Two Faces of Responsibility for Beliefs

Abstract

The conception of responsibility for beliefs typically assumed in the literature mirrors the practices of accountability for actions. In this paper, I argue that this trend leaves a part of what it is to be responsible unduly neglected, namely the practices of attributability. After offering a diagnosis for this neglect, I bring these practices into focus and develop a virtue-theoretic framework to vindicate them. I then investigate the specificity of the belief case and conclude by resisting two challenges, namely that attributability cannot amount to genuine responsibility and that it can be reduced to a sort of accountability.

There is little agreement on the conditions that someone must meet to be responsible for their beliefs.¹ Yet, the literature is quite consistent in characterising the target phenomenon: someone is responsible for a belief insofar as they exercise some control over it and have some capacity or fair opportunity to revise it. We hold believers responsible and blame them when, in believing, they violate (without excuse) some norms that we could expect and demand them not to violate, and with which they could have complied; they are praiseworthy if they comply with them exemplarily. The relevant norms are typically epistemic, and epistemic is the nature of believers’ responsibility. Thus, paradigmatic blameworthy beliefs are false, unjustified, incorrect.² Recently, some have argued that we can also violate moral norms in believing (Basu 2018; Basu and Schroeder 2019) but this leaves the general understanding of what it is to be responsible unchallenged.

In this paper, I argue that this understanding is moulded on the ‘practices of accountability’ and that these do not exhaust the explanandum: the ‘practices of attributability’ identify another, unduly neglected way of being responsible. My overarching aim is to bring these practices into focus. Not every time we deem someone responsible do we hold them responsible, and not every time we criticise (or commend) someone in responsibility-involving ways for their beliefs do we do so insofar as they fail to abide by (or extraordinarily comply with) some (epistemic) norms.

I begin by offering a diagnosis: the practices of attributability have been neglected because of a rather silent shift of reference of ‘accountability’ and ‘attributability’ from kinds of practices, as Gary Watson (1996) originally meant them, to explanatory frameworks (section 1). In section 2, I look at the practices of attributability. The suggestion I explore is that they involve participant but third-personal reactive attitudes and that someone is attributability-responsible for what displays their values, insofar as it does. ³ I develop these ideas within a virtue-theoretic framework. Section 3 examines the specificity of the belief case. The challenge is to clarify how theoretical reasons can show anything of our values since, unlike practical reasons, they don’t derive from the value of the belief in favour of which they count. I consider a case study and argue that the results can be sufficiently generalised. Section 4 and section 5 defend the proposal from two key objections: that attributability cannot be an integral part of responsibility and that it is reducible to accountability. However compelling one may find the positive view suggested, I hope to convince the reader that this uncharted area calls for greater attention.

1. A diagnosis for the neglect

The distinction between ‘attributability’ and ‘accountability’ is not unfamiliar to the philosophical discourse, and many correctly acknowledge its paternity to Gary Watson (1996). However, whilst for Watson they identify two kinds of responsibility, over time they have been taken as referring to competing theories. This occurred especially since Neil Levy’s introduction of ‘volitionism’ and ‘attributionism’ as “two contending accounts” of moral responsibility for actions (Levy 2005, 2). ⁴ Admittedly, Levy does acknowledge that Watson’s distinction refers to two distinct and mutually irreducible kinds of responsibility but dismisses the point rather quickly: “[v]olitionists and thoroughgoing attributionists … dissent from Watson’s claim” ⁵ and “responsibility as understood by attributionism and responsibility as understood by volitionism” “map neatly” “onto Watson’s distinction” (3). Because of this decisive but subtle shift of reference, the other kind of responsibility, namely the practices of attributability, have lost their place in the conceptual landscape. Most contemporary discussions of ‘attributability’ figure within discussions of attributionism.

Let me briefly retrace the genealogy of this neglect and clarify my take on attributionist proposals.

1.a Volitionism and attributionism

Both volitionists and attributionists agree that an agent is morally responsible for something when “the Strawsonian reactive attitudes [especially praise and blame] are justified in relation to her with regard to [it] (Strawson 1962)” (Levy 2005, 2). Whether someone is praise- or blameworthy depends on their compliance with the relevant norms, which we can legitimately expect and demand them not to violate. The disagreement between volitionists and attributionists concerns what makes these attitudes fit. ⁶

On the one hand, volitionists impose strong control and epistemic conditions: “An agent is responsible for something … just in case that agent has—directly or indirectly—chosen that thing” (Levy 2005, 1). Someone can be blame- or praiseworthy for φing if only if they exercise direct or indirect control over φing and they are aware of φing (and, usually, of the relevant norms). This vindicates the centrality of demands, expectations, sanctions, and fairness: we can fairly demand and expect that S φs better, and sanction them otherwise, if and only if they φed controllably and consciously.

Attributionists rebut: reactive attitudes are responses to ‘who someone is’ as it is revealed by their actions. Control matters only to the extent that lack of it can block this ‘revelation.’ ⁷ The interpretations of ‘who someone is’ differ: some talk of someone’s “emotional reactions, spontaneous attitudes, and values,” which show their “moral personality” (Bernecker 2018); others focus on their ‘character’ (Nottelmann and Fessenbecker 2020; Wieland 2017); their ‘deep selves’ (Brownstein 2016; Wolf 1990); ‘what the agent values’ (van Oeveren and Wieland 2017); or the “quality of their will.”

One familiar objection to attributionism is that reactive attitudes can surely not be justified for everything that displays someone’s self. How could we blame and praise someone, and perhaps even sanction and award them, for something they cannot control? Reactive attitudes require more stringent—control and epistemic—conditions; volitionists have a stronger case. Attributionists, in response, argue that only some things truly express who we are, e.g., those that we evaluatively control (Hieronymi 2006, 2014), or that are in rational relations with our underlying evaluative commitments (Smith 2005, 2008). ⁸

Yet, the game may be rigged: perhaps responsibility is outlined in a way that best suits a volitionist framework, and this is why volitionists have an easier time explaining the target phenomenon. Let me turn then to the practices of accountability.

1.b Accountability

I said that for Watson accountability and attributability are different practices. On the one hand, the practices of accountability are essentially interpersonal: we hold others accountable and are always accountable to someone. Also, we always hold others accountable against norms to which we all are subjected as members of the same community. Because of this shared subjugation, every member is “entitled (in principle) to react in various ways” (Watson 1996, 230) and has the authority to demand that others respect those norms or, otherwise, make up for their violations (by way of apologies, compensations, sanctions). ⁹ Because “the practice[s] of moral accountability … involve the imposition of demands on people,” “they raise issues of fairness” that “underlie the requirement of control (or avoidability) as a condition of moral accountability” (235). ¹⁰ In other words, we can fairly demand and expect that S φs better, and sanction them otherwise, only if there is one right way of φing, i.e., if φing is regulated by prescriptive norms, and only if S can φ, i.e., if they are capable, in control, and aware both of what they are doing and of the relevant norms.

To reiterate: attributionists set out to explain this phenomenon. Smith explicitly argues that there is just one kind of responsibility, namely accountability. Fritz claims that attributionists and volitionists agree that reactive attitudes apply only in relation to obligations, although they understand obligations differently (Fritz 2018, 849). For Bjornsson, ‘quality of will’ is understood in relation to what can be properly demanded of agents (2017). And even Shoemaker, who supports pluralism, thinks that attributability-reactive attitudes are similar to accountability-reactive attitudes since both refer “to failures with respect to some norm” (2018, 1000). ¹¹

Notice also that the literature on responsibility for beliefs assumes that accountability exhausts the phenomenon. Of course, the need for some tailoring is acknowledged: since we cannot believe at will, control becomes either voluntary but indirect or nonvoluntary; we can’t talk of ‘volitionism’ strictu sensu; and the relevant norms are epistemic. Nevertheless, believers’ responsibility is taken to depend on their capacities to somehow control and revise their beliefs, and their openness to criticism or commendation on their compliance with relevant norms (McHugh 2013, 132).

1.c Attributability

For Watson, the practices of attributability capture another face of responsibility. These practices are “important to issues about what it is to lead a life … and about the quality and character of that life” (1996, 229). ¹² They are not essentially interpersonal as accountability is: when I am attributability-responsible, I am not responsible to you; you cannot hold me attributability-responsible; my conduct carries no offense. ¹³

To illustrate: suppose that a friend “betrays her ideals by choosing a dull but secure occupation in favor of a riskier but potentially more enriching one” (231). In saying that she’s acted “badly—cowardly, self-indulgently, at least unwisely … we are not thereby holding her responsible” (231). To do so, “we would have to think that she is accountable to us or to others, whereas in many cases we suppose that such behavior is ‘nobody’s business’” (231). Attributability-criticism does not refer to flouted, shared prescriptive norms, and therefore no one can demand that your friend choose otherwise; she doesn’t owe compensation to anyone; her practical identity is “nobody’s business.”

Nevertheless, in a strong sense she is responsible for how she leads her lives, for ‘who she is.’ To flesh this out, Watson writes:

conduct can be attributable … to an individual as its agent and is open to appraisal that is therefore appraisal of the individual as an adopter of ends. (229)

And this is because:

if what I do flows from my values and ends, there is a stronger sense in which my activities are inescapably my own: I am committed to them. As declarations of my adopted ends, they express what I’m about, my identity as an agent. They can be evaluated in distinctive ways (not just as welcome or unwelcome) because they themselves are exercises of my evaluative capacities. (233)

When we deem someone attributability-responsible, we appraise them aretaically. These aretaic appraisals target individuals qua adopters of ends. Since my intentional actions “flow” from my values and ends, they are my actions; I do not carry them out intentionally by accident.

Watson’s remarks remain more underdeveloped than one may hope, but here’s a way of reconstructing the thought:

1. 1. Practical reasons derive from the value of the action for which they are reasons. Someone will take themselves to have a reason to φ only if they find φing somehow worth doing. [premise]

2. 2. Finding something to be of value generates practical reasons to pursue or instantiate it. [from 1]

3. 3. Having a certain value involves finding certain things to be of value. [premise] ¹⁴

4. 4. Having a certain value is to (take oneself to) have certain practical reasons. [from 2, 3]

5. 5. One’s adopted ends reflect what they take to be conclusive practical reasons. [premise]

6. 6. If all goes well, one’s intentional actions reflect one’s adopted ends. [premise]

C. If all goes well, one’s intentional actions reflect the values one has. [from 4, 5, 6]

What I do stands for a “declaration” of my adopted ends because it displays what I consider conclusive practical reasons, and so what is for me worth doing, “what I am about.” And for that, I am responsible.

In effect, Watson seems to think that adopted ends result from some act of adopting them—he says that one “commits oneself,” “stands for” and “declares” one’s values (233–34). Clearly, if this were the marker of attributability, it would be hard to grant attributability-responsibility for beliefs: believing that p cannot instantiate or result from an act of adopting an end. ¹⁵ However, we can take his reference to adopted ends as distinctive of the action-case and his general thesis to concern the conceptual connection between attributability and what he calls “an individual’s fundamental evaluative orientation” (234):

Attributability: Someone is attributability-responsible for what manifests something of their evaluative orientation, insofar as it does.

Whether someone is open to criticism or commendation will depend on the quality of their evaluative orientation.

I believe that one promising way of illuminating the ‘quality of one’s evaluative orientation’ and unifying the practices of attributability for actions and beliefs is to develop a virtue-theoretic framework. Although the stability of my main claim does not depend on this, it may be helpful to gesture at one. Define ‘evaluative orientation’ as someone’s sensitivity to both practical and theoretical reasons, oriented by their values. ¹⁶ My evaluative orientation is the way in which I, as a distinctive individual, see things. What I find “worth’ choosing, doing or believing identifies just one way, i.e., the way of practical reasons, in which my values are displayed. Next: whilst manifesting my evaluative orientation identifies what I am attributability-responsible for, the quality of the manifested evaluative orientation determines whether I am open to commendation or criticism. Watson calls these evaluations ‘aretaic appraisals’ because they are “made from the aretaic perspective” (231). And “many of these appraisals concern the agent’s excellences and faults—or virtues and vices—as manifested in thought and action” (231). Generalising: someone is commendable for their actions and beliefs that display a virtue; they are criticisable for those displaying a vice. ¹⁷ More precisely: if, as John McDowell’s puts it, virtues are those “states of character whose possessor arrives at right answers to a certain range of questions about how to behave” (1979, 331), ¹⁸ then someone will be criticisable for getting things wrong, for seeing things, believing and reacting cowardly, dishonestly, foolishly. Someone is commendable for getting things right, for seeing things the right way, for being a person who believes, acts, feels, fears well. Drawing on this, we can think of virtues and vices as distinctive sensitivities to both theoretical and practical reasons: in most situations there is one way in which the ideally (and exclusively) kind or dishonest person would see situations. The ideally kind person will be attentive and react to certain aspects of each situation that are essentially different from the ones salient to the paradigmatic honest or callous person. However, no virtue or vice exhausts someone’s evaluative orientation, which identifies the distinctive way in which I see things precisely because it incorporates more than a trait, and thus no trait in its purest form. ¹⁹ The ways in which different traits are integrated in my persona explains what sorts of facts count for me as good reasons in any given situation. Furthermore, I take it that virtues and vices do not exhaust the set of things over which ‘my values’ can range. I can have values that orient my way of seeing things consistently but without issuing obviously wrong or right answers to the questions what to do and believe. For instance, valuing etiquette, being old-fashioned or an optimist can figure in both commendations and criticisms, whilst being deemed honest will always ²⁰ be a form of commendation and dishonest of criticism. ²¹

All this sheds some light on the kinds of things for which we can be attributability-responsible, i.e., those displaying something of our evaluative orientation. Let me develop a more detailed picture of how/what the practices of attributability look like before turning to exploring the specificity of the belief case (section 3).

2. The practices of attributability

In Freedom and Resentment, Strawson famously distinguishes between two stances one might take in relation to other subjects, a third-personal ‘objective stance’ and a second-personal ‘participant stance.’ Responsibility occurs within the latter. This is because when we treat someone objectively, we treat them “as an object of social policy; as a subject for what, in a wide range of sense, might be called treatment; as something certainly to be taken account, perhaps precautionary account, of; to be managed or handled or cured or trained” (Strawson 1962, 5). The one who’s treated objectively “is not … seen as a morally responsible agent, as a term of moral relationships, as a member of the moral community” (10), or as someone with whom we can have “ordinary adult human relationships” (6). But being responsible is to be open to reactive attitudes that are issued from within those very relationships. The participant attitudes “rest on, and reflect, an expectation of, and demand for, the manifestation of a certain degree of goodwill … or at least … for, an absence of the manifestation of active ill will or indifferent disregard” (8). ²²

I agree with Strawson that the participant stance is the marker of responsibility. Yet, Strawson himself admits having “presented nothing more than a schema, using sometimes a crude opposition of phrase where we have a great intricacy of phenomena” (15), thus “neglecting the ever-interesting varieties of cases.” Pressing this line, my suggestion is that participant attitudes are not always second personal: the practices of attributability exist somewhere in between the second-personal participant stance (of accountability) and the third-personal objective stance (of mere grading evaluations). They belong to a third-personal participant stance. Let me elaborate.

On the one hand, aretaic appraisals are not second personal. When we deem someone attributability-responsible, we do not take ourselves to have the right to demand that they live by different values. And the conduct of those who are attributability-criticisable does not display lack of regard for others’ right to be undamaged (harmed or misinformed). However, attributability attitudes are not issued from an objective stance either: when we appraise someone aretaically, we surely see them as a term of a social relationship, as a member of the same, ethical community—someone with values, virtues, and vices, a personality and distinctive individuality. We don’t treat them as an object to manipulate, one “posing problems simply of intellectual understanding, management, treatment, and control” (10). Their actions and beliefs are intelligible to us, although we may disagree with their correctness or importance. Further indication that attributability-attitudes are not issued from the objective stance comes from the fact that there’s a clear sense of injustice attached to false ascriptions of vices: being falsely deemed boring or lazy has a certain kind of “force” or “sting” (Hieronymi 2004, 117) that being falsely deemed a Gemini or Taurus hasn’t. It is about who we are; it touches us, it matters to us. In general, the psychological and emotional mechanisms attached to the practices of attributability occur only within relationships between persons, even though with respect to something we are not entitled to demand be different.

What’s the nature of these third-personal participant attitudes? If the reactive attitudes that belong to moral accountability can be placed on a scale that goes roughly from resentment to gratitude (or from regret to pride in the first person), and the ones belonging to epistemic accountability may be measured in terms of epistemic trust, I suggest that attitudes of attributability belong to the family of esteem. To hold someone in high esteem is to consider them good qua person, to look up to them. They manifest what it is to be good as a human being—to excel in some type of trait, to have a virtue. The other end of the spectrum splits into two: those who are open to attributability-criticism are so either insofar as they are unremarkable, unimpressive, or insofar as they are dishonourable, unprincipled. A helpful way of thinking about attributability-attitudes is to look at the attitudes attached to ascriptions of several virtues and vices. ²³ When we deem someone wise, kind, generous, or lazy, cunning, foolish, our attitudes towards them include revulsion, disapproval, mockery, veneration, contempt, disgust, admiration, annoyance, boredom, disinterest, sense of superiority or inferiority.

This fits well with the fact that only accountability-responses draw on considerations that involve someone else besides the assessed agent. Suppose that a friend broke a promise without excuse and another wastes their talent. Something like disappointment fits both situations. But in the former you would say, “I trusted them”; “they hurt me”’; “they gave me their word”; “I thought we were friends,” whilst in the latter: “such a lack of ambition, what a pity!”; “I think they could do more with their life”; “so sad, they are so timid and fearful!” The latter set comes from a place of (some) detachment. ²⁴

Finally, although from the third-personal participant stance we cannot demand that someone act or believe differently, a certain kind of normative expectation does apply. This is captured by sentences like “I thought you were different,” or “Oh, that’s my boy!” In saying things like these, we expect someone to deem certain reasons conclusive, have certain preferences, put a certain amount of effort in this rather than that. If they do not, we are disappointed in them. What a pity that they are not who we thought; what a joy that they are. Put another way, although no ought to do applies within attributability, perhaps some sort of ought to be does. ²⁵

A final point. I agree that holding someone accountable for what they cannot help doing or believing is inappropriate. But so long as an action or belief displays something of someone’s evaluative orientation, appraising them aretaically is neither unfitting nor unfair. First, aretaic appraisals do not involve sanctions. Second, lack of a virtue is often fertile soil for the flourishing of another: someone’s lack of courage may be precisely what makes them so precise and reflective. Finally, lack of control, education, or alternative opportunities does affect what attributability-attitudes are fitting insofar as it determines what trait a certain action or belief is displaying: due to different upbringings, the same belief may display kindness in one, simple-mindedness in another. Similarly, the fact that children and psychopaths exercise limited control over something does affect what attitudes are fitting, although it doesn’t determine their suspension from the game of responsibility tout court—as long as they are assessed for what displays something of their evaluative orientation.

Even more: believers can be open to negative aretaic appraisals also for beliefs held for what are for them good (theoretical) reasons. Believing rationally does not get us off all the hooks, since there are more vices than those concerned with failures of rationality. For instance, a believer may be dull, boring, unstylish. Giving central stage to attributability is precisely a way to highlight that beliefs can be normatively assessed in richer and more varied ways than merely in terms of what one ought to believe, or in terms of truth, justification, or knowledge.

3. Attributability for beliefs

Suppose that this is convincing. Still, some will say, how could we be attributability-responsible for our beliefs? How can they manifest anything about us, our values, and evaluative orientation? ²⁶ Surely beliefs don’t express one’s ‘adopted end’ of believing that p; unlike practical reasons, theoretical reasons don’t derive from the value of the belief in favour of which they count. ²⁷

3.a A case study: Norpois

To get a better sense of the ways in which theoretical reasons can display someone’s values, consider a case study from Proust’s Madame Swann at Home. The narrator’s parents chitchat about a diplomat, the Marquis de Norpois:

Meanwhile, my father, so as to forestall any criticism that we might feel tempted to make of our guest, said to my mother: “I admit that old Norpois was rather ‘trite’ [poncif], as you call it, this evening, when he said that it would not have been ‘seemly’ to ask the Comte de Paris a question, I was afraid you might burst out laughing.”

“Not at all!” answered my mother. “I was delighted to see a man of his standing and age keep that sort of simplicity, which is really a sign of straightforwardness and good breeding.”

“I should think so, indeed! That does not prevent him from being shrewd and discerning; I know him well [ … ].” (Proust [1918] 2015, 61)

To begin with, Norpois is not appraised for saying something, as when a statement is considered polite, imprudent, or aggressive. What he is said to demonstrate—simplicity, straightforwardness, and good breeding—are not evaluations of his utterance: he is banal or candid not in virtue of voicing his beliefs, but because he believes that asking certain questions to the Comte is inappropriate. The narrator’s parents are evaluating Norpois himself, e.g., he “was rather ‘trite,’” in the light of the belief underlying his utterance.

Because Norpois believes that posing certain questions to a Comte would be unseemly, he is deemed a ‘poncif,’ a French noun without adjectival form that denotes ‘cliché,’ stereotype. And Norpois would remain a poncif even if he did go on and pester the Comte, so long as he kept considering it inappropriate: the truth of their assessments does not depend on whether Norpois acts on his belief. The point is: Norpois’s belief displays some of his values. For instance, honour, dignity, elegance, and a deep respect for formalities. There’s a proper way to behave in society, and it is important to live by it. That Norpois values these things makes him sensitive to situations in a certain, distinctive way. He is particularly attentive to the way other people present themselves, and he demonstrates pronounced social intelligence. His responsiveness to the minimal indicators of someone’s discomfort or embarrassment makes him foresee potential triggers and successfully avoid conflicts. As he dislikes discord, he eschews strong emotions and disapproves of public outpourings. By the same token, he is jealous of his private thoughts and careful not to overexpose his stance; he leads rather stilted, impersonal—though flawless—conversations. The result is at once courteous and contrived. Whilst he is always well-groomed and mindful, his playing by the book makes him open to some criticism: he is conventional, predictable, old-fashioned. His confidence in his ability to read situations gives him an air of pompous superiority. Not only does he care about doing things “well,” he is also convinced to be the best, who knows best. But although he is criticisable for being so affected, the appropriate criticism does not involve the kind of opprobrium, disgust, or contempt that is apt when someone is ill willed. In effect, he even triggers some hilarity. We can see the narrator’s parents smirking. The mother admits being amused by Norpois’s naivety (“naïveté”), which, she glosses, indicates a certain integrity and candor (“honnêteté”).

Norpois is open to these evaluations because of what he believes. These are deep evaluations: deeming him old-fashioned is not like judging him short—it goes to the core of his identity. If we allow for attributability-responsibility, we have an easy time vindicating our reactions to his beliefs: Norpois is open to aretaic appraisals, positive and negative third-personal participant attitudes. ²⁸

By contrast, if we assume that responsibility just is accountability, then either Norpois and his beliefs are merely graded or he is held accountable for them. No third way is given, but neither of these two is suitable. The former option should leave us unsatisfied: the whole participant, nuanced, normative dimension of the interaction between the narrator’s parents and Norpois is lost. The latter is false: Norpois’s belief does not meet the conditions of blameworthiness. For one thing, he violates no epistemic or moral norm. He may well believe for good reasons; mention considerations that bear on the truth of his belief; put himself in the best possible position to assess its truth; have evidence that the Comte would be bothered; have been told by reliable sources; or even know it. Also, he does not show disregard for anyone’s well-being, nor does his belief constitute any wrongdoing. Furthermore, Norpois’ attachment to conventions is no one’s business. We lack the entitlement to demand that he is more progressive.

3.b Generalising Norpois

Norpois is not an isolated case: although not every belief displays something of the believer’s values, quite a wide range of (evaluative and nonevaluative) beliefs do. Attributability-responsibility is sufficiently broad a phenomenon.

To begin with, consider cases of disagreement. It is rather common that two believers, who consider whether p and are presented with the same body of evidence, can disagree. ²⁹ The same electoral program convinces you that the candidate would be a great president, whilst I conclude that they would be an awful leader. Presented with the same dinner invitation, you believe someone is nice; I think they are trying to win favour with me. Seeing a plant on the floor is enough evidence for me to conclude that my housemate was clumsy, whilst you suspend judgement.

Our disagreement does not prove, in and by itself, that one of us must be irrational. We can fill in the details so that our doxastic states are rational responses to the information available to each of us; we disagree because we have different background beliefs and evaluative standards. ³⁰ Drawing on Nelson’s denial that we have positive duties to believe (Nelson 2010), Willard-Kyle has argued that evidence cannot be evaluated apart from standards that depend largely on believers’ inclinations and are not themselves part of the evidence. Thus: ³¹

different cognitive interests or different epistemic weightings of the theoretic virtues can result in different rational attitudes even when there is no difference in evidence. (Willard-Kyle 2017, 1013)

But if it is possible for believers to disagree without irrationality, then exactly what belief someone holds will display something of their evaluative orientation; their system of values and beliefs; what kind of person they are. And this happens all the time. As Joseph Raz stresses, in normal circumstances, there is more than one emotion, belief, or action that is supported by reasons, and none to a higher degree than any of the others. And yet,

[n]eedless to say, in all situations our actual response is more definite than that. That is it goes beyond what reasons require of us. And that is the key to the understanding of the limits of reasons, and of the richness of the sources of our responses to the world, which include very much more than our rational capacities. (Raz 2011, 5)

Rationality is not the only relevant dimension of assessment when we consider beliefs, and aretaic appraisals track the “richness” of our responses to the world; they assess someone on the basis of exactly what values their (permissible and rational) beliefs display. ³²

Next: consider two believers who hold the same belief but for different truth-related reasons. For instance, we both might believe that that man is married. However, you believe so because of his wedding ring, whilst I do because his shirt is accurately, but not professionally, ironed. Once again, exactly what considerations we consider salient and conclusive reveals something important about us. Accordingly, our beliefs bear different aretaic appraisals insofar as they express different evaluative orientations. Here, although we hold the same (maybe true) belief, only I can be deemed, say, old-fashioned for it.

Finally, believers settle not only questions like “Is p true?” but also like “What is true?” Thus, something of them is made manifest also by what propositions they think are conclusively supported. Suppose that I believe the man before me is married, the old woman a widow, and that girl a doctor. You believe that trains are not cleaned regularly, that man looks suspicious and likely doesn’t have a ticket, and it might start to rain soon. Suppose that all our beliefs are true. For one thing, it is hard to say that there is any norm that dictates that we ought to form or refrain from forming these beliefs. Nor is it obvious that we ought to attend or refrain from attending to one or the other of these features of the world. Thus, we cannot be held accountability-blameworthy (or praiseworthy) for our doxastic practice here. And yet: the fact that I formed those beliefs and you didn’t—regardless of what beliefs we could have formed—does say something about us. The fact that I hold beliefs about people might indicate that I am interested in others more than you are; because of this, I may be deemed curious, attentive, observant. Yet, the fact that I focus on their marital or social status might show that I am old-fashioned, prudish, nosy. ³³ You, instead, tend to see the glass half empty. You are a complainer, a pessimist. We can think that I value marriage or a secure job, or maybe that I am sensitive to related factors because I despise both. ³⁴ And we can say that you value cleanliness, safety, and transparency, and because of this you have a lower threshold of what counts as “suspicious and fishy.” Because you value safety, you are sensitive to what you think endangers it.

Someone may insist: What about beliefs that are clearly nonevaluative, e.g., that it’s raining, or that Napoleon was French? I see three possible replies. Either these express something of one’s evaluative orientation that is so widely shared so as not to be distinctive of anyone’s. Or they display something of one’s values that is too shallow to ground aretaic appraisals. Or else, they do not display anything relevant for ascriptions of responsibility at all. But even if the latter were the case, it wouldn’t prove that no belief can display values. And my aim here is only to suggest that we should make room for the practices of attributability for beliefs, not that we need be attributability-responsible for all of them.

I want to conclude by resisting two objections. First, that what I described does not really capture a face of responsibility (section 4). Second, that the cases I have considered can be seen as belonging to the practices of accountability (section 5).

4. Responsibility enough?

Many deny that attributability captures a face of responsibility because it cannot distinguish, they say, responsibility-involving blame from responsibility-free mere criticism (Levy 2005; McKenna 2012). Because of the absence of strong control and epistemic conditions, aretaic appraisals collapse onto mere grading evaluations and thus cannot be an integral part of a genuinely normative phenomenon. I offer two lines of response: a negative and a positive one.

4.a A good worry?

It is not exactly clear how to hear the worry. I consider four ways of specifying it and argue that none hold.

First, it may be an indiscernibility worry: attributability cannot but conflate ascriptions of responsibility and mere ‘grading evaluations.’ ³⁵ The distinction is rather hard to capture (noncommittally ³⁶) but the idea is that ascribing responsibility involves something more than registering how something stands with respect to some standard, and that the aretaic perspective does not respect this asymmetry—someone is bad in the aretaic sense as a car is defective or damaged. But this is false. Watson does offer quite a strong criterion against such conflation: “conduct can be attributable … to an individual as its agent and is open to appraisal that is therefore appraisal of the individual as an adopter of ends” (1996, 229). Someone is responsible for their action in the attributability sense for what displays the ends they adopted or, more generally, for what displays their evaluative orientation, which their bone structure does not express. ³⁷

My opponent may remain unsatisfied: Why does the fact that my actions express my values make them something for which I am responsible? But in a sense, this question is disingenuous. It asks too much. It is like asking the accountability sympathiser why I should be responsible for what is under my control. They would respond: “Well, that is precisely why; it’s under your control!” Similarly, I am tempted to say I can be responsible for my actions precisely because they express my values. We reach the bedrock. When we evaluate someone for their beliefs, we assess who they are as a person. We commend or criticise them for being someone who’s kind, dishonest, envious, arrogant. When we do so, we are not simply admiring their thick hair or high cheekbones. We are evaluating the quality of their values, how they see the world, what moves them.

A second way of reading the worry concerns overinclusivity: people will be responsible for more things than we would ordinarily accept. For instance, Smith insists that there are “many, many other psychological features of agents that quite clearly help to explain their motives and attitudes in normative domains but for which we clearly do not regard them as responsible in any sense” (Smith 2012, 584). Suppose that someone is affected by obsessive-compulsive disorder (OCD), whereby they have uncontrollable, reoccurring thoughts and/or behaviours that they feel the urge to repeat. This may explain why they see the world a certain way. Smith stresses that we cannot deem them responsible for it, whilst views like mine seem committed to saying otherwise. In fact, I agree with Smith that subjects are not responsible for their OCD behaviour and beliefs. ³⁸ But I insist that I am not committed to saying otherwise: allowing for attributability-responsibility does not entail that one is attributability-responsible for anything that can be predicated of them. Quite the opposite: ascriptions of attributability-responsibility are fit only in reference to what displays someone’s evaluative orientation, which need not include all the features of their psychology that contribute to explain their conduct. And actions and beliefs caused by OCD say nothing of one’s values and of what one sees as good reasons. ³⁹

The third detractor will press a blamelessness worry: within the practices of attributability there is no room for a distinctive sense of full-blown blame as it is typically understood in the accountability literature. Clearly, this is troublesome only under the assumption that (accountability-)blame characterises responsibility tout court. Some would simply deny this; Watson himself talks of ‘aretaic blame.’ But even if we agree that ‘full-blown blame’ is intelligible only in the accountability sense and that “in order for someone to be the legitimate target of [blame], it is not sufficient that her attitudes and actions reflect her judgment” (Smith 2008, 376), it’s unclear how this proves that attributability cannot be a kind of responsibility.

Finally, suspicion may derive from ideals of metaphysical economy: we should avoid dividing responsibility into kinds. Metaphysical theses based on aesthetic ideals are always hard to assess. But note that most accounts of responsibility for beliefs begin with the disclaimer that they won’t be concerned with legal or merely causal kinds of responsibility. And if we accept this but not the attributability/accountability distinction, we shall give an explanation or face charges of arbitrariness.

4.b Responsibility enough!

There is also a more positive way of rebutting the objection that attributability is not a kind of responsibility, which draws on my account of the practices of attributability in section 2. Understood as I have suggested, not only do they not collapse into mere grading evaluations, they are also sufficiently similar to the Strawsonian paradigm to count as an integral part of the phenomenon of responsibility. As Shoemaker puts it, aretaic appraisals are “a range of emotional responses that, in terms of their force, feel, aim, and spark, are sufficiently akin to the other sorts of responses theorists agree are responsibility responses” (Shoemaker 2018, 1000). And on my proposed view, attributability responses are such that:

1. a. certain kinds of esteem-related attitudes are fit;

2. b. they are issued from the participative stance, and occur only between fellow human beings;

3. c. they involve certain kinds of normative expectations;

4. d. they affect social relationships;

5. e. inaccurate assessments are (not merely false but) unjust.

If we accept that what Strawson describes is full-blown responsibility, then it is hard to deny that the practices of attributability are too.

5. Irreducibility

To conclude, some worries concern the relationships between attributability and values on the one hand and accountability and norms on the other. ⁴⁰ I’ve claimed that there are two ways of being (and deeming others) responsible. In some cases, we are accountability-responsible. When controllably, consciously, and without excuses, we do something impermissible, others can blame us, expect and demand that we do not do what’s impermissible or make up for it, and sometimes sanction us. In other cases, we are attributability-responsible: we are assessed aretaically for who we are, what we value, how we lead our lives, and the quality of our character. Yet, some insist, it may seem that the cases I have considered can be explained by reference to norms after all. For instance, it could be that when we criticise someone for being dogmatic, we are but registering that they violated norms of belief-revision. I have a few things to say in response.

First, I am not committed to saying that viciousness never involves norm-violation. Nor need I deny that in many cases one will be responsible (as in, accountable) for that violation. My point is simply that holding someone accountable for that violation doesn’t give the full story: it also matters that they took a particular (putative) fact to be a good reason to do or believe something rather than something else, where both options were permissible. So if someone is dogmatic, we may well hold them epistemically accountable for failing to revise their belief correctly and appraise them aretaically for taking this and this (putative) fact to be an excellent reason. In fact, not all dogmatists are the same, and we do not react to the antivaxxer, the orthodox Christian and the hippie with the same (third-personal participant) attitudes.

Second, and this is a far larger issue than I can hope to do justice to here, viciousness cannot be reduced to norm-violation. After all, the revival of virtue theory since Anscombe (1958) has been meant to provide a radical alternative to deontologism, and we should take the suggestion seriously before declaring the project failed. Acting or believing viciously is not to do something impermissible. It is to get it wrong, miss the point, see things cowardly, dishonestly, foolishly. When one is criticised aretaically, one is not criticised for having flouted a prescription but for being a certain way, which grounds and explains why they see things a certain way and why they flout certain norms and not others, in certain ways and not others. It seems to me that this gets something right about our practices. When we say things like “You shouldn’t believe that!” we don’t mean “You failed to comply with a rule of belief-formation!” Most frequently, we are saying: “How can that strike you as true?,” “How can you think that that is a good reason?,” where these don’t point at a malfunctioning of someone’s belief-formation mechanism as much as at the quality of the reasons-responsiveness that is displayed by their doxastic attitudes. ⁴¹

A satisfactory defence of these thoughts would deserve a much longer treatment, which I must leave for another occasion. Here, my goal was to highlight that importing a certain understanding of responsibility from the case of actions requires more care than it is often acknowledged and that, most importantly, it may have made us lose sight of another, integral part of the phenomenon we were after.