Why are emotions epistemically indispensable?

ABSTRACT

Contemporary philosophers are attracted by the Indispensability Claim, according to which emotions are indispensable in acquiring knowledge of some important values. The truth of this claim is often thought to depend on that of Emotional Dogmatism, the view that emotions justify evaluative judgements because they (seem to) make us aware of the relevant values. The aim of this paper is to show that the Indispensability Claim does not stand or fall with Emotional Dogmatism and that there is actually an attractive alternative way of developing it. We first argue that the Indispensability Claim requires a deflationary account of the cognitive bases of emotions, which we call the Simple View and defend against a variety of criticisms. Armed with the Simple View, we show that the Indispensability Claim is made true by three roles of the emotions: they help us access the relevant evidence, provide a needed hook-up between evidence and judgement, and contribute to evaluative understanding.

KEYWORDS:

• Emotion
• epistemology
• value

It is often observed that emotions have privileged connections to specific values: amusement to the funny, anger to the offensive, sadness to loss, indignation to injustice or viciousness, admiration to excellence, pride to success, contempt to the debased, shame to the shameful, etc.

On the strength of this observation, philosophers of mind have typically claimed that emotions appraise the environment in evaluative terms. In indignation, the environment is appraised in terms of viciousness, in amusement, in terms of funniness (Kenny 1963; Lyons 1980). What unifies all the specific approaches that endorse this claim is the idea that emotions have evaluative correctness conditions: they are correct if and only if the way they appraise the environment corresponds to the facts, incorrect otherwise. Torture is a correct object of indignation, for example.

For their part, epistemologists have insisted on the epistemic contribution of emotions to evaluative knowledge. The fundamental intuition here is, for example, that what we know about losing what is dear to us somehow derives from our capacity to feel sadness, that what we know about the funny somehow derives from our capacity to feel amused and that what we know about excellence somehow comes from our capacity to feel pride and admiration. In the literature, this intuition takes the form of what we will call the Indispensability Claim: emotions are indispensable in acquiring knowledge of the relevant values (D’Arms and Jacobson 2010; De Sousa 1987; Goldie 2004; Johnston 2001; Tappolet 2016). Assessing the Indispensability Claim is obviously important to understand the role of emotions in generating knowledge about values that are central to our lives. It is potentially even more significant if we think that these ‘emotional values’ constitute foundational building blocks for our more general evaluative practices and, ultimately, for ethics.

The Indispensability Claim concerns the relations between emotions and the acquisition of knowledge about specific values. In line with a venerable tradition, the knowledge at issue is conceived roughly to consist in a true and justified evaluative judgement. The epistemologists’ focus of interest has been on justification, as it is commonly thought to be the prime epistemological ingredient in knowledge (Ichikawa and Steup 2017). The simplest way of conveying what this notion of justification amounts to is in terms of evidence in favour of a judgement. More precisely, justification requires that the judgement is or would be rationally permissible in light of this evidence (Pollock and Cruz 1999).

While justification is distinct from truth or correctness (e.g. Echeverri 2019), these two properties do not of course float free of one another. One recurrent thought is for instance that it is insofar as emotions have evaluative correctness conditions that they are apt to justify the relevant evaluative judgements. It is because pride and admiration are correct responses to the display of excellence, say, that they are apt to justify judgements of excellence. In the contemporary literature, the Indispensability Claim has led many philosophers to take inspiration from the epistemology of perception and to adopt a form of so-called dogmatism. According to perceptual dogmatism, perceptual experiences justify perceptual judgements because they give us access to the relevant features of the environment (Huemer 2013; Pryor 2000). Emotional Dogmatism claims that the same is true of the relations between emotions and evaluative judgements. For the dogmatist, emotions justify evaluative judgements because they make us aware, or seem to make us aware of the evaluative features of the environment that make the relevant evaluative judgements true. The emotions would provide foundations for evaluative knowledge in a way that parallels the foundational role of perceptual experiences vis-à-vis perceptual knowledge.

The aim of this paper is to show that the Indispensability Claim does not stand or fall with Emotional Dogmatism and that there is actually an attractive way of developing it that is not wedded to Dogmatism. The discussion is structured as follows. In Section 1, we briefly review familiar reasons for thinking that Emotional Dogmatism cannot be made to stand in light of the dependence of emotions on their cognitive bases. The fact that emotions, as opposed to perceptual experiences, are states that can be justified or unjustified requires that we assess the prospects of the Indispensability Claim in light of two issues: the upstream issue – which concerns the relation of emotions to their cognitive bases – and the downstream issue – which concerns the relation of emotions to evaluative judgements.

Section 2 is devoted to the upstream issue, i.e. whether the fact that emotions depend on cognitive bases for their justification threatens the Indispensability Claim. The threat goes as follows. Suppose that hearing about a crime justifies your indignation only if, prior to the indignation, you judge that the crime reveals viciousness (cognitive base). This would threaten the Indispensability Claim, as it would lead us to conclude that indignation is dispensable in the acquisition of knowledge about viciousness. In short, one would need to assess whether the situation displays viciousness prior to the occurrence of indignation. We argue that the Indispensability Claim is not immediately threatened if we adopt another, deflationary account of the cognitive bases of emotions, which we call the Simple View. The bulk of the section is devoted to a defence of the Simple View against a variety of criticisms.

Suppose that this defence of the Simple View is successful. This does nothing to show that (justified) emotions make an indispensable contribution to evaluative knowledge. Section 3 is concerned with assessing whether the fact that emotions are themselves justified or unjustified, and because of that epistemologically redundant, undermines the Indispensability Claim. We argue that the answer depends on one’s position about the upstream issue, and that emotions may still shape evaluative knowledge in at least three ways: they help us access the relevant evidence, provide a needed hook-up between evidence and judgement, and contribute to evaluative understanding.

1. Emotional dogmatism

As we have seen, the Indispensability Claim has often led philosophers to defend Emotional Dogmatism. In this section, we lay out this form of dogmatism and briefly rehearse some of the main reasons for thinking that it falls short. This will put us in a position to explore the prospects of the Indispensability Claim independently of Emotional Dogmatism in the next sections.

In the epistemology of perception, dogmatism is a form of foundationalism according to which perceptual experiences provide evidence apt to justify the relevant perceptual judgements (Huemer 2013). They do so in virtue of the fact that they disclose (or seem to disclose) to the subject the way her surrounding environment is (Chudnoff 2013). The subject is (seemingly or really) aware of what makes the relevant judgement true, and this is why her judgement has a distinctive epistemological status that puts an end to the need for justification. It is for instance because my visual experience discloses the redness of a vase that it provides evidence that justifies the judgement that there is a red vase in front of me. The justification at stake is thought to be of a privileged sort in the sense that it is immediate. This is to say that the evidence justifies without the subject’s having to rely on any inference (Morretti and Piazza 2013; Millar 2019). Although immediate, the justification provided by perceptual experience can be defeated: it is prima facie (Pollock 1974; Pryor 2000). Your visual experience as of a red vase may justify your judgement that there is a red vase, but this justification is defeated once you realise that the environment is full of holograms. A judgement is prima facie justified when it is grounded in evidence that makes it rationally permissible in the absence of countervailing considerations.

Emotional Dogmatism is the application of these ideas to the relations between emotions and the relevant evaluative judgements. Emotions would justify evaluative judgements in virtue of the fact that they disclose (or seem to disclose) the evaluative features of the environment that make the relevant evaluative judgements true (Cowan 2018; Döring 2007; Johnston 2001; Pelser 2014; Tappolet 2016). Our judgement that the situation is threatening is justified because fear is an experience of the threatening aspect of the environment. Suppose that Sam is aware of an approaching dog, which has big teeth and is displaying erratic behaviour. He is afraid. According to Emotional Dogmatism, the fact that his fear presents the dog as a threat explains why Sam is in an ideal epistemic position to judge that the dog is threatening. Similarly to what is claimed by dogmatism about perceptual experience, Emotional Dogmatism has it that emotions provide immediate and prima facie evidence for the relevant evaluative judgements. Fear of the lion justifies your judgement that the lion is threating without inference. But the evidence fear provides may be defeated, e.g. if you learn that the lion is tame, old and well-fed.

Emotional Dogmatism has been under heavy fire in recent years (Brady 2013; Deonna and Teroni 2012a, 2012b; Harrison 2020). Two worries have occupied centre-stage. The first concerns the way Emotional Dogmatism constrains our understanding of emotional experience. In order to play the epistemological role of disclosing the relevant evaluative aspect of the surroundings, emotional experience must be understood as a presentation of these aspects in a way that is identical, or at least a close parallel, to the way perceptual experience presents its objects. Many are convinced that this is an unattractive approach to emotional experience: emotions feel like reactions to what is presented to us, and not as presentations (Brogaard and Chudnoff 2016; Deonna and Teroni 2020; Müller 2019). We shall not insist on this worry here.¹

The second worry, the one we wish to insist upon here, targets the foundationalist dimension of Emotional Dogmatism, i.e. the claim that reference to the relevant emotions ends the need for justification. Its starting point is the fact that the emotions themselves are assessed in light of evidence in their favour, just in the same way as judgements are (Brady 2013; Deonna and Teroni 2012a). According to Emotional Dogmatism, the occurrence of an emotion (anger at a remark, say) can justify an evaluative judgement (the judgement that the remark was offensive). But we are also prone to enquire into the justification for the relevant emotion (why he is angry). If we ask someone why he judged that a remark was offensive and he replies that the remark got him quite angry, we are likely to further enquire as to why the remark angered him (in which context was it made? What was its exact content? How was it conveyed? etc.). The answer we expect is supposed to bring forward considerations in favour of this emotion.² Suppose that someone asks you why you feel indignation. You are expected to mention aspects of the situation that make it an appropriate object of indignation. In light of these ‘why-questions’ about the emotions, it should come as no surprise that the justification of the relevant evaluative judgements presupposes the justification of the emotions that elicit them (Brady 2013; Deonna and Teroni 2012b).³ It seems to us that the existence of ‘why-questions’ constitutes a fundamental constraint on any account of the epistemological role of emotions: it manifests that fact that we do not pre-theoretically conceive of them as putting an end to the need for justification.⁴

What emerges from the foregoing observations is that emotions play a distinctive mediating role. On the one hand, they react to aspects of the environment the subject apprehends in one way or another. It is customary to describe the mental states to which emotions react as their cognitive bases. These come in many varieties – doxastic, mnesic and imaginative states, perceptual experiences, etc. – and their contents may have different formats. On the other hand, emotions typically elicit evaluative judgements. For example, one reads about a crime (this is the cognitive base), reacts with indignation (the emotion) and judges that the perpetrator is vicious (the evaluative judgement).

This mediating role of emotions raises two issues regarding justification that will interest us – they concern respectively what happens upstream and downstream of the emotions. The upstream issue concerns whether the existence of cognitive bases for the emotions undermines the Indispensability Claim. In Section 2, we argue that we can develop a deflationary account of cognitive bases apt to justify emotions, the Simple View, that is fully compatible with the Indispensability Claim.

The downstream issue concerns the epistemological relation between emotions conceived as reactions to cognitive bases and the relevant evaluative judgements. As we shall see, the truth of the Simple View may lead one to think that emotions cannot constitute a source of evidence apt to generate evaluative knowledge – the evidence in light of which emotions are themselves justified does also justify the relevant evaluative judgements. In a nutshell, emotions become epistemologically redundant and, therefore, dispensable. In Section 3, we argue that the final verdict is much more nuanced than this line of thought suggests and that the Indispensability Claim is true despite the epistemological redundancy of emotions.

2. Emotional justification: the upstream issue

a. The simple view

We have observed that emotions play a mediating role: they react to what the subject apprehends and typically elicit evaluative judgements. The upstream issue concerns whether the fact that emotions have cognitive bases threatens the idea that emotions play an indispensable role in our acquiring evaluative knowledge (the Indispensability Claim).

To recall, the threat is the following. If cognitive bases take the shape of an evaluative judgement about our surroundings, this threatens the idea that emotions play an indispensable role in acquiring evaluative knowledge. Suppose that hearing the joke justifies your amusement only if, prior to the amusement, you judge that the joke is funny (cognitive base). This would undermine the Indispensability Claim, as it would lead to the conclusion that amusement is dispensable in the acquisition of knowledge about the funny. One would need to assess whether the joke is funny prior to the occurrence of amusement. In short, the mediating role of emotions would rather show that evaluative knowledge is presupposed by justified emotions.

While the threat is particularly vivid when cognitive bases are evaluative judgements, it is also present if the relevant upstream evaluative uptake has a less demanding shape. In the literature, there has been talk of feelings of value (Massin 2021; Mulligan 2010), perceptions of value (McGrath 2018), evaluative construals (Müller 2019) or value intuitions playing the role of the emotions’ cognitive bases. Here is not the place to discuss these suggestions, but it is easy to see that they threaten the Indispensability Claim: if emotions are justified by feelings of value, for instance, what role is left for the emotions in the acquisition of evaluative knowledge?⁵

The challenge is then to show that there are evaluative uptakes apt to justify emotions without threatening the Indispensability Claim. On the one hand, the evaluative uptake must not be so substantive that the acquisition of evaluative knowledge is over and done with before emotions enter the stage. On the other hand, the uptake must somehow pick up on evaluative information. The evidence provided by cognitive bases must indeed be substantive enough to justify emotions, i.e. reactions that have, as we observed, evaluative correctness conditions. In what follows, we argue that what we call the Simple View successfully meets this challenge.

To introduce the Simple View, consider the following well-known example. Suppose that you witness teenagers setting a cat on fire, laughing in the process (Harman 1977). You are indignant. If your indignation is justified, what makes it the case? The Simple View champions a straightforward and attractive response: awareness of the teenagers setting a cat on fire for fun can be sufficient to justify indignation (Deonna and Teroni 2012a; Goldie 2004; Setiya 2012). To be more precise: awareness of this evidence constitutes immediate prima facie justification for this emotion. More generally, according to the Simple View, an emotion is justified if, in the situation in which the subject finds herself, the properties she is (or seems to be) aware of and to which her emotion is a reaction constitute (or would constitute) an instance of the evaluative property that features in the correctness conditions of the emotion she undergoes.⁶ Observe that the Simple View is deflationary in the sense that it does not require, as opposed to more substantive conceptions, that the evaluative uptake sufficient to justify emotions apprehends the situation in evaluative terms. Being aware of a cat set on fire for fun, where this presupposes neither ‘construing’ the situation as vicious, nor ‘feeling’ or ‘intuiting’ its viciousness, is sufficient to justify indignation (more on this below).

Before we examine the prospects of the Simple View as an account of the justification of emotions, let us pause to state clearly where we stand dialectically. If the Simple View is true, we seem to be in a position to reconcile the existence of cognitive bases with the Indispensability View. This raises two questions. First, can the Simple View be made to stand? The rest of this section answers this question by exploring a variety of worries the view is likely to raise. Second, what is the residual epistemological role of emotions once the Simple view is established and is it enough to support the Indispensability claim? Section 3 is devoted to this issue.

b. The looseness gap

Is registering an instance of the evaluative property that features in the correctness conditions of a given emotion sufficient to justify it? For instance, is registering the cat’s being deliberately set on fire sufficient to justify indignation? If indignation is a reaction to viciousness, then awareness of the cat’s being deliberately set on fire must qualify as registering viciousness. There are various reasons why one may think that it does not qualify, and therefore that there is a gap between the cognitive bases privileged by the Simple View and justified emotions.

A first worry runs like this. As a general rule, emotions have different cognitive bases – indignation may be elicited by seeing the burning cat, by learning about discrimination, by reading about political corruption, etc. There is nothing in common between these states of affairs, except that they may exemplify the same value, viciousness. Conversely, any situation can cause different emotions in different subjects (or in the same subject at different times). Upon witnessing the situation, you may feel despair at the human condition, or compassion for the cat. All this makes for a rather loose relation between cognitive bases and emotions (Echeverri 2019; Mulligan 1998; Scherer 2001). Shouldn’t we tighten this connection and insist that, in order for indignation to be justified, it must be preceded by a more substantial awareness of viciousness?

A first attempt at resisting a positive answer to this question is to point to the role of emotional sensitivities (D’Arms and Jacobson 2010; Deonna and Teroni 2009; Echeverri 2019; Jones 2006; Keller and Ombrato 2022). Emotional sensitivities are more or less ingrained aspects of our psychology that underscore our individual and sometimes idiosyncratic patterns of emotional responses. Indignation is the manifestation of one’s sense of fairness, say. Similarly, there is a sense of danger, a sense of loss, a sense of humour, a sense of honour, etc. Such sensitivities, which have been discussed since Aristotle, are essential determinants of emotional responses by ensuring that attention is drawn to the relevant aspects of situations, and we shall wonder as to their epistemological import throughout our discussion. For now, observe that there is a complex developmental story to be told about them (Kristjánsson 2015). You would for instance not react with indignation if you did not understand that cats suffer, that being burned is painful, etc. The learning process is more complex for indignation at discrimination or corruption, where the emotion has been progressively calibrated to more complex contents (Prinz 2004).

Can emotional sensitivities bridge the looseness gap? The relation between cognitive bases and emotions is certainly tighter once sensitivities are factored in, since they determine how our attention needs to be deployed in varied and complex circumstances. What is more, emotional sensitivities may bridge the gap in a way congenial to the Simple View. In the case of the burning cat, the view is that indignation is justified by the situation with the cat. The key idea behind the Simple View is that emotions, which you will recall have evaluative correctness conditions, can be justified by awareness of non-evaluative evidence (Deonna and Teroni 2012a, 2012b; Goldie 2004). This deflationary approach is vindicated in light of the intimate, constitutive relation between non-evaluative and evaluative properties (Dancy 2004). The viciousness of a situation may be constituted by deliberately burning a cat for fun, for instance. Again, understanding which non-evaluative features constitute an evaluative property may require a complex learning process. But this learning process and the resulting emotional sensitivity do not look like additional evidence for the emotion: they may simply equip the subject with a sensitivity to what justifies indignation. Similarly, we have to learn a lot to be able to apply observational concepts – for example, we have to learn that the concept ‘green’ refers to a given portion of the spectrum. Yet, once this is in place, a perceptual experience as of a green object is sufficient to justify the judgement that it is green (Lyons 2011).

Does this really address the looseness gap at its core? Perhaps the gap between cognitive bases and emotions remains even when emotional sensitivities are in place. One may insist that someone who reacts emotionally does not simply register a situation, she also interprets it and these interpretations or modes of presentation are amongst the main determinants of emotional responses (Lutz and Döring 2014). ‘A cat burning alive’ may elicit compassion, ‘intentionally burning a cat alive’ may elicit indignation and ‘intentionally burning a cat alive in order to appease the gods’ may elicit hope. In short, a situation elicits different emotions depending on the way it is interpreted (Lazarus 1991; Scherer 2001) – and whatever developmental explanation there is to the fact that a situation is interpreted in a given way, the actual way of interpreting it is a key determinant in justification.

This rings true. Emotions are not only elicited by awareness of specific aspects of a situation, the specific nature of this awareness also impacts the emotions’ justification. Our claim is that this is primarily a matter of how various factors modulate attention – through their impact on attention, goals, moods, recent past events and, obviously, emotional sensitivities all contribute to make some aspects of the environment more salient than others. All this is well documented and uncontroversial (Brady 2013; Ledoux 1996; Vuilleumier, Armony, and Dolan 2003). What is controversial is whether these modes of presentation have to be evaluative – in order for indignation to be justified, is attention to the features that constitute the relevant value enough, or should it be supplemented by awareness of these features in evaluative terms? Is awareness of a cat being set on fire for fun enough, or should the situation be apprehended in terms of viciousness? It is not clear that we have been given a reason to adopt the more substantive answer.⁷

Moreover, what gain is there in claiming that an emotion is justified only if it is preceded by a substantive awareness of the situation in evaluative terms? After all, if what drives the requirement that justified indignation be preceded by awareness of viciousness is the fact that indignation is correct if and only if the situation is vicious, why not require the same of the prior awareness of viciousness? A regress looms, which must at a certain point be stopped. A very plausible story in this regard has it that a psychological state with evaluative correctness conditions must be justified by non-evaluative evidence. Why not claim this of the emotions themselves?

Appeal to emotional sensitivities and attention has helped us further articulate the Simple View and repelled the most immediate threats to the Indispensability Claim.

c. The generality gap

The idea that there is a gap between the cognitive bases at the core of the Simple View and emotions may come from another direction, however. We started with the idea that emotions appraise the environment in evaluative terms. The worry is that the Simple View cannot accommodate this fact. Let us see why one may think so.

Obviously, a response with evaluative correctness conditions is justified only if it is properly modulated by the evaluative facts. The aversive reaction elicited by the burning of a cat, for example, would fail to qualify as a response to viciousness as such if it was, say, a primitive, hard-wired response to pain. In order to qualify as a response to viciousness (and not to pain), it must manifest a more general sensitivity to viciousness, i.e. it must track it in many circumstances that may show much variation in their non-evaluative properties. After all, viciousness is, like all evaluative properties, multiply realisable.

Again, it may appear that evaluative judgements must occur upstream of emotions. The thought may be that, in order for an emotion to be modulated by the evaluative facts, the subject must be aware of a general principle that governs the application of the relevant evaluative concept to the situation (Dokic and Lemaire 2013). Indignation is justified only if the subject is aware of a principle to the effect that, say, whenever there is gratuitous imposition of suffering, there is viciousness, a principle he applies to the situation with the burning cat.⁸ This of course goes against the Simple View, as the claim is that awareness of non-evaluative facts must be complemented by the application of general principles in order to generate justification.

The advocate of the Simple View may not fret: she will likely insist that laying emphasis on the awareness of general principles is adopting the wrong order of explanation (Dancy 2004). If a principle is to contribute to justification, it must be justifiably believed. So, what justifies one in endorsing the relevant principles? One appealing answer, which is at the core of the Simple View, refers to the concrete instances of viciousness: it is only if one can appreciate that burning (freezing, drowning, etc.) a cat (a dog, a mouse, etc.) for fun is vicious that one may end up being justified in endorsing the principle. This suggests that conceiving of justified indignation as dependent on the principle has things backwards.

It is important to appreciate how different the case at hand is to other cases in which appeal to general principles is very plausible. Visual experiences as of water may justify judging that there is H₂O only if the subject can rely on the principle ‘transparent and liquid stuff in this environment is H₂O.’ (Echeverri 2019) Here, the rationale for thinking that the principle contributes to justification is the existence of an inferential gap between the contents of visual perception and of belief, a gap which required centuries of experimentation to bridge.

The Simple View insists that the relation between awareness of the intentional burning of a cat and indignation is different. This claim is backed up by the so-called Fitting Attitude analysis of value, a long tradition according to which the concepts of the vicious, the sad, the comical, etc. are anchored in concepts of emotions (Brentano 1889; Ewing 1947). The concept of the vicious, for instance, is the concept of what merits indignation and it is anchored in primitive scenarios of the burning cat variety. Someone grasps what viciousness is in virtue of responding with indignation to such types of situations. We have observed that emotional sensitivities are key determinants of our emotions. We can now add that they are shaped through our responses to such paradigmatic scenarios (De Sousa 1987) and that we learn to extend our use of the relevant concepts by understanding how situations that are, as opposed to the molecular structure of water, directly available to us resemble these scenarios (D’Arms and Jacobson 2010). If something along these lines is correct, no gap needs to be bridged by an inference and there is here no threat to the Simple View (Deonna and Teroni 2021).

We have again appealed to emotional sensitivities to bridge a gap. This may be too optimistic, however, given that sensitivities are the object of endemic scepticism. Situationists have expressed doubts as to whether emotional responses can manifest general or wide-scope sensitivities (Doris 2002). So, it is important to insist that justified emotions need not rely on the wide-scope sensitivities that have taken the brunt of the situationist attack, i.e. sensitivities that track values across the whole range or a substantial subset of the relevant situations. It might be that such sensitivities are rare. But this fact has no clear impact on the sorts of emotional sensitivities that are needed to bridge the generality gap. True, we would not ascribe a sense of humour to someone who happens to be amused by one blonde joke and left cold by all the rest. It is also true that anyone’s sense of humour is likely to be calibrated to certain forms of the comical, perhaps only a few of them. But the fact that British understatement leaves one cold, say, does not speak against the justification of amusement at Gallic graphic spoonerism.

All in all, this suggests that there is no need to anchor the justification of emotions in wide-scope sensitivities. The advocate of the Simple View can rest content with narrow-scope sensitivities.

d. The robustness gap

The idea that justified emotions are modulated by the evaluative facts can be recruited to identify a final potential gap between cognitive bases and emotions within the Simple view, a gap which trades on the distinction between propositional and attitudinal justification (Turri 2010). Attitudinal justification requires that emotions be grounded in awareness of the relevant evidence. Recall the individual who has reasons to feel indignation upon learning that there has been a horrendous crime. What makes it the case that his indignation is a response to the crime and not, say, to the fact that the criminal is a foreigner? The gap here is that between awareness of evidence and having an emotion because of this evidence. Simple awareness of the features around which the Simple View is structured does not guarantee that this gap is bridged (Echeverri 2019), and it is bridged only by tightening the link between awareness of the relevant features and emotional responses.

The situation is typical of any purported causal relation. There is an antecedent condition with various features and its effect. What we want to know is which feature(s) of the antecedent condition cause(s) the effect. We know when we determine which counterfactuals of the form ‘the effect would have occurred if F was not there’, ‘it would not have occurred if G did not occur’, etc. hold. Similarly, feeling indignation because of the viciousness of a situation depends on an emotional sensitivity with the right counterfactual profile. For example, indignation should still have occurred if the cat was grey instead of black, if it was bigger, or if a mouse was set on fire. And it should not occur if the subject knows that it’s a toy cat, for instance. In a nutshell, emotions must be counterfactually robust.⁹

Can emotions be counterfactually robust in the required way? One first reason to doubt that this is the case emerges when we appreciate that emotional sensitivities are often unreliable, because they are easily disrupted by irrelevant factors (Monteleone 2017). Someone’s sense of humour may vary widely depending on the time of the day, the amount of alcohol, the level of fatigue, etc. Your capacity to be indignant might be hostage to similar factors. The result is that the profile of responses to actual and counterfactual situations goes against the idea that emotions are modulated in the required way by the evaluative facts.

There is an important empirical component to any informed reply to this worry, but we shall rest content with some remarks regarding its epistemological import. In order to assess the epistemological credentials of an emotion, the reliability of the relevant psychological processes must certainly be taken into account. The first thing to remember is that the justification under discussion is prima facie – and unreliability may sometimes simply defeat prima facie justification (Pelser 2014). Beyond that, the consequences of the fact that some factors are detrimental to the accuracy of our emotional responses are far from clear. Is it really the case that absence of amusement when fatigued or its prevalence when under the influence of alcohol prevents amusement when fresh and sober being justified? These are instances of general epistemological problems that have received extensive treatment (Conee 1998). What should be emphasised is that if we are alarmed by these cases to the point of concluding that emotions are never justified, this is likely to foster a comprehensive scepticism: evaluative judgements are after all as likely as emotions to be disrupted by irrelevant factors (Arpaly 2002). Any psychological state is the product of cognitive processes that can derail when background conditions are unfavourable.

The idea that emotions cannot be modulated by the evaluative facts may be developed in a different form, however. The focus is not anymore on the distribution of responses across situations, but rather on any concrete emotion. Being quick and dirty responses, emotions are alleged to have a kind of rigidity that prevents them from being attuned to the relevant features. It is not rare for anger or jealousy to persist once we have realised that the evidence that seemed to speak in their favour is defeated (Goldie 2008). More generally, many have insisted on the recalcitrance of emotions in the face of new evidence (Döring 2015; Helm 2015).

These are again reasons to think that emotions cannot be justified in the manner of the Simple View. Once more, the temptation is to appeal to evaluative judgements in order to control the deliverances of emotional sensitivities: evaluative judgements may have a flexibility that allows them to be modulated by the evaluative facts more finely than emotions. It is easy to come up with examples in which evaluative judgements disconfirm emotions, which are for that reason unjustified (Döring 2007). A recalcitrant bout of anger is overruled by the judgement that there was no offence, for instance. In many cases, this just shows that anger has prima facie justification or that it is unjustified. While these considerations have led some to insist on the necessity of putting emotional sensitivities under the monitoring of reasoning capacities (Jones 2006), we think we should regard this solution with suspicion. To start with, we should not overplay the rigidity of emotions – we have seen that they are constantly modulated by changes in how attention is distributed and how situations are apprehended. Moreover, many scholars have provided examples of so-called inverse akrasia in which it is the subject’s emotions rather than her judgements that are attuned to the evaluative facts (Arpaly 2007; Tappolet 2016). There is no clear reason to claim that rigidity is a characteristic of emotion, and flexibility one of judgement – this may have to be settled on a case-by-case basis.

We have distinguished two questions regarding the prospects of the Indispensability View. The first question is whether an attractive deflationary understanding of the emotions’ cognitive bases, one that does not immediately threaten the Indispensability View, is within reach. In this section, we have argued that the Simple View provides such an understanding. The second question is how exactly the Simple View constrains the role of emotions in the acquisition of evaluative knowledge. This is the question to which we now turn.

3. Justified evaluative judgements: the downstream question

a. The threat of redundancy

Our defence of the Simple View must now confront the question regarding which epistemological role emotions can play vis-à-vis evaluative knowledge and whether this role is enough to support the Indispensability Claim.

The first lesson we should draw from the foregoing is that the prospects for Emotional Dogmatism do not look better than they seemed at the outset of our discussion. We have begun with the observation that emotions are subject to why-questions, and that this is a problem for Emotional Dogmatism. True, we have adopted a deflationary account of the relevant cognitive bases in the shape of the Simple View, but even the Simple View is difficult to square with Emotional Dogmatism. Let us see why.

We have seen that emotions play a mediating role. On the one hand, they react to and are potentially justified by their cognitive bases. On the other hand, they elicit and potentially justify evaluative judgements. Obviously, only justified emotions have the potential to justify evaluative judgements. The Simple View has it that justified emotions need not be based on substantive cognitive bases; awareness of relevant constellations of natural properties is sufficient. This is enough to put pressure on Emotional Dogmatism. After all, if awareness of constellations of natural properties is sufficient to justify emotions, why is it not sufficient to justify the corresponding evaluative judgements?

According to this line of thought, emotions at best transmit the evidence in favour of evaluative judgements that is given by the cognitive bases – they do not provide additional evidence in favour of these judgements (Brady 2013; Deonna and Teroni 2012a; Goldie 2004; Müller 2019). Emotions function in the same way as a messenger transmitting a piece of news: the messenger herself transmits the message as well as whatever evidence the sender may have in favour of what the message contains, but she does not add any new evidence it its favour. The point just made grounds what is known in other contexts as a form of ‘buck passing’ regarding the relation between value judgements and what justifies them. The central claim of the buck passing account is that value (and value awareness) does not provide any additional justification to the (awareness of the) relevant constellation of natural properties. In the words of Scanlon (1998), who fathered the buck passing account of value, we should avoid double-counting. Here, we should avoid saying that the relevant evaluative judgements are justified by both the cognitive bases and the emotions these cognitive bases justify. If emotions are to play an indispensable role in the acquisition of evaluative knowledge, this role cannot consist in providing evidence in favour of the relevant judgements. We have to look elsewhere.

b. Access, grounding and understanding

One indispensable role that emotions do play vis-à-vis evaluative knowledge concerns their capacity to make us aware of the relevant evidence. Let us come back to our example. You are aware of the teenagers setting a cat on fire for fun. We have acknowledged that indignation does not provide additional evidence in favour of the judgement that the teenagers are cruel. Still, we may wonder whether you would have been aware of the situation if you were not equipped with the proper emotional sensitivity, i.e. a sense of basic humanity that gives rise to indignation in the relevant circumstances.

We are constantly bombarded with a multitude of potentially significant stimuli and, as emphasised above, emotional sensitivities play a crucial role in drawing our attention to a limited subset of these. It is well documented that the absence or disruption of these sensitivities (e.g. because of cerebral damage), has dramatic consequences for the capacity to detect salient evaluative aspects of the environment (Damasio 1994; De Sousa 1987). In this light, it is safe to assume that, while affective sensitivities and their manifestations in emotions are not evidence for evaluative judgements, they are indispensable means of accessing this evidence.

This suggests that emotions are in the business of initiating epistemic searches aimed at uncovering evidence for (or perhaps against) them and the corresponding evaluative judgements (Brady 2013). On this model, (many) emotions function a bit like ‘alarm bells’ which signal the potential presence of something harmful or beneficial to the subject and motivate her to check whether this initial evaluative verdict is appropriate or not. When everything goes well, the subject unearths evidence that puts her in a position to support her evaluative judgement.

In this connection, we should also insist on a second role emotions may play in the acquisition of evaluative knowledge: they contribute to attitudinal justification in grounding evaluative judgements in the relevant evidence. The best way to approach this role is to ask the following question: if you did not feel indignation, could your evaluative judgement that the teenagers are cruel be grounded in your awareness that they set a cat on fire for fun? The observations in Section 2 regarding the role of emotions in modulating attention and drawing it to the relevant features of the situation can be directly exploited here. The idea is that, in the absence of emotionally driven attentional processes, evaluative judgements would often fail to hook up on the relevant features in complex and unfolding situations (Brady 2013; Deonna and Teroni 2012b). Sometimes, it is thanks to emotions that judgements display a counterfactual profile that allows them to respond to evaluative evidence.

The third and final role that the emotions play in the acquisition of evaluative knowledge concerns the understanding of value. Knowing that p for any evaluative proposition presupposes that one understands the evaluative concept that features in the justified and true judgement that we make. We shall now see why emotions play an indispensable role in this regard.

Judging that something has a given value requires deploying the relevant evaluative concept. Possession of this evaluative concept in turn requires that one be capable of applying it to a variety of objects. Concepts go hand in hand with categorisations, and categorisation means treating different things as the same (Margolis and Stephen 2019). What is more, it is natural to think that a subject can meet this constraint on concept possession only if she is somehow guided in her categorisations. Now, according to the Simple View, justified emotions are justified by constellations of natural properties that change significantly from one instance of a given emotion to the next – this is again the multi-realizability of evaluative properties. This means that the subject cannot be guided in the application of her evaluative concept by the constellations of natural properties she is aware of.

The natural thing to say at this juncture is that it is the emotions themselves that provide the needed guidance: the way these various constellations of natural properties are experienced in the emotions, i.e. under the same evaluative guise, is what provides unity to these various constellations and guides the application of evaluative concepts. For instance, it is the subject’s anger, i.e. her experience of offence, which guides her application of the concept of offence. Evaluative categorisations would be essentially emotion driven (D’Arms and Jacobson 2010; McDowell, 1985; Wiggins, 1987).

As you will recall, this is the idea at the centre of the Fitting Attitude analysis of value, which builds upon the intuitively compelling thought that categorising an object as funny or shameful is not detachable from the subject’s understanding that there is evidence in favour of the relevant emotion. And we might wonder what sort of understanding of there being such evidence we would preserve were we deprived of emotions: what provides unity to different constellations of natural properties is the fact that they merit the same emotional response. Emotions, as we have just emphasised, are intimately interwoven with evaluative concepts: it is by experiencing the former that we come to understand what we mean by the latter (Deonna and Teroni 2021; Peacocke 1996; Zagzebski 2003).

The force of this point comes to light if we imagine a creature deprived of emotions who has been able to get a handle on our evaluative practices because, say, she has learned to recognise the responses of others. This creature is thus linking her application of evaluative concepts to responses she can discriminate correctly in others. But does she understand the evaluative judgements she makes? For her, judging that something is amusing, degrading, or offensive consists in realising that the object’s properties justify the amusement, shame, or anger of others. While her evaluative verdicts are in line with ours, the conditions of application of her evaluative concepts are radically different from ours. She no doubt understands something, but not, we may think, the point of our evaluative practices. Her lack of emotional responses means that she cannot experience objects as giving her reasons to react in various and distinctive ways. Being deprived of the capacity to experience for herself situations as offensive, shameful, or amusing, the sense in which we may think of her as animated by emotional sensitivities, such as a sense of decency, of honour, or of humour, is elusive to say the least. Although she might succeed in blending in, as it were, such sensitivities could only be those of the people on whose responses she models her evaluative competence. If there is any point for her in making evaluative judgements, it is simply not the same as ours.

4. Conclusion

Emotions, it turns out, are indispensable for the acquisition of evaluative knowledge, since they play at least three fundamental roles vis-à-vis evaluative knowledge: they draw our attention to evidence, help constitute the counterfactual dependence of evaluative judgements on this evidence, and are key determinants in our understanding of the evaluative concepts that are deployed in evaluative judgements. The Indispensability Claim is true, albeit not in the form of Emotional Dogmatism, as none of these roles consists in the emotions being themselves evidence for the judgements they typically elicit.

5. Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 See Poellner 2016 and Mitchell 2017 for worthy attempts to defend Emotional Dogmatism in this regard.

2 Considerations for attitudes come in different guises (D’Arms and Jacobson 2000). Amusement at a joke may be epistemically justified – the joke is witty and corrosive – despite being morally unjustified – it is cruel – and/or prudentially unjustified – it targets a person of power and could backfire. We shall be exclusively interested in epistemic justification. As is apparent in this example, epistemic justification is distinctive in being indexed on the values that feature in the correctness conditions of the emotions: for amusement, it is indexed on funniness.

3 The claim here is not that a subject must answer (or even be capable of answering) these why-questions in order to be justified. This would clearly be too demanding. The claim is much more modestly that the existence of why-questions helps reveal the pre-theoretical need for evidence in favour of emotions.

4 We sometimes ask why-questions about perceptual experiences. To take a well-known example (Markie 2005), one may ask why a gold digger has a visual experience as of gold and receive the answer that this is because he craves for riches. However one wants to interpret them, the existence of such limiting cases does in our opinion nothing to establish that perceptual experiences respond to evidence in the way emotions do. But see Mitchell 2017 for a contrasting view.

5 The merits of these candidates are discussed in Deonna and Teroni (2022) and Milona and Naar (2020).

6 We are interested in how cognitive bases justify emotions, not in how cognitive bases are themselves justified. Clearly, emotions are sometimes justified by cognitive bases – such as beliefs – that are also in need of justification. We will not touch on this issue here.

7 This conclusion is supported by a thorough discussion in Mitchell (2019, 2021: chap.3). For a criticism, see Müller (forthcoming).

8 This idea is less implausible than it may seem initially, as the application of general principles can be understood in dispositional terms (Döring 2007; Jones 2006).

9 Observe that what needs to be secured is the relevant counterfactual profile, and that this profile may be secured by factors that take place outside the subject’s ken (Bergmann 2006). The counterfactual profile needs not be the result of the subject’s consciously applying general principles to the situation at hand – the gap at issue here is thus different from the generality gap discussed above.