What Perceptualists Can Say About Reasons for Emotion

Abstract

Perceptualism is a prominent theory analyzing emotions as perceptual experiences of value. A long-standing challenge to perceptualism says that emotions cannot be perceptual because they are subject to normative assessments in terms of reasons and rationality, while perceptual experiences are not. I defend perceptualism from this charge. My argument begins by distinguishing two forms of normative assessment: fundamental and non-fundamental. Perceptualism is compatible with the latter (i.e., non-fundamental reasons and rationality); even sensory experiences are so assessable. I next argue that emotions are only non-fundamentally assessable. Following this argument, I outline a perceptualist-friendly theory of emotions as non-fundamentally normatively assessable.

1. Introduction

Are emotional experiences a type of perceptual experience? Numerous parallels with paradigm perceptual experiences (i.e., sensory experiences) suggest that they might be. ¹ Both exhibit intentionality and have robust phenomenologies. Both likewise regularly conflict with belief. For example, one may hear a ventriloquist’s puppet as talking despite believing otherwise or fear flying despite believing it safe. There is arguably even an epistemological similarity, namely that emotional experiences (henceforth, emotions) justify evaluative beliefs much as sensory experiences justify empirical beliefs. These and other purported likenesses have led some philosophers, often called perceptualists, to answer “yes” to the opening question (e.g., Roberts, 2013; Tappolet, 2016). ² More precisely, what perceptualists claim is that emotions are perceptual experiences of value. Here I use “value” broadly as inclusive of the normative (e.g., reasons, ought) and thus allow for substantial variation in terms of how perceptualists can construe the values that particular emotions are about.

Perceptualism faces important challenges. This paper offers a solution to one of the most prominent: that emotions cannot be perceptual experiences because they are normatively assessable in a way that perceptual experiences are not. For example, while we might say that fearing a puppy is irrational, we wouldn’t, the objection goes, say that a sensory illusion is (e.g., Grzankowski, 2020; Helm, 2001; Salmela, 2011). This asymmetry is arguably explained by the fact that while there are reasons for emotions, there aren’t for perceptual experiences. So whatever emotions are, they must be understood as capable of assessment in terms of rationality and reasons (e.g., Brady, 2011; Deonna & Teroni, 2012; Müller, 2017). This is the normative assessability challenge.

Perceptualists have several options. One is to argue that emotions are a special type of perceptual experience subject to normative assessments not applicable to paradigm perceptual experiences (e.g., Tappolet, 2012). Or they may weaken their view to a mere analogy with perceptual experiences (e.g., Brady, 2007, p. 282n1). A danger with these responses is that, if emotions are subject to different normative assessments, this threatens the foundational role perceptualists often assign to emotions in value epistemology (see Brady, 2011; Cowan, 2018). Yet denying that emotions are ever proper subjects of normative assessments (cf. Döring, 2014) reduces perceptualism’s attraction. For it does seem a pretheoretical datapoint that emotions are at least sometimes properly assessable, even if this isn’t as obvious and incontestable as much of recent emotion theory would seem to suggest (more on this in Section 3). ³

This paper takes the following approach. After more carefully reviewing the challenge (Section 1), I observe a distinction between fundamental and non-fundamental normative assessability (Section 2). ⁴ If emotions are only non-fundamentally normatively assessable, then perceptualism escapes the challenge; this is because sensory perceptual experiences are also non-fundamentally assessable. I next argue against fundamental reasons for emotion (Section 3). This argument is rooted in the possibility and significance of scenarios in which an emotion is normatively unimpeachable despite the emoter’s awareness that the emotion is misleading. In the wake of this argument, I present the perceptualist-friendly theory according to which reasons for emotion are fundamentally for intention (Section 4). I close by blunting the force of two related objections, namely that the way in which emotions arise on the basis of reasoning (Section 5) or for motivating reasons (Section 6) indicates that they are fundamentally assessable.

2. The Normative Assessability Challenge

The normative assessability challenge hinges on an apparent asymmetry between emotion and sensory experience. The purported asymmetry is that emotions are assessable in ways paradigm perceptual experiences (i.e., sensory experiences) are not. The challenge is often presented by centering cases in which emotions conflict with evaluative beliefs, typically called recalcitrant emotions. Here is how Michael Brady puts it:

An obvious difference between perceptions and emotions is that at least some emotions can be assessed for rationality as well as accuracy. For instance, it is common to think of ‘recalcitrant’ emotions as in some sense irrational: it is irrational to be afraid of house spiders that one judges harmless, or feel guilt when one believes one has done nothing wrong. This marks a contrast with perceptual experiences. Even though perceptual experience can diverge from belief, it is not irrational to perceive the lines in the Müller-Lyer illusion as having different lengths even when one knows that they are the same length (2011, p. 139).

One dimension of the normative assessability of emotions, then, is rationality. Brady explains rationality in terms of normative reasons:

[O]ne explanation of this difference in assessment is that, intuitively at least, emotions can be responses to features of objects or events which constitute reasons for those responses, and hence to which a subject will respond emotionally insofar as they are rational. It makes little sense, however, to talk of features of objects or events as reasons to literally see or hear something in a particular way (Brady, 2011, p. 139; see also Deonna & Teroni, 2012, p. 93).

Here is a summary of the challenge. The first premise is that emotions are sometimes normatively assessable. I stipulate that by “normatively assessable” I mean subject to assessment in terms of reasons and rationality. The second premise is that perceptual experiences are never normatively assessable. It follows that emotions aren’t perceptual experiences and that perceptualism is false.

I emphasize that Brady’s way of putting the challenge links rationality and reasons; he takes the latter to explain the former. I follow him in this respect. Specifically, I assume that rationality is a matter of being supported by sufficient reasons (cf. Schroeder, 2015, pp. 239–242). A given emotion, then, is rational just in case it is responsive to sufficient reasons in favor of it (and to some extent irrational just in case it is responsive to actual/apparent reasons that are not). ^{5 , 6} This picture about the relationship between rationality and reasons means that normative assessability, and so the normative assessability challenge, is ultimately about reasons. It offers a more manageable path forward, for it means that there aren’t two distinct challenges, one about rationality and another about reasons; there is a unified challenge about both. ⁷

Additionally, while Brady focuses on the case of emotional recalcitrance, the challenge is broader. To see this, imagine a basketball coach angry at a player for missing what would have been a game-winning shot. Further suppose that this anger is caused by certain beliefs. In particular, the coach believes that the player didn’t care enough about winning, a belief they take to be supported by the fact that the player didn’t practice as much as the others. But the coach’s reasoning is irrational. They know, or at any rate should know, that this player is recovering from injury, and resting best positioned them to contribute. Despite doing all that could reasonably have been expected, they missed the shot. We may be tempted to say that the coach’s anger is irrational and inadequately supported by reasons. By contrast, one doesn’t (the thought goes) arrive at a perceptual experience by any process of reasoning, shoddy or otherwise. For example, one doesn’t reason to a visual experience of an orange ball. One simply decides where to look and then has (or doesn’t have) the experience.

As the forgoing discussion indicates, the two premises of the normative assessability challenge tend to be supported in two compatible ways. The first appeals to everyday discourse, namely that it’s familiar to hear normative assessments of emotions but not sensory experiences. The second reflects on the way in which emotions, in contrast with sensory experiences, arise on the basis of reasoning and/or for motivating reasons (see Naar, 2019, pp. 574–579 ff). ⁸

As we proceed, some readers might suspect that the issue for perceptualism is more precisely about fittingness, namely that emotions can be fitting/unfitting but perceptual experiences cannot. I don’t start from that perspective, however; it risks making the challenge seem of more restricted concern, for not all construals of fittingness create a problem for perceptualism. For example, Justin D’Arms and Daniel Jacobson (2000) characterize a fitting emotion as one that “accurately presents its object as having certain evaluative features” (p. 65; see also Svavarsdóttir, 2014, p. 101; Naar, 2021). Because perceptual experiences can be accurate, they can be fitting in this sense. Of course, competing characterizations of fittingness appear to threaten perceptualism. For example, one might think of fittingness as entailing “right-kind” reasons, and it isn’t obvious that there can be such reasons for perceptual experiences. On this construal, the challenge to perceptualism is about reasons, but ones of a particular variety. I return to right-kind reasons below (Section 4).

3. Assessability: Fundamental and Non-Fundamental

My defense of perceptualism draws a distinction between fundamental and non-fundamental normative assessability. Given my assumptions about how reasons and rationality relate, this distinction is ultimately a matter of fundamental and non-fundamental reasons. This section draws on Conor McHugh and Jonathan Way’s (2022) characterization of this contrast. I do so with an eye toward Section 3’s effort to shelter perceptualism from the challenge by arguing against fundamental reasons for emotion.

One inroad to the distinction between fundamental and non-fundamental reasons considers the relationship between reasons for action and reasons for intention. Take Chunhua’s:

1. 1) Chunhua’s reason to visit her grandfather is that she hasn’t seen him recently.

2. 2) Chunhua’s reason to intend to visit her grandfather is that she hasn’t seen him recently.

Both statements might be true. But consider: when such reasons exist, are there really two distinct reasons, or might there in some sense only be one? Many hold the latter position (see McHugh and Way (2022, pp. 152–156) for numerous references). That is, some hold that reasons for intending are nothing more than reasons for acting (e.g., Dancy, 2018, p. 4), while others maintain that reasons for acting are nothing more than reasons for intending (e.g., Scanlon, 1998, p. 21). Generalizing a proposal from McHugh and Way (2022, p. 152), I’ll say that reasons for φ-ing (e.g., intending, acting) are fundamental reasons for ψ-ing if reasons for φ-ing are “nothing over and above” reasons for ψ-ing (2022, p. 152). By contrast, reasons for φ-ing are fundamental insofar as they aren’t constituted by reasons of any other variety (2022, pp. 152–153). Suppose, then, that reasons for acting just are reasons for intending, or vice versa. In terms of (1) and (2) above, what this means is that both statements are ultimately made true by the same relation. ⁹ If reasons for action are fundamental, this will be a favoring relation holding between the consideration that Chunhua hasn’t seen her grandfather recently and the act of visiting him; or if instead reasons for intention are fundamental, this will be one holding between that consideration and intending to visit him. ¹⁰

Taking certain reasons to be more fundamental than others is helpful; it allows us to vindicate much of our everyday reasons talk. Consider the statement “Jake has a reason to be at work this evening.” Such a reason doesn’t seem as if it could stand on its own. ¹¹ This is because normative reasons appear to be the kinds of things to which we can respond (i.e., be motivated by). Reasons in some fashion favor a response. But being located somewhere isn’t a response. This is because persons, just like any other sort of entity (e.g., Jake’s coffee mug), can be at certain locations without being motivated by any normative reasons to be there. But simply because being located somewhere isn’t a response doesn’t necessarily mean that we need to deny that there are ever such reasons. The fundamental/non-fundamental distinction can be wielded to avoid such a counterintuitive result. ¹² After all, reasons to be at certain locations may fundamentally be reasons for what is a response. Here intentions and actions are attractive candidates. Jake’s reason to be at work this evening may be nothing over and above a reason for action (e.g., to go to work this evening) or, alternatively, a reason for intention (e.g., to intend to go to work this evening).

Understanding reasons to be at certain locations in terms of reasons for action/intention is attractive for another reason. Consider that such reasons systematically covary with reasons for acting/intending. That is, one only ever has a reason to be somewhere if they also have a reason (to intend) to do something that brings this about. If reasons of the former sort are fundamental reasons of the latter sort, this systematic covariation is easily explained. ¹³ In sum, the distinction between fundamental and non-fundamental reasons is helpful. It preserves much of our ordinary reasons talk and likewise explains why certain reasons systematically covary.

Going forward, I flag my agreement with those who treat reasons for action as fundamental reasons for intention. ¹⁴ This is because it seems that one can respond to a reason for action only insofar as they intend to do it (McHugh & Way, 2022, pp. 157–159). Actions thus qualify as responses only insofar as they are done intentionally (cf. Scanlon, 1998, p. 21). This is why my eventual proposal about reasons for emotion is in terms of intention rather than action. Those who prefer to treat reasons for action as fundamental are welcome to translate accordingly.

Here is a key dialectical point. Suppose an advocate of the normative assessability challenge rejects the distinction between fundamental and non-fundamental reasons. Going this route betrays comfort in dismissing large swathes of everyday reasons talk as metaphysically confused. Just consider the following: such a theorist is likely to reject the existence of reasons for action or (alternatively) reasons for intention. And they’re almost certain to deny that there are ever reasons to be at locations. But then if such theorists aren’t very impressed by everyday ways of talking about reasons, they shouldn’t put much stock in everyday talk about reasons for emotion (i.e., as evidence that there are such reasons). Insofar as such ways of talking are an animating source of the challenge, there is pressure for its advocates to acknowledge the distinction between fundamental and non-fundamental reasons.

4. A Perceptualist Starting Point

Much of the contemporary literature on emotions treats their normative assessability as largely uncontroversial. ¹⁵ Although I ultimately agree that they are, I find this state of play surprising. After all, people say things like, “This is just how I feel” or “You feel how you feel.” Such remarks seem intended to situate emotions (at least some of them) outside the bounds of normative assessment (see Jones, 2004, p. 333; Mitchell, 2017, p. 66). Of course, as we have seen, people also talk as if some emotions were normatively assessable. Yet if we agree that some aren’t (e.g., Naar, 2019), the desirability of a unified theory makes it tempting to explain such talk away (just as those who think emotions are normatively assessable may be tempted to explain away the relative hesitancy to critique emotions in contrast with actions/intentions). For example, talking of reasons for emotions might be explained by the fact that it is useful for changing emotional dispositions (cf. Tappolet, 2012), even if more hedged/circumspect discourse (i.e., that speaks of reasons to do what is likely to change one’s emotional dispositions) would be more accurate.

However, recognizing the distinction between fundamental and non-fundamental assessability reveals the question of whether there are reasons for emotion as oversimplified. It becomes difficult to deny that emotions are sometimes normatively assessable, a point that I hope becomes apparent. But this isn’t necessarily a problem for perceptualism, since emotions may only ever be non-fundamentally assessable. I argue that we should ultimately doubt fundamental reasons for emotions on account of the possibility and importance of what I call clear-eyed distortion. Although I have no illusions that what I say is conclusive, my ambition is modest; it is to put the perceptualist-friendly theory to follow on reasonable footing, to show it merits taking seriously.

I begin with a familiar sort of example, one involving a recalcitrant emotion, labeled Skywalk ¹⁶:

Adrian stands on the threshold of what appears to be a precipitous cliff. The wind swirls, almost causing him to stumble. His tour guide smiles reassuringly, encouraging him to step off the ledge. Doing so would be safe. There is a stable glass floor stretching from the ledge. But regardless of what his trusty tour guide tells him, and despite the fact that Adrian knows that he is safe, he nevertheless experiences intense fear. The fear persists even as he does his best to fix his attention elsewhere. Adrian nevertheless steps onto the bridge.

If Adrian’s fear involves a representation of danger, as perceptualists and some others claim (e.g., judgmentalists), then his fear counts as inaccurate. But despite this mismatch, there is a sense in which it is appropriate for Adrian to fear stepping off the ledge.

We can illustrate the way in which Adrian’s fear is appropriate by considering an analogy to ventriloquism. People observing a skilled ventriloquist ordinarily perceptually experience the puppet as speaking despite believing it isn’t. But such mismatches don’t indicate anything awry. In fact, they indicate all is well. This is principally because such illusory experiences are the output of sensory systems with the following feature: being such as to reliably generate accurate perceptual experiences in typical contexts. So experiencing the illusion is in that sense appropriate. The same is true with Adrian’s fear. In general, it is incredibly dangerous to step off a high cliff if it looks like there is nothing but air between you and the ground. Adrian’s fear is the output of emotional dispositions that tend to produce accurate representations of danger in typical contexts. This example highlights a sense of “appropriateness” that pairs naturally with perceptualism, one concerning whether the emotion is the output of well-functioning emotional dispositions. I refer to this as functional appropriateness. ¹⁷

Whether an emotion is functionally appropriate is an evaluative question concerning the emotional dispositions giving rise to it. ¹⁸ It seems to me that if one is invited to eat chocolate ice cream made to look like feces (as customers are at Toronto’s Poop Café), it is appropriate to feel a measure of disgust. Like other emotions, disgust is especially sensitive to information provided by the senses, and, in this case, visual cues strongly indicate the presence of something that would ordinarily be contaminating. Unless a person’s history includes regularly ingesting feces-appearing ice cream, a measure of disgust indicates their disgust dispositions are in good order. Other examples have an ethical flavor. For instance, if a truck driver strikes a child who suddenly appears from behind some shrubs, they haven’t necessarily done anything wrong (see Williams, 1981, p. 28). But if they’re guilt-free in the moments (or perhaps even months) after causing such a tragedy, we might be skeptical that their dispositions to experience guilt are in good order.

Turn now to the interface between functionally appropriate emotions and fundamental/non-fundamental reasons for emotion. I argue that cases of clear-eyed distortion are possible: (i) the emoter is fully aware that their emotion is inaccurate, (ii) the emotion is functionally appropriate by virtue of arising from well-functioning emotional dispositions, and (iii) the emoter has, at that moment, no decisive fundamental reasons to emote otherwise. Given my understanding of how reasons and rationality relate (Section 2), this last point amounts to the claim that the emoter is not to any extent irrational. ¹⁹ Clear-eyed distortion is possible because reasons/rationality are tied to criticism. Here is how Errol Lord puts it:

The sort of credit or blame one is open to when one is rational or irrational is…a very personal evaluation. To react rationally is to show good sense; to react in a fully irrational way is to be stupid or crazy. One can react in a less than fully rational way without being crazy, but one will always be less than reasonable or sensible when one reacts in ways that fall short of full rationality (Lord, 2018, p. 4). ²⁰

The mismatch between Adrian’s affective representation of danger and belief that he is not in danger indicates a psychology in good order. Adrian is not to any extent criticizable.

To further support the possibility of clear-eyed distortion, consider that evaluations of a given mental state should be sensitive to that state’s nature, or the kind of state that it is (see D’Arms & Jacobson, 2023, pp. 13–14). A notable feature of emotions is that they arise from information processing that is to an extent modular. The central feature of modularity, and what I focus on here, is information encapsulation. As Jerry Fodor puts it, a cognitive faculty is informationally encapsulated insofar as it “has access, in the course of its computations, to less than all of the information at the disposal of the organism whose cognitive faculty it is” (Fodor, 1987, p. 139). Emotions specifically appear to exhibit what Raamy Majeed calls synchronic encapsulation, meaning that they arise from dispositions that are “insensitive to information presently stored outside the module” (2019a, p. 291). ²¹ This is consistent with the possibility of diachronic revision (ibid.,p. 291). And certainly, our emotional dispositions do seem to change based on a wide variety of personal and environmental factors (Tappolet, 2012, pp. 210–212). ²²

Synchronic encapsulation and diachronic flexibility are both important. The latter allows Adrian’s guide to do their job without constant terror, for example. The former, by contrast, facilitates speedy responses. As Majeed puts it, “Some ecological problems we face come with significant time constraints, and the speed which is purchased by encapsulated processing systems ensures that we can emotionally respond to such situations in a quick and efficient manner” (2019a, p. 281; see also Fodor, 1987, p. 140). Furthermore, synchronic encapsulation plays a role even with time for reflection. Consider the familiar idea that emotions help to identify certain considerations as salient (de Sousa, 1987). ²³ For perceptualists, emotions manage this task through affective representations of value (Milona, 2022; cf. Majeed, 2019b, pp. 57–58). Synchronic encapsulation ensures that these representations tend to persist even when one believes that the emotion is misleading. By continuing to make something salient, the emotion invites the emoter to reconsider what they may be getting disastrously, even fatally, wrong (Majeed, 2019b, pp. 68–69).

The cost of such encapsulation is that some atypical environments will be hostile to emotional reliability (Majeed, 2019a, pp. 288–289). The hostility consists in decoupling ordinarily highly reliable indicators of a given value (e.g., danger) from the value itself. But emoters don’t necessarily merit critique when they find themselves in such an environment and recognize their emotion to be misleading. This is because well-functioning emotions have a measure of inertia (i.e., synchronic encapsulation), a feature that, while often helpful, leaves one exposed to potential error. This is why I have proposed that in Skywalk it would be a sign of something awry if Adrian weren’t afraid of venturing across the walkway. Part of the point of such attractions is that they sever the link between an ordinarily reliable indicator of danger and danger itself. But readers needn’t agree about Skywalk to agree that clear-eyed distortions are possible. Given synchronic encapsulation, any agent with well-functioning emotional dispositions may experience a functionally appropriate emotion in an environment that is, at least for a time, an emotional “house of mirrors.” ²⁴

For readers unpersuaded by Skywalk, other candidates for making the possibility of clear-eyed distortion vivid feature emotions in the context of immersive role-playing games. Such games regularly offer players the option to perform “evil” in-game actions. For example, one may play a character who betrays their lover for financial gain, or one who murders a brother to take the throne. These may be fine choices to further a narrative, and they certainly haven’t wronged anyone. This becomes especially apparent if we limit our focus to single-player video games whereby one’s choices only impact programmed entities in a virtual world. But given how immersive such games can be, it strikes me that a measure of guilt for fictional wrongdoings can nonetheless signal well-functioning emotional dispositions, and insofar as guilt is functionally appropriate, no criticism is merited even given awareness of inaccuracy. ²⁵ And if no criticism is merited, there is no failure to respond to any decisive fundamental reasons against the emotion. In sum, taking clear-eyed distortion to be possible is a reasonable stance, given reflection on cases as well as synchronic encapsulation.

The possibility of clear-eyed distortion is a crucial step in my argument against fundamental reasons for emotion. But it doesn’t straightaway secure that result. An advocate of fundamental reasons for emotion may adopt the following general account of such reasons: considerations that the emoter’s dispositions to experience the relevant emotion would respond to if those dispositions were well-functioning. This view allows that agents undergoing clear-eyed distortion are responding to sufficiently weighty fundamental reasons for the relevant emotion, even while being aware that the emotion is misleading.

But this isn’t a promising story about cases of clear-eyed distortion. To begin, consider that belief and intention never work in this way. A person who believes something despite being aware of considerations in virtue of which their belief is inaccurate, and who, moreover, recognizes these considerations as such, is always criticizable. ²⁶ A person is likewise criticizable if they intend to do something while being aware of decisive considerations in virtue of which they should not, at least insofar as they likewise recognize these considerations as such. It seems, then, that we evaluate beliefs and intentions in ways that disallow anything like the first-personally manifest gaps between accuracy and functional appropriateness that can arise with emotions. What’s more, sensory experiences do allow for such gaps. And yet we don’t say that in such cases (e.g., the ventriloquism case) the illusory experiences are fundamentally rational. Or at least the advocate of the normative assessability challenge had better not say this. So while there are considerations that a person’s sensory dispositions will be sensitive to if they are well-functioning, these are not fundamental reasons for the sensory experience.

In general, then, given that evaluations of emotion pattern similarly to evaluations of sensory experience—both allow for first-personally manifest gaps between accuracy and functional appropriateness in a way that beliefs and intentions never do—the default assumption should be that the same evaluation is present in each case. Clear-eyed distortion thus reveals serious grounds to doubt fundamental reasons for emotion, for it isn’t clear what such reasons could plausibly be in those cases. One may be tempted to add: because emotions are (synchronically) encapsulated, they are not sufficiently connected to a person’s agency, even if they reflect a person’s self/character, in contrast with intentions and beliefs, to be subject to fundamental reasons. This reflects the common view that emotions are passive in a manner analogous to perceptual experience whereby the world impinges on us rather than vice versa. ²⁷ Of course, perhaps a compelling argument will appear requiring us to graft another layer of evaluation onto functionally appropriate emotions (but not sensory experiences), namely an evaluation in terms of fundamental reasons. I speak to some of these considerations below (Sections 5 and 6).

5. Reasons for Emotion as Non-fundamental

This section sketches an account of reasons for emotion in the wake of the foregoing doubts about fundamental reasons for emotion. I call it the perceptualist-friendly theory. The aim is not to argue that this is the best account of reasons for emotion, a task beyond the scope of this paper, but rather that it is worth taking seriously and emerges naturally out of consideration of reasons for sensory experiences. Note that, while emotion theorists sometimes seem to suggest that any talk of reasons for sensory experience is confused (e.g., Brady, 2011; Deonna & Teroni, 2012), this isn’t true. There’s nothing problematic about saying, for example, “Maria has a reason to see the Grand Canyon.” But this tells us almost nothing about the nature of visual experiences. After all, Maria also has a reason to be at the Grand Canyon next summer, and being somewhere isn’t even a mental state. Trouble only sets in if we theorize such reasons as fundamental. The present proposal, however, says that such reasons are non-fundamental.

The central claim of the perceptualist-friendly theory is that reasons for emotion are fundamental reasons for intention. To illustrate, start with a reason for sensory experience: Maria’s reason to see the Grand Canyon. This is plausibly nothing more than a reason for her to intend to visit the Grand Canyon to direct her attention across the landscape to have a visual experience of it. Similarly, Reynold’s reason to be angry at the politician for lying is fundamentally a reason for Reynold to intend to do what will bring about such anger. This can involve numerous different emotion regulation techniques (see Gross, 1998). Most obviously, it might be a reason to intend to focus on the politician’s lie to bring about anger at him for lying. Another possibility is that Reynold’s reason to be angry is a reason to intend to put himself in a situation (e.g., a political protest) that will generate such anger. In general, the idea is that reasons for emotion and sensory experience are like reasons to be located somewhere. Although there aren’t reasons that fundamentally govern being at certain locations, our ability to exercise agency to be at certain locations makes possible non-fundamental reasons of this variety. The perceptualist-friendly view treats reasons for emotion/sensory experience as being brought into the realm of reasons (and so rationality) through fundamental reasons to intend. Furthermore, I want to be open about what explains such reasons. But to make the proposal more concrete, I am attracted to a view appealing to the value of the relevant experiences. This might be because the experience is intrinsically valuable, a possibility about which I’ll say more below, or because it brings about some benefit (e.g., being able to brag about having seen something, being able to join in solidarity with others who are justly angry).

One might complain that, even if there are reasons for sensory experiences and emotions of the sort just described, they fit the mold of what are sometimes called “wrong-kind” rather than “right-kind” reasons (see Gertken & Kiesewetter, 2017). ²⁸ What this distinction amounts to is notoriously controversial, but here is one common characterization: right-kind reasons are those emerging from an attitude’s constitutive standards of correctness, while wrong-kind reasons are those that do not (Lord, 2023, p. 253). For example, a father may have a reason to admire his daughter’s performance in the school play, which is that it will improve her self-esteem. This is a wrong-kind reason; it has no bearing on whether such admiration would be correct (in some intuitive but as yet unspecified notion of “correct”). By contrast, a right-kind reason for admiration, one bearing on admiration’s correctness, is that the child never missed a line. The worry is that while the perceptualist-friendly account might be capable of explaining reasons that fit the mold of wrong-kind reasons, it isn’t suited for capturing anything like right-kind reasons.

However, the perceptualist-friendly view can arguably capture this distinction. Starting with right-kind reasons, take the father’s reason to admire his daughter’s performance in the school play: that she never missed a line. The present view requires that this fundamentally be a reason for intention. What I propose is that it is a reason to intend the following: to focus on this fact (i.e., that she never missed a line) as grounds for admiring her performance. To unpack what this amounts to, remember that perceptualists conceive of admiration as a kind of experience of value. Let’s say, just to fix ideas, that admiration is more precisely an experience of excellence. The relevant fundamental reason can thus also be glossed as a reason to intend to focus on the fact that his daughter never missed a line as grounds for experiencing the excellence of her performance. Focusing on her performance in this way, if admiration follows, makes for well-functioning admiration. On the present proposal, then, what makes a given reason a right-kind reason for emotion is that it is fundamentally a reason to intend to focus on what would be constitutive of a functionally appropriate emotion of the relevant sort. ²⁹ Emotional “correctness” is thus understood more precisely in terms of emotional well-functioningness.

The perceptualist-friendly theory thus draws a contrast between functionally appropriate emotions and ones responsive to right-kind reasons. This is easy to see in shifting to nonhuman emotions. Imagine a gazelle coming to fear a cheetah upon observing the predator prowling nearby. The gazelle has not intentionally focused on the cheetah as grounds for experiencing danger. It may not even have the capacity to do so. Assuming it doesn’t, the perceptualist-friendly theory should say that some creatures capable of well-functioning emotions (e.g., infants, many nonhuman animals) lack right-kind reasons for them. If one thinks that (many) nonhuman animals aren’t normatively assessable at all in virtue of their emotions (see Naar, 2019), this result of the perceptualist-friendly theory is a happy one. ³⁰ Returning to the father-daughter example, the father may similarly have well-functioning admiration for his daughter’s performance on account of her never missing a line, without forming a relevant intention to focus in a way that brings about such an experience. Here his emotion is caused by what is in fact a right-kind reason, namely that his daughter never missed a line, but he hasn’t exercised any rational capacities in doing so. ³¹ I thus return to a point noted above: emotions are brought into the realm of normative assessability (i.e., the realm of reasons and rationality) by virtue of emoters having the capacity to contemplate their own experiences and, in particular, to form intentions of the sort described.

Now consider wrong-kind reasons. Suppose the father’s reason to admire his child’s performance in the school play is that it will improve her self-esteem. What qualifies this reason as a wrong-kind reason, on the perceptualist-friendly view, is that it doesn’t meet the standards of being a right-kind reason; focusing on this consideration would not be constitutive of well-functioning admiration. That is, if the father’s admiration consisted in experiencing the fact that it would improve their child’s self-esteem as indicative of the excellence of the child’s performance, this would not be functionally appropriate admiration. The explanation for this is simple: that it would improve one’s self-esteem to be admired for something doesn’t reliably make for or track the excellence of that something. In sum, perceptualists have a reasonable way to distinguish right-kind and wrong-kind reasons, making use of the notions of non-fundamental/fundamental reasons and functionally appropriate emotions. But as the right-kind/wrong-kind reason distinction is so contentious, this proposal should be seen as preliminary, awaiting further exploration and comparison with alternatives.

Nevertheless, the following question is dialectically pressing: does something like the right-kind/wrong-kind distinction arise for sensory experiences? I propose that, at least for some such experiences, it arguably does. The relevant experiences are those whereby the experience of one property is (synchronically) dependent on the experience of others. To illustrate, consider a hiker observing a flower that is in fact a white dryad. Since this hiker is trained in botany, they also have a visual experience as of a white dryad. ³² This experience is made possible by their experience of other properties, including color and shape properties, which are from their perspective white-dryad-indicating. Here I assume that this is a well-functioning white dryad experience. Now consider whether they have a reason for this experience. Specifically, I am asking whether the colors and shapes of the flower give them reason to experience it as a white dryad, a reason that would parallel the right-kind reasons for emotion described above. Plausibly they do have such a reason that is fundamentally a reason for the hiker to intend the following: to focus on those colors and shapes as grounds for an experience of the flower as a white dryad. ³³ So long as we allow that this reason is explained by the well-functioningness of the experience in question, it works just like the above proposal for right-kind reasons for emotion. Furthermore, they may also have a wrong-kind reason to see the flower as a white dryad, whether it would be functionally appropriate to do so, because doing so offers some benefit.

This proposal about sensory experience makes salient the question of why well-functioningness, whether sensory or emotional, would ever be a source of reasons, i.e., considerations that carry any normative authority/weight. A tentative suggestion is that there is intrinsic value in functionally appropriate emotional/sensory experiences. ³⁴ This is easier to see in the case of emotions since they are oriented to values: it seems plausible that there is intrinsic value in the well-functioning exercise of evaluative capacities. This strikes me as true even when the result is an inaccurate representation of value. By way of comparison, an agent whose action arises from ethically virtuous characteristics or habits may sometimes do the objectively wrong thing, perhaps because they have limited information or find themselves in an atypical situation that requires quick decision; yet it strikes me as plausible that there is nonetheless intrinsic value in the exercise of those capacities. Furthermore, and more tentatively, it may seem as if there is intrinsic value in the well-functioning exercise of sensory/empirical capacities, even independently of the accuracy of the resulting experiences (granting that well-functioning experiences will be accurate in most typical contexts). Such inaccurate experiences can nonetheless be an expression of a kind of sensory virtue, or excellence. ³⁵

6. Emotions and Reasoning

This section and the next clarify the perceptualist-friendly approach to the normative assessability of emotions by addressing two related objections. The first says that the possibility of reasoning to emotions, in contrast with sensory experiences, indicates that there are fundamental reasons for emotion (see Naar, 2019, pp. 574–579 ff.). The most straightforward way to motivate this purported asymmetry is by way of examples. Consider the following, which I call Priorities Hoping:

A recent college graduate is meditating on their hopes for the future. They are tempted both by the offer of a high-paying job in their hometown as well as an offer to pursue graduate school abroad. As they weigh the different possibilities, their competing hopes wax and wane.

By contrast, we don’t seem to similarly arrive at or adjust sensory experiences. For example, if one is looking in the direction of a blue car, if all goes well, one simply experiences the car as blue.

But matters are not so straightforward. The contrast between sensory experiences and emotions here is plausibly one of frequency. Sensory experiences can be sensitive to reasoning in ways that are, at least on the surface, quite similar to Priorities Hoping. Consider a person having a visual experience as of another human in the distance. But then a shift occurs: what they were experiencing as a human now appears as a statue. The shift is precipitated by a thought process, including about how unlikely it is that someone could be so perfectly still in such windy weather. Cases like this call for caution in drawing contrasts between emotion and perception on the basis of cases such as Priorities Hoping. Simply because an experience shifts in response to reasoning, we shouldn’t immediately assume that this is a (fundamentally) normatively assessable transition rather than a merely causal one.

Perceptualism leads to a simple story about how emotions relate to reasoning. Here I adapt an insight from Michael Milona and Hichem Naar (2020, pp. 3081–3083), whose aim is to investigate what perceptualists should say about the relationship between emotions and the representations on which they depend. For example, one cannot fear a snarling dog unless one somehow represents a snarling dog, and the emotion itself doesn’t contribute that representation. As Milona and Naar point out, it is natural for perceptualists to view this relationship as analogous to that between sensory representations of empirical properties and the proximal stimulations that give rise to them. For example, just as certain light arrays on one’s retina may generate a visual representation as of a green cube, so too may certain mental representations (e.g., perceptions with certain content) give rise to an experience of fear (i.e., an affective representation of danger). ³⁶ Furthermore, the relationship between proximal stimulations and sensory representations is merely causal; there is no (direct/unmediated) normatively assessable relation. The fact that emotions can be caused by other mental states, including those featured in complex chains of reasoning, may make it seem as if emotions were normatively assessable in a way sensory experiences are not. But barring further argument, this may be a misleading appearance rooted in the typical causes of emotion in contrast with sensory experiences.

The above analogy notwithstanding, an opponent of perceptualism may find it suspicious that emotions paradigmatically, or perhaps even by their nature, respond to other mental states, and it may seem that mental states that by their nature respond to other mental states are the best candidates for being fundamentally assessable. I am not sure how seriously we should take this suspicion, but the perceptualist can helpfully speak to it by noting that certain sensory experiences seem to require other mental states much as emotions do (cf. Faraci, 2015, p. 2065). For example, if a person is watching a skilled ventriloquist, and their sensory capacities are functioning well, then they will have a multimodal sensory experience as of the puppet talking. Such an experience is rooted in visual experiences as of the puppet’s mouth moving in certain ways and an auditory experience as of sound coming from that direction. That is, if it weren’t for the latter visual and auditory representations, the multimodal representation wouldn’t be possible. So any attempt to argue that there are fundamental reasons for emotions on the basis of the idea that we can reason to emotions needs to motivate an explanation of how emotions arise from reasoning in cases like Priorities Hoping (such that it grounds fundamental reasons for emotions) that is unlike the way in which sensory experiences sometimes arise from other such experiences. For my part, I’m not optimistic that any compelling explanations of this sort are forthcoming. But one may attempt to offer one in terms of motivating reasons, and it is to that possibility I turn.

7. Motivating Reasons

It is sometimes said that emotions, in contrast with sensory experiences, are had for motivating reasons. And if emotions are had for motivating reasons, then presumably these reasons can be good or bad, and good and bad motivating reasons perhaps amount to (or are strong evidence of) real and apparent fundamental normative reasons. ³⁷ Although I don’t have space to fully explore each step of this reasoning, I aim to blunt its force by exploring the sense in which there are motivating reasons for emotion.

The thought that emotions are had for motivating reasons can be supported by considering different “why” questions that we ask about emotions (see Müller, 2017, pp. 285–286). For example, the question, “Why are you [Reynold] angry at the politician?” might be probing causal reasons, normative reasons, or (apparently) motivating reasons. If the question were understood as about causal reasons, then Reynold might answer by appealing to his irritable mood (or even his upbringing, biological makeup, etc.). If understood in terms of normative reasons, Reynold might point to the politician’s voting to start a war. This isn’t to suggest that the vote isn’t also a cause, but here Reynold would be pointing to it as justification for his anger. The question may also be understood as probing the features of the situation in light of which the emoter is angry. These features might be glossed as the emoter’s motivating reasons for their emotion. In Reynold’s case, the answer here might also appeal to the politician’s vote. But Reynold could answer the question truthfully even while he doesn’t think his motivating reasons are good ones (i.e., motivating reasons that are also normative reasons).

My response on behalf of perceptualism has two parts. The first is to notice that, when we ask people why-questions about their emotions, we are often probing special sorts of causes that are, from a theoretician’s perspective, easily confused with motivating reasons. We can see this by revisiting the analogy with sensory experience and the ventriloquism example. In that case, the sensory experience of the puppet as talking is rooted in other sensory experiences. One would not be able to experience the puppet as talking if not for seeing the puppet’s mouth as moving in a way that coordinates with a voice coming from that direction. Now imagine someone asking, “Why did you see the puppet as talking?” Or, to put the question more naturally, “What about the puppet leads you to see it as talking?” We can imagine a child asking such a question of a parent, the latter having just seen a ventriloquist. The parent could give a causal answer that entirely misses the special sort of cause that interests the child. For instance, the parent might note that they had just earned a bonus at work and so indulged in some local entertainment. The relevant causes, however, are the sensory experiences that brought about the multimodal illusion. Once this point is recognized, it’s easy to conjure similar examples. One might ask an ornithologist, “Why do you see that bird as a wren?” Here the questioner doesn’t want just any cause. They want to know about the lower-level sensory cues in light of which the ornithologist experienced it as a wren. In general, then, we sometimes ask “why” questions about sensory experiences that probe a special sort of phenomenologically salient experiential grounds for the experience.

When it comes to emotions, we similarly ask “why” questions that probe certain special causes. For example, return to the question “Why are you [Reynold] angry at the politician?” This may be asking about a cause, but not just any cause. We might not care, for instance, about Reynold’s irritable mood (or his upbringing or biology) even though that is indeed among the causes of his anger. We likely want to know about what Reynold takes the politician to have done that leads him to anger. As I have already noted, if emotional experiences are experiences of value (whether understood along a perceptual model or otherwise), then there will ordinarily be something of which the emoter is consciously aware that is the putative source of value. ³⁸ For instance, Reynold’s anger might involve taking the politician to have betrayed his constituents, including Reynold, by reneging on a campaign promise not to vote in favor of new wars. So when we ask people questions such as, “Why are you angry?” or “In light of what are you angry?” we may not be probing motivating reasons but rather certain causes that form part of the conscious basis for the emotional attribution of value.

Of course, we can ask people explicitly about what motivates their emotions. For example, one might ask Reynold, “What motivates your anger toward the politician?” or “Why are you motivated to be angry at the politician?” The second part of my response, then, is to adapt the fundamental/non-fundamental distinction to the case of motivating reasons. For example, suppose we ask, “Why aren’t you hungry for dinner this evening?” This might be a question about motivating reasons. ³⁹ But motivating reasons to be hungry (or not) are presumably nothing over and above the reasons for which one intends to act that predictably lead to being hungry (or not). To illustrate, one might answer, “A colleague baked cookies for the office and I just couldn’t resist.” The perceptualist, then, naturally understands motivating reasons for emotion in terms of motivating reasons for intentions (including intentions to focus on certain things).

A proponent of the objection may nonetheless insist that there is a sensible question—one which doesn’t rest on a confusion—about why someone emotes that cannot be understood as probing special causes or non-fundamental motivating reasons and that can only be interpreted as about fundamental motivating reasons (cf. Hieronymi, 2005, p. 454). Yet in everyday life, the different possibilities for what people are querying don’t typically come so neatly disentangled. ⁴⁰ In theorizing emotion, then, I caution against treating the possibility of fundamental motivating reasons as a pretheoretical fixed point or default assumption. Furthermore, and stepping back a bit, we should recall an important assumption of the present objection, namely that fundamental motivating reasons and fundamental normative reasons come as a package. Given this assumption, Section 3’s doubts about the latter raise doubts about the former. In sum, then, perceptualists can reasonably be skeptical about fundamental motivating reasons for emotions.

8. Conclusion

In sum, the normative assessability challenge for perceptualism is unpersuasive. This is because it requires that reasons for emotion be fundamental, and we should be skeptical that there are such reasons. I supported such skepticism by defending the possibility and importance of cleared-eyed distortion. Clear-eyed distortions reveal a crucial way in which the evaluation of emotions is importantly akin to sensory experience, for which there are no fundamental reasons, and unlike belief and intention, for which there are. From a vantage point of skepticism about fundamental reasons for emotion, I developed the perceptualist-friendly theory of reasons for emotion as fundamentally for intention. Furthermore, apparent problems for this theory emerging from assumptions about how emotions can arise, either from processes of reasoning or on the basis of reasons, are much less threatening than first appears. Going forward, it should not be taken for granted that the normative assessability of emotions is a problem for perceptualism; the perceptualist-friendly theory merits serious consideration.