What is happiness made of? A fundamental intuition is that emotions and, more generally, affective states, contribute, for better or for worse, to the value of our lives. There are surely several reasons why emotions have value and, more specifically, why they have what is usually called prudential value—that is, the kind of value that allows them to contribute positively or negatively to the happiness of those who experience them. For example, emotions often inform us about the things that matter to us, prepare us to negotiate challenges that we face, and are central aspects of the relationships of love and friendship, but also of hate and rejection, at the heart of our existences (Deonna and Teroni, 2013; Haybron, 2008; Rossi and Tappolet, 2015). In what follows, we shall leave aside many contributions of emotions to happiness so as to focus on the traditional idea that emotions affect happiness because they are states of pleasure and displeasure. To this extent, the intuition that constitutes our starting point is hedonistic.[1] On the basis of this intuition, we wish to address two questions that have preoccupied hedonists and their critics.

First, what constraints must emotions satisfy in order to contribute hedonically to happiness? In the first part of our discussion, we start from the idea that the constituents of a subject’s happiness must somehow engage him or her: they cannot leave the subject totally cold or be imposed on him or her from the outside. Let us call this the nonalienationconstraint (Fletcher, 2013; Railton, 1986). On this basis, we shall try to show that the defence of the intuition that emotions have prudential value presupposes that they are forms of engagement, a condition that only some approaches to the emotions satisfy.

The second question we wish to address, which follows from the idea that different emotions are distinct types of engagement, is that of the unity and variety of states of (dis)pleasure (Clark, 2005; Feldman, 2004). If we acknowledge the existence of a great variety of (dis)pleasures, we still need to understand what makes all of them (dis)pleasures—otherwise the attractiveness of hedonism would diminish. To what extent are (dis)pleasures unified? Let us call this the unity constraint. In the second part of our discussion, we shall argue that conceiving of the emotions as forms of engagement allows us to resolve in an attractive way the difficulties concerning the variety and unity of (dis)pleasures that weigh on traditional hedonism.

In the last part, we shall defend the form of affective hedonism that will have emerged. In particular, we shall argue that the approach we advocate can be extended from emotions to other affective states and that the central role we give to action tendencies in our conception of affectivity does not call into question the idea that emotions contribute to happiness because of their hedonic value.

If something contributes to subjects’ happiness, it cannot leave them totally cold and fail to engage them. As Peter Railton aptly puts it, “it would be an intolerably alienated conception of someone’s good to imagine that it could fail in any way to engage him” (1986/2003, p. 47; see also Fletcher, 2013). In the hedonist framework we presuppose here, letting chocolate melt in one’s mouth, reading a novel, listening to an opera, riding a bicycle up a mountain pass, and so on, contribute to happiness only to the extent that one takes pleasure or displeasure in such activities.

The same observation seems to apply to any bodily sensation: the sensation of chocolate melting in your mouth contributes positively to your happiness only if you take pleasure in it.[2] Consider the ascetic, a figure who is often said to have an unfavourable attitude toward pleasures. To the extent that the relevant sensations do not please the ascetic, as opposed to what is the case for most of us, these sensations do not contribute positively to the ascetic’s happiness. The challenge for hedonism is to preserve the intuition that (dis)pleasures contribute to happiness, which requires the development of a conception of taking (dis)pleasure that engages the subject. This is the nonalienation constraint on the constituents of happiness.[3]

If we accept our initial intuition that emotions are constituents of happiness, this constraint must be met: emotions must constitute forms of commitment on the subject’s part. Which approaches to the emotions make them relevant forms of commitment? We shall first argue that two classical theories of the emotions do not fit the bill: the Jamesian theory and the theory that conceives emotions as perceptions of values. On this basis, we shall argue that the nonalienation constraint supports theories that emphasize the close links between emotions and actions.

Let us first consider the Jamesian theory (James, 1890; see also Prinz, 2004). According to this theory, emotions are configurations of bodily sensations. These sensations are underpinned by changes in a multitude of body systems: in facial and skeletal muscles, in the autonomic nervous system (regulating heartbeat, sweating, breathing, digestion), and in the endocrine system (regulating adrenalin, for example). The central idea of the Jamesian theory is that these changes are rich enough to distinguish through the resulting sensations different types of emotions broadly corresponding to distinctions encountered in ordinary language (Deonna and Teroni, 2017). Whether one considers this program of emotion individuation through profiles of bodily sensations promising or not, the fact remains that these sensations do not seem to directly affect the happiness of the subject experiencing them. The challenge posed by the ascetic takes its full significance here: if any sensation can leave us cold—that is, can fail to satisfy the nonalienation constraint—then this is true for the configurations of bodily sensations recruited by James. This means that the Jamesian theory cannot satisfy the intuition that emotions are, as such, hedonic constituents of happiness.[4]

The way in which we have just used the distinction between bodily sensations, on the one hand, and the fact that a subject takes pleasure in them, on the other hand, points in the direction of a conception of the emotions that would make it possible to avoid the difficulties faced by the Jamesian theory. The latter is indeed often criticized for its inability to account for the intentionality of emotions: bodily sensations are at best about bodily changes, whereas fear is about the approaching dog; anger, about a colleague’s remark; amusement, about a joke, and so on (of course emotions can sometimes be about sensations, as when we are looking forward to a forthcoming caress). This is why emotions should rather correspond to the idea of “taking (dis)pleasure,” which we have just seen at work in the criticism of the Jamesian theory: sensations, as well as reading a novel, listening to an opera, and all other things, should move us positively or negatively in order to contribute to our happiness. According to a popular approach, we should analyze the phenomenon of “taking (dis)pleasure” by means of evaluative representations. To take pleasure in something is to represent it positively; to take displeasure in it is to represent it negatively. In contemporary philosophy of emotions, the most influential theory of this type conceives emotions as perceptual experiences of specific “thick” values.[5] To be frightened by a dog is to have a perceptual experience of its dangerousness; to get angry at a remark is to have an experience of its offensiveness; to be amused by the joke is to have an experience of its comic character (Tappolet, 2000, 2016; Döring, 2007).

Do we have here the necessary ingredients to support the claim that emotions engage us as such and thus to satisfy the nonalienation constraint? At first sight, yes: speaking of value representation or evaluation seems to implicate the subject in the relevant way. On closer inspection, however, representing the positive or negative value of an object is not yet being committed in any way. Consider the perceptual experiences on the model of which the theory in question understands emotions—for example, the visual experience of a pink poppy. Such an experience may fail to engage the subject in any way; at most it can be said to involve an “impression of truth or presence” of that poppy (Tucker, 2013). The same is true of perceptual experiences regardless of the properties they represent, and this includes the thick values represented by emotions according to the theory at issue. Indeed, it is not clear why the mere fact of representing a remark as offensive would transform the characteristic relation the subject has to his or her environment in perceptual experience into a more substantial kind of engagement. The nonalienation constraint is thus still not met, and hence the intuition that emotions as such contribute to happiness cannot be satisfied by a perceptual theory of the emotions.

The lesson to be drawn is that “taking (dis)pleasure in” cannot be reduced to perceptual experiences of values. In order to find the kind of engagement required, we must try to substitute the psychological mode of “perceiving” or “representing” at the heart of representational theories with another mode that can do justice to the intimate links between emotions and actions, links that are already suggested by the very term “emotion.”

A number of theories of the emotions satisfy the nonalienation constraint by insisting on the close links between emotions, motivation, and action.[6] We shall focus here on the attitudinal theory (Deonna and Teroni, 2012, 2015, 2022), which seems to us to illustrate particularly well the type of engagement suited to meet this constraint. How does this theory understand the emotions? Where representational theories, such as the perceptual theory, appeal to only one psychological mode, that of perceiving or representing, directed towards a content in which the relevant thick values feature, the attitudinal theory redistributes the cards: different types of emotions are so many distinct evaluative modes directed towards contents in which the relevant thick values do not necessarily feature. The perceptual theory asserts that to be angry at a remark is to have the perceptual experience of this remark’s offensiveness. The attitudinal theory, on the other hand, asserts that to be angry is to be in a psychological mode specific to anger directed towards that same remark. How to characterize this psychological mode? While various options are open, the version of the attitudinal theory that we prefer conceives of these psychological modes as specific types of engagement with the relevant objects in a way that is directly relevant to our purpose. Specifically, the theory appeals to feelings of action readiness directed toward some aspects of the environment (Deonna and Teroni, 2012, 2015).

The central idea is that we grasp the sense in which emotions are evaluative attitudes when we realize that they are experiences of the body disposed or ready to act in a differentiated way towards a given object or situation. These felt bodily attitudes are typically directed towards objects outside the body of the subject who feels them. Moreover, the idea of attitude must be understood in a broad sense to include tendencies to move away from, to approach, or to go against an object, to focus on it, to submit to it, to see oneself attracted by it, to disengage from it, or even to suspend all interaction with it. Consider anger. In an episode of anger, we feel our bodies prepare for a form of active hostility. This idea is happily applied to other types of emotions. When we are afraid, we feel our bodies mobilize to neutralize something; when we are ashamed, we feel a readiness to move away from the gaze of the person who is causing this emotion; and when we are sad, we feel symptoms of withdrawal from the object whose loss is causing this emotion. Finally, it is important to appreciate that these emotional attitudes harness the subject’s attentional resources and direct them to prioritize processing of the situation towards which the emotion is directed (Scarantino, 2014).

Through these examples, which all underline the agential dimension of the feelings characteristic of emotions, we can appreciate how they are evaluative attitudes. Having such an attitude towards a given object is subject to a norm of appropriateness or correctness referring to the relevant thick value (D’Arms and Jacobson, 2000). The attitude in anger—namely, a bodily preparation for hostility—is an attitude that it is appropriate or correct to have toward a given person if and only if he or she has committed an offence. Fear of a dog is another evaluative attitude, which consists in the experience of the body preparing to limit an object’s impact (flight, preemptive attack, immobility, etc.), an attitude that it is correct to adopt if and only if that object is dangerous. And shame is an experience of self-degradation, insofar as we feel our bodies ready to disappear underground or out of sight of those who denigrate us, a correct attitude when we are indeed degraded. These are just a few examples of an approach to emotions that seems attractive to us—its fruitfulness depends of course on the possibility of describing all emotions in a rich and convincing way in terms of felt bodily attitudes.[7]

If we adopt this attitudinal approach to the emotions, the immediate benefit for our purposes is that they possess the right profile to satisfy the nonalienation constraint and, hence, to be hedonic constituents of the happiness of the subject who feels them. Indeed, the types of preparation to act recruited by the attitudinal theory are attitudes that the subject takes towards what is represented. According to this theory, the subject undergoing an emotion does not remain cold in relation to what it represents, be it the ascent of a mountain pass, the experiencing of an opera, the tasting of a chocolate bar, the experiencing of a bodily sensation, and so on. We thus escape the difficulties raised by both the Jamesian and perceptual theories.

Is this really the case? Doesn’t the challenge posed by the ascetic resurface in a slightly different way, but with the same unfortunate consequences, for emotions conceived in line with the attitudinal theory? Consider a situation in which you enjoy a chocolate bar, but are, because of your diet, angry with yourself and feel slightly guilty. Or consider a case of morbid curiosity in which you feel a great disgust for an object yet are interested in exploring it further. In such situations, should we not conclude that the emotion of delight or disgust itself does not engage you, your engagement being manifested instead in your anger, guilt, or interest? Doesn't this show that subjects’ engagement in their emotions is always optional? No, these verdicts would be too hasty. What these situations suggest is at most that the type of implication of the subject characteristic of the emotions is, under certain conditions, opposed or even cancelled. The cases we have just presented do not call into question the idea that we are engaged in the joy of tasting the chocolate bar or in the disgust that an object arouses in us. Rather, they call into question the kind of engagement on our part that these emotions manifest in complex situations involving other attitudes and, in particular, involving emotions that are in tension with them.

How to deal with these relatively common cases in our investigation of the contribution of emotions to happiness? There is no obvious answer, but here are a few thoughts. When a given situation arouses two contradictory emotions, as in morbid curiosity, then their respective contributions will remain distinct: the contribution of curiosity will be that of a pleasure; that of disgust, a displeasure.[8] Things get more complicated when the situation features “embedded” emotions, as when joy leads to anger or guilt. Here, two options are available. First, one might go for a “winner takes all” strategy and say that the higher-order emotion fixes the only contribution of the emotional set at issue to happiness. If your joy makes you angry, this is a displeasure that contributes negatively to your happiness. Alternatively, we can apply the recipe suggested for morbid curiosity and say that, in such a situation, we are engaged in two different ways, each making its own contribution to our happiness: the pleasure of rejoicing contributes positively; the displeasure of anger and guilt, negatively. Whether we choose one or the other of these treatments of complex emotional situations, they do not call into question the attitudinal approach recommended here.

In this section, we have seen how an attitudinal approach to the emotions can meet the nonalienation constraint and, hence, stand by the intuition that emotions are hedonic components of happiness. However, by admitting that emotions are distinct attitudes of preparation for action, have we not called into question the unity characteristic of states of (dis)pleasure? It is to this question that we now turn.

In the hedonistic perspective that we have endorsed from the beginning of our discussion, mental states contribute to happiness in that they are (dis)pleasures. However, exploration of the nonalienation constraint has led us to adopt a specific type of attitudinal theory. And it is not easy to see how, on this theory, emotions qualify as pleasures and displeasures. Let’s take a step back.

Traditionally, a (dis)pleasure is a hedonic quale that is often conceived as a primitive given. Understood in this way, all (dis)pleasures are constituted by qualia of this nature—even if these qualia can of course be combined with other aspects of the experience of the situation or activity that generates the subject’s pleasure. This conception is often criticized for not being able to respect the heterogeneity of (dis)pleasures. Compare the pleasures of savouring chocolate, listening to opera, reading a novel, riding a bike up a mountain, and winning a competition (Clark, 2005). The conclusion that many draw is that such a comparison does not support the idea that these different pleasures are pleasures because they all contain the same hedonic ingredient. In light of this, the strategy that is often seen as the most promising is to admit a great phenomenological diversity in the experiences or, more broadly, situations or activities that give us (dis)pleasure, and to explain what makes them (dis)pleasures by claiming that they are all accompanied by an attitude of taking pleasure or displeasure in the relevant experiences, situations, or activities (Feldman, 2004, but see Bramble, 2013 and Labukt, 2012).

The viability of this strategy must, of course, be measured by what we can say about the nature of these two attitudes. If, for example, we were led to appeal once again to irreducible hedonic qualia at the level of these two attitudes, then we would again encounter the problem of heterogeneity. In fact, the most influential advocate of this approach, Fred Feldman, does not say much about the nature of these attitudes, and he even insists that they do not necessarily have any phenomenal properties. The claim that some pleasures are free of any hedonic phenomenology is, it seems to us, an overreaction to the problem of heterogeneity.[9]

An attractive way to navigate these two pitfalls—uniform phenomenology and absence of phenomenology—is to summon the distinction between a determinable and its determinates. A determinable property is a property that an object cannot exemplify as such; an object can exemplify a determinable property only indirectly—that is, by virtue of its exemplification of a determinate property (Johnson, 1921). Colours provide a good illustration of the connection between a determinable property and its determinates. The properties of being coloured and of being red are determinable properties, because an object cannot be coloured or red tout court: it is always coloured or red in virtue of the exemplification of a more determinate property—for example, a given shade of vermilion. Having this shade of vermilion is one of many ways of being red or coloured (which is why being vermilion is not the same as being red or coloured), and an object must have a shade in the relevant range to be red or coloured. Moreover, at the level of phenomenology, these links between the determinable “being coloured” and the determinate colours that we perceive allow us to clearly identify the relations between the chromatic properties that we perceive. The perceived colours are unified in that they are determinates of a single determinable, and they are distinct from one another in that they are different determinates. In addition, it should be stressed that the similarities among these various determinates can be arranged in a space structured around three axes that are provided by saturation, tonality, and luminosity.

This distinction between a determinable and its determinates has been applied to (dis)pleasures by Roger Crisp (2006) in order to address the problem of heterogeneity. Regardless of the specific way in which Crisp applies this idea, it is easy to see how it helps avoid the two pitfalls mentioned above: the relation between different (dis)pleasures is not that between different exemplifications of one and the same property, but rather that between different determinates of a given determinable. The way in which this idea should be applied to (dis)pleasures depends of course on their psychological nature. In the foregoing, we have understood emotions as evaluative attitudes that are correct as a function of whether their object exemplifies a given thick value (the threatening, the amusing, etc.). The idea is that these different evaluative attitudes are determinates of two determinables: taking pleasure and taking displeasure. There is not one single attitude of taking pleasure and another single attitude of taking displeasure. Taking pleasure and taking displeasure are abstractions. What is psychologically real are the different determinate ways of taking pleasure or displeasure: to have fun, to rejoice, to be proud, to admire, to be amazed, and so on, on the one hand, and to be angry, to be afraid, to be sad, to be ashamed, and so on, on the other hand. The problem of heterogeneity is thus solved, since there are in fact different ways of taking pleasure and displeasure, each with its own phenomenology.

The next question is of course whether these emotional attitudes are not too different at the phenomenological level for it to make sense to put them all under the same umbrella. In the case of colours, this difficulty is avoided by speaking of (dis)similarities among the perceived colours, (dis)similarities that generate a colour space structured around three axes. What can we say about emotional attitudes if we conceive them as action tendencies? Here, too, different dimensions of (dis)similarity can be highlighted within the relevant experiences. The degree of bodily arousal or relaxation, approach or avoidance, and attentional involvement (local or global, intrinsic or instrumental), and the sense of dominance or submission are all dimensions along which different emotional attitudes may be more or less similar. For example, the characteristic attitudes of fear and anger are similar in that they both involve a high degree of bodily arousal and sustained attention to their object, but they differ in that anger is characterized by a sense of dominating the situation; fear, by a sense of being dominated by it. We can also look at the phenomenology of these attitudes from a more local perspective, asking what kinds of bodily sensations they include (muscle tension, heart rate, localization of sensations in this or that part of the body, etc.) (Deonna and Teroni, 2017, 2020; Numenmaa et al., 2014). Thus, two emotional attitudes can also be compared in light of these more localized sensations.

Furthermore, the different emotional attitudes constitute determinates of the determinables “taking pleasure” and “taking displeasure” in that they constitute different ways of accepting or rejecting a given value. An emotional attitude welcomes a value when it favours, promotes, supports, or protects it. It rejects a value when it neglects, avoids, or suppresses it. As far as emotional attitudes can be classified in these ways, they constitute determinates of these two determinables. This leaves open the possibility of admitting, as we shall see in the next section, the existence of borderline cases.

If we subscribe to such an approach to (dis)pleasures in terms of evaluative attitudes, what are its consequences for the question of the respective contributions of different pleasures and displeasures to happiness? One of the promises of traditional hedonism, which we also find in the attitudinal hedonism defended by Feldman, is the possibility of quantifying (dis)pleasures, of comparing them to each other, and of calculating happiness by adding up pleasures and subtracting displeasures. This promise seems to presuppose an approach in terms of hedonic qualia, or in terms of two attitudes that can vary in intensity and duration: one of pleasure and another one of displeasure. This approach is much more difficult to defend when the attitudes of taking pleasure and taking displeasure are conceived as determinables that are psychologically realized by determinate attitudes such as enjoying, being amused, being afraid, or being disgusted. How indeed can we go about comparing an attitude of rejoicing at degree n for fifteen seconds with an attitude of feeling pride at degree n* for eight seconds—for example, in order to determine their respective contributions to happiness? The heterogeneity of the relevant attitudes, as they are portrayed in the approach we have put forward, suggests that such comparisons are at best difficult, if not meaningless. Is this a serious problem? Clearly, the implication of our approach is that it is difficult to quantify how a rich and varied emotional life contributes to happiness. This is a problem if one considers that the quantitative ideal promoted by traditional hedonism is appropriate to the domain. However one wishes to position oneself in relation to this issue, it seems to us that our approach is exactly in line with the doubtful or even dazed reaction that many of us have to attempts to measure the respective impact of different emotions on happiness.

In light of what we have seen in this section, the attitudinal approach to emotions seems to navigate the two pitfalls of adopting a uniform hedonic phenomenology and of claiming that the nature of (dis)pleasures is not to be located in their phenomenology: emotional attitudes possess a rich phenomenology that allows us to anchor our intuitions concerning both the unity and the diversity of the hedonic domain.

In discussing the contribution of emotions to well-being, we have emphasized the advantages of satisfying constraints of nonalienation and unity, and we have sought to show that an attitudinal approach to emotions can meet these constraints. Yet we may think that this approach encounters difficulties related to the way in which it accounts for the hedonist intuitions from which we have started. Doesn’t adopting an attitudinal approach amount to abandoning de facto the hedonistic perspective? This difficulty comes in the form of necessity or sufficiency. On the one hand, one can ask whether there is really a necessary link between the hedonic components of happiness and tendencies to act. On the other hand, one may ask whether a tendency to act is sufficient in itself to make a hedonic contribution to happiness. This last section focuses on these two worries.

Are the hedonic components of happiness necessarily action tendencies? Have we not given up on our intuitions about the (un)pleasantness of emotions by conceiving them as psychological states constituted by felt tendencies to act? Two types of mental states—some bodily sensations and moods—call into question the necessity of such a relation, since they intuitively appear to be important hedonic components of happiness. What consequence should we draw from this? Should we abandon the necessity of the relationship and thus a unified approach to the hedonic components of happiness, or should we seek to apply an attitudinal approach to the sensations and moods at issue?

As far as bodily sensations are concerned, the worry may seem serious. Indeed, do we really want to say that all bodily sensations are (dis)pleasures because of the presence of an evaluative attitude, rather than because of their exemplification of a hedonic quale? The intuition that we should go for the second option is particularly strong with respect to pain sensations. Yet there is good reason to think that the lesson we have drawn at the beginning of our discussion concerning the “pleasures” of the ascetic applies equally to the “displeasures” of pain sensations. In both cases, we must avoid returning to the idea that there is a positive or negative hedonic quale shared by all (dis)pleasures. The fact that this is true even for pain sensations can be seen in the phenomenon of pain asymbolia, a dissociative state well documented by clinical studies (see, for example, Grahek, 2007). What is it?

Asymbolic patients can refer to and recognize their pains, but they are not moved by them at all. When an experimenter pierces their skin with a needle, they tend to smile and say something like “It hurts, but I don’t care.” In our view, we should understand pain asymbolia as follows: asymbolic patients refer to distinct bodily sensations that happen not to displease them (Grahek, 2007; Teroni, 2019). Given that asymbolic patients describe their sensations as painful, one might initially think that this conception of their condition is inconsistent with the evidence. When they say that their sensations are painful, doesn’t that mean that they experience the archetype of displeasure? And, since this displeasure is not accompanied by any negative attitude, should we not abandon our attitudinal approach in favour of a more traditional strategy in terms of hedonic qualia? No. In fact, pain sensations differ from other sensations in that they represent bodily harm (e.g., Tye, 1995). This opens the possibility that asymbolic patients use the term “painful” to describe bodily sensations that represent damage. The peculiarity of their condition is that their pains (i.e., sensations that represent parts of their body as damaged) are not displeasures. What is disconcerting about asymbolic patients is that their proprioceptive representations of bodily damage (i.e., their pains) always “leave them cold.” That is, their pain sensations are not accompanied by the kind of evaluative attitude of rejection that is familiar to us and that makes them displeasures. The idea is that the representation of bodily harm specific to pain sensations does not explain what makes them displeasures—what explains it is rather the presence of an attitude of rejection toward them. This allows for a unified approach to unpleasant bodily sensations. Many other unpleasant sensations are after all not painful—that is, they do not represent bodily harm (tickling, itching, etc.). What makes all these sensations unpleasant is that they are accompanied by attitudes of rejection. This is typically the case for most of us most of the time, but it is not a necessity. The same strategy applies to “positive” bodily sensations, such as those generated by certain caresses, which are often pleasurable, but may cease to be so.

Let us now consider moods. Are moods hedonic components of happiness because they are attitudes? The answer is tricky in that contemporary theories of moods are highly polarized. The most salient dividing line separates theories that make them intentional attitudes whose objects are simply more “global” than those of emotions (Tappolet, 2018) from theories that make them sets of diffuse and enduring bodily sensations (Deonna and Teroni, 2009). Whether one subscribes to one or the other of these theories, we can apply the ideas that we have put forward to moods. If we adopt the first theory, moods, such as anxiety around exam time, should be seen as (dis)pleasures by virtue of the fact that they are rejecting or welcoming attitudes on the model of emotions defended above. “Moody” attitudes will simply be more global than emotional attitudes, and their duration typically longer. If we adopt the second theory, moods—for example, the bodily tension I feel before drinking my first morning coffee—are (dis)pleasures in that they are made up of (dis)pleasurable bodily sensations, which we understand on the model sketched above. In either case, it is not obvious that moods are a serious threat to the approach we advocate.

We now turn to the worry regarding sufficiency: is every felt tendency to act really a hedonic component of happiness? After all, some felt tendencies do not seem to contribute to happiness. Consider, for example, the carpenter who is about to hit the head of a nail, or the tennis player who is preparing to receive her opponent’s serve. The action tendencies in question are clearly felt, but are not (un)pleasant. This is indeed the case, and the approach that we have advocated above is able to account for what singles out the relevant action tendencies. A felt tendency to act is a (dis)pleasure taken in a given situation onlywhen it is an evaluative attitude. What does this mean? An evaluative attitude is an attitude that it is correct to adopt with respect to a situation if and only if it exemplifies a certain value. It must be stressed that these correctness conditions are not imposed from the outside but derive from the very nature of these felt preparations to act: it is because fear is a specific attitude of preparation to act that it is correct if and only if there is a threat.[10] It is only when an attitude has such correctness conditions that it makes sense to speak of an evaluative attitude, an attitude that constitutes a (dis)pleasure and thus contributes hedonically to happiness.

Will critics be convinced by this answer? Perhaps not. How, they will insist, does the fact that emotions are evaluative attitudes give them the status of hedonic components of happiness? This residual worry can target all emotions or certain types of emotions only.

If it targets all emotions, the concern is probably that the phenomenology of welcoming or rejecting a situation in a certain way is not enough to account for the fact that emotions are (dis)pleasures. Does our approach have the resources to alleviate this worry? Yes, once we realize that the concern is motivated by disproportionate attention to (un)pleasurable bodily sensations and to their salience within emotions of a certain intensity. For it is undoubtedly the importance given to unpleasant sensations that feeds the idea that the mere presence of an attitude of welcoming or rejecting is not enough to account for hedonic phenomenology. Of course, it must be admitted that, when they are present, these sensations also contribute to (un)happiness. But, in the light of an attitudinal approach to (dis)pleasures, this admission should not lead us astray: as we have observed, the bodily sensations present during emotions are themselves (dis)pleasurable only because of evaluative attitudes directed towards them. And, if emotions are mostly not directed towards sensations, they also are evaluative attitudes and they are (un)pleasant because of this fact. The concern at issue is therefore misleading, because it does not pay sufficient attention to the fact that sensations are (un)pleasant only when they are accompanied by evaluative attitudes. This being admitted, we can maintain that evaluative attitudes directed towards something other than sensations are (dis)pleasurable for the same reason that (dis)pleasurable sensations are.

If we now seek to apply the worry in a more nuanced way only to certain types of emotions, the diagnosis made by an attitudinal approach will be different and will proceed case by case. Let us consider here anger, one of the borderline cases between taking pleasure and taking displeasure to which we alluded above. One might think that anger, if it is a paradigmatic example of a felt tendency to act with evaluative correctness conditions, is not clearly a component of (un)happiness. The proponent of an attitudinal approach considers that this idea arises not from the fact that anger is a hedonically “silent” attitude, but rather from the fact that it is hedonically ambivalent. Our intuitions about anger are not clear-cut, for episodes of anger—and especially those that can be given free rein—often combine the displeasure of witnessing or remembering an offence with the anticipated pleasure of revenge. Since these two aspects of anger tip it on both sides of the hedonic scale, its overall hedonic contribution to happiness is unclear. If we can apply such a strategy to all the emotions that constitute borderline cases, one may conclude that this is a significant asset of the approach we advocate.

We have argued in this section that an attitudinal approach has the resources to appease the worries related to the necessity and sufficiency of action tendencies to constitute hedonic components of happiness.

Starting from the intuition that emotions contribute to happiness, we have explored how hedonism can accommodate this intuition. Through the nonalienation constraint, we realized that this intuition calls for a conception of (dis)pleasures as states that engage the subject in a way that is irreconcilable with the traditional quale approach to (dis)pleasures. The conception of emotions that emerged from this discussion equates them with feelings of action tendencies. In light of the unity constraint, we then argued that, while emotions are so many preparations to act, they nevertheless constitute a unified category of mental phenomena consisting in welcoming or rejecting something. Finally, we have responded to some concerns raised by this approach. On the one hand, we have explained how it can be extended to other affective states. On the other hand, we have argued that the role played by action tendencies in our approach to affectivity does not call into question the idea that emotions contribute to happiness because of their hedonic aspect. This is not to say that hedonism has no further and important challenges to meet, or even that we should favour this approach to well-being. Still, we hope to have contributed to recent attempts to make hedonism more attractive.