Affective Reason

Abstract

This paper contributes to the recent explosion of literature on the epistemological role of emotions and other affective states by defending two claims. First, affective states might do more than position us to receive evidence or function as evidence. Affective states might be thought to appraise evidence, in the sense that affective states influence what doxastic state is rational for someone given a body of evidence. The second claim is that affective evidentialism, the view that affective states function rationally in this way, is not just possible but plausible and fruitful. We offer two arguments in favor of affective evidentialism.

Over the past quarter century or so, the philosophical literature on emotions has exploded. There are debates about what emotions are, how emotions relate to other affective states, what role emotions do and ought to play in our lives, and on and on. This is a paper about the role of emotions and other affective states in rational belief.

We defend two claims. First, the recent debate on the roles of emotions and other affective states in our epistemic lives has missed an important possibility. We substantiate this claim by articulating three grades of affective involvement in epistemic rationality. An affective state is epistemically relevant in the first grade if it functions as a positional influence on rational belief or knowledge. An affective state is second-grade epistemically relevant if it functions as evidence (or some other evidence-like warrant-relevant thing). Grades one and two have multiple, prominent defenders. But an affective state is third-grade epistemically relevant if it appraises evidence (in a sense we will specify). Our purpose in defending this first claim is to display this third possible relationship between affective states and rational belief, not to defend the first two grades. The second claim we defend is that the third grade of affective reason ought to be acknowledged, explored, and even embraced. We argue third-grade affective reason follows from prominent contemporary philosophical positions given minor and plausible auxiliary assumptions.

1. Two Grades of Affective Reason

We will say that reason is affective in some sense or other if some affective state – whether emotion, mood, sentiment, etc. – makes a difference as to whether a judgment has some substantively normative, positive epistemic feature.¹ We'll focus on epistemic rationality understood in a broadly evidentialist setting, but that is non-essential. One could make our case directly about knowledge, for example, especially if one thinks rational belief is necessary for knowledge. We do admit that radically externalist epistemologies – pure reliabilist views, for example – are not obviously amenable to the arguments we make here. But for anyone who thinks there is something epistemically interesting or important about rational belief, even if rational belief is not necessary for knowledge, this paper is for you.

There is a fairly substantial tradition according to which affective reason fails to exist. Maybe affective reason is even impossible or borderline incoherent. According to this tradition, reason is a cold, emotion-less attention to truth, a passionless pursuit of reality. Insofar as our affective life intersects with reason, affect does nothing but cause trouble. So one finds DeSousa (1987: 142) summing up such views: “By and large, common sense holds that emotions are typically both subjective and irrational.”

In the past few decades, however, emotions have come to play a more prominent positive role in epistemology.² The foci of the conversation about the positive contributions emotions make to human reason and knowledge have been two: emotions position us evidentially and emotions are evidence. These are the first two grades of affective reason.

1.1. First Grade Affective Reason: Positional Influence

On the first grade of affective reason, affective states are positional influences on rational belief. Positional influences “put us in a particular position, rather than some other, to possess certain reasons” or evidence.³ Desires are often positional influences, sometimes in ways that aren't even intimately connected to the evidence one acquires in their wake. For example, one might open up the New York Times one morning because one wants to find the weather forecast but find oneself confronted with a headline about the latest scandal involving a recently fallen-from-grace political figure. One's desire to know the weather positioned one to acquire evidence about that politician. A little reflection reveals that positional influences can of course be more subtle.

Affective states are widely recognized as positional influences in both philosophy and psychology, and their influence is of a more subtle form. Brady (2013), for example, contends that “emotions direct and focus our attention onto objects and events that are potentially significant” (p. 90) and that emotions “capture and consume attention” (p. 92). This attentional influence of emotion is valuable because it can “enhance our representation of potentially significant objects and events” (p. 93). Similarly, Deonna and Terroni (2012) suggest that emotions are “intimately connected with types of action readiness” (p. 79) that cover “the tendency to attend to an object, to submit or be drawn to it, to disengage from it,” etc. (p. 80). More picturesquely, DeSousa (1987) claims that “emotions set the agenda for beliefs and desires” (p. 196) and that an emotion can be “a disposition to persist in asking certain questions” (p. 199). In psychology, Frederickson's (2004) ‘broaden-and-build’ theory of positive emotions claims that “positive emotions broaden peoples' momentary thought-action repertoires, widening the array of the thoughts and actions that come to mind” (p. 1369). Negative emotions are widely regarded as narrowing one's attentional focus. Whether broadening or narrowing, emotions shape what one is primed to notice or attend to in one's environment. This is a paradigmatically positional influence on epistemic rationality. Each of these authors representing multiple fields and approaches to the emotions agree that affective states are positional influences on epistemic rationality. Many others are among their ranks.

Importantly, this is a somewhat insubstantial role for affective states in epistemic rationality. By ‘substantial’ here we don't mean important or essential. Rather, we mean something more like constitutive. Positional influences are vitally important, of course, but they are only indirectly relevant to rationality. Positional influences are not factors on which rational deliberation turns, nor do positional influences impact the structure of rationality. Positional influences are, as it were, precursors to rational deliberation, rather than being part of rational deliberation itself. This is not a criticism, but it serves to contrast first grade affective reason with affective reason of the second and third grades.

1.2. Second Grade Affective Reason: Evidence

On the second grade of affective reason, affective states are evidence. Affective states function, on this grade, as inputs to the process of judgment formation.

Whether emotions are evidence has, perhaps, been the central epistemological question in the recent literature on emotions. In particular, a number of sophisticated philosophical positions on emotions undergird the idea that emotions rightly function as an important class of evidence for propositions about the normative domain. Typically this view stems from the idea that emotional experiences are perceptual experiences of the normative domain. This sort of account is, at the very least, a force to be reckoned with (Brady 2013). For example, in various places Robert Roberts argues that emotions are a kind of non-sensory perceptual state, a state he often describes as “concern-based construals”.⁴ The details of Roberts's views need not preoccupy us. This perceptual theory of emotion is one view that prominently figures in defenses of the idea that emotional experiences are evidence. As many have rightly noted, including Roberts himself, it's a natural consequence of thinking of emotions perceptually that emotions supply evidence for certain judgments.⁵ After all, perceptual experiences of ordinary macrophysical objects are plausibly the best evidence we have that those objects are as we perceive them to be. The same would seem to hold for perceptions of the normative realm.

This role for emotions situates the affective domain more centrally in rationality than first grade affective reason. Evidence is a constitutive aspect of rationality, whereas positional influences are mere attendants to rationality.

2. Third Grade Affective Reason: Appraisers of Evidence

We devote more time to the third grade of affective reason, on which affective states are appraisers of evidence.⁶ The possibility of third grade affective reason is largely unnoticed and therefore undeveloped in the contemporary literature. We hope to rectify that lack here. A central burden of this section will be to say what exactly it means to say that members of some class are appraisers of evidence. We tackle that, along with some important background, in the first subsection. Then we turn to clarifying the idea that affective states are appraisers.

2.1. The Structure of Epistemic Rationality

Recent conversations about epistemic rationality, driven in no small part by reflection on issues about higher-order evidence and the rational force of testimony and disagreement, have revealed a schism to do with the relationship between evidence and rational doxastic states. This schism is between those who think that, given a particular body of evidence, a believing agent is rationally required to occupy a particular doxastic state or one among a constellation of doxastic states, and those who think that rationality involves more than just bodies of evidence and doxastic states.

This schism is not about uniqueness. Uniqueness says, roughly, that anyone with some particular body of evidence is rationally required to adopt a particular doxastic stance with respect to a given proposition.⁷ That doxastic stance might be a specific sharp credence, a mushy credence, or one of withholding or belief or denial. But there is a doxastic stance that is required given that evidence, and everyone with that evidence is rationally required to take that stance. Anyone who occupies the “we need more than evidence to determine rational requirements” side of the schism must deny uniqueness. But the denial of uniqueness is compatible with occupying the “evidence is all you need to determine rational requirements” side of the schism. Therefore, the schism is not about uniqueness.

On the other hand, the source of the schism does concern an issue in the neighborhood of uniqueness. In particular, philosophers who deny uniqueness often justify their denial by staking their claim on the “we need more than evidence to determine rational requirements” side of the schism. More on that below.

Evidentialism about rationality is typically characterized along the following lines:

Evidentialism. A subject S is rational to adopt doxastic state D iff S's evidence supports S's being in D.⁸

Different versions of evidentialism emerge as one specifies what counts as evidence, what it is for a subject to possess evidence, and the nature of evidential support. None of that matters for our purposes. What matters here is that evidentialism, so understood, characterizes rationality as a matter that is settled by evidence alone. Rationality is understood as a relation between bodies of evidence and doxastic states. No other factors play a role in rationality. Therefore, the rationality relation is binary: it holds between bodies of evidence and doxastic states.

Evidentialism does not entail uniqueness. Uniqueness requires that rationality is (analogous to) a function from (something like) bodies of evidence to (some constellation of) doxastic states.⁹ Evidentialism, as stated, is compatible with the relation of rationality failing to determine a function from bodies of evidence to doxastic states because evidentialism leaves open the possibility that some bodies of evidence support multiple doxastic states.

However, there is some pressure on evidentialists to embrace uniqueness. This pressure emerges from the fact that evidence is meant to indicate the truth. Failures of uniqueness seem to violate this role for evidence. If a body of evidence appears to indicate ambiguously among distinct individual options, it would instead indicate uniquely a single ambiguous option: one does not have a choice among a set of incompatible doxastic states, but instead ought to evenly distribute one's confidence over the variously indicated “options”. So, for example, if some body of evidence putatively indicates believing some proposition no more or less than disbelieving it, then it would seem that body of evidence really uniquely requires that one withhold on that proposition. To think otherwise would be to implicitly embrace the idea that rationality is plagued with an arbitrariness analogous to that of coin flips.¹⁰

On the other hand, uniqueness seems subject to counterexample. Some evidential situations seem permissive. Permissive cases are ones in which two subjects with similar-enough evidential positions can rationally occupy different doxastic states. Standard permissive cases are well known, and the territory is well worn. We make no claims to add to the conversation about the truth of uniqueness or the possibility of permissive cases. We simply assume that there are permissive cases, especially where high-level theoretical questions are in view, including politics, religion, and so on.

The arbitrariness worry sketched above is significant for evidentialists who reject uniqueness. One reaction to the arbitrariness worry is to move away from the idea that rationality is solely a matter of evidence but in a way that retains the core of the evidentialist project. If successful, this strategy promises to handle permissive cases without giving up on the spirit of evidentialism while avoiding arbitrariness in rationality. We unpack this idea presently.

Evidentialism as we have defined it makes rationality a binary matter: there are just bodies of evidence and their corresponding doxastic state(s). A general strategy for evidentialists who deny uniqueness is to jettison that binary in favor of non-binary evidentialism:

Non-Binary Evidentialism. A subject S is rational to adopt doxastic state D iff S's evidence supports S's being in D given X.

Non-binary evidentialism involves the idea that rational doxastic states are determined by more than bodies of evidence. Epistemic rationality is relative to whatever one specifies for X. Call states or factors occupying or functioning in the X-role in non-binary evidentialism ‘appraisers (of evidence)’. Different forms of non-binary evidentialism will differ with respect to the class of states that are appropriate appraisers. And given some particular specification of a class of appraisers, different subjects might be rationally permitted distinct doxastic states even though they have the same evidence. Thus, such states are ‘appraisers’ because, intuitively, they influence the probative force of the evidence for particular subjects. To get clear on the idea here, we turn to examples of non-binary evidentialism.

Perhaps the most prominently proposed class of appraisers are rational standards. Schoenfield (2014: 199), for example, suggests that “what makes it permissible for agents to have different doxastic attitudes [despite having the same evidence] is that different attitudes may be prescribed by their different standards”. A set of standards, for Schoenfield, are rules that a subject relies on to translate bodies of evidence into doxastic states that they take to be truth conducive, where that means the subject “has high confidence that forming opinions using her standards will result in her having high confidence in truths and low confidence in falsehoods” (2014: 199). Schoenfield floats the idea that one's prior and conditional probability functions constitute one's rational standards. Kelly (2014) appeals to a believing subject's need to balance the cognitive goals of believing truths and avoiding falsehoods. We might think of this in terms of cognitive risk-aversion. The more cognitively risk-averse one is, the more evidence one will demand before one believes. Cognitive risk-aversion is, in this sense, a species of rational standards.

Schoenfield and Kelly agree that there are appraisers of evidence and that appraisers are something like rational standards for moving from bodies of evidence to doxastic states. Depending on one's prior and conditional probability functions or one's level of cognitive risk-aversion, different doxastic states might be required for two subjects who possess the same body of evidence.

Pragmatic encroachment epistemologies can also be glossed as forms of non-binary evidentialism according to which appraisers emerge from a subject's pragmatic setting.¹¹ For example, practical adequacy pragmatic encroachers claim that a subject's credence in some proposition p is sufficient for rational belief that p only if that credence is practically adequate. A credence that p is practically adequate for a subject if and only if any contextually salient action that is decision-theoretically rational given that subject's actual credences is also decision-theoretically rational conditional on p.¹² On this sort of view, practical factors appraise evidence by shifting the threshold credence required for rational belief. Colloquially, the idea is that practical factors in part determine whether some body of evidence is good enough to warrant a judgment about a particular question.

There are differences among these pragmatic encroachment views. For example, practical adequacy pragmatic encroachers do not think appraisers impact rational credence, while rational standards evidentialists may. Also, rational standards evidentialists seem to think that appraisers constitute a relatively stable factor for particular individuals, whereas pragmatic encroachers are open to significant and perhaps rapid variability in an individual's stock of appraisers. This list could be extended. But all parties agree that there is some constellation of doxastic states the rationality of which is impacted by appraisers of one type or another.

Importantly, all of these non-binary evidentialisms retain a tight connection between evidence and the normative demands of epistemic rationality. Evidence plays a crucial and ineliminable role. Consider quasi-uniqueness: anyone with some particular body of evidence and some particular stock of appraisers is rationally required to adopt a particular doxastic stance with respect to a given proposition. (The earlier clarifications from the discussion around uniqueness regarding which sorts of doxastic states are required still apply.) Non-binary evidentialism suggests but does not imply quasi-uniqueness, just as binary evidentialism suggests but does not imply uniqueness.

A commitment to quasi-uniqueness assists non-binary evidentialists in dealing with the arbitrariness worry highlighted earlier. From the point of view of epistemic rationality, non-binary evidentialism can insist that the demands of rationality for two subjects in a permissive case are different. In a permissive case, two subjects share a body of evidence. But this is compatible with those subjects differing with respect to appraisers. So long as there are different appraisers for the different subjects, rationality can accommodate differing rational requirements with no arbitrariness whatever.

2.2. Emotions as Appraisers

What does all this have to do with emotions? The view we are interested to defend is that version of non-binary evidentialism on which affective states are appraisers of evidence:

Affective Evidentialism. A subject S is rational to adopt doxastic state D iff S's evidence supports S's being in D given S's affective profile.

A person's affective profile is that more or less stable constellation of affective dispositions, including especially their emotional dispositions. By design, this is as yet not particularly precise. We have more to say about this presently, but there will be significant work left to do to develop and clarify affective evidentialism in other work. Our case for affective evidentialism will not turn on these details; in fact, below we will consider different implementations of affective evidentialism that make different demands on this front. Therefore, we will clarify which emotions function as appraisers as we discuss those different implementations.

Before we move to our arguments for affective evidentialism, one clarificatory comment is in order. One might be tempted to think affective evidentialism is just a form of virtue epistemology. The emotional dispositions to which we appeal below are in the neighborhood of character traits. And virtue epistemology is, at its core, the idea that certain character traits at least in part determine which doxastic states of a subject are rational.

Affective evidentialism is not virtue epistemology. Virtue epistemologies do not make rationality relative to affective profiles (in the form of virtues, especially intellectual virtues). Rather, virtue epistemologies make rationality dependent on affective profiles. Notice first that virtue epistemologies are difficult to construe as evidentialist. Second, affective evidentialism is not a claim about virtue in general, much less intellectual virtue. Affective evidentialism is concerned with affective profiles much more broadly. Third, insofar as evidence and virtues interact for virtue epistemologies, virtues either open one to evidence or help one attend to one's evidence or allow one to acquire certain evidence or some such.¹³ These are positional influences implicating first grade affective reason or emotions-as-evidence implicating second grade affective reason, not third grade affective reason where affective profiles are appraisers of evidence. Finally, most importantly, and building on the first three points, virtue epistemologies hold that a subject is only epistemically rational insofar as they form doxastic states out of virtues. So one cannot believe rationally and be intellectually vicious. Affective evidentialism denies this. On affective evidentialism, a subject can be perfectly epistemically rational even if intellectually vicious; indeed, intellectually vicious subjects could be epistemically irrational if their judgment matches the intellectually virtuous. Again, epistemic rationality is dependent on affective profiles within virtue epistemology but relative to affective profiles within affective evidentialism. This emphasizes the sense in which virtues do not appraise evidence for virtue epistemologists.

3. Arguing for Affective Evidentialism

Above, we clarified the idea of affective evidentialism through attention to one's own rational life. Such introspective investigations, undertaken with care and commitment, lend some credibility to affective evidentialism all on their own. We have not, however, produced any arguments in favor of the claim that affective states are appraisers. It is to this task we turn. We offer two positive arguments. The first argument couples pragmatic intellectualism, understood as a version of non-binary evidentialism, to a widely recognized cognitive role of emotions in order to defend the claim that emotions shape evidential thresholds for certain normative judgments. This is a warm-up argument, in that it is designed to show that affective evidentialism seems unsurprising given certain epistemological commitments. The second argument is more ambitious and suggests a more thoroughgoing version of affective evidentialism. This argument defends the claim that affective profiles are fit candidates to appraise the probative force of bodies of evidence. The theoretical utility of affective evidentialism will emerge as we develop this argument, and we will articulate some open empirical questions that must be settled to round out the case for affective evidentialism.

3.1. From Pragmatic Intellectualism to Affective Evidentialism

The first argument for affective evidentialism launches from Mark Schroeder's ‘pragmatic intellectualism’.¹⁴ The advantage for this argument is that it does not require any claims about rationality generally. Instead it requires, supposing pragmatic intellectualism is true, that emotions supply non-evidential reasons to withhold particular judgments in particular circumstances. We defend this claim after we explain pragmatic intellectualism.

Pragmatic intellectualism is, roughly, the view that, while evidence is the only kind of epistemic reason to believe that p, for any proposition p, pragmatic factors can give one reason to fail to believe that p (not to believe that not-p!). Pragmatic factors might supply reason to withhold judgment about some proposition. Pragmatic factors, on this view, make a difference to what evidence is sufficient to justify belief. Consider, for example, a case in which one has some evidence in favor of a proposition but could very easily come by much better, even decisive evidence. Maybe, for example, you vaguely recall having received a booster shot for a particular vaccination. It's the kind of memory that you trust when asked about, say, unimportant details of virtual department meetings over dinner with your family. At a dinner party, your vague recollection is enough to believe you had the booster. But in a different practical environment, that evidence isn't enough. Perhaps whether you got the vaccine has become practically important because there's an outbreak of the horrific disease the vaccine flawlessly protects against. You could easily check your health record on an app. In such a practical environment, it would be epistemically irrational to believe you had your booster given just your vague recollection. The presence of the outbreak gives one reason to withhold on the question whether you had the booster, and your vague recollection isn't enough to overcome that reason. So that's pragmatic intellectualism: practical factors sometimes give one reason to withhold judgment about some salient proposition.

The central conditional we must defend is this: if pragmatic intellectualism is true, then affective evidentialism plausibly follows. We make no attempt to defend pragmatic intellectualism. Our interest is just in what follows if it is true. If pragmatic intellectualism is true, then pragmatic factors are appraisers of evidence. Pragmatic factors in part dictate when evidence is good enough for belief. In that sense, pragmatic factors appraise a subject's evidence. The target conditional is therefore true if affective states are sometimes among the practical factors that give one reason to withhold judgment about some salient proposition. We turn to a defense of this claim.

Plausibly, emotions can function in a way analogous to the outbreak of the horrific disease in the case two paragraphs above. Consider being in the presence of a lion. Whether the lion's enclosure is adequate to restrain the lion has a different pragmatic significance depending on the behavior of the lion. If it is lazing in the heat of the day, and as an observer one is correspondingly calm and in awe, one doesn't think to check the integrity of the glass separating oneself from the lion. If the lion were snarling through the glass, or charging toward the onlooking zoo visitors, and you rightly felt fear or anxiety, you might feel compelled to check for cracks or other signs of weakness in the glass. Of course, the behavior of the lion is not evidence regarding the integrity of the lion's enclosure. One's fear seems the relevant difference, and that fear demands better evidence to maintain one's judgment that the enclosure is adequate to restrain the lion.

This sort of effect is seen in more mundane cases as well, cases in which the evidential sameness is more obvious despite affective differences. Most of us, while at home, feel more or less safe. Remembering having locked the doors and closed the windows before bed is sufficient to feel safe when one wakes in the middle of the night. But consider a case in which you wake in the middle of the night experiencing fear. You have no idea whether a suspicious noise woke you or whether you simply startled awake for no reason at all. In such a case, you give up your judgment that you are safe. This shift from belief to non-belief is perfectly rational. But as with the lion case, the fear on its own is not evidence that you failed to lock the doors, or that there is someone outside your window, or whatever. Instead, you give up your judgment that you are safe because the amount of evidence you demand given the fear is greater than when you feel safe. Your memory of having locked the doors and closed the windows, say, is no longer enough.

Another case, this time involving positive emotion. Suppose you oversee an organization run mainly by a half dozen overtaxed volunteers. Dependence on unpaid help makes staff turnover commonplace and leaves the organization fragile. Two staffers have recently resigned. On the basis of your evidence you judge that the organization will not survive the next two years. While responding to the resignation of one of your volunteers, you express, and then experience, gratitude for the contributions the person has made to the organization. In the context of this gratitude, you find yourself hopeful that the organization will survive this round of trials, as it has in the past. In this new emotional state, you give up your judgment that the organization will fail (perhaps even coming to believe that it will persist).

This shift seems rational, and on the version of affective evidentialism we are considering it is rational. Hope does not provide evidence that the organization will persist, nor do the past contributions of the volunteer that one recalls as one writes the letter of gratitude. Instead, on the view we are considering, hope gives one non-evidential reason to withhold on the proposition that the organization will fail. In the presence of the positive emotion of hope, you demand more evidence for that negative judgment. In the context of hope, the current trials cannot outweigh your evidence regarding the resiliency of the organization.¹⁵

Thus, just as ‘negative’ emotions like fear demand better evidence to make ‘positive’ judgments like ‘I am safe’, so ‘positive’ emotions like hope demand better evidence to make ‘negative’ judgments like ‘this organization will fail’. We invite the reader to construct additional cases.

The upshot is this. Pragmatic intellectualism requires that certain epistemic but non-evidential factors give one reason to withhold judgment about certain propositions in certain situations. The cases above show that emotions sometimes supply such reasons to withhold judgment. If that is right, then emotions are sometimes appraisers because they can alter the strength of evidence required to rationally judge that certain propositions are true. In other words, if pragmatic intellectualism is true, then affective evidentialism is as well.

This appraisal effect of emotions, we think, explains why emotions are positional influences. It is precisely because emotions supply reasons to withhold and therefore impact the threshold of evidence required for rational belief that emotions prompt the hunt for evidence.¹⁶ In the cases highlighted above, suppose the negative emotions drive up the threshold of evidence needed to foreclose a question. This disrupts one's belief in a corresponding proposition about which one is deeply concerned to have answered. Thus, one hunts for further evidence. So not only does pragmatic intellectualism support affective evidentialism, it also reveals why affective states plausibly function as positional influences.

This revelation allows us to fill a lacuna in Brady's (2013) treatment of emotions as positional influences. Brady suggests that emotions raise justificatory questions by “moving us to seek out considerations that bear on whether this object or event really is as it emotionally appears to be” (2013: 97). But Brady offers no distinctively epistemic explanation for why this is so. Pragmatic affective evidentialism answers this why-question: emotions can unsettle our judgments about questions of concern without themselves functioning as evidence by pushing around evidential thresholds. Because our emotions, like perceptual states, are generally helpful responses to the environment, this unsettling will generally be practically relevant in a way that demands inquiry. This role for emotions is, then, useful, so long as our emotions remain appropriately connected to real-world features. (Note that this can happen even if one is vicious, intellectually or otherwise!)

3.2. From Rational Standards Evidentialism to Affective Evidentialism

The second argument for affective evidentialism assumes both rational standards evidentialism and the possibility of permissive cases regarding politics, religion, and the like. Both assumptions are forcefully defended elsewhere; we do not rehearse those defenses. The implementation of affective evidentialism we defend here takes its cues from rational standards evidentialism rather than pragmatic intellectualism. For rational standards evidentialism, appraisers are, as it were, background states of rationality. Affective appraisers, likewise, must be affective states that color one's entire rational life, or at least the high-level aspects of one's rational life that shape one's perspectives on politics, religion, and so on. These affective states must, therefore, function in a background role in one's cognitive approach to these big-picture questions of life, meaning, and morality.

To set up the more direct argument we're interested in pursuing, it is worth noting that there is a long tradition in philosophy on which one's affective profile positively shapes one's epistemic life. Many philosophers have maintained that there are background conditions that shape epistemology, and many of those seem to think that states that have at least an affective dimension are among those background conditions.

Plato famously suggested in the Symposium that love drives us toward knowledge (cf. esp. 210a–211b). Thomas Aquinas argued for a complex interrelation between affection, desire, will, and intellect. The role Aquinas carved out for affection has clear import for the workings of the intellect. Cates (2009: 10), for example, argues that on Aquinas's view emotions are “interior motions that have a complex relationship to a set of powers or capabilities by which we receive and process information”. This complex relationship entails, at least according to Miner (2009: 99), that “if one does not love a science or a person, the knowledge that one has of that science or person will inevitably be superficial”. We do not claim these historical figures were committed to affective evidentialism.¹⁷ But their thought is at least suggestive. Other thinkers have come still closer to affective evidentialism.

To take just one underappreciated example, consider Jonathan Edwards. Edwards argued extensively for the idea that a benevolent heart was necessary for one's rational faculties to understand certain parts of reality, especially God.¹⁸ Edwards contends that the virtue necessary for religious understanding is a “universally benevolent frame of mind” (1749 [1987]: 620). This true benevolence is a spiritual sense constituted by a general benevolent orientation toward Being, which is in Edwards's theological tradition just God. Intriguingly, Edwards claims that “this new spiritual sense is not a new faculty of understanding, but it is a new foundation laid in the nature of the soul, for a new kind of exercises of the same faculty of understanding” (1754 [1959]: 206). This is at least suggestive of affective evidentialism. This “new foundation” is constituted by affective states of an individual and reorients that individual's understanding in order to “see” that which they could not previously see. Indeed, Wainwright's (1995) treatment of Edwards perhaps comes closer to affirming affective evidentialism than any contemporary author: “The promptings of true benevolence in this context just are the assessments of the force of a body of evidence” (1995: 52, emphasis in original). The idea of “assessments” of evidential “force” is very near that of appraisers, and Wainwright is clear that these assessments are robustly affective for Edwards. There is more to say here, but at the very least affective evidentialism is a promising interpretation of Edwards's religious epistemology.

Further, affective evidentialism can be thought of as an amplification of contemporary epistemological theories that recognize the role of robustly embodied social experience in shaping one's reception of future bits of evidence.

Consider standpoint epistemology, for example. In general, standpoint epistemology is characterized by two central claims. First, knowledge is only achieved within particular “standpoints”, that is, social locations that influence one's experiences, concepts, and so on. Second, the standpoints of individuals within marginalized or oppressed groups are better off than those not within those groups.¹⁹ Certain developments of standpoint epistemology at least suggest affective evidentialism. Standpoint epistemology tends to emphasize the role of robustly embodied experience in shaping belief and knowledge, especially experiences characteristic of marginalized or oppressed groups. And standpoint epistemologists often emphasize that this role of embodied experience goes beyond what we might think of as the role of third-personal, “objective” evidence. No amount of attention or evidence collection can overcome the epistemic difficulty for a member of a dominant class. A robust exploration of how affective evidentialism interacts with standpoint epistemology requires separate treatment, but we can gesture here at one important component of that engagement.

If affective states appraise evidence, standpoint epistemology's emphasis on embodied experience can be readily embraced. Since our affective lives are robustly embodied, affect is a plausible candidate domain of experience for standpoint epistemologists to exploit. Relatedly, affective evidentialism offers a characterization of at least some aspects of the mechanism by which embodied social experience becomes a standpoint, namely, through the production of affective profiles that in turn function as appraisers of evidence. Further, because attention and evidence collection cannot overcome the epistemic difficulty for members of a dominant class, first- and second-grade affective reason are insufficient to produce a standpoint without resorting to relativism about truth. Affective evidentialism is a helpful way to understand this point. What is needed is a change at the level of appraisers, and that cannot occur without certain affective experiences. These experiences may simply be unavailable to those within a dominant class. And so there is a natural explanation for the importance of careful attention to the testimony of those who are epistemically privileged but socially marginalized or oppressed.

So there is historical precedent from strikingly diverse intellectual locations pointing in the direction of affective evidentialism. It is important to be clear about what this does and does not show. We have not endeavored to show that affective evidentialism is entailed by or entails these other views, much less that these other views are true. Instead, we only mean to show that the idea that affective profiles might play some important background role in epistemic rationality is not unheard of; indeed, there are varying strands of philosophical thought representing a wide array of views that are committed to such a role for the affective domain.

So we proceed to the more direct route promised earlier: identifying the affective states that putatively function as appraisers and defending the idea that they are fit to do the work of appraisers.

The most initially plausible candidates for such affective states are long-standing moods and world-directed emotional dispositions. We have in mind not the grumpy mood one finds oneself in when one's football team loses to their arch rivals, or the chipper mood that attends the early stages of romantic love. Rather, we have in mind the sort of emotional dispositions or moods characteristic of long-running depression or joy, or perhaps optimism and pessimism. Religious individuals might include sentiments such as awe or gratitude toward their preferred ultimate reality, whether that is a God, the universe, or something else. These affective states are more characterological rather than episodic, though such characterological realities have episodic consequences. A different tack toward the same idea takes off from the fact that proper engagement with stories involves affect. In order to read a story as a tragedy, for example, one must experience the story in a certain affective mode. Further, whether consciously or subconsciously, persons think of themselves as occupying a place within a cosmic story. (More on this below.) More bluntly: we structure our lives according to a weltanschauung – a religion, a worldview, an ideology, a social imaginary, a standpoint – and these weltanschauung can be expressed as stories. That story might be comedic or tragic. But one's preferred cosmic story will carry with it an affective dimension like any other story, and that dimension will plausibly be captured by the long-standing moods and emotional dispositions we are hoping to circumscribe. Affective evidentialism, then, embodies the idea that the constituents of the affective dimension of one's weltanschauung function as appraisers.

Given both non-binary evidentialism and the possibility of permissive cases, and targeting this sort of affective evidentialism, our central burden is to garner considerations in favor of the idea that affective profiles of this sort are better suited than rational standards to be appraisers.

There are two demands here. First, such affective profiles must be up for the relevant appraisal work. Second, affective profiles must be better appraiser candidates than rational standards. We will use the moods that attend one's worldview as a test case. Presently, to ease the prose, we will call these simply ‘moods’. So the task becomes to show that such moods are both pervasive and plentiful. This makes our task more difficult, since affective evidentialists might appeal to both moods and other affective states when specifying which affective states function as appraisers. Readers dissatisfied with our defense of affective evidentialism given this limitation are invited to consider whether the gaps they sense in our defense can be filled by expanding the scope of affective appraisers beyond worldview-implicating moods.

Take the second demand first: moods are better appraisers than rational standards. First, notice that affective evidentialism is simply a step further along the path cleared by rational standards evidentialists. Rational standards evidentialists, recall, assert that one's balance of Jamesian goals or abductive virtues or one's prior probabilities function as appraisers. Affective states plausibly contribute to these domains. Emotions are standardly taken to be relevant to discerning value-laden aspects of reality, and one type of value is aesthetic.²⁰ And rational standards plausibly have an aesthetic dimension. Lipton (2004), for example, defends at length the idea that inference to the best explanation is best understood as inference to the loveliest explanation. Strikingly in the present context, he also notes that judgments of loveliness are bound up with certain things operating in the background of one's rational life.²¹

Emotion theorists have come very close to affirming the principles needed to bridge between rational standards evidentialism and affective evidentialism. De Sousa (1987: 191–2), for example, floats though does not develop the idea that emotions are part of the solution to “the problem of the choice of significance level: how probable must it be that your hypothesis is true on the evidence … before it is rational to accept it?”. Significance level in this sense is clearly connected to the issue of balancing the Jamesian goals of believing truths and avoiding falsehoods. Perhaps more strikingly, Tappolet (2017: 170) argues at length for the claim that “moods are perceptual experiences of the evaluative possibilities that are taken to be likely” and “have the function of signaling the likely possibility of evaluative properties being instantiated” (2017: 182). This characterization of moods suggests that moods are something like representations of prior probabilities or perhaps function like priors in rationality, since taken abstractly one's priors describe a probability space populated by possibilities you take to be live. To be clear, neither de Sousa nor Tappolet explicitly endorse non-binary evidentialism, much less affective evidentialism. But what they do say does offer indications of the move from one view to the other.

Affective evidentialism also has resources to make progress on an arbitrariness problem plaguing rational standards evidentialism. One might think that rational standards evidentialism doesn't solve the arbitrariness worry put to deniers of uniqueness. (The worry for uniqueness deniers was discussed above.) At best, rational standards evidentialism merely pushes the problem back to the rational standards themselves. Schoenfield (2014), for one, recognizes this concern. Her response is simple: “Whether we are permissivists or not, we can never give reasons for why we weight the evidence in one way rather than another that are independent of everything else” (2014: 202). In other words, everyone has that problem. Even those who embrace uniqueness and a singular set of appropriate rational standards cannot, independently of those standards, justify those standards.

Plausibly, however, there are substantive constraints on what affective profiles are rational, and these constraints are independent of epistemic rationality while sharing vital features of epistemic rationality.²² Affective states are subject to a form of rational evaluation. This form of rational evaluation is world-directed: affective states are rational if they are responses to situations involving particular normative properties. One who is disgusted by a central Texas sunset or who is overjoyed at the thought of human trafficking, is feeling wrongly. One dimension of that wrongness is not moral; it is very close to epistemic: the emotion doesn't fit reality. This lack of fit is something like the way illusory visual experiences don't fit reality. But it is closer to evidential fit than we commonly think applies to visual experience. Illusory perceptual states aren't meaningfully irrational; sideways emotional reactions are irrational. This is because our emotional responses, like our judgments, are indirectly under our control. At any rate, this world-directed normativity means that affective profiles are evaluable on a rational dimension other than the dimension they in part constitute. Not so with rational standards.

Affective evidentialism can, therefore, ground the epistemic rationality of disparate doxastic states in permissive cases by appealing to distinct affective profiles without sacrificing further inquiry into the rationality of those doxastic states. Further, and maybe more importantly, this view makes clear why certain affective profiles are more rational than others. It is because they better fit the world as a whole. Rational standards evidentialism has no such resources. Rational standards cannot be evaluated on a dimension of rationality independent of that which those rational standards in part constitute.

To be clear, the practical usefulness of this is minimal, and this is why affective evidentialism implies that there are permissive cases. The dimension of rational evaluation to which affective profiles are subject is, though distinct from straightforward epistemic rationality, nevertheless pervasive within one's rational life. Therefore, one cannot, in any practical sense, evaluate the rationality of the affective profiles themselves without settling factual questions about the world to which those affective profiles are responses. And the settling of those factual questions will happen from within a rational stance in part constituted by one's affective profile. The circle here is broadened, but not eliminated entirely. This seems to us a virtue, for there is something right in the idea that, whatever a particular non-binary evidentialism offers as appraisers, evaluating the rationality of appraisers should be impossible in this practical sense.

However, there is hope for a form of rational persuasion that is not based on evidence. For one can invite someone to empathetically occupy an alternative affective stance. This might happen through storytelling. Or one might invite another to alter their affective stance through openness or attachment to an object or individual, a group of objects or individuals, or some combination of these. This means that affective evidentialism inherits all the benefits of rational standards evidentialism when it comes to areas like the problem of disagreement while offering prospects for progress unavailable to rational standards evidentialism.

So let us return to the first demand: moods must be up for the relevant appraisal work. We have already seen that affective evidentialism follows in the spirit of important historical figures and certain contemporary epistemologies. And indeed, the affective profiles exploited by both Edwards and standpoint epistemology are in the neighborhood of the sorts of moods presently in view. That lends some credibility to the contention that moods are capable appraisers. Further, we again invite readers to consider their own rational lives.

There is more work needed here, however. Affective profiles must be both pervasive and plentiful. Affective profiles are pervasive if everyone has one. It would be bad for affective evidentialism if the structure of epistemic rationality required a factor that not all rational subjects possessed. Affective profiles are plentiful if there are sufficient types of such profiles to explain the relevant permissive cases. Explaining the prevalence of permissive cases is a central motivation for non-binary evidentialism. If affective profiles are non-plentiful, therefore, affective evidentialism cannot serve one of its core purposes.

The claims that affective profiles are both pervasive and plentiful require empirical support. Plausibly, affective profiles are at least pervasive given that it seems that everyone has a worldview and that worldviews involve affective profiles. But a merely a priori defense of these claims is insufficient. Empirical work to confirm especially the idea that worldviews involve affective profiles is crucial. Likewise, empirical work is necessary to confirm or disconfirm the idea that affective profiles are plentiful. Here, the question is whether distinct worldviews involve correspondingly distinct affective profiles. A new and promising empirical research program on primal world beliefs may provide evidence that affective profiles are both pervasive and plentiful.²³ Primals are conceptualized as maximally general beliefs characterizing the world as a whole. Primals extend along a continuum of opposites, for example, “the world is safe/the world is dangerous”. The affective dimension has yet to be isolated or studied, but seems to be a natural extension of the current conceptualization of primals. We, therefore, mention this as investigating that needs to be done in order to complete this first argument for affective evidentialism. At this point, affective evidentialism motivated in this way must rest on its theoretical utility in explaining reasonable disagreement, etc.

It is worth noting, however, that rational standards evidentialists are no better off on this score. They infer differences in rational standards from other differences, usually differences in first order judgments together with the plausibility of the claim that such differences aren't rooted in irrationality. Affective evidentialism can explain these differences in an analogous way but with the possibility of making empirical progress settling whether the proffered explanation of those differences is real.

The upshot is that anyone who is attracted to rational standards evidentialism should be interested in the prospects for affective evidentialism.

We have sought to clarify three, not just two, grades of affective reason and to defend third grade affective reason using two independent arguments. The first begins with pragmatic intellectualism, the second with rational standards evidentialism. Along the way, we hope we have illuminated the fruitfulness of thinking that affective states function as appraisers of evidence. At minimum, affective evidentialism deserves sustained development and critique.²⁴