1.1 Feelings as non-cognitive affective appraisals

According to Kant, feelings are kinds of sensations. Although his account of sensation is too complex to do it justice here, what is directly relevant to my argument is that sensation is ‘a modification of [the subject’s] state’ in inner sense that results from her being affected (CPR, A320/B376). But sensation can be either ‘objective sensation, as perception of an object of sense’ or ‘subjective sensation, through which no object is represented’, namely feeling (CJ, 5: 206).Footnote ⁴ Both types of sensation are an awareness of the self in the sense that they manifest the state of the subject as she is affected by the world or by herself.Footnote ⁵ By contrast with objective sensation, however, in feeling, the object is considered only ‘as an object of satisfaction (which is not a cognition of it)’ (CJ, 5: 206). Feeling is thus an evaluative mental state, what I call an affective appraisal.Footnote ⁶ This feature of feeling is manifested by the fact that it presents as pleasure or pain – it is valenced – whereas objective sensations do not present in this way; they have no valence.Footnote ⁷ Feeling and objective sensation are thus two distinct and irreducible ways in which the subject is aware of herself as being affected: the former is evaluative and the latter is not. As a result, contrary to Merritt’s objection, the fact that the subject and the feeling that appraises her are somewhat independent does not turn the awareness she has of herself through feeling into some kind of self-cognition. For as I have argued, the awareness we have of ourselves through feeling is different in kind from the awareness we have of ourselves through cognitive states: the former is valenced while the latter is not.

It is because of the central role I assign to the valence of feeling that my interpretation cannot accommodate Merritt’s own proposal. She suggests that the best way of understanding feeling as a manifestation of the self is to align it with apperception, for ‘both are ways in which “self” is manifest as a subject, rather than as some kind of object’ (p. 000). While she does not provide much detail on the nature of this alignment, she relies on a passage from the Anthropology (A, 7: 127) where Kant points to a felt mode of self-manifestation in children that precedes the way adults think themselves as subjects (i.e. apperception). She takes the former to be an affective form of self-consciousness that accompanies our activity and that enables us to feel ourselves as subjects just as we think ourselves as subjects. If correct, this would involve modelling the relation between feeling and agency on the relation between apperception and thought, thus making it an internal relation rather than an external one, as in my account. But while it may well be the case that we have such a capacity for affective self-consciousness, I simply do not see how such feelings can belong to the faculty of feeling, a faculty Kant repeatedly calls ‘feeling of pleasure and displeasure’ (e.g. CJ, 5: 197–8, 20: 245–6).Footnote ⁸ For as Merritt herself notes, the feelings she describes are ‘not obviously specifiable as any kind of pleasure or pain’ (p. 000). In this respect, her proposal is at odds with Kant’s very definition of the faculty of feeling.

1.2 What is orientation?

Merritt’s second objection also pertains to the fact that, on my account, the subject and the feeling that appraises it are somewhat separate. She argues that, if the affective awareness afforded by feeling is independent of the activity it appraises, feeling cannot fulfil its orientational function. For it is unable to provide the subject with the guidance she needs in order to act. I believe that Merritt’s objection arises because we hold different interpretations of Kant’s notion of orientation. To show why this is the case, I follow Merritt’s lead and compare feeling with proprioception.

Merritt thinks that proprioception is not a good model for understanding the orientational function of feeling because it alone cannot give us direction: it cannot orient us, in her sense of ‘orientation’ (p. 000). In my sense, by contrast, orientation consists in determining where I stand in relation to myself and the world. Proprioception is thus an excellent model for understanding the orientational function of feeling. For it is the felt awareness of my body’s position and movement – it makes me aware of how I am moving as I am moving without needing to think about my movement.Footnote ⁹ On my account, feeling plays the same role for the mind: it is the affective awareness of myself and my activity as I am acting.Footnote ¹⁰ Both enable me to be immediately aware of myself and my activity. That is, both provide me with an awareness of my activity that does not require me to think about it. And I believe that it is this kind of felt awareness of myself that is necessary for orienting myself. To support this claim, I will show that moving purposefully requires orientation, and orientation requires proprioception to provide a felt sense of myself. And the same goes for agency: acting requires orientation and orientation requires feeling of pleasure and pain understood as an affective sense of myself.

To understand why orientation requires feeling, note that, according to Kant, orienting myself in space is not simply locating my position in the world; it involves determining where I stand in relation to objects in the world (e.g. whether the south is on my left or my right). To do so, Kant argues, objective landmarks are not sufficient; I need another point of reference, one that pertains to me rather than to the world. As he writes, ‘I orient myself geographically only through a subjective ground of differentiation’, namely ‘the feeling of a difference in my own subject’ (e.g. ‘the mere feeling of a difference between my two sides’: WOT, 8: 134–5).Footnote ¹¹ Although a lot could be said about Kant’s account, what matters for the purpose of my argument is that spatial orientation requires feeling to provide a subjective point of reference, a felt sense of myself that is akin to proprioception. My claim is that the same goes for agential orientation. It is not simply a matter of determining what I am doing or where I am going (i.e. what activity I am engaged in and what its ends are); it involves determining how I am doing (i.e. how my activity is going, including how it is going in relation to its ends).Footnote ¹² Doing so requires a subjective point of reference, one that pertains to the sense I have of myself as an agent, that is as a being who acts rather than merely behaves and whose ongoing activity needs to be evaluated in light of my ends. On my account, this sense of myself is provided by feelings: they enable my awareness of how I am and how I am doing. This is why, just as spatial orientation requires proprioception, agential orientation requires feeling.

Yet one may worry that the analogy between proprioception and feeling is actually a non-starter because there is a fundamental difference between them. While proprioception enables me to distinguish up from down, left from right, clockwise from counter-clockwise, it does not present any particular direction as to-be-pursued or to-be-avoided. Feelings are not neutral in this way. Through their valence, they present objects or activities as to-be-pursued or to-be-avoided (CJ, 20: 230–1; CJ, 5: 220). Does this not mean that feelings do more than proprioception? That instead of just situating us, they do in fact guide us, as Merritt suggests? I do not think so. For on my account, the valence of feeling should not be taken to be prescriptive. I can always, and often ought to, choose pain over pleasure. After all, Kant’s account of agency is not hedonistic.Footnote ¹³ In this sense, it is true that feeling and proprioception differ in the way just noted. Whereas proprioception is the felt awareness of the position and movement of my body (‘what my body is doing’), feeling is an evaluative manifestation of myself and my activity (‘how I am’ or ‘how I am doing’). But despite this difference, their orientational role is the same: they enable the affective awareness of myself and my activity that is necessary for orienting myself. However, while it may well be true that proprioception is the only way of being aware of my body and its movement in this way, it is less clear that feeling is the only way of making me aware of how my activity is going as it is going on. While the claim that agential orientation requires self-appraisals is plausible enough, why think that feeling alone can provide them? Why can they not be done cognitively, through introspection and reflection?

Kant supports the claim that feeling is necessary for orientation by showing that it captures something about myself that cannot be articulated cognitively and yet that is necessary for orienting myself. As he argues, I have to orient myself in space on the basis of a felt sense of the difference between my left and right because this difference cannot be captured conceptually (WOT, 8: 135).Footnote ¹⁴ In the same way, he shows that, when reason leaves the bounds of experience to explore the supersensible,

The line of argument from geographical orientation to metaphysical orientation is extremely rich, and I cannot possibly do it justice here.Footnote ¹⁵ What is relevant to my argument, however, is that I believe that the same goes for feelings of pleasure and pain: they capture something about me that cannot be articulated cognitively, namely how I am doing and how my activity is going as it is going – an immediate affective appraisal of myself and my activity.

Of course, one could object that these appraisals can also be achieved cognitively through introspection. However, introspecting my activity as it is going on is at best inconvenient and at worst disruptive, which makes real-time cognitive appraisal problematic if not impossible. Not only is introspection cognitively very costly and thus potentially too slow, turning my attention to myself as I am acting instead of focusing on my activity as it is going on interferes with the activity, and at worst interrupts it all together. By contrast, feelings are immediate appraisals that neither distract nor interrupt the activity they appraise.Footnote ¹⁶ It is in this respect that they are the best if not the only way of appraising and thus orienting our activity as it is going on.Footnote ¹⁷ Of course, a lot more needs to be said to defend this claim fully. But within the remit of this article, it suffices to show that, contrary to Merritt, feelings are necessary to orient ourselves despite the fact that they are independent of the activity they appraise.

Where Merritt is right, however, is that, on my account, feelings should not be taken to orient us in and of themselves. For I cannot know what kind of feeling I am feeling simply by feeling it. Judgement is necessary to interpret the feeling as the kind of feeling that it is in light of its context. In this sense, feelings are necessary but not sufficient to orient my activity: orientation also requires judgement. Merritt worries that, if so, ‘it is no longer clear that feeling does the distinct work of orienting activity’ (p. 000), which I take to mean that it now looks like it is judgement that performs the role of orienting us, not feeling.

To see in what sense the function of feeling can be said to be orientational despite the fact that it requires the exercise of judgement to fulfil its function, let us return once more to Kant’s argument about the necessary role of feeling for orientation. On my reading, it is not meant to show that feelings orient in and of themselves, but rather that, and more importantly why, they are necessary for orientation. Take his account of orientation in space (WOT, 8: 135). What it shows is that, while the feeling of left and right is indeed necessary to orient myself, it is not sufficient to do so; spatial orientation also requires judgement. For judgement is necessary to relate the felt awareness of my body to its spatial environment, judge the distance between my body and the objects that surround it, draw inferences to situate it relative to them, and so on. However, the fact that spatial orientation requires judgement as well as feeling does not entail that feeling thereby loses its distinctive orientational function, which, recall, consists in providing a subjective point of reference, a point of reference that pertains to the subject (e.g. ‘the feeling of a difference in my own subject’, ‘the sense of the difference between my left and my right’: WOT, 8: 134). In this sense, what does the orienting is feeling, not judgement. In the same way, it is the feeling of pleasure and pain that does the work of orientation by enabling my awareness of how I am doing; judgement simply interprets the feeling in light of its context – a role I will return to in section 2.2.

In conclusion, I have argued that Merritt and I disagree on the orientational role feeling plays for agency because we hold different and incompatible interpretations of the meaning of orientation.Footnote ¹⁸ However, not only do we agree that feeling is supposed to play a significant role in orientation, we share a more fundamental belief, namely that there is for Kant a deep connection between feeling and agency.

2.1 Emotion versus feeling

As is now well-known, Kant did not have a concept of emotion.Footnote ¹⁹ One may thus worry that any Kantian account of emotion will be de facto anachronistic. However, this is not Eran’s worry. His concern is rather that my identification of emotion with Kantian feeling lacks justification since, he argues, my account is premised on a particular view of emotion that I fail to argue for (p. 000). And as Eran rightly points out, what constitutes an emotion is a matter of debate. Not only are there disagreements about whether individual mental states should be considered emotions, there is no consensus about the very nature of emotion and whether it constitutes a unified class.Footnote ²⁰ Some hold that emotions are irreducible mental states; others that emotions can be reduced to other mental states.Footnote ²¹ Eran is thus undoubtedly correct to draw attention to the fact that whether one thinks that emotions can be identified with Kantian feelings will depend on the view of emotion one holds (p. 000). Where he is incorrect, however, is in interpreting my account as presupposing a particular view of emotion. What it does presuppose is Kant’s faculty psychology.

Readers of Kant have tended not to take his faculty psychology seriously or give it much weight.Footnote ²² But it is the ‘textual bedrock’ and a guiding-thread of my interpretation, as Merritt rightly notes (p. 000). In fact, I take one of Kant’s most fundamental insights to be that there is a faculty distinct from the faculty of cognition and the faculty of desire – the faculty Kant calls ‘the feeling of pleasure and pain’. On my reading, this entails that, since the function of the faculty of cognition is to know the world and the function of the faculty of desire is to change it, it follows that feeling has a function that is both distinct from and irreducible to the mind’s cognitive and conative functions. Furthermore, since ‘there is always a great difference between representations belonging to’ each faculty (CJ, 20: 206; see also A, 7: 153), it follows that feelings are distinct mental states that differ from and are irreducible to conative and cognitive states.

Why, then, do I take what we now call emotions to be Kantian feelings? To begin with, my proposal would be entirely unmotivated if some significant proportion of the states we now call emotions did not count as Kantian feelings. But as it happens, a significant proportion does, although as Eran rightly points out, there are a number of cases that do not work. Notwithstanding, my view does not entail that we should identify the set of all the states we now call emotions with the set of states that Kant calls feelings. On the contrary, it entails that some of the states we now call emotions are not Kantian feelings and thus should not be considered emotions; and that mental states that we do not call emotions are in fact Kantian feelings and so should in fact be considered emotions. In this respect, bodily feelings of pleasure and pain, what Kant calls ‘sensuous pleasures’ (A, 7: 230), present an interesting case.

Although there is no space to discuss them in detail, for Kant, they are full-blown feelings, as is confirmed by the fact that he defines them in the same way as he defines non-sensuous ones (A, 7: 230–1). Thus, as Eran points out, I am committed to the claim that these sensuous feelings are emotions. Eran is unconvinced by this claim, and he cites Kant’s examples of pleasure in wine and laughter (CJ, 5: 212, A, 7: 262) as cases of sensuous pleasures that are implausible as emotions on the ground that they are too primitive (p. 000). But I do not share his reservations, for not only are they clearly affective appraisals of the self, they are not as primitive as Eran takes them to be. They could be interpreted in many different ways and thus manifest rather complex states of the subject (e.g. a pleasure in wine could be interpreted as a manifestation of snobbery, the fulfilment of the aim to get drunk or a personal taste for fruity wines). But more generally, many contemporary accounts of emotion include bodily feelings, so Kant’s account is not particularly outlandish in this respect.Footnote ²³

Another type of mental state that is of particular concern to Eran is desire. He rightly notes that, insofar as I suggest that emotions be understood as Kantian feelings, and the faculty of feeling be distinct from the faculty of desire, it follows that on my account nothing that Kant calls a desire falls under the category of emotions. Eran questions this claim, arguing that at least some of the mental states that belong to Kant’s faculty of desire should be considered emotions (p. 000). So let me conclude this section by showing why I disagree with Eran on this point.

Eran finds my claim that we should not conceive of desires either as emotions or as constitutive parts of emotions particularly problematic in light of what he calls ‘the reasonable and generally accepted assumption that what we call “emotions” refers to mental states of the faculty of feeling and the faculty of desire’ (p. 000). I believe that this assumption is motivated by an intuition that strikes me as right, namely that emotions are connected with action in a way that suggests that they have motivational powers. Categorizing desires as emotions or parts of emotions makes sense of this intuition: emotions motivate us to act in certain ways because they include desires. But emotions do not need to comprise desires to have motivational powers. To see why, let me turn briefly to Kant’s account of the relationship between feeling and desire.

On my reading of Kant, he holds a causal view of the connection between feelings and desires according to which some feelings cause desires and some feelings follow desires as their effect. Most of our feelings are causally connected to desires in this way – they are what Kant calls ‘practical pleasures’.

The fact that many feelings necessarily cause desires explains the widely shared intuition that emotions are motivational, but it does so without requiring the inclusion of desires into the category of emotion.Footnote ²⁴

In conclusion, if my claim that Kantian feelings, and only feelings, are emotions is ultimately grounded on a presupposition, it is not the one Eran identifies. Rather, it is grounded on a functionalist reading of Kant’s faculty psychology. It is on the basis of my reading of Kant’s account of the faculty of feeling that I interpret Kant as putting forward a non-assimilative theory of the emotions according to which they are distinct from, and irreducible to, both beliefs and desires.Footnote ²⁵ This interpretation is undoubtedly rather controversial, and it may well put off Kantians and non-Kantians alike. But what I hope to have shown is that Kantian feelings offer a credible model for thinking about what we now call emotions.

2.2 The intentionality of feeling and the role of reflective judgement

As Eran rightly observes, my account of feeling tries to make sense of and hold on to Kant’s insistence that in feeling ‘nothing at all in the object is designated’ (CJ, 5: 203). However, he objects that I have ‘an unnecessarily strong reading’ of this claim, for I take it to entail that feelings have no intrinsic intentionality (p. 000). To show in what sense this description of my view is not quite right, I clarify my account of the intentionality of feeling and how I understand judgement’s contribution to it.

On my account, I can be aware of a feeling without it making sense to me, that is without knowing exactly what I am feeling. Feelings need to be reflected upon and interpreted in order to acquire a particular meaning. Since such reflection and interpretation is the work of judgement, judgement is the condition of the intentionality of feelings: we cannot understand what they mean unless we reflect on them in light of their external context (what triggered them, their causal history, but also the way the world is more generally) as well as their internal context (my past experience, my emotional history, my personality and so on).Footnote ²⁶ But what Eran puts pressure on – fairly, since I do not discuss this issue in much detail in my original paper – is the status of feelings sans judgement. Do they lack intentionality entirely? I believe that my account commits me to denying this. For the fact that feelings are affective appraisals entails that they have the form of proto-judgements: first because they are appraisals, and second because their appraisals manifest through their valence.Footnote ²⁷ In this sense, feelings alone already present a form of aboutness: they are an awareness of something.Footnote ²⁸ But what are they an awareness of?

I would like to suggest that feeling alone, sans judgment, is an awareness of subjective purposiveness. While this may seem like a strange claim, there are in fact numerous passages where Kant seems to identify feeling with subjective purposiveness.Footnote ²⁹ For instance, he writes that ‘the concept of the formal but subjective purposiveness of the object … is fundamentally identical with the feeling of pleasure’ (CJ, 20: 230). He also talks about ‘the empirical consciousness of subjective purposiveness, or the feeling of pleasure in certain objects’ (CJ, 20: 228), suggesting that subjective purposiveness and pleasure are connected in some fundamental way.Footnote ³⁰ To shed light on the nature of this connection, recall that, insofar as feeling is the appraisal of the way in which I am affected by an object or an activity, pleasure is in effect the manifestation of its purposiveness for me (and vice versa for the counter-purposiveness of pain).Footnote ³¹ So when Kant defines pleasure as ‘the representation of the agreement of an object or of an action with the subjective conditions of life’ (CPrR, 5: 9n.), what I take him to mean is that pleasure is the manifestation of the fact that the way in which I am affected “agrees” with me (whether it is with my ends, my desires, my interests, my nature and so on). On this basis, Kant’s identification of pleasure and subjective purposiveness can be interpreted as follows: for a rational embodied subject, to feel pleasure is to be aware of something as ‘purposive for her’. As a result, contrary to Eran’s reading of my account, feelings alone do have a form of aboutness. What I am aware of when I feel pleasure is that I am being affected in a way that agrees with me. But it is an indeterminate form of aboutness since, as I hinted at in section 1.2, making sense of a feeling beyond the fact that it is pleasurable or painful in certain ways requires the exercise of judgement.

To support this claim, it is crucial to note that, according to Kant, the power of judgement contains the a priori principle of feeling (CJ, 5: 196). While commentators have often struggled to understand this statement, on my interpretation it makes perfect sense. For reflective judgement confers its particular meaning to a feeling by applying its own principle, the principle of subjective purposiveness, to the feeling.Footnote ³² More precisely, it is by reflecting on the feeling guided by the principle of subjective purposiveness that reflective judgement is able to interpret the feeling either as the manifestation of the attainment of a particular end (in the case of pleasure), or as the manifestation of the failure to attain an end (in the case of pain). Thereby, the subject becomes aware of the feeling as the kind of feeling that it is (e.g. jealousy rather than mere pain). In the case of lower feelings, for instance, insofar as they are appraisals of the hedonic potential of objects for our lower desires, judgement interprets agreeableness as the manifestation of the purposiveness of certain objects for us (e.g. this wine suits my taste). Higher feelings function in the same way. Judgement makes sense of them by interpreting them as the manifestation of the activity of our faculties, and in particular their progress (or lack thereof) towards their ends. For instance, I am aware of a feeling of pleasure as a feeling of moral contentment because judgement interprets it as manifesting the agreement between my faculty of desire and practical reason: what I have chosen to do is in agreement with what I ought to do.Footnote ³³ While this agreement manifests as a feeling of pleasure, it is judgement that enables my awareness of it as the kind of feeling that it is (i.e. moral contentment) rather than simply as pleasure. In this sense, the power of judgement is transcendentally necessary for feeling as the faculty that enables the subject to be affectively aware of its own state.

So to sum up, Eran and I share a lot more than he suggests in his paper. For I agree with him that Kantian feelings do have a form of intrinsic aboutness. For us to become aware of them as the kinds of feelings that they are, however, requires the exercise of judgement guided by its principle.