1 1. Introduction

Emotions are taken by some authors as a kind of mental state epistemically akin to perception. Emotional experience (e.g., anger about the assassination of an innocent) is said to justify evaluative judgments (e.g., the judgment that this is unjust) and actions (e.g., retaliation) in the same way in which perceptual experiences (e.g., seeing a red sphere-like object in your surroundings) justify existential beliefs (e.g., the belief that there is an apple in this room) and/or actions (e.g., extending your arm to pick up the apple).

Unlike perceptual phenomenology, which allows being treated dogmatically (Huemer, 2001; Pryor, 2000), emotional phenomenology is puzzling in the following respect. When you feel an emotion, you feel an urge to act, you feel, among other things, your body’s action readiness (Deonna & Teroni, 2012; Frijda, 2007; Scarantino, 2014). On the other hand, at least sometimes, you are aware that an emotion by itself is not a sufficient reason to justify an evaluative judgment and/or an action, not even prima facie (Brady, 2013; Brogaard & Chudnoff, 2017). Emotional phenomenology seems to be at once dogmatic and hypothetic. The meaning of “dogmatic” is taken from the view in epistemology called phenomenal dogmatism (Huemer, 2001; Pryor, 2000). As we will see, phenomenal dogmatism is the view that, absent defeaters, if it seems to you that P, then you have prima facie justification for believing that P. By contrast, when I talk about a type of phenomenology being epistemically hypothetic what I mean is that it lacks the phenomenal character required to dogmatically justify a judgement/action. This is the new puzzle addressed in this paper. New, because, to my knowledge, it has not been articulated yet in the literature.

Mental states are rationally causally related to other mental states (Davidson, 1980; Zangwill, 1998). If you desire that P, and you believe that P, only if Q, then that tends to rationally cause your instrumental desire of Q. It seems that emotions also have those rational functional roles. If you see a predator and you feel fear, it seems to you that that tends to rationally cause the formation of a conscious evaluative judgment and/or an avoidance behavior.Footnote 1 I will call the former, the justificatory role of emotion for forming conscious evaluative judgments. I will call the latter, the guiding role of emotion for action and evaluative judgements. Some authors have claimed that certain mental states’ guiding roles are constituted by their justificatory roles. For instance, Elijah Chudnoff (2013) claims that intuition’s guiding role for mental actions, like inference (e.g., inferring the consequent), is grounded in intuition’s justificatory role (e.g., to know that the antecedent is true and “see” how its truth necessitates the truth of the consequent). The same seems plausible mutatis mutandis for perception. However, it seems that emotions alone fall short of the justifying role in which their guiding role would be grounded. Unless we are willing to claim that emotions cause action blindly (i.e., not rationally), we need an account of the distinctive epistemic role of emotion that renders its guidance role rational, but not in a way similar to perception or intuition.

Some caveats about the nature of the puzzle are in place. It should be noted that when talking about the realization that one’s emotions are not sufficient reason to justify one’s evaluative judgments/actions, we are not talking about a second order mental state that has emotion as its content and predicates of it that it is not sufficient reason to justify other mental state/action. Rather, what is being claimed is that emotional phenomenology lacks some property P, possibly present in other phenomenal states, that would render it able to immediately justify other mental states/actions. The absence of this property can sometimes be felt, and then, the puzzle is triggered: your emotion compels you to act while you know the emotion alone is not sufficient for justifying your action. It may sound paradoxical that one can feel an absence, but it is not. One may or may not be aware that emotional phenomenology lacks property P. When one is aware of it, we have the puzzle and there are ways in which one can be aware of this absence. Consider that perceptual phenomenology has a phenomenal justificatory property, as some authors claim (Chudnoff, 2013; Huemer, 2001; Pryor, 2000). Now imagine that one is comparing their emotions to their perceptions and “seeing” that perception justifies existential beliefs in virtue of its phenomenology, but emotion cannot justify evaluative beliefs in virtue of its phenomenology alone. In this case one is aware that emotional experience lacks something perceptual experience has. One is aware of this absence in comparing their phenomenal perceptual experience with their phenomenal emotional one.

Another caveat: perceptual and emotional experiences are both fallible sources of factual and evaluative knowledge respectively. However, when I say that perceptual experiences allow being treated dogmatically, whereas emotional experiences don’t, I am referring to the phenomenal character of their epistemic commitment. Perceptual experience is felt as epistemically dogmatic whereas emotional experience is not. As we will see in the lines to come, this may be explainable in terms of perceptual experience’s possession of presentational phenomenology and emotional experience’s lack of it (Brogaard and Chufnoff, 2017). My claim is that, irrespectively of their fallibility, the phenomenology of perception is dogmatic whereas the phenomenology of emotion is not or not entirely dogmatic. Emotional phenomenology would be open, by its very nature, to criticism. That is, perceptual phenomenology has a dogmatic qualitative character (it feels dogmatic so to speak) whereas emotional phenomenology is, besides dogmatic in a certain sense, qualitatively epistemically hypothetic (it feels epistemically hypothetic). Thus, both perception and emotion are fallible, though perceptual phenomenology has a phenomenal dogmatic character, in an epistemic sense, which emotional phenomenology lacks.

As a final caveat, it is important to note that this non-dogmatic aspect characteristic of emotional phenomenology has nothing to do with the introspective ability of the subject of the emotion to know in which emotional state they are. Pace psychoanalysis, the subject of the emotion can justifiably be dogmatic about their justification to form a belief about the emotional state in which they are on the basis of introspecting their emotion. However, what the subject cannot be dogmatic about is the information conveyed by the emotion about the axiological properties of the world. The subject may be and feel infallible regarding their emotions, but their emotions are not, nor are always felt as, infallible sources of axiological information.

The plan of the paper is the following. Once the puzzle has been outlined, the time comes to diagnose why it has gone unnoticed so far. I will outline the puzzle and diagnose the reasons why it went unnoticed in Section 1. Section 2 develops the consequences of the puzzle and addresses its impact in the philosophical discussion around the nature and epistemology of emotion, especially for the perceptual theory of emotion. Finally, Section 3 aims to offer the articulation of some tentative but, I believe, promising ideas to start (dis)solving the problem, plus some objections to them and my answers to those objections.

2 The puzzle and why it has gone unnoticed

In this section, I will try to develop the folk, scientific and philosophical considerations that support the idea that emotional phenomenology seems to be Janus-Faced, both manifestly unable to justify action but strongly motivating. I will also explain why this new puzzle has gone unnoticed in the literature so far.

2.1 Laws in tension?

Let’s start with Nico Frijda’s laws of emotion. Frijda’s considerations about emotion are synthesized in the inductive generalizations that constitute his laws of emotion (Frijda, 2007). One of the main contributions of Frijda’s account is the idea that emotions are patterns of felt bodily action readiness to react in the face of emotionally relevant objects. These patterns have a law-like behavior, summarized in the mentioned laws. What is significant for the question under discussion is that there is a tension between central laws. My claim is that this tension manifests our puzzle. Here is Frijda’s Law of Closure:

Emotions tend to be closed to considerations that its aims may be of relative and passing importance. They are closed to the requirements of interests other than those of their own aims. They claim top priority and are absolute with regard to appraisals of urgency and necessity of action, and to control over action.

(Frijda, Ibid., p. 15).

Now consider the Law of Care for Consequence:

Emotions are not always as absolute as just sketched. Emotions do manifest deliberation, calculation, or consideration. Infatuation can be stingy, and anger can be prudent. Although closure reflects the basic shape of emotions, that basic shape may run into opposite tendencies, usually caught under the heading of emotional regulation or emotion control. These manifest the law of care for consequence. Every emotional impulse elicits a secondary impulse that tends to modify it in the view of its possible consequences.

(Frijda, Ibid., p. 17).

These two laws seem to be in acknowledged tension. The law of closure seems to capture the dogmatically motivational role emotions play in the evaluative life of their subjects. By contrast, the law of care for consequence seems to capture the openness of emotion to criticism and consideration which reflects their non-dogmatic phenomenal aspect. The apparent tension between these two laws can be taken as an expression of our puzzle.

2.2 Degrees

While it is true that perceptual contents can vary in their degree of vivacity (e.g. this shade of green is brighter than that one), they do not vary in their degree of epistemic commitment. While one subject can experience the apple as more reddish than other subject perceives it, if both subjects really undergone an episode of perceptual phenomenology then, irrespectively of the vivacity of the phenomenal properties instantiated by the perceived objects, both represent-as-existent the apple (Kriegel, 2015). Thus, if one has perceptual experiences and other things being equal, one has sufficient reason to move to a corresponding perceptual belief, irrespectively of the vivacity with which the perceptual contents are presented. Perceptual experience is constituted by a dogmatic attitude, it enables you to form a justified perceptual belief, other things being equal.

By contrast, emotional phenomenology is subjected to degrees of epistemic commitment. One can emotionally represent a particular object (e.g., an insult) as instantiating some evaluative property (e.g., offence) to a greater or lesser extent. Thus, the intensity of an emotion, or in the current terminology, the degree of epistemic commitment of an emotion, can vary. This property of emotions is akin to hypothetical or conjectural thought. To illustrate this point, consider Frijda’s Law of Apparent Reality:

Emotions are elicited by events with meanings appraised as real, and their intensity corresponds to the degree to which this is the case.

Events that are taken to be real elicit emotions. Threats deploying in front of one’s eyes, opportunities that are for the taking with smells, gestures or glances that caress and evoke terror or tenderness or mellowness within. What does not impress as true and unavoidable elicits no emotion, or a weaker one.

(Frijda, Ibid., p. 8).

2.3 Problems with emotional dogmatism

Phenomenal dogmatism is the view that, absent defeaters, if it seems to you that P, then you have prima facie justification for believing that P (Huemer, 2001; Pryor, 2000). Phenomenal dogmatists tend to understand the term “appearance” in terms of a phenomenal experience akin to a propositionally structured representation. The propositional content of, for instance, perceptual experience allows it to enter into logical relations with perceptual belief, such that perceptual experience’s propositional contents can justify perceptual beliefs’ propositional content. Phenomenal dogmatism is a very plausible view to account for the epistemic relation between perception and perceptual beliefs. If it perceptually seems to me that there is a rat under the table, this is prima facie sufficient reason to justify my belief that there is a rat under the table. One of the reasons for the plausibility of phenomenal dogmatism for the case of belief is due to the fact that, pace skeptics about the external world, perceptual experience is considered one of the privileged epistemic accesses to the external world.

Is phenomenal dogmatism about emotion a plausible option? I do not think so. Michael Brady (2010, 2011, 2013) has given three related reasons to doubt that emotional dogmatism is a plausible view. First, unlike perceptual experience, emotional experience seems to be subject to reasons. Whereas it would not make sense to ask a normative why-question regarding why you are having a certain perceptual experience, it seems to make sense to ask such a question regarding emotional experiences (see also Deonna & Teroni, 2012). The only type of why question that seems appropriate for perception is the causal one. The fact that perceptual experiences do not accept normative why-questions is also related to the fact that perceptual experiences are taken as justificatory bedrock when it comes to perceptual existential beliefs. The same cannot be said about emotional experience. To make this reasoning more vivid, take the following passage by Pryor and try to devise similar considerations for emotion. I bet that you would find it difficult:

When asked, “what justifies you in believing there are hands?” one is likely to respond, “I can simply see that there are hands”. One might be wrong: one might not really be seeing a hand. But it seems like the mere fact that one has a visual experience of that phenomenological sort is enough to make it reasonable for one to believe that there are hands.

(Pryor, 2000, p. 536).

Second, if phenomenal dogmatism about emotional experience were true, then, as a rule, every emotional appearance undefeated at the moment would constitute sufficient reason for the corresponding evaluative judgments and, plausibly enough, for an action intelligible in the light of that emotion. However, Brady shows us how this would lead to undesired consequences. The following is a reconstruction of Brady’s argument.

  1. (1)

    Emotions and evaluative judgments have the same set of reasons. That is, something is a reason R for an emotion iff R is a reason for the corresponding evaluative judgement.

  2. (2)

    By interdefinibility of the biconditional in (1), it follows that: If R is a reason for an emotion, then R is a reason for the corresponding evaluative judgement and if R is a reason for an evaluative judgement, then R is a reason for the corresponding emotion.

  3. (3)

    By elimination of conjunction in (2) we obtain: If R is a reason for an evaluative judgement, then R is a reason for the corresponding emotion.

  4. (4)

    Perceptualist/Dogmatist claim: an emotion is a reason R for the corresponding evaluative judgement, other things being equal.

(C) An emotion is a reason R for the corresponding emotion.

Premises (1), (2) and (3) seem very plausible: the set of appropriate reasons for an episode of fear about object O, is the same set of reasons for the corresponding evaluative judgement about the dangerousness of O. Premise (4) is the perceptualist claim as defended by authors like Christine Tappolet (2016). (C) follows if one jointly accepts all those premises. Now, premises (1), (2) and (3) are very plausible, so one should reject (4) in order to avoid the undesired consequence (C).

Third and narrowly linked with the previous consideration: phenomenal dogmatism about emotional experience collides with an independently plausible theory about the metaphysics and semantics of evaluative discourse, namely, neo-sentimentalism (D’Arms & Jacobson, 2005; McDowell, 1985; Wiggins, 1987). Neo-sentimentalism is the view that our appropriate deployment of evaluative concepts entails the recognition that the object of which we predicate the evaluative concept merits certain emotional reactions. Thus, if I predicate the evaluative concept < dangerous > of a certain object, that entails recognising fear about that object as a fitting response. Now, epistemic emotional dogmatism is equivalent to this conditional: if I have an emotional episode in normal circumstances, then the corresponding evaluative judgment is prima facie justified by the emotion. If we put together emotional dogmatism and neo-sentimentalism we obtain, by transitivity, that having an emotional episode in normal circumstances is sufficient for recognising that emotion as an appropriate or fitting response to the relevant object. Again, we obtain the undesired and circular upshot that emotions can justify themselves. Given the independent plausibility of neo-sentimentalism about evaluative concepts, we have reasons to reject phenomenal dogmatism about emotional experience.

If Brady is right, we have reasons to reject phenomenal dogmatism for emotional experience. In the next section, I will address some recent criticism of Brady’s arguments by Jonathan Mitchell (2017). For current purposes, it is sufficient that we acknowledge that once we reject phenomenal dogmatism for emotions, we have our puzzle. If phenomenal dogmatism about emotional experience were true, we could easily explain the action readiness of emotion: it seems to us that we are directly perceiving a value and for that reason we feel the need to act correspondingly. However, if Brady is right, then we cannot feel emotional experiences as direct perceptions of value. However, and here is the puzzle, we behave as if we were experiencing emotions that way.

Now someone can ask: how does this paper go beyond Brady’s proposal? Footnote 2For starters, my problem is a problem for Brady himself: if Brady is right, we cannot take emotional experiences as direct perceptions of value, but we behave and feel as if we were experiencing emotions that way. How is it possible to keep the rationality of emotions when they have such a contradictory phenomenology? Perception’s guiding role seems to be grounded in perception’s justifying role (e.g., I avoid the car, perception’s guiding role, because I see it coming towards me, perception’s justifying role), but if emotions lack a similar justifying role, then what guarantees that their guiding role is rational? Unless we are willing to assume that emotions are irrational causes of our behavior, we, including Brady need a solution to the puzzle here presented.

2.4 Why has the puzzle gone unnoticed?

If the previous considerations are on the right track, we have confirmed a new puzzle for the philosophy of emotion. The natural question now is: why has this puzzle remained unnoticed in the literature so far? The reason is to be found in the puzzle itself.

Emotional experiences are pervasive, and they play if not an essential at least a typical role in our evaluative mental economy. They seem to be present in high cognition when, for instance, we experience awe contemplating the vastness of the universe. They also seem to be present in low cognition when, for instance, we experience fear at the sight of a spider in our bathroom. Emotion’s pervasiveness makes our puzzle really pressing, because, if they are present in the etiology of most of our actions and they are not rational, then that could have the consequence that most of our actions were infected with the purported irrationality of their base. That is why we desperately need a solution to this puzzle.Footnote 3 Granted the importance of the puzzle, let me now explain why it has gone unnoticed so far.

It seems that emotions directly guide our actions. Following the previous examples: we start studying physics to understand the universe, we run away from the bathroom, etc. In all those cases our emotions are not isolated, they are accompanied by their usual cognitive bases (perceptions, imaginations, beliefs, etc.) but also by other mental states like evaluative judgements (for instance, the judgement that ordered complexity is beautiful), memories about past events or recollection of information (for instance, about spiders’ venom), etc. The evaluative molecular net of mental states, of which the emotions is a part, contributes as a whole to the guidance of our behaviour. However, given its affective nature, emotion captures almost all the attention of their subjects and it seems to them that they are acting solely on the basis of their emotional experience. It is similar to how a painful experience in your foot captures your attention partially occluding the feeling of the feet against the shoe. Of course, this interpretation depends on having a rich enough conception of phenomenal consciousness (Schwitzgebel, 2007). But the richness required is not so rich. In the end we are talking about contiguous experiences in an evaluative process (i.e., emotions, judgements, recognition of reasons for emotions and judgements, etc.) and not experiences in different modalities all the time (i.e., visual experiences, cognitive experiences, visceral experiences, tactile experiences, imaginative experiences, etc.). It is not a coincidence that in the analogy we used, the pressure of the shoe and the pain occur in the same foot.

My diagnosis, then, is that the scandalous phenomenal character of emotional experiences tends to capture almost all the attention of their subjects, partially occluding other phenomenal phenomena in the vicinity, like evaluative judgements and the conscious recognition of reasons for them. This has led to the overestimation of the epistemic and guiding roles of emotion in giving access to the evaluative aspect of the world, in justifying our evaluative judgements and in guiding our behavior. This partially explains as well, the attractiveness of perceptual theories of emotion, which will be examined in the light of our new puzzle in the following section.

3 The puzzle and the nature and epistemology of emotional experience

When people have talked about the epistemic role of emotional experience assimilating it to perceptual experience’s epistemic role, two accounts have been considered: the reliabilist and the phenomenalist as pointed out by Cowan (2016) and Kurtz (2022). In a nutshell: reliabilists claim that emotions justify evaluative judgements and actions in virtue of external reliable processes that correlate emotional experience and the presence of evaluative properties. The locus of justification is not in the emotional experience but in its reliable correlation with evaluative properties, orchestrated, for instance, by natural selection. By contrast, phenomenalists claim that the locus of justification is in the emotional experience itself. Our problem is more pressing for the phenomenalists than for the reliabilist. The former considers that an emotion justifies an evaluative judgement if and only if it has the right kind of phenomenology, whereas the latter thinks that an emotion justifies an evaluative judgement if and only if the relevant emotion reliably tracks evaluative properties, independently of its phenomenal character. Our problem is a problem with first-personal character and hence eminently phenomenological. Why? Because if the reliabilist is right, then, the emotional subject need not be aware of the reliable process, it would be sufficient for them to be a part of that process, that is, to be the subject whose emotional processes correlate with evaluative properties. Thus, they would be externally justified in following the dogmatic pull to action of their emotion, irrespectively of whether their emotional experience lacks a justifying role akin to perceptual experience’s one. By contrast, our problem is that our emotional subject is phenomenally aware that their emotional reactions alone cannot be sufficient for justifying her judgements and actions but, in a way, she feels as if her emotional experiences were able to do so due to the characteristic action readiness that accompanies those emotional experiences. Thus, from now on, when I speak of epistemic perceptual theories of emotion, I will be referring to phenomenalist perceptual theories of emotion.

In the subsections to come I will show that the perceptual phenomenalist about emotion’s epistemic role cannot give an account of emotional phenomenology that makes it capable of overcoming the difficulty outlined in Section 1. Even though they fail, they have identified an important aspect of emotional phenomenology: it plays an epistemic role in our mental economy when it comes to evaluation. Emotions are essential for evaluation. However, the role of emotion regarding evaluation is not properly captured by the analogy with perception and perceptual existential belief. In Section 3 I will try to offer some ideas to capture the essential role of emotion for evaluative thought and knowledge without relying on the perceptual analogy.

The argument in this section is going to follow this path. First, I am going to point out that the feature of perceptual experience that enables it to have the epistemic justificatory role it has is lacking in emotional experience. As we will see, emotional experience lacks what Chudnoff (2013) has called presentational phenomenology. We will examine how emotional experience’s absence of presentational phenomenology can combine with Brady’s arguments, examined in the previous section, ruling out the perceptual theory of emotion as a plausible account of the epistemic role of emotional experience. After that, we will examine a recent defense of the perceptual theory against Brady’s objections by Mitchell (2017). I will show that Mitchell’s defense is not successful. The upshot of this section will be the following: emotional experience’ epistemic role cannot be modeled in perceptual experience’s one, as Brady’s arguments plus considerations on emotional experience’s lack of presentational phenomenology show. However, once we embrace this result, we need to provide an account of how emotions can have a rational guiding role while lacking a justifying role analogous to perception’s one. We need a solution to our problem, which is also a problem for Brady and accounts like his account. Besides, the inability of the perceptual theory of emotion for formulating this problem, which would be impossible given its assumptions, also tells against it. Finally, I will provide a sketch of an initial solution in the next section, Section 3.

3.1 Emotional experience lacks justificatory phenomenology

Chudnoff (2013) identifies the justificatory phenomenology the phenomenalist is looking for with presentational phenomenology, which is the phenomenology characteristic of perception and intuition. Those mental attitudes make a subject aware of a content, say p, and make it seem to the subject as if they make them aware of what makes p true. Here is Chudnoff (2013, p. 18):

An experience has presentational phenomenology with respect to the proposition that p when it not only makes it seem to its subject that p, but also makes it seem to its subject as if it makes him or her aware of the very chunk of reality that makes p true.

Thus, when you have a perceptual experience of a red apple, the experience makes it seem to you that there is a red apple in your surroundings and that this precise experience is what makes you aware of the truth maker of the proposition that there is a red apple in your surroundings. The same goes for the phenomenology of intuition, whose contents are abstract rather than concrete. When you have the arithmetic intuitive experience of the proposition that 2 + 2 = 4, the experience makes it seem to you that 2 + 2 = 4 and this precise experience is what makes you aware of the truth-maker of the proposition that 2 + 2 = 4.

The experiences that have presentational phenomenology in the above-described way, rationally allow their subjects to be dogmatists about the information conveyed by them. If I have a perceptual experience, I am, absent defeaters, rationally justified in taking at face value the information conveyed by it. Mutatis mutandis for intuitive experiences. As we saw in the previous section, emotional phenomenology alone does not allow their subjects to take at face value the information conveyed by it. Now we know why: emotional phenomenology lacks presentational phenomenology. When you are angry about your in-law’s remark at the family dinner, it may be that your emotional experience makes you, together with auditory experience, (more) aware of the remark. However, your emotional experience does not make it seem to you as if it makes you aware of the chunk of reality that makes the experience veridical (Brogaard & Chudnoff, 2017). The emotion just presents the incident under a certain evaluative light but it does not seem to present the “chunk of reality” that makes that impression veridical. That is, unlike perceptual or intuitive phenomenology, emotion requires further evidence. It needs justification.

If we put together the previous considerations about emotional experience’s lack of presentational phenomenology with Brady’s considerations against modeling the epistemic role of emotion on perception, then we have a strong case against the emotional dogmatist. However, and here our puzzle reappears, we still need an account of the epistemic role of emotion in the plausible rational guidance of our actions. At least if we want to preserve the intuition that emotions are rational, something that I take as a desideratum for a plausible theory of emotion.

3.2 The wreck of perceptual/intuitionist theories of emotional experience and their theoretical treasure

If the above reasoning is on the right track, then theories of emotion that consider that they are epistemically akin to perception are doomed to fail. In this subsection I will consider reasons to reject perceptual theories of emotion. I will conclude that even if perceptual theories cannot succeed in explaining the epistemic role of emotion, they have left us an important theoretical treasure: the intuition that emotions are different from beliefs and desires and rationally guide action. I will devote the final part of the paper to try to explain and secure that intuition in non-perceptual terms. In the next subsection I will examine a recent defense of perceptual theories by Mitchell (2017).

Let’s start with the core theses of perceptual theories of emotion. Consider the following quotes by leading perceptual theorists of emotion. You can see that they conceive of emotions as a dogmatic experience along the lines of perception.

Johnston (2001, p. 189):

I maintain the following: it is because affect can be the disclosure of the appeal of other things and other people that it can have authority in the matter of what we should desire and do. By ‘the authority of affect’ I mean not to refer to its sheer effectiveness as a source of desire or action, but rather to the fact that the presence of the affect can make the desire or action especially intelligible to the agent himself. It can make the desire or act seem apt or fitting in a way that silences any demand for justification.

Döring (2007, p. 386):

In judging that she ought to take action against the punishment the person has not only a reason but is also motivated to do so. This is so because the chain of reasoning which leads to that judgement starts from an emotion. As we have seen, emotions are capable of both rationalizing and motivating, although their representational content is not that of belief, nor is their motivational force that of desire. The person’s judgement that she ought to take action against the punishment has motivational force due to the justifying relation holding between her judgement and her emotion. This relation forms a link in the form of necessary connection, and, because of that link, the emotion’s motivational force is transmitted to the judgement.Footnote 4

Tappolet (2016, p. 170):

Quite generally, the justification of an evaluative belief would turn on the absence of any reason to believe that the emotion on which the belief is based is inappropriate. Consider the belief that a friend of yours is admirable. What I suggest is that your belief is justified on condition it is based on your admiration and you have no reason to distrust your admiration.

Now, if everything said about presentational phenomenology is true, that is, if it is true that emotion lacks the mark of justificatory phenomenology, then it seems that the above approaches to capture the nature of emotion are just false. Indeed, those approaches exploit the hot/categorical face of emotion while ignoring the hypothetical one. Those authors are not aware, for the reasons given above, of our problem and, if they were, they would be unable to solve it. Their theory of emotion is unilaterally based on the more phenomenologically salient face of emotion, the pain in the foot, ignoring the no less real hypothetical face, the sensation of the shoe against the foot.

What is more, the point of departure of these theories makes them incapable of even acknowledging our problem and hence precludes any hope for solving it within them. If you are convinced that our problem is real and that perceptualist theories of emotion cannot solve it, nor even formulate it, then you should conclude with me that they must be abandoned. Thus, if the identification of the problem is correct, then perceptual theories of emotion should be discarded, because there is nothing in the phenomenology of emotion that justifies being dogmatist about it, unlike the phenomenology of perception, and, hence, treating emotions like perceptions is just a bad theoretical move. However, there is a kernel of truth, a theoretical treasure, in perceptual theories of emotion: the idea that emotions are not as cognitively sophisticated as judgements and are something hotter, shared by beasts and babies, that orient our navigation through a world full of value and disvalue playing a significant rational motivational role. When I try to sketch a solution to the problem in the final section, I will emphasize how important it is that we reject dogmatism about emotional phenomenology being able to keep their motivational power and claiming that such power is, indeed, rational and not blind. That would start (dis)solving satisfactorily our new problem.

3.3 Mitchell’s attempt to save the Perceptual View

Before abandoning the perceptual theories and stealing their treasure, it is advisable to consider some reasons offered by Mitchell (2017) to counteract skepticism about them. One of the criticisms made by Brady and reinforced by emotions’ lack of presentational phenomenology is that emotions are not justificatory bedrock but instead are in need of justification themselves. That is, it is in general appropriate to ask normative why-questions regarding emotional experiences, but it is not so in the case of perception. Mitchell claims that there are some cases in which normative why-questions for emotions are not appropriate, because the emotion is, in those cases, as epistemically fundamental as a perception. He offers the case of someone in a funeral who is sad about the loss of a loved one. Mitchell claims that it would be inappropriate, and not for conventional or social reasons but for epistemic ones, to ask that person why she is sad.

Mitchell’s clever example is interesting. However, I do not think it is able to make a strong case for phenomenal dogmatism about emotional experience. First, Mitchell’s example is a special case, not the norm. So, generalizing from it would be theoretically suspicious. Second, the example can be analyzed differently and without the need to suppose that emotional experience is epistemically like perceptual experience in that case. Let’s agree with Mitchell that the reasons for the inappropriateness of the normative why-question is intrinsically epistemic. However, the relevant epistemic aspect is not emotion’s epistemic role in analogy with perception’s epistemic role, but the relation between justified emotion and justified evaluative judgement. Let me elaborate. It is not appropriate to ask why the subject is sad because she knows that we know that she knows, given our shared knowledge of the human condition and our shared knowledge of our knowledge of the human condition, that she has suffered a personal loss. Given that, as Brady notes (2013), justified emotions and corresponding justified evaluative judgements have the same set of reasons, if one knows that the judgement is justified, one knows that the corresponding emotion is justified and, hence, it does not make sense to ask for its justification. Asking why she is sad would be like asking why she believes that normal people run away from fire. In that case we know that we know that human flesh can be damaged by fire and that that is painful. So, asking why would be either theoretically insulant or a reason to suppose that we belong to another planet. This kind of psychologically iterative explanation, in which for instance we know that we know that we know what the human condition is, is not strange in human interaction. Indeed, the Gricean tradition has made its case for its definition of meaning in terms of them (Grice, 1957), and for some Gricean philosophers, mutual knowledge is constitutive of human communicative exchanges (Loar, 1981; Schiffer, 1972).

Another argument deployed by Mitchell in the defense of emotional perceptualism consists in arguing that, contra Brady, emotions can include a sense of epistemic appropriateness from the inside. That is, there is a way in which, at a minimum, emotions can be and be felt as appropriate without entailing, on the assumption that Neo-sentimentalism is true, problematic self-justification. Here it is Mitchell:

It is plausible that for an evaluative property to be experienced as qualifying the particular object of an emotion, the emotion must necessarily (re)present its object’s evaluative standing as independent of that particular experience of it. In this way a certain kind of content externalism implies phenomenal objectivity with regard to that evaluative standing, and this is necessary if at least some emotional experiences have value properties as part of their intentional content.

(Mitchell, 2017, p. 77).

Though Mitchell is right about the idea of a minimum epistemic appropriateness and a minimum feeling of epistemic appropriateness, this sense of appropriateness is not sufficient to claim that, epistemically speaking, emotions are to evaluative beliefs what perceptions are to existential empirical beliefs. It is true that emotions are outward directed and that they sometimes inform us about axiological aspects of the world, but they are far from having the objective character perception has, they lack robust presentational phenomenology. Emotions are more akin to hypothetical thought and, as such, they are about the external world and, sometimes, they get it right. However, this is not strong enough to posit an emotional epistemic role analogous to perception. In the following line I elaborate on these considerations further.

The idea that emotions involve a strong sense of objectivity akin to the sense of objectivity characteristic of perception is really controversial. As we saw, perceptual experience exhibits presentational phenomenology (Chudnoff, 2013), whereas emotional experience lacks it (Brogaard & Chudnoff, 2017). The idea of presentational phenomenology seems to imply some version of the transparency thesis in the sense that the experience which has presentational phenomenology seems to present directly the chunk of reality that makes that precise experience true or veridical. That is, the experience does not present itself presenting reality but seem to present (a chunk of) reality directly, transparently. Thus, perceptual experience exhibits at least some sort of transparency, as has already been claimed by some authors (Harman, 1990; Salmela, 2011).

Mitchell acknowledges that emotions are not transparent in the perceptual sense but still insists that this does not prevent that emotion plays an epistemic role similar to that of perceptual experiences. It seems to me that this is not plausible, since it trivializes the distinctive epistemic role of perception and makes the notion of robust presentational phenomenology and the transparency entailed by it epistemically irrelevant. Let’s consider the transparency thesis for perception. It seems that when you describe your perceptual experience you just refer to objects and properties which happen to be the correctness conditions of your perceptual experience (Searle, 2015). By contrast, it seems that you cannot describe your emotional experience without mentioning the experience itself. This is a point acknowledged by Mitchell himself: he says that emotions have reflexive (in fear, something in presented as dangerous for me) and valence opacity (fear feels bad to me). Note the tension between this acknowledgement of emotional opacity and the above quoted passage. The transparency of perceptual experience, as opposed to the opacity of emotional experience, makes, for sure, an important epistemic difference between these two mental states. The former, but not the latter, seems to put you in immediate contact with the chunk of reality that would make it true. The former, but not the latter, has a robust presentational phenomenology.

If the reasoning above is right, we can learn what perceptual theories can teach us, but we should not follow them until their ultimate consequences.

3.4 Our new problem and theories of emotion apart from the perceptual ones

The existence of the new problem together with the considerations against emotional dogmatism are fatal for perceptualist theories of emotion or for theories of emotion that epistemically equate them with intuition. For the remaining theories of emotion, like attitudinal theories (Deonna & Teroni, 2012) or motivational theories (Scarantino, 2014), it is just a challenge, though a formidable one. Concretely, it is a challenge that every theory of emotional phenomenology should address. If emotional phenomenology is not presentational in the sense in which perceptual or cognitive phenomenology are, then how can we characterize emotional phenomenology while preserving the intuition that it plays a rational motivating role. Indeed, it is a problem for Brady himself: once we reject that emotional experience is epistemically akin to perceptual experience, we cannot appeal to an emotional justifying role, as it is conceived by perceptualists, to explain emotions’ guiding role. In the next and final section I offer the first brushes of a solution in which emotions have a fundamental epistemic role which is not akin to perception.

4 Sketching a solution to the problem

4.1 The sketch

It seems that the best way of characterizing emotional phenomenology is as a sui generis attitude (Álvarez-González, 2021; Dewalque, 2017; Mitchell, 2020). Uriah Kriegel has attempted to provide an account of the primitive phenomenal attitudes indispensable for explaining the stream of consciousness. Kriegel’s basic elements of phenomenology are attitudinal phenomenal types of mental states. These phenomenal attitudes contrast with their content, namely with the object, proposition or state of affairs presented by them. Thus, those types of phenomenology refer to the what-it-is-like of the mode in which the mental state presents its content rather than to the what-it-is-like of the content itself. Take, as an example, cognitive phenomenology. According to Kriegel, cognitive phenomenology presents-as-true its contents, for instance, that 2 + 2 = 4. Thus, it is the phenomenology of the mode of presentation or mental attitude (i.e. presenting-as-true), instead of the phenomenology of the content (i.e. that 2 + 2 = 4), that is relevant here. Given that emotional phenomenology is not transparent, it is reasonable to suppose that its intentional contribution is not only in terms of its contents but also in terms of its distinctive phenomenal attitude (Kriegel, 2019). Given this, I am going to assume Kriegel’s framework in order to develop a characterization of emotional phenomenology able to start (dis)solving our problem.

I present my proposal and then I indicate how it bears on the puzzle under discussion. My proposal: emotional phenomenology presents-as-worth-attending its contents. More specifically, emotional phenomenology presents-as-worth-attending-relative-to-a-specific-meaning its contents. The relativization to a specific meaning is because emotional experience is determined by primitive ways of presentation of their contents relative to specific given meanings or formal objects (Deonna & Teroni, 2012; Kenny, 1963; Lazarus, 1999). Thus, for instance, fear presents-as-worth-attending-relative-to-danger its contents and anger presents-as-worth-attending-relative-to-offense its contents. The relativity to a specific meaning also explains why emotions are or are not justified. If emotions are needed for a dedicated representation of what is worth attending relative to a specific meaning, then we can stablish some primitive connections of appropriateness. Anger is appropriate iff its contents are worth-attending-relative-to-offense; fear is appropriate iff its contents are worth-attending-relative-to-danger…Note that some objects may be harmless but worth attending relative to danger. I can rationally fear a dog I do not know yet even if, once I come to know him, I realize that he is harmless. In that case, my fear of the dog at that time was fitting and justified Footnote 5because, for all I knew, the dog could have been dangerous and, hence, it was-worth-attending-relative-to-danger. The same goes, mutatis mutandis, for the other emotions.

Emotional phenomenology presents-as-worth-attending-relative-to-a-specific-meaning its contents. This is not the same as judging or perceiving that its contents instantiate the meaning under which they are being considered. Rather, emotional phenomenology presents a particular object under some evaluative light based on an universal concern of the emotional subject specifiable in terms of the formal object of the emotion. Fear presents as-worthy-of-attention-relative-to-danger its contents, which is not the same as-presenting their contents as dangerous. Something similar is said by Roberts (2003). According to him emotions are concern-based construals. That is, emotions would be “ways of seeing” their particular object analogously to how one can appreciate the rabbit-duck image as either a rabbit or a duck. Thus, when one has an emotion, one contrues the object of the emotion under the particular guise of the concern characteristic of that emotion: in fear, one construes the particular object of the emotion as dangerous; in anger, as offensive; in joy, as-goal-congruent, etc. Roberts’ account is criticized because it cannot account for the emotionality of emotions (Deonna & Teroni, 2012; Morag, 2016; Müller, 2019).

That criticism can be overcome by means of reconstructing Robert’s proposal in more sophisticated terms. Thus, Roberts’ approach can be complemented with the idea that the construal is emotional because of the relevance the formal object of the emotion has for the emoter. This relevance makes intelligible the fact that the emoter feels an urgent categorical pull to inquire (which is a form of acting) whether it is the case that the particular object of the emotion instantiates the formal one.

We now have the ingredients for the solution. Emotional phenomenology presents-as-worth-attending-relative-to-a-specific meaning its contents. This is not the same as judging or perceiving that its objects instantiate the meaning under which they are being considered. The emotional experience is not epistemically dogmatic. However, one feels this categorical pull to action because of the meaningfulness for one’s life the formal object of the emotion has. That’s what explains the urgent, hot, character of emotion and is also what explains why the emotional experience is essentially linked to action, why it guides action. However, the emotion does not guide action through a justificatory role analogous to perception’s one. Instead, emotions guide attention towards their intentional object through which the corresponding action is guided. Thus, emotions guide action mediately through the guidance of attention. Brady (2013) emphasized the important link of emotion, its epistemic role and attention. For him, emotions promote evaluative understanding, instead of making us aware of evaluative properties, by means of capturing and consuming attention.

Emotions fix and keep the attention fixed on what is important for the subject, guiding the processing of information, which includes perceptions, beliefs, motor representations, etc. Emotional appraisal is not dogmatic, it is hypothetical. However, given the importance of the formal object of emotion, the need to fix and keep the attention fixed until one reaches a resolution, is dogmatic. Emotional phenomenology is then epistemically non-dogmatic but practically dogmatic. Indeed, Wendy Wilutzky (2015) has conceived of emotions as epistemic actions that help the individual to uncover axiological information in a process of continuous appraisal and re-appraisal whose end is precisely to change the individual epistemic status in regards to that information.Footnote 6

It seems to us that emotional phenomenology is epistemically dogmatic, that we perceive the instantiation of a value in some natural properties and that that explains the characteristic action readiness we feel when we are emotional. However, we are misattributing the dogmatic character: it is not the emotional appraisal that is dogmatic, but rather the need to confirm it and to keep the attention fixed until this happens, and keep the attention fixed subsequentially if needed (e.g., in order to fight with the predator, in order to hide from it, etc.). Of course, the scandalous phenomenal character of the need to keep such important values tracked, values like dangerousness (fear), toxicity (disgust), goal congruence (joy), injustice (anger), etc., occlude the epistemically hypothetical character of our initial emotional appraisals and, often, lead us to act before confirmation.

This is understandable since false positives are less costly than false negatives. However, this strategy of optimization does not entail that our emotions are dogmatic. It just entails that we have prudential reasons to take them as such, especially when we lack the time to do better. We are just justifiably betting. Indeed, reappraising initial emotional appraisals seems to be constitutive of emotions as Lazarus pointed out (1994, 1999); not only of human emotions but also of animal ones. It seems indispensable for coming up with the best strategy of coping with the situation targeted by the emotion. Here is Lazarus:

Although the initial appraisal may be hasty and limited, if the opportunity for further investigation of what is happening presents itself, it would be a strange creature that let things drop before a full functional understanding has been achieved.

(Lazarus, 1994, p. 215)

The constitutive role of secondary appraisal should convince us that emotions are not essentially dogmatic, though the values represented by them recommend that sometimes we act dogmatically.

4.2 Some objections to the sketch

The first objection comes from considering the so-called factive emotions, emotions like, regret-that, anger-that or delight-that (Dietz, 2007; Gordon, 1987; Unger, 1975; Williamson, 2000). Unlike fear and hope, these kinds of emotions always admit a factive reading. Take regret-that. Imagine that you regret that you made a silly comment in the seminar. That you know that you made a silly comment in the seminar is a necessary condition for regretting it. That is, if you regret that you made a silly comment in the seminar, then you know that you made a silly comment in the seminar, and if you know that you made a silly comment in the seminar, then you made a silly comment in the seminar. In case you did not make a silly comment in the seminar, then you cannot regret it, though you may think that you regret it or you feel as if you regret it. It seems, then, that when we attribute the state of regretting-that to us or someone else, we are not in need of reasons for considering that the object of our regret is the case (either we know it is the case or we think we know it is the case). Someone could argue, then, that we do not need to confirm any emotional hypothesis in those cases. However, from the factivity of these emotions it does not follow that our general claim about the phenomenology of emotions is mistaken. What emotions like regret-that p entail is that we know that p and therefore that p is the case, not that p is worthy of regret. That is, we can regret that p, and, then, p is true, but that does not entail that we should regret p. That we regret p, entails that we know that p, but it does not entail that we know that p is shameful, worthy of regret, etc. If I regret p, it is clear, I know that p and, hence, p is the case. However, whether p instantiates the axiological properties that would made it worthy of regret is another question.

The second objection comes from emotions that seem to make us axiologically infallible, like disgust. It seems that if I feel disgust toward object o, then o is disgusting, and, hence, the question of whether o is disgusting or not is already settled. It seems to me that this objection relies on a radical subjectivist view on emotions like disgust. That I feel disgust towards some object is not sufficient for conferring to it the property of being disgusting. My episodes of disgust can be mistaken. There is some objective property, perhaps grounded on human psychology or evolutionary history, that make certain things disgusting but not others. That is, what is disgusting would be a mind-dependent property, dependent on the species, but not single-mind-dependent, that is, not dependent on the mind of a particular individual. Hence, if I feel disgust, the question of whether what disgusts me is indeed disgusting is still open, until I am able to provide sufficient reasons for ratifying or rectifying my initial emotional appraisal.

The third objection targets our characterization of emotion’s phenomenal attitudinal profile. Above we said that emotions present-as-worth-attending-relatively-to-a-specific-meaning its contents. Someone could say that there is something phenomenologically inaccurate with this characterization: boring objects and disgusting objects are not worthy of attention. When we are in the face of a boring or disgusting object what we want, precisely, is to stop paying attention to it. This objection is partially correct, we do not want to pay attention to boring or disgusting objects because they are aversive. However, this same aversiveness is what demands our attention. That we do not want to pay attention to something does not mean that we do not pay attention to something. We need to pay attention to boring or disgusting objects to get rid of them, in order to avoid them. The objector may reply, then, that they are not intrinsically worthy of attention (only instrumentally) and, hence, that our characterization errs: we only pay attention to them to stop paying attention to them. However, this is not a good interpretation of our proposal. Disgusting and boring objects/events are intrinsically worthy of attention, though negatively valenced, because they reveal our axiological predicament and that is something we cannot be more interested in.

The fourth, last and more important objection is inspired by Uriah Kriegel (2015). According to him, there is “(i) a constitutive connection between attention and importance, and (ii) a contingent connection between emotion and attention” (p. 155). Thus, our characterization of emotional phenomenology’s attitudinal profile as that phenomenology which presents-as-worth-attending-relatively-to-a-specific-meaning its contents would just refer to a contingent and not a constitutive connection between emotions and the values that they seem to track. In that case, our characterization is best understood as the characterization of attention’s phenomenal profile. We would just need to substitute “presenting as-worth-attending” for “presenting-as-important”. Here I just can sketch a way out of this difficulty which I think is correct. In the consideration quoted above, it seems that Kriegel reverses the order of explanation and misconceives of attention. I can focus on unimportant things. That is, I can have attention without considerations of importance. You, too: if I ask you to pay attention to the sensations in your nose while the air go through, that is something you can do, even if there is no interest in doing so at this particular moment. You have attention without appraisals of importance, that is, appraisals of importance are not a necessary condition for attention and, for that reason, importance cannot be constitutive of attention. Now, it seems unconceivable to have emotions without a sense of importance and, hence, a sense of importance seems to be constitutive of emotions. Our emotions determine what is worth-attending, because they have an essential connection with what is important. So, I suggest we modify Kriegel’s thesis: there is a constitutive connection between emotion and importance and, for that reason, a constitutive connection between emotion and what is worthy of attention.

5 Conclusion

If the previous reasonings are correct, then I take myself as having given a plausible characterization of emotional phenomenology that should be considered when one theorizes about the emotions and their epistemic role. Otherwise, one should provide an alternative account to face the problem presented here or just endorse the undesirable conclusion that our emotions are irrational forces.