will restate the account of epistemic worth as follows:

This account gets the right result in Lily’s case. While she is right to believe that her ticket lost, she is not creditable for this, because the evidence on the basis of which she believes is unsafe.

Since the notion of safe evidence is factive in the above sense, the notion of epistemic worth this principle captures is also factive: a belief is epistemically worthy only if it is true. One might object: surely false beliefs can be epistemically worthy. Consider again Macie who receives a text from her friend telling her that she will be at the cinema at 7pm. If on that basis Macie believes that her friend will be at the cinema at 7pm, she might deserve praise from an epistemic point of view, even if it turns out that, for reasons beyond her ken, the friend is unavoidably delayed.

To respond in a way that echoes the discussion of moral worth, being epistemically worthy is here (a label for) being creditable for believing what is right, hence, true. No doubt we can credit a subject like Macie with a great deal—including regard for the truth—but we cannot credit her with believing what is right since, given Truth, she does not do so.

Similar remarks apply to a case in which a person believes what is true on the basis of unsafe evidence that nonetheless appears (in some sense) to her to be safe. Her belief, I claim, lacks epistemic worth. That is not to deny that her belief might reflect well on her in some way. The claim is only that, since it is a happy accident that her concern led her on this occasion to believe what is true, she does not deserve credit for believing what is right, even if she deserves credit for other things. Hence, as I use the label here, her belief lacks epistemic worth.[27]

To see that the inclusion of ‘as such’ in Epistemic Worth is well motivated, consider an epistemic analogue of the revised shopkeeper case.[28] Pangloss knows that the bottle is half-empty. This is safe evidence that the glass is half-full; that is, in the nearby worlds in which the bottle is half-empty, the glass is half-full. However, Pangloss is an incurable optimist. On the basis of any evidence whatsoever, no matter how flimsy or fragile, he concludes that the glass is half-full. Had the bottle been empty, Pangloss would still have believed on that basis that the glass is half-full. In this case, Pangloss’s belief is not epistemically worthy—he does not deserve credit for believing truly that the glass is half-full.

This verdict accords with Epistemic Worth. While Pangloss believes on the basis of what happens to be safe evidence for the truth of his belief, he does not believe on that basis because or in virtue of the fact that it is safe evidence.

5. Epistemic Assessment

I now return to Littlejohn’s challenge. His objection, recall, is that the ‘truth-first’ view cannot explain why epistemic assessment has its ‘inward-looking focus’, that is, why such assessment focuses on the reasons for which a person believes. By appeal to Epistemic Worth, a proponent of Truth can explain this, as I will now show.

Epistemic assessment, like ethical assessment, has both an ‘outward-looking’ dimension, which focuses on the relationship between a person’s belief and the facts, and an ‘inward-looking’ dimension, which focuses on the relationship between a person’s belief and her evidence. Truth captures the outward-looking dimension; according to it, a person is right to believe a proposition just in case there is a correspondence between what she believes and the facts. Epistemic Worth, I suggest, captures the inward-looking dimension.

Miyuki, Lily, and Pangloss in the above cases all believe the truth. Nonetheless, they are criticisable. The criticism in each case does not concern the rightness of their beliefs but their worthiness—the subjects believe rightly but do not deserve credit for doing so. Either they do not believe on the basis of evidence, or they believe on the basis of unsafe evidence, or they believe on the basis of safe evidence but not because it is safe evidence. In each case, and in different ways, it is a matter of luck or accident that the subject believes what it is right to believe, that is, what is true.

Consider instead Agatha. She knows that Smith’s fingerprints are on the gun. This is safe evidence that Smith is the murderer; that is, in the nearby worlds in which Smith’s fingerprints are on the gun, Smith is the murderer. On the basis of her evidence, and because it is safe to do so, Agatha believes that Smith is the murderer. In this case, given Epistemic Worth, Agatha’s belief has epistemic worth; she deserves credit for believing the right thing. In believing what she does and as she does, Agatha manifests a concern for the truth.

So, the proponent of Truth can not only accept that there is an inward-looking dimension to epistemic assessment but also explain the standard that delivers that assessment by combining her first-order substantive commitment concerning what makes for right belief with a suitably refined meta-normative account of worthiness. In this way, by appeal to Epistemic Worth, a proponent of Truth can meet Littlejohn’s explanatory challenge. In doing so, she blocks a case for thinking that Knowledge, not Truth, is the fundamental norm for belief.[29]

I take the preceding, not only to provide a response to the challenge Littlejohn raises for the ‘truth-first’ view, but to reveal that view, as I have developed it, to be an attractive one, insofar as it delivers an attractive account of epistemic assessment, one which reveals it to mirror ethical assessment (in structure, not substance). In this way, the preceding amounts to (part of) a positive case for thinking that Truth really is the fundamental norm for belief.

One might object to the view I put forward on the following grounds. According to it, there are two standards to which belief is subject: Truth and Epistemic Worth. One might satisfy the former without satisfying the latter. In that case, don’t the norms conflict?

They do not, since they concern different statuses. The first is a deontic status—which I call rightness—and the second is an evaluative status—which I call worthiness. If a person believes in an epistemically worthy fashion, she is appropriately oriented toward the truth, which is good (cf. Sylvan 2012). If a person believes in an epistemically unworthy fashion, she is not so oriented, which is not good, or at least not as good.

However, one might press the concern by asking how these norms could guide deliberation. Don’t they pull in different directions? This assumes a connection between general principles such as Truth and Epistemic Worth, on the one hand, and deliberation, on the other, which the proponent of those principles need not accept. She can—I think, should—suggest that what guides deliberation as to whether to φ are considerations which provide evidence of features of the situation which make φing right, not considerations bearing on whether φing is right or, for that matter, worthy.

A different way to object is to point out that, while the view I have developed speaks to Littlejohn’s challenge as originally presented, that is, in the remarks quoted at the outset, Littlejohn states that challenge in other terms and it is far from clear that I have spoken to that version of it. Consider:

Above, I agreed that all is not well, epistemically speaking, if a person believes a proposition for bad reasons or for no reasons, even when that proposition is true. Such a person is criticisable from the epistemic point of view. However, Littlejohn’s thought seems really to be that it is simply not right, but instead wrong, for a person to believe a proposition unless she believes it for good reasons, that such a person has violated her ‘epistemic obligations’. His challenge to the defender of the idea that Truth is the fundamental norm for belief is to account for this.

It is clear that the appeal to the notion of epistemic worth does not answer this version of the challenge. After all, the point of that appeal is to explain in what way a person might be criticisable for believing, even though she is right (by the lights of Truth) to do so. However, I do not think this version of the challenge deserves to be taken so seriously.

Merely to assert that epistemic assessments of rightness have an inward-looking focus is to beg the question against Truth, which states otherwise. One might think that it is just intuitive to think that it is not right to believe a proposition merely when that proposition is true. But I take it that advocates of Truth, myself included, do not share that intuition. Moreover, to the extent that a person possesses these intuitions, I can explain them away as tracking, not rightness, but worthiness, a distinction that is already widely recognized as holding in the moral domain.

Of course, proponents of the idea that Knowledge is more fundamental than Truth, including Littlejohn, have presented various lines of thought which might seem to support their view. However, those lines of thought directly challenge Truth. In that case, they do not serve to bolster Littlejohn’s challenge, so much as to supplant it. Be that as it may, in the next section I will explain how, by combining Truth with an account of epistemic worth, it is possible to account for the data which might otherwise seem to speak against Truth and for Knowledge.

6. Worthy Belief as Knowledge

An influential argument against Truth, due to Williamson (2000: 244–249), appeals to lottery cases.[30] He argues that, if a person flat-out believes that her ticket lost, she is criticisable. A proponent of Truth, Williamson suggests, cannot explain this. After all, her belief might be true. Even if the defender of Truth suggests that that norm generates some derivative evidential norm, he continues, she cannot explain why lottery beliefs are criticisable. After all, on the evidence, such beliefs are highly likely to be true.

This line of thought is intended, not only to undermine Truth, but to support Knowledge. On the assumption that a person cannot know the lottery proposition, Knowledge straightforwardly explains why she is criticisable for believing it: she is wrong to do so.

Given Epistemic Worth, a proponent of Truth can explain why subjects are criticisable for believing lottery propositions. As noted above, such subjects are not creditable for believing what is true, hence, for believing what it is right to believe. Their beliefs lack epistemic worth. So, reflection on lottery cases does not count against the view that Truth is the fundamental norm for belief.

Williamson makes two additional points in support of Knowledge and against Truth. First, it is commonplace to criticize a person’s beliefs in terms of knowledge (Williamson 2000: 256). If Miyuki expresses her belief that Coppola directed Rumble Fish, Nishi might respond by saying, ‘You don’t know that!’ or ‘How do you know that?’ If Truth is the norm for belief, one might think, it is hard to understand why these remarks constitute challenges. If Knowledge is the norm for belief, in contrast, this is easy to understand.

Second, consider Moorean propositions (Williamson 2000: 253–254), such as:[31]

It is absurd to believe Moore.[32] But Truth cannot explain this. After all, both conjuncts might be true. So, as far as Truth is concerned, I might be right to believe Moore. In contrast, Knowledge can account for the absurdity here. It is not possible to believe Moore while satisfying the norm to which belief is subject. If I know the first conjunct, I do not know the second. If I know the second conjunct, I do not know the first. So, given Knowledge, it is never right to believe Moore.

A full discussion of these points is beyond the scope of this paper.[33] But I do want to outline a line of response. Consider the view that safe belief is knowledge.[34] Here is one—not the only—way to capture this idea:

Given Epistemic Worth, S⇒K is equivalent to:[35]

On this view, a person knows a proposition when she believes the truth and is creditable for doing so.[36]

To accept S⇒K is not to commit to the view that knowledge can be analysed in terms of believing on safe evidence (cf. Williamson 2000). In a similar fashion, one can consistently advance EW⇒K and maintain that knowledge is not analysable in terms of epistemically worthy belief.[37] It is a well-worn point that entailment relations need not be explanatory relations.

Given EW⇒K, one can explain the data Williamson adduces. First, in saying, ‘How do you know?’ or ‘You don’t know that’, Nishi questions whether Miyuki’s belief has epistemic worth. Second, it is not possible to believe Moore in an epistemically worthy fashion. If my belief in the second conjunct possesses epistemic worth, my belief in the first conjunct lacks it, and vice versa.

The more general point here is that, armed with S⇒K, hence, EW⇒K, a proponent of Truth can agree with her opponent that knowledge provides a standard for believing, by reference to which one might criticize or challenge a person’s beliefs. But, she will insist, knowledge is a standard, not of right belief, but of worthy belief.

Of course, EW⇒K is controversial, at least insofar as S⇒K is controversial. The post-Gettier (1963) literature serves as a reminder of the difficulties facing attempts to specify sufficient conditions for knowledge. It might seem audacious to make such an attempt here.

Needless to say, it is neither feasible nor desirable to consider here all possible Gettier-style counterexamples to S⇒K, hence, EW⇒K. But I will offer some reasons for thinking that S⇒K is not vulnerable to them.

One might think that one can generate a counterexample to safety-based accounts of knowledge in general, and S⇒K in particular, simply by taking a standard Gettier case and making it modally robust (cf. Lackey 2006: 289).[38] Consider this version of Gettier’s (1963) original example:

In this case, Smith’s belief is true but not knowledge. However, one might think, Smith’s evidence is safe. In nearby worlds in which Smith’s boss says that Jones will get the job, the person who will get the job has ten coins in their pocket. The impish pickpocket ensures this.

In this case, while Smith believes on the basis of what happens to be safe evidence, she does not do so because or in virtue of the fact that it is safe evidence. After all, the feature of the situation which makes the evidence safe, the presence of the pickpocket, is not a feature to which Smith is sensitive. Smith would have formed the same belief on the same basis had the pickpocket been absent or disinclined to interfere. So, the above case is not a counterexample to S⇒K, hence, to EW⇒K.

If the aim were to give an account of the nature of knowledge, one might think that it is incumbent on me to give an account of what it is to respond to safe evidence as such. I am not sure about that but, in any case and once more, that is not the aim. The aim, instead, is to show how, by appeal to a connection between the status of knowledge and that of epistemically worthy belief, one might defend Truth and resist Knowledge. For that purpose, all that matters is that we have a grip on the idea of a person, not merely responding to safe evidence of truth, but responding to its being safe evidence. For all I say here, it might be that grasp of this depends on an independent grasp of what it is know a proposition.

However, the reader might not press for an analysis of the notion expressed by ‘as such’ but simply some gloss on it to inform judgements about cases like the one above. This is a reasonable request.

A promising way to think about responding to safe evidence because it is safe evidence is in terms of the exercise of abilities (powers, capacities).[39] The idea is that a person might have the ability to form a belief on a certain basis when—and only when—it is safe evidence for the truth of that belief. Crucially, this is a competence the manifestation condition of which is the presence of safe evidence—its exercise is trigged only by safe evidence.

To return to the above case, it is clear that Smith does not manifest such an ability. While she believes on the basis of what is in fact safe evidence, she would have made the inference on the same basis had the pickpocket not been involved, hence, had the evidence been unsafe. Insofar as it is dependent on the intervention of the pickpocket, it is not the expression of a competence on Smith’s part. Since Smith is not manifesting the ability to believe only in response to safe evidence, she is not believing on the basis of safe evidence as such.

Before moving on, I will consider a final objection. Why not take the standard of epistemic worth to be fundamental and the standard of truth to be derivative? As one might put it, rather than think of worthy belief as a matter of right belief plus certain conditions, one might think of right belief as worthy belief minus certain conditions.

I suggest that we cannot settle this issue while focusing only on the epistemic domain. We need to think about how an account of epistemic assessment fits into a broader view that encompasses assessment in other domains, for example, moral.[40] If moral rightness is explanatory prior to moral worthiness, then we should expect epistemic rightness to be explanatorily prior to epistemic worthiness. And, indeed, it is the dominant view in ethics that moral rightness comes first in the order of explanation.[41] This is highly plausible. We think well of Ananya in the above case because she exhibits a sensitivity to what fundamentally matters morally: wellbeing. Similarly, I suggest, we think well of Agatha in the above case because she exhibits a sensitivity to what fundamentally matters epistemically: truth.

This might not convince. Nonetheless, first, someone suggesting that the order of explanation runs in the other direction in the epistemic domain owes an account of how exactly the standard of worthiness generates the standard of rightness. Of course, worthy belief as characterized here entails true belief. But why should (mere) true belief be of normative significance? Compare: worthy belief entails belief, but (mere) belief does not enjoy a positive epistemic status. Second, they then need to defend non-orthodox views about the order of explanation in the moral domain or explain why we should not find it puzzling that there is a structural mismatch between the epistemic and moral domains.

In view of these points, and pending further details from the proponent of the opposing view, there is a presumption in favour of thinking that epistemic rightness has explanatory priority over epistemic worthiness.

7. Moral Worth Revisited

The notion of safe evidence is familiar in epistemology. It is unfamiliar in ethics. Before concluding, I will briefly raise the prospect that that notion might have a place in an account of moral worth. Recall:

One might ask what the evidence on the basis of which a person acts must be like for her action to have moral worth. Once again, a tempting answer is that the evidence must make it highly likely that the relevant right-making feature obtains. Consider, however, a practical analogue of the lottery case. Macie makes a promise. She knows that, if she pulls a certain lever, it is highly probable (say, 0.999 on her evidence) that it will result in keeping the promise, although there is a small chance (0.001) that pulling the lever will result in breaking the promise.[42] On the basis of what she knows, Macie pulls the lever and keeps the promise.

In this case, Macie does the right thing—she keeps her promise—and does so on the basis of evidence which makes this highly probable. However, her action lacks moral worth. More fully, Macie does not deserve credit for keeping her promise. After all, she could very easily have done the same thing on the same basis and failed to keep her promise. Indeed, one might think that Macie is morally criticisable: she manifests a lack of concern for fidelity. In the relevant sense, it is a matter of luck or accident that she keeps her promise.

Cases like this suggest that, for an action based on certain evidence to possess moral worth, there must be a modally robust connection between that evidence and whatever makes that action right. This in turn suggests that the notion of safe evidence might figure in an account of moral worth, just as it does in an account of epistemic worth.

Since my primary concern lies with matters epistemic, and since an adequate development and assessment of a safety-based account of moral worth is beyond the scope of this paper, I leave this as a pointer for future research.

8. Conclusion

Whether it is morally right for a person to act depends on the facts, for example, on whether what she does alleviates suffering or keeps a promise. Likewise, whether it is epistemically right for a person to believe depends on the facts, in particular, on whether what she believes is true. Whether a person’s action has moral worth, that is, whether she deserves credit for acting rightly, depends on the basis on which she acts. Likewise, whether a person’s belief has epistemic worth, that is, whether she deserves credit for believing rightly, depends on the basis on which she believes. In each case, and more specifically, worthiness depends on whether the basis on which a person acts or believes stands in a non-accidental—specifically, safe—connection to the features of her situation which make so acting or so believing right. Arguably, an epistemically worthy belief, so understood, is knowledge. Be that as it may, the important point is that to think that truth is fundamental when it comes to epistemic assessment is not to think that it exhausts epistemic assessment.

Acknowledgements

Thanks to D Black, Clayton Littlejohn, Neil Mehta, Mark Schroeder, two anonymous referees for this journal, and audiences at the University of Southampton, the University of Melbourne, Saarland University, Yale-NUS College, and Stockholm University for feedback on earlier versions of this material and discussion of the issues it concerns. Extra thanks to Åsa Wikforss for her response to my presentation in Stockholm.

References

• Adler, Jonathan (2002). Belief’s Own Ethics. MIT Press.
• Arpaly, Nomy (2002). Moral Worth. Journal of Philosophy, 99(5), 223–245.
• Arpaly, Nomy and Timothy Schroeder (2014). In Praise of Desire. Oxford University Press. https://doi.org/10.1086/678364
• Baumann, Peter (2008). Is Knowledge Safe? American Philosophical Quarterly, 45(1), 19–30.
• Bird, Alexander (2007). Justified Judging. Philosophy and Phenomenological Research, 74(1), 81–110.
• Boghossian, Paul (2008). Content and Justification. Oxford University Press.
• Broome, John (2004). Reasons. In Philip Pettit, Samuel Scheffler, Michael Smith, and R. Jay Wallace (Eds.), Reasons and Value (28–55). Oxford University Press.
• Brown, Jessica (2010). Knowledge and Assertion. Philosophy and Phenomenological Research, 81(3), 549–566. https://doi.org/10.1111/j.1933-1592.2010.00369.x
• Bykvist, Krister and Anandi Hattiangadi (2007). Does Thought Imply Ought? Analysis, 67(296), 277–285. https://doi.org/10.1111/j.1467-8284.2007.00691.x
• Bykvist, Krister and Anandi Hattiangadi (2013). Belief, Truth, and Blindspots. In Timothy Chan (Ed.), The Aim of Belief (100–122). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199672134.003.0006
• Comesaña, Juan (2005). Unsafe Knowledge. Synthese, 146(3), 393–402.
• Dreier, James (2000). Dispositions and Fetishes: Externalist Models of Moral Motivation. Philosophy and Phenomenological Research, 61(3), 619–638. https://doi.org/10.2307/2653615
• Engel, Pascal (2005). Truth and the Aim of Belief. In Donald Gillies (Ed.), Laws and Models in Science (77–97). King’s College Publications.
• Fassio, Davide (2011). Belief, Correctness and Normativity. Logique et Analyse, 54(216), 471–486.
• Feldman, Richard (1988). Epistemic Obligations. Philosophical Perspectives, 2, 235–256.
• Gettier, Edmund (1963). Is Justified True Belief Knowledge? Analysis, 23(6), 121–123. https://doi.org/10.1093/analys/23.6.121
• Ghijsen, Harmen, Christoph Kelp, and Mona Simion (2016). Norms of Belief. Philosophical Issues, 26(1), 375–392. https://doi.org/10.1111/phis.12077
• Glüer, Kathrin and Åsa Wikforss (2009). Against Content Normativity. Mind, 118(469), 31–70. https://doi.org/10.1093/mind/fzn154
• Glüer, Kathrin and Åsa Wikforss (2013). Against Belief Normativity. In Timothy Chan (Ed.), The Aim of Belief (80–99). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199672134.003.0005
• Graham, Peter (2010). In Defense of Objectivism about Moral Obligation. Ethics, 121(1), 88–115. https://doi.org/10.1086/656328
• Greco, John (2004). Knowledge as Credit for True Belief. In Michael DePaul and Linda Zagzebski (Eds.), Intellectual Virtue (111–134). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199252732.003.0006
• Greco, John (2010). Achieving Knowledge. Oxford University Press.
• Greenberg, Alexander (in press). Should I Believe All the Truths? Synthese.
• Hawthorne, John and Ofra Magidor (2018). Reflections on the Ideology of Reasons. In Daniel Star (Ed.), Oxford Handbook of Reasons and Normativity (113–139). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780199657889.013.6
• Herman, Barbara (1981). On the Value of Acting from the Motive of Duty. Philosophical Review, 90(3), 359–382. https://doi.org/10.2307/2184978
• Hurka, Thomas (2001). Virtue, Vice, and Value. Oxford University Press.
• Johnson King, Zoë (in press). Praiseworthy Motivations. Noûs.
• Kant, Immanuel. (1997). Groundwork of the Metaphysics of Morals. Mary Gregor (Ed.). Cambridge University Press. Original work published 1785. https://doi.org/10.1017/cbo9780511809590.004
• Kearns, Stephen and Daniel Star (2009). Reasons as Evidence. In Russ Shafer-Landau (Ed.), Oxford Studies in Metaethics: Volume 4 (215–242). Oxford University Press.
• Kelp, Cristoph (2009). Knowledge and Safety. Journal of Philosophical Research, 34, 21–31.
• Lackey, Jennifer (2006). Pritchard’s Epistemic Luck. Philosophical Quarterly, 56(223), 284–289. https://doi.org/10.1111/j.1467-9213.2006.00443.x
• Lasonen-Aarnio, Maria (2010). Unreasonable Knowledge. Philosophical Perspectives 24(1), 1–21.
• Lewis, David (1973). Counterfactuals. Blackwell.
• Lewis, David (1979). Counterfactual Dependence and Time’s Arrow. Noûs 13(4), 455–476. https://doi.org/10.2307/2215339
• Littlejohn, Clayton (2012). Justification and the Truth-Connection. Cambridge University Press.
• Littlejohn, Clayton (2013). The Russellian Retreat. Proceedings of the Aristotelian Society, 113(3), 293–320. https://doi.org/10.1111/j.1467-9264.2013.00356.x
• Littlejohn, Clayton (2017). How and Why Knowledge Is First. In J. Adam Carter, Emma Gordon, and Benjamin Jarvis (Eds.), Knowledge First (19–45). Oxford University Press.
• Lord, Errol (2017). On the Intellectual Conditions for Responsibility. Philosophy and Phenomenological Research, 95(2), 436–464
• Lord, Errol and Kurt Sylvan (in press). Believing for Normative Reasons: Prime, Not Composite. In Patrick Bondy and J. Adam Carter (Eds.), Well-Founded Belief. Routledge.
• Lynch, Michael (2004). True to Life. Harvard University Press.
• Mantel, Susanne (2013). Acting for Reasons, Apt Action, and Knowledge. Synthese, 190(17), 3865–3888. https://doi.org/10.1007/s11229-012-0230-8
• Mantel, Susanne (2017). Three Cheers for Dispositions. Erkenntnis, 82(3), 561–582.
• Markovits, Julia (2010). Acting for the Right Reasons. Philosophical Review, 119(2), 201–242.
• McGlynn, Aidan (2013). Believing Things Unknown. Noûs, 47(2), 385–407. https://doi.org/10.1111/j.1468-0068.2011.00843.x
• McHugh, Conor (2014). Fitting Belief. Proceedings of the Aristotelian Society, 114(2), 167–187.
• Mehta, Neil (2016). Knowledge and Other Norms for Assertion, Action, and Belief. Philosophy and Phenomenological Research, 93(3), 681–705. https://doi.org/10.1111/phpr.12222
• Millar, Alan (2009). How Reasons for Action Differ from Reasons for Belief. In Simon Robertson (Ed.), Spheres of Reason (140–163). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199572939.003.0006
• Neta, Ram and Guy Rohrbaugh (2004). Luminosity and the Safety of Knowledge. Pacific Philosophical Quarterly, 85(4), 396–406. https://doi.org/10.1111/j.1468-0114.2004.00207.x
• Pritchard, Duncan (2005). Epistemic Luck. Oxford University Press.
• Pritchard, Duncan (2012). Anti-Luck Virtue Epistemology. Journal of Philosophy, 109(3), 247–279. https://doi.org/10.5840/jphil201210939
• Raz, Joseph (2011). From Normativity to Responsibility. Oxford University Press.
• Roush, Sherrilyn (2007). Tracking Truth. Oxford University Press.
• Sainsbury, R. Mark (1997). Easy Possibilities. Philosophy and Phenomenological Research, 57(4), 907–919. https://doi.org/10.2307/2953809
• Schroeder, Mark (2015a). In Defense of the Kantian Account of Knowledge: Reply to Whiting. Logos and Episteme, 6(3), 371–382. https://doi.org/10.5840/logos-episteme20156328
• Schroeder, Mark (2015b). Knowledge Is Belief for Sufficient (Objective and Subjective) Reason. In Tamar Szabó Gendler and John Hawthorne (Eds.), Oxford Studies in Epistemology: Volume 5 (226–252). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780198722762.003.0008
• Schroeder, Mark (2018). Believing Well. In Conor McHugh, Jonathan Way, and Daniel Whiting (Eds.), Metaepistemology (196–212). Oxford University Press. https://doi.org/10.1093/oso/9780198805366.003.0012
• Schroeter, Laura and François Schroeter (2009). Reasons as Right-Makers. Philosophical Explorations, 12(3), 279–296. https://doi.org/10.1080/13869790903067691
• Shah, Nishi and J. David Velleman (2005). Doxastic Deliberation. Philosophical Review, 114(4), 497–534. https://doi.org/10.1215/00318108-114-4-497
• Sliwa, Paulina (2015). Praise without Perfection. American Philosophical Quarterly, 52(2), 171–185.
• Sliwa, Paulina (2016). Moral Worth and Moral Knowledge. Philosophy and Phenomenological Research, 93(2) 393–418. https://doi.org/10.1111/phpr.12195
• Smith, Martin (2016). Between Probability and Certainty. Oxford University Press.
• Smith, Michael (1994). The Moral Problem. Blackwell.
• Sosa, Ernest (1999). How Must Knowledge Be Modally Related to What Is Known? Philosophical Topics, 26(1&2), 373–384. https://doi.org/10.5840/philtopics1999261/229
• Sosa, Ernest (2011). A Virtue Epistemology: Apt Belief and Reflective Knowledge, Volume I. Oxford University Press. https://doi.org/10.1353/cjp.0.0084
• Steglich-Petersen, Asbjørn (2010). The Truth Norm and Guidance: A Reply to Glüer and Wikforss. Mind, 119(475), 749–755. https://doi.org/10.1093/mind/fzq063
• Stratton-Lake, Philip (2000). Kant, Duty and Moral Worth. Routledge.
• Sutton, Jonathan (2007). Without Justification. MIT Press.
• Svavarsdottir, Sigrun (1999). Moral Cognitivism and Motivation. Philosophical Review 108(2), 161–219. https://doi.org/10.2307/2998300
• Sylvan, Kurt (2012). Truth Monism without Teleology. Thought, 1(3), 161–169. https://doi.org/10.1002/tht3.26
• Thomson, Judith Jarvis (2008). Normativity. Open Court.
• Turri, John (2010). On the Relationship between Propositional and Doxastic Justification. Philosophy and Phenomenological Research, 80(2), 312–326. https://doi.org/10.1111/j.1933-1592.2010.00331.x
• Way, Jonathan (2017). Creditworthiness and Matching Principles. In Mark Timmons (Ed.), Oxford Studies in Normative Ethics: Volume 7 (208–227). Oxford University Press.
• Way, Jonathan and Daniel Whiting (2016). Reasons and Guidance (or, Surprise Parties and Ice Cream). Analytic Philosophy, 57(3), 214–235. https://doi.org/10.1111/phib.12086
• Wedgwood, Ralph (2002). The Aim of Belief. Philosophical Perspectives, 36(16), 267–297.
• Wedgwood, Ralph (2006). The Normative Force of Reasoning. Noûs, 40(4), 660–686. https://doi.org/10.1111/j.1468-0068.2006.00628.x
• Whiting, Daniel (2010). Should I Believe the Truth? Dialectica, 64(2), 213–225.
• Whiting, Daniel (2012). Does Belief Aim (Only) at the Truth? Pacific Philosophical Quarterly, 93(2), 279–300. https://doi.org/10.1111/j.1468-0114.2012.01421.x
• Whiting, Daniel (2013a). Nothing but the Truth. In Timothy Chan (Ed.), The Aim of Belief (184–203). Oxford University Press.
• Whiting, Daniel (2013b). Truth: The Aim and Norm of Belief. Teorema, 13(3), 121–135.
• Whiting, Daniel (2015). Knowledge Is Not Belief for Sufficient (Objective and Subjective) Reason. Logos and Episteme, 6(2), 237–243. https://doi.org/10.5840/logos-episteme20156216
• Whiting, Daniel (2018). Right in Some Respects: Reasons as Evidence. Philosophical Studies, 175(9), 2191–2208. https://doi.org/10.1007/s11098-017-0954-x
• Williamson, Timothy (2000). Knowledge and its Limits. Oxford University Press.
• Zagzebski, Linda (1996). Virtues of the Mind. Cambridge University Press.

Notes

1. See Boghossian (2008), Fassio (2011), Greenberg (in press), McHugh (2014), Lynch (2004), Millar (2009), Shah and Velleman (2005), Wedgwood (2002). The above do not all formulate the norm in the same way or in the same terms. For an influential challenge to the view that belief is subject to a norm of truth, see Bykvist and Hattiangadi (2007) and (2013). I address that challenge in Whiting (2010) and (2013b).

2. In previous work (e.g., his 2012), Littlejohn defends the idea that truth is the fundamental norm for belief. In Littlejohn (2013), he rejects his earlier explanation for the ‘inward-looking focus’ of epistemic assessment.

3. This differs in a harmless fashion from Littlejohn’s formulation. Other proponents of principles along the lines of Knowledge include Adler (2002), Bird (2007), Engel (2005), and Sutton (2007). Ghijsen, Kelp, and Simion (2016) and Mehta (2016) defend the view that knowledge is the fundamental norm for belief, though they deny that the relevant norm is deontic.

4. To say that one can explain the norm of knowledge by appeal to the norm of truth is not to take a stand on whether knowledge itself is analysable in any way. One can consistently maintain both that knowledge is explanatorily basic (with, among others, Williamson 2000) and that norms in which it figures are explanatorily derivative.

5. See, e.g., Boghossian (2008), Lynch (2004), Wedgwood (2002). For critical discussion, see Glüer and Wikforss (2013), Littlejohn (2013), Whiting (2013a).

6. On guidance, see the exchange between Glüer and Wikforss (2009) and Steglich-Petersen (2010).

7. As I discuss below, epistemic worth is factive in the sense that, if a belief is epistemically worthy, it is true. Many, though not all, consider justification to be non-factive.

8. The shopkeeper’s action might have prudential worth—i.e., it might be worthy from a prudential point of view. To keep things simple, here I stick to moral worth.

9. Influential proponents of views along these lines include Arpaly (2002) and Markovits (2010), though their considered views are more nuanced. A leading alternative is the view that moral worth requires moral knowledge. See Sliwa (2015) and (2016). For criticism, see Arpaly (2002), Arpaly and Schroeder (2014: 176–187), Way (2017).

10. One can ask both whether an action is morally worthy and how morally worthy it is. I will focus on what it is for an action to have worth (full stop), not what it takes for an action to have some degree of worth.

11. Similar remarks apply to cases in which a person does the right thing for the wrong reasons due to blameless ignorance. Consider a version of the case involving Macie in which it is right for her to go the cinema but not for the reason that it keeps a promise.

12. I do make the substantive assumption that what it is right to do depends on the facts, rather than a person’s evidence or beliefs about the facts (cf. Graham 2010). Some reject this ‘objectivist’ view. However, those who maintain that rightness is evidence- or belief-relative need not, and typically do not, take it to be ‘inward-looking’ in Littlejohn’s (2013) sense. So, the debate between objectivists and their opponents is orthogonal to the issue at hand. To put the same point differently, even if moral rightness is ‘subjective’, one can still distinguish right and worthy action (cf. Markovits 2010).

13. For similar examples in the service of similar points, see Arpaly (2002: 225), Arpaly and Schroeder (2014: 59–61), Lord (2017), Mantel (2017), Stratton-Lake (2000: 66–67), Markovits (2010: 230–237), Way (2017).

14. For views in this ballpark, see Arpaly (2002), Arpaly and Schroeder (2014), Lord (2017), Markovits (2010), Way (2017).

15. Inspired by a case Kearns and Star discuss (2009). Markovits (2010: 219) anticipates the point to follow. Her response differs from mine. For criticism, see Sliwa (2015) and (2016).

16. A different challenge to Moral Worth₁ and its descendants comes from habitual actions. One might think that when a person acts from habit, she does not act for a reason, though her action might have moral worth. For discussion, see Arpaly and Schroeder (2014: 80–86).

17. Moral Worth₃ is neutral with respect to an issue central to the recent debate between proponents of explanation-based and evidence-based accounts of normative reasons. (For my contribution to that debate, see Whiting 2018.) Suppose that evidence of a right-making feature is not a reason for acting (cf. Broome 2004; Schroeter & Schroeter 2009). In that case, an action might have moral worth even if it is not done for a reason. Suppose instead that evidence of a right-making feature is a reason for acting (cf. Kearns & Star 2009; Thomson 2008). In that case, an action has moral worth only if it is done for a reason.

18. To say this is not to endorse Michael Smith’s (1994) argument for judgement internalism, roughly, the view that, if a person judges that it is right for her to act, she is thereby motivated to do so. Arguably, the opponent of internalism need not ascribe to moral agents an exclusive concern for the right, so conceived (cf. Dreier 2000; Svavarsdottir 1999). For a recent challenge to the idea that such a concern is anyway objectionable, see Johnson King (in press). It would distract from the main thread of this paper to respond to that challenge here. For present purposes, it is enough to note that one could revise the account as follows: a person’s action is morally worthy in some way just in case she acts on evidence that it is morally right in some way to do so or evidence of features that make it morally right in some way to do so. Typically, evidence of the former is evidence of the latter, and vice versa.

19. It is, of course, not uncontroversial that assessments of rightness in ethics are wholly outward-looking. Those who defend the doctrine of double effect, e.g., might claim that whether it is right for a person to act in a way that causes harm depends on whether that harm is intended. However, both the doctrine and its formulation are contentious issues, which I set aside. The important point is that, even when what it is right for an agent to do is independent of her motives or reasons for acting, there is a further dimension of ethical assessment which concerns such things.

20. In passing, Markovits suggests generalising Moral Worth₂ in a similar fashion (2010: 214), although she does not develop the proposal. Markovits’s suggestion is that epistemic justification is the analogue of moral worth. As noted above, I take no stand here on the relationship between worth and justification.

21. One might try to explain what it is to be creditable for believing rightly in virtue-theoretic terms (cf. Greco 2010; Sosa 2011). For an interesting attempt to explain the notion of acting for a reason in virtue-theoretic terms, see Mantel (2013).

22. Raz, in a different context, writes, ‘Epistemic reasons are governed by one concern: determination whether the belief for which they are reasons is or is not true. Reasons for a single action may, and typically are, governed by many concerns’ (2011: 41).

23. Insofar as Epistemic Worth₁ is neutral with respect to first-order disputes, it is available to a proponent of Knowledge. On the view that results, a belief is not worthy unless it is held on the basis of evidence that it is knowledge, i.e., on the basis of evidence that the belief is true, justified, non-Gettiered, etc. This strikes me as false, but I will not pursue the point here.

24. I assume that it is not necessary that the evidence make the truth of the proposition certain.

25. Several suggest that knowledge is subject to a safety condition (see Williamson 2000; Pritchard 2005; Sainsbury 1997; Sosa 1999). I return to this below. There are many ways to formulate the safety condition, which need not be in competition. The notion of safe evidence I work with here is due to Martin Smith (2016).

26. Nearness here is understood, following Lewis (1973; 1979), in terms of similarity in relevant respects.

27. If one prefers a different label for being-creditable-for-believing-what-is-right, that is okay with me. What matters for present purposes is the status, not the terminology.

28. For similar cases, see Arpaly and Schroeder (2014: 59–61), Lord and Sylvan (in press), Turri (2010), Way (2017), Way and Whiting (2016), Wedgwood (2006). Note that these authors are concerned, not with what it takes for a belief to have epistemic worth, but with what it is for a belief to be justified or properly based. As noted at the outset, I take no stand here on what justified belief involves.

29. At least, the case Littlejohn presents in his (2013). For a different argument for thinking that Knowledge is the fundamental norm for belief, see Littlejohn (2017). I hope to address this elsewhere.

30. More fully, Williamson uses this point to motivate the idea that knowledge is the norm for assertion, but it is supposed to carry across to belief. As he says, ‘we may project the account of assertion back onto its mental counterpart, judgement (or belief)’ (2000: 10). For reservations about this strategy, see Whiting (2013a).

31. Following Williamson, Littlejohn (2013) appeals to Moorean propositions in support of Knowledge.

32. In earlier work (Whiting 2013a), I questioned this. See also McGlynn (2013).

33. As others note (e.g., Brown 2010: 553–554), the observations are in danger of proving too much. In relation to Miyuki’s belief that Coppola directed Rumble Fish, Nishi might say, ‘You don’t know that for certain!’ And, as with Moore, it seems absurd for me to believe that it is raining but I am not sure that it is raining. Parity of reasoning, then, would suggest that certainty is the norm for belief, a conclusion that neither party to the present dispute would welcome.

34. For the view that knowledge is, or entails, safe belief, see Pritchard (2005), Sainsbury (1997), Sosa (1999), Williamson (2000). The tendency is to suggest that safety is a necessary condition for knowledge; to account for the data in support of Knowledge, I suggest that it is a sufficient condition for knowledge. I take no stand here on the necessity claim. For objections to it, see Baumann (2008), Comesaña (2005), Kelp (2009), Neta and Rohrbaugh (2004). By the same token, since I allow for unworthy knowledge, I make no claim to solve the Meno problem, i.e., to explain why knowledge is better than true belief or, more generally, belief that falls short of knowledge.

35. Schroeder (2018) defends the view that knowing is having an epistemically worthy belief or, in his terms, believing well. While his primary concern is not to defend the idea that a norm of truth fundamentally governs belief, he does agree that ‘rightness for belief is truth’. Clearly, we are in different ways exploring similar ideas. One important difference is that Schroeder does not offer an account of epistemic worth in terms of safe evidence; indeed, he does not advance any account of what believing well consists in. In Schroeder (2015b), he defends the view that knowing, hence, believing well, is just a matter of believing a proposition on the basis of sufficient (objective and subjective) evidence. The case of Pangloss might cast doubt on this. For criticism, see Whiting (2015). For a reply, see Schroeder (2015a),

36. Again, an alternative way to develop this idea is in virtue-theoretic terms.

37. Schroeder (2018) presents his view as an account of the nature of knowledge.

38. For further Gettier-style counterexamples to safety-based accounts of knowledge, see Roush (2005: 122–123). Lasonen-Aarnio (2010) considers and responds to challenges of a different sort to the idea that safe belief suffices for knowledge.

39. For different ways of developing this idea, see Mantel (2013), Mantel (2017), Lord (2017), Lord and Sylvan (in press), Way (2017), Wedgwood (2006).

40. Of course, it would be instructive to consider other domains too, such as the aesthetic or prudential.

41. This is common ground among participants in recent debates on moral worth, e.g., Arpaly (2002), Arpaly and Schroeder (2014), Johnson King (in press), Markovits (2010), Silwa (2016). See Schroeder (2018) for an argument that the order of explanation must go from (what I call) rightness to worthiness. It is noteworthy that extant attempts in epistemology—taking virtue ethics as the model—to treat a notion like worth (virtue) as prior in the order of explanation to rightness do not derive anything like Truth (see, e.g., Zagzebski 1996).

42. Assume also that pulling the lever is not the only means of keeping the promise.