2. The Distinction between Rational and Natural Sympathy

Kant draws an explicit distinction between what I am calling rational and natural sympathy in at least five passages, ranging over at least twenty years.Footnote ² His descriptions of the distinction involve variegated terminology, but I think we can discern the same distinction in all five places.Footnote ³ It is only Kant’s elision of this distinction in the Groundwork which produces the apparent conflict we saw above. I will label the rational side of the distinction with ‘(a)’ and the natural side with ‘(b)’.

Friedländer’s 1775–6 Anthropology notes draw a distinction between (a) ‘reason’s sympathy’ (Antheil der Vernunft; Anth-F, 25: 610) and (b) ‘physical sympathy’ (physicalischen Sympathie; 25: 607). Reason’s sympathy is ‘in accordance with ideas’ (25: 607). Physical sympathy is ‘based not on deliberation, but on animality’ (25: 607): ‘as soon as I am not the master of it, but am placed in it against my will, then it is an affect’ (25: 611). Friedländer’s notes define affect as ‘[t]hat degree of sensation that makes us unable to estimate and compare the object with the sum total of all our sensation’: one example is ‘joy … if one is pleased with an object which has no noticeable influence on the whole of our well-being’; another is ‘if one becomes angry about a dish having been broken in two’, which Kant says likewise has no noticeable impact on our well-being as a whole (25: 589). This suggests that sympathetic affect can cause us to arbitrarily focus on particular features of others’ experience in ways that exaggerate their impact on their happiness as a whole and diminish the accuracy of our sympathy. Kant also comments on the problem that sympathy (Antheil) can ‘rise into an affect, or rather degenerate into it’ at MM, 6: 409. There he describes affects as feelings which make ‘reflection … impossible or more difficult’ (6: 407) and dispose us to act in ways that we ought not to act.

The same or perhaps a somewhat more general distinction, which includes our distinction as one of its main species, is drawn in Anthropology from a Pragmatic Point of View (1798) and Mrongovius’ Anthropology notes (1784–5, demonstrating that Kant was lecturing about the distinction either shortly before or simultaneously with the publication of the Groundwork in 1785). Kant uses the terms (a) ‘sensitivity’ (Empfindsamkeit) and (b) ‘sentiment’ or ‘sentimentality’ (Empfindelei) (Anth-Mr, 25: 1320–1; Anth, 7: 235–6), which is also called ‘touchiness’ (Empfindlichkeit) at Anth-Mr, 25: 1320. According to Mrongovius, sensitivity ‘is the faculty of being able to have a sensation of the agreeable and disagreeable’. It is a ‘strength’ which allows us to ‘choose for others what they will enjoy’ and ‘does not come from the senses, but from concepts’. Sentimentality, on the other hand, is a ‘weakness’ that makes it possible to be ‘easily carried away by every sensation’ and prevents ‘rational reflection’ (Anth-Mr, 25: 1320–1). According to the Anthropology, ‘Sensitivity … is a faculty and a power which either permits or prevents both the state of pleasure as well as displeasure from entering the mind, and thus it possesses choice’ (Anth, 7: 235–6). The Anthropology explains the idea of choosing for others what they will enjoy in more detail (though regrettably with misogyny): ‘Sensitivity is manly; for the man who wants to spare his wife or children difficulties or pain must possess such delicate feeling as is necessary in order to judge their sensation not by his own strength but rather by their weakness’ (7: 236). The connection between the concepts of sensitivity/sentimentality and sympathy is made explicit when Kant says that sentimentality ‘is a weakness by which we can be affected, even against our will, by sympathy (Theilnehmung) for others’ condition who, so to speak, can play at will on the organ of the sentimentalist’ (7: 236).

The most famous passage addressing the distinction is Doctrine of Virtue §§34–5, titled ‘Sympathetic Feeling is Generally a Duty’ in Gregor’s translation (MM, 6: 456–8, also cited above). Gregor’s translation of this section has received criticism from commentators who think she errs in translating as ‘sympathy’ terms which should be translated differently, and a brief discussion of some of this controversy will help explain my motivation for providing an account of how we rationally sympathize. All commentators agree that Kant is drawing a distinction between (a) humanitas practica, to which we have an obligation, and (b) humanitas aesthetica, to which we have no obligation. According to Gregor’s translation, humanitas practica is the ‘capacity and the will to share in others’ feelings’, which is ‘free’, and based on ‘practical reason’, while humanitas aesthetica is ‘unfree’ – it is ‘the receptivity, given by nature itself, to the feeling of joy and sadness in common with others’, which ‘can be called communicable … like receptivity to warmth or contagious diseases … since it spreads naturally’.

Rudolf Makkreel emphasizes the crucial distinction between passive and active feeling which I too emphasize, and his view is thus an important antecedent of the view I advance here (Makkreel Reference Makkreel2012: 111–12). He points out that the term in the section title which Gregor translates as ‘sympathetic feeling’, theilnehmende Empfindung, can be translated as ‘participatory feeling’ (111). He argues that his translation is preferable, because he thinks that if we begin the section with the title ‘Sympathetic Feeling is Generally a Duty’, as Gregor does, we start by claiming we have a duty to experience feelings that sound like humanitas aesthetica, which Kant goes on claim we do not have a duty to experience, rather than humanitas practica, with respect to which we do have a duty (ibid.).

There is, however, strong support for Gregor’s translation in the Friedländer Anthropology lecture notes, which include the following passage:

The key word übersetzt here is also rendered as ‘translated’ elsewhere in the Cambridge translation (CPrR, 5: 60; MM, 6: 237). So this passage amounts to Kant’s own translation advice on this controversy: he tells us to translate Sympathie as Theilnehmung. I think this provides strong evidence that Kant sees no difference between the feelings to which these terms refer. I will therefore translate Theilnehmung as ‘sympathy’ frequently below, though I will sometimes use ‘participation’ because of the way it evokes the idea of perspective-taking, which (as I will explain below) is crucial to understanding sympathy.

However, Makkreel’s motivation for wanting a translation which distinguishes Sympathie and Theilnehmung is to address a problem about how we are to understand Kant’s contrast between activity and passivity in Kant’s theory of sympathy, which remains to be resolved even after we adopt Kant’s translation advice. Makkreel sees sympathy as an inherently passive feeling, and theilnehmende Empfindung as a fundamentally different feeling which is ‘a more active counterpart to sympathy’ and is ‘not passively received, but a spontaneous expression of “practical humanity”’ (Makkreel Reference Makkreel2012: 111), which has the ‘moral import to cultivate a “love of human beings”’ (112). By contrast, according to my interpretation, the passive and the active kinds of feeling are both kinds of sympathy. However, in view of the remarks we have seen in which Kant associates sympathy with passivity, it would be unreasonable to simply stipulate that there is an active kind too. An account of what it means for it to be active is also required. Makkreel’s strategy for resolving this problem is to argue that the active feeling is not sympathy at all.

Melissa Seymour Fahmy’s interpretation is also an important antecedent for my account, and she resolves this problem about activity and passivity in yet another way. I will follow some of the threads in her interpretation to get to the path to the alternative resolution I want to propose here. Fahmy highlights Kant’s original German description of humanitas practica:

Fahmy argues that to accurately translate this, we should not give Gregor’s version, ‘humanity can be located … in the capacity and the will to share in others’ feelings’, but instead ‘This [humanity] can be located in the capacity and will to communicate with each other in view of (with respect to) one’s feelings’. She concludes that the duty of humanitas practica is not a duty to have or share feelings, but rather to communicate about whatever sympathetic feelings we have (Fahmy Reference Fahmy2009: 35). She acknowledges Kant’s clear reference to a duty to ‘cultivate the compassionate natural … feelings in us’ (MM, 6: 457), but interprets this as a duty to strive to have these feelings, which we can fulfil even if we do not succeed (Fahmy Reference Fahmy2010: 321–2; 2019: 418–19).

Fahmy’s interpretation is textually well-grounded if we focus just on MM, §§34–5. But we often talk of communication and feeling without the prepositional mediation of expressions like in Ansehung (with respect to) – communicating feelings can mean that one person conveys feelings to another in such a way that the other experiences the feelings too. In §34 Kant is clearly using this unmediated sense of ‘communication’ in characterizing humanitas aesthetica (the natural side of the distinction) – he says it ‘can be called communicable (mittheilend) … like receptivity to warmth’. In my view, despite the presence of in Ansehung in the description of humanitas practica, we cannot be sure that Kant means to rule out the unmediated sense of communication there – he may think that the rational side of the distinction involves communication about feelings which prompts communication of feelings.

Evidence that this is indeed what he thinks is provided in a discussion of friendship and communication in Vigilantius’ Ethics notes (1793–4) from just a few years before the Metaphysics of Morals, which I take to distinguish (a) ‘moral’ sympathy and (b) ‘instinctual’ sympathy:

In this passage, we have three references to communication of feelings without any prepositional mediation like in Ansehung, all of which are clearly meant to characterize the moral side of the distinction. While Kant also emphasizes the communication of thoughts, it is clear that the purpose of communicating thoughts here is to accurately convey feelings to another, so that the other can have the feelings too, and by virtue of having them, express moral sympathy.

The conclusion I draw is that, here too, we see that Kant’s lecture notes provide support for Gregor’s translation, and so we should think that he really is telling us that sympathetic feeling is generally a duty (MM, 6: 456), and that fulfilling it involves the capacity and will to share in others’ feelings (MM, 6: 456). However, as noted above, embracing the idea that morality requires us to have sympathetic feelings, as I aim to do here, poses its own interpretative challenge. In what sense can we actively or spontaneously prompt sympathetic feelings in ourselves?

As Makkreel and Fahmy both emphasize, Kant claims that we cannot will ourselves to have feelings of love: ‘Love is a matter of feeling, not of willing, and I cannot love because I will to’ (MM, 6: 401; Fahmy Reference Fahmy2010: 313; Makkreel Reference Makkreel2012: 114). It is natural to think of sympathy as a kind or aspect of love, so this claim about love presumably extends to sympathy. But Kant also holds that we can come to love people through the practice of acting in beneficent ways: ‘Beneficence is a duty. If someone practices (ausübt) it often … he eventually comes actually to love the person he has helped’ (MM, 6: 402). According to the Collins Ethics notes, ‘if I love others from obligation … by practice (Uebung) it becomes love from inclination’ (Eth-C, 27: 419).

The connection between the concepts of action, practice and the prompting of feeling in these passages makes it noteworthy that Kant’s account of rational sympathy also includes the notions of action and practice. I interpret the ability to rationally sympathize as a skill in performing mental actions which prompt feelings, and Kant implies that rational sympathy is a skill that we can acquire by practice:

If we can acquire a skill, then it is deontically consistent for reason to tell us that we ought to acquire it, and if we can acquire it by practice, it is consistent for reason to tell us that we ought to practise until we acquire it. When we have acquired a skill, we have volitional control over the task we are skilled in. This passage implies that rational sympathy is such a skill, and that when we have acquired it, we can volitionally (mentally) act in ways that prompt sympathetic feeling.

As I will discuss in greater detail below, the ‘power to transpose the I’ (ibid.) is a skill of imagination, and the locus of activity (volitional control) in rational sympathy is in imagining, not in feeling. That is, the voluntary mental actions which occur in the process of rational sympathy are actions of imaginatively adopting another’s subjective viewpoint on the world, and imagining content to furnish that viewpoint. We acquire the sympathetic feelings themselves in response to these reason-guided imaginings, not by willing the feelings to spontaneously spring forth. In this way, Makkreel’s point about the passivity of sympathetic feeling is, in an important sense, preserved in my account.

Before I explain how this skill works in more detail, let me comment on why I think reason requires it, that is, on why morality requires sympathy. At MM, 6: 399, Kant lists a set of feelings which include moral feeling, respect, conscience and love, which he claims are ‘conditions of receptivity^t (Empfänglichkeit)’ to the concept of duty. Sympathy is an aspect of love, and I take Kant’s reference to Empfänglichkeit at MM, 6: 456, to imply that sympathy is included in that set of feelings. I think rational sympathy is required for receptivity to others’ permissible ends, and that we need such receptivity if we are to adopt such ends as our own, which we must do to fulfil the duty of beneficence (G, 4: 430; MM, 6: 450). (Some formulations of the duty of beneficence refer to others’ happiness (MM, 6: 452; also see CPrR, 5: 34), and happiness is an end (G, 4: 396; CPJ, 5: 437; MM, 6: 388) which is the ‘sum’ of our other, more particular permissible ends (CPJ, 5: 531; TP, 8: 282–3).) We need rational sympathy for receptivity to others’ ends because receptivity to an end requires a desire to pursue it, and desire requires feeling (G, 4: 460; CPrR, 5: 9n.; MM, 6: 212–13), not merely as a psychological force, but as an incentive which can be incorporated as a motivational ground. Ends are individuated by concepts which motivationally engage with agents’ feelings (CPJ, 5: 220; MM, 6: 384–5). Concepts with marks of law (for example, obligation for perfect duty, and merit for imperfect duty) engage the feelings of all rational agents (that is just what it is to be a rational agent), and such concepts individuate objective ends (G, 4: 427). But concepts without marks of law only engage agents with sensibilities of a specific contingent constitution, and thus individuate subjective ends (4: 428). Agents’ concepts of their permissible ends are often of the latter sort. I can only adopt those same ends under the guidance of practical reason if I can be voluntarily moved by the concepts that move them, which I can only do if I can voluntarily conform my contingent feelings to theirs, and this is the role of rational sympathy. I can promote others’ permissible ends by taking them as means to further ends, for example, improving my reputation or fulfilling my imperfect duty of beneficence. In both these cases I pursue a different end than the other’s: in the first case, an end individuated by a concept with a mark of reputation; in the second, with a mark of merit (see MM, 6: 388, for an argument which I think implies this). It may well be worthwhile to promote others’ ends as means, but there is a distinctive non-consequentialist value involved in adopting them as ends, not out of desire for reputation or even my own moral excellence, and I think rational sympathy is required for this. More argument is required to defend this view of why morality requires sympathy, which I cannot offer here due to constraints of space. But the evidence for thinking that the distinction between natural and rational sympathy is important should make scholars interested in an account of how rational sympathy works even if they are not persuaded by this sketch of why morality requires it. The goal of this paper is to focus on the how rather than the why.

3. Sympathy and the Imagination

Kant repeatedly connects sympathy with the imagination, and here too his view is clear despite varying terminology. He says that sympathy is ‘an effect of imagination’ (MM, 6: 321n.; also see 6: 457), and he refers to the ‘sympathetic power of imagination’ (Anth, 7: 179; also see 7: 238). If sympathizing is a way of imagining, then the difference between rational and natural sympathy must be a difference between two ways of imagining.

Imagination is a fundamental power in Kant’s theory of mind, one of two ‘parts’ of sensibility, the other of which is ‘sense’ (Anth, 7: 153). Sense is the ‘faculty of intuition in the presence of an object’, while imagination is ‘intuition even without the presence of an object’ (Anth, 7: 153; also see B151). Another key function of imagination is to make connections which are different from the logical relations between concepts, but which are nonetheless necessary for experience.

In the first Critique’s Transcendental Deduction, an aspect of imagination which Kant calls a priori productive imagination plays a role in spontaneously structuring sensibility as a part of the transcendental synthesis which makes a priori cognition possible (see e.g. B151–2; Anth, 7: 167, 174). In this role, the imagination is responsible for what Kant calls the figurative synthesis of the manifold of sensible intuition (B151–4), one aspect of which is the establishment of a priori connections among times to form the schemata necessary for the application of a priori concepts of the understanding, that is, the categories of the understanding (A142–5/B182–5).

The imagination plays a crucial role a posteriori as well, including providing schemata for empirical rather than a priori concepts (B179–81/A140–2) and playing a crucial role in the generation of empirical concepts (a point to which I will return below). Its activities make up much of the mental content of our waking and dream lives. Passively received sensible content is due to sense, and transitions between mental states are the work of the understanding or reason insofar as they are guided by active reasoning, but Kant thinks that the rest of the contents of intuition and transitions between mental states are the work of the imagination.

Even when reasoning is involved in the functioning of the imagination, Kant seems to think that the role of reasoning is merely to direct the imagination. Friedländer’s Anthropology notes say that ‘[t]he power of choice can only do something insofar as it gives direction to the imagination and then it straightaway runs according to its new direction, like water in the stream’ (Anth-F, 25: 515), and that when the imagination is not ‘subjugated to the power of choice’ it is ‘often the path of many vices’ (ibid.). Mrongovius’ notes say that, if we do not choose the direction of the imagination, ‘[t]he imagination directs itself according to the inclinations’ – for example, ‘[i]f one feels hatred, then the imagination shows everything from its most detestable side’ (Anth-Mr, 25: 1260). The Collins Ethics notes relate our ability to choose the direction of the imagination to a power compared to a kind of ‘monarchy’:

These ideas, along with the points about the hazards of affect in natural sympathy which we saw above, make it reasonable to think of natural sympathy as an aspect of the incessant involuntary churn of the imagination. The idea that it is constantly at work in us in a way that is not inherently subject to our will could help explain Kant’s thought above that it is ‘like receptivity to warmth or contagious diseases’ (MM, 6: 409). We are not telepaths, so feeling cannot be directly conveyed from the mind of the other; however, if sympathy is part of the constant involuntary roving of our imaginations, it would make sense to think it would have a phenomenology such that sympathetic feelings seep into our sensibility unbidden, like sensations of temperature or disease symptoms.Footnote ⁶ This suggests that we can in turn understand rational sympathy as the voluntary guiding of the imaginative associations we make while sympathizing, so that we avoid affects that make it difficult to think clearly, and sympathies that dispose us to act wrongly.Footnote ⁷

Contrary to what readers familiar only with the discussion of imagination in the first Critique might gather, Kant distinguishes between two kinds of productive imagination: an a priori and an a posteriori kind. The first Critique emphasizes a priori productive imagination, which (as mentioned above) is responsible for the transcendental figurative synthesis of the sensible manifold (B151–4). However, a posteriori imagination has both reproductive and productive faculties, and both are important in understanding sympathy. Reproductive imagination, which is exclusively a posteriori, ‘brings back to the mind an empirical intuition that it had previously’ (Anth, 7: 167; also see B152, Anth-F, 25: 512 and Anth-Mr, 25: 1257). A posteriori productive imagination possesses a kind of spontaneity which reproductive imagination lacks, and which is different from the kind of spontaneity possessed by a priori productive imagination. Kant says a posteriori imagination is ‘inventive’, though ‘not exactly creative’ (Anth, 7: 168) – while it does not merely bring previous empirical intuition back to the mind, in the way reproductive imagination does, ‘it is not capable of producing a sense representation that was never given to our faculty of sense. One can always furnish evidence of the material of its ideas’ (Anth, 7: 168). It is nonetheless ‘very powerful in creating, as it were, another nature, out of the material which the real one gives it’ (CPJ, 5: 314).Footnote ⁸ The fact that a posteriori productive imagination must draw on representations previously given to sense may suggest that it is a kind of spontaneity which is (as it were) of a lower order than the spontaneity of a priori productive imagination, since the latter involves determining the spatiality and temporality of things in sensibility which in themselves have no temporality or spatiality, and thus arguably adding a more fundamentally novel contribution to human experience than is added by a posteriori productive imagination. On the other hand, there is no scope in the activity of the a priori productive imagination for the distinctive kind of spontaneity which is transcendentally free moral action, and on my view, the a posteriori productive imagination involves that kind of spontaneity when it is guided by practical reason, as is the case in rational sympathy.Footnote ⁹ To keep things concise, ‘productive imagination’ is used henceforth just to refer to a posteriori productive imagination.

Kant’s description of the distinction between reproductive and productive imagination requires us to assume that sympathy involves the productive imagination (though as we will shortly see the text directly supports this as well). Representing some of our feelings as shared or like another’s requires us to represent them as something more than mere recapitulations of our own experiences. However, since we are not telepaths, we can only aim at having feelings like another’s by drawing on our own previous experiences in creative ways. But as we will shortly see in more detail, sympathy requires the functions of reproductive imagination too.

Productive imagination can function both involuntarily and voluntarily (Anth, 7: 174; Anth-Mr, 25: 1257). Involuntary productive imagination is called ‘fantasy’ (Phantasie) (Anth, 7: 167, 175; also see Anth-Mr, 25: 1258; Met-Mr, 29: 884–5), and in connection with this term Kant makes an explicit connection to sympathy:

The term translated here as ‘oversensitivity’ (Empfindelei) is the same as the term we saw translated as ‘sentimentality’ earlier. Since we saw earlier that the kind of sympathy involved in Empfindelei is natural sympathy, the connection Kant makes at CPJ, 5: 273, between Empfindelei, Theilnehmung and Phantasie implies that natural sympathy is a species of fantasy, and is thus a species of involuntary productive imagination.

We can maintain the symmetry in the distinction between rational and natural sympathy in the texts considered thus far if we suppose that rational sympathy is a species of the rational counterpart to fantasy, which Kant calls ‘disciplined fantasy <phantasia subacta>’ at Met-Mr, 29: 885.Footnote ¹⁰ This makes rational sympathy a species of voluntary productive imagination. It may give us pause to see that Kant appears to use ‘disciplined fantasy’ as equivalent to terms he uses for the productive imagination of artists – that is, ‘fabrication’ (Anth, 7: 175, Erfindung; Met-Mr, 29: 885, Erdichtung), and ‘composition (Composition)’ (Anth, 7: 175). However, as we will see in more detail later, Kant thinks that art and sympathy involve the imagination in similar ways.

Productive imagination must also draw on the powers of reproductive imagination. Reproductive imagination is entirely governed by what Kant calls the ‘law’ (or sometimes ‘laws’) of ‘association’, and the productive imagination is not (CPJ, 5: 240, 269, 314; Anth-F, 25: 512; Anth-Mr, 25: 1272). The law of association is that ‘empirical ideas that have frequently followed one another produce a habit in the mind such that when one idea is produced, the other also comes into being’ (Anth, 7: 176). Through this law, ‘ideas that were often connected with present ones … are produced’ (Anth-Mr, 25: 1273). The point that the productive imagination can organize itself according to this law is implicitly established by the fact that Kant states the law of association in the Anthropology in a section titled ‘On Sensibility’s Productive Faculty of Association’ (7: 176).

Let us approach the question of how the law of association functions in productive imagination by first considering how it functions in the reproductive imagination. This occurs when the current, passively received contents of one’s consciousness involuntarily prompt recapitulations of previous intuition:

These habitual associations constitute the foundation in empirical psychology of our capacities to inductively generate new beliefs about empirical laws. Kant famously claims Hume goes wrong in thinking that these associations generate our concept of cause as well – he thinks Hume fails to realize that the possibility of temporal order itself depends upon a transcendentally prior pure synthesis of imagination (A766–7/B794–5; P, 4: 257–8). Reproductive imagination and its law of association is not limited to temporal order. Kant says it also associates intuitions based on ‘contiguity’ (‘unity of place’: Anth-F, 25: 513), which creates the subjective unity of space which, along with the subjective unity of time, subjectively unifies the empirical form of intuition. Last but not least, it associates intuition based on ‘similarity’ (Anth-F, 25: 513) or ‘affinity’ (A766/B794), and this kind of association underwrites our empirical-psychological capacity to generate empirical concepts, as well as our ability to call up imagined intuitive content to accompany concepts.

The associations made by the reproductive imagination produce subjective ‘unity of given representations’ which we then incorporate into ‘objective unity’ by testing them with judgements according to the categories of the understanding, to ensure that ‘representations are combined in the object … regardless of any difference in the condition of the subject’ (B141–2). The subjective unity produced by the reproductive imagination and the objective unity synthesized according to the categories differ in that the subjective unity is unified from the first-person standpoint of the particular human subject who does the unifying, while the objective unity is one which would be cognized by any human subject who synthesizes experience according to the categories.

As noted above, the productive imagination is not entirely governed by the law of association in the way the reproductive imagination is. The productive imagination’s freedom from the law of association is both hazardous and valuable. It can lead not only to inadequately disciplined fantasy which prompts affect, as it does in natural sympathy, but also to a ‘ruleless fantasy’ which ‘approaches madness, where fantasy plays completely with the human being and the unfortunate victim has no control at all over the course of his representations’ – its inventions cannot ‘find their place in a possible world’, ‘because they are self-contradictory’ (Anth, 7: 181). On the other hand, freedom from the law of association also allows disciplined fantasy, one species of which we have identified as rational sympathy. In disciplined fantasy, the imagination is not inherently governed by the law of association, but it must still regulate itself according to laws for its productions to be possible in imaginary worlds. On this point it is helpful to consider at greater length a passage quoted earlier:

When Kant says productive imagination allows us to ‘transform the material which the real’ nature gives us according to ‘analogous laws’, it is not immediately clear whether the analogous laws he has in mind are analogous laws of nature, or of association – he refers to nature (though not natural laws) in the previous sentence, and he refers to ‘the law of association’ later in the same sentence. It is plausible to think that laws analogous to both of these play a role in productive imagination’s transformation of material from real nature into another one. The context for this passage is a discussion of art. It makes sense to think disciplined fantasy would produce fictional worlds of literature with natural laws different from, but analogous to, our own, though presumably not too different: fairy tales are salient examples, and Kant objects to them because they strain children’s imagination, so he seems to regard them as examples of insufficiently disciplined fantasy (Ped, 9: 476; also see OFBS, 2: 214). Fictional worlds of literature typically differ most from the actual world in the invention of fictional initiating events, from which the plot proceeds according to laws not discernibly different from the actual laws. However, when we draw on disciplined fantasy for the purpose of sympathizing with actual people, we are presumably required to imagine their lives as governed by the laws of the actual world, so the notion of analogous natural laws seems to add no useful detail to our picture.

It is plausible to think that a disciplined fictional world must also provide, through the perspective of characters within it, an imagined subjective unity like that which the reproductive imagination creates when it follows the law of association. Just as we can build a fictional world by creatively imagining fictional events from which the story proceeds according to laws much like those of our own world, we can build a fictional world by imaginatively stepping into the first-person standpoint of a fictional character, and making associations according to principles of association much like those we apply from our own first-person perspectives, as we empirically synthesize our own subjective unities. This activity of productive imagination makes sense to posit not only in relation to fictional characters, but also in relation to actual people with whom we sympathize.

4. Putting Ourselves in Others’ Places

Though Kant never explicitly connects sympathy with the subjective synthesis of the imagination, there is substantial evidence to think it must be his view. He remarks in many places that sympathy is the power to imaginatively put ourselves ‘in the other’s place’ (Anth-F, 25: 575; for similar language referring to sympathy and others’ places, see MM, 6: 321n.; Eth-H, 27: 58, 65; Anth-F, 25: 607).Footnote ¹¹ The reproductive imagination is what provides the subjective unity which puts me in my own place, so it is reasonable to suppose that it is the reproductive imagination’s capacities placed in the harness of productive imagination which allow me to put myself in another’s place. Kant claims that it is this imaginative projection which allows Theilnehmung, the sympathy/participation in others’ lives discussed above through which we share their feelings (OFBS, 2: 215; MM, 6: 457, 469; Anth-F, 25: 476, 606; Eth-V, 27: 685–6), and which finds its most complete form in the ideal of friendship, an ‘ideal of each sympathizing with and communicating about the other’s wellbeing^t (Ideal der Theilnehmung und Mittheilung an dem Wohl eines jeden)’ (MM, 6: 469; also see Eth-V, 27: 677–8, quoted above) which guides us to strive toward a ‘maximum’ (MM, 6: 469) of sympathy in which ‘each mutually shares in every situation of the other, as if it were encountered by himself’ (Eth-V, 27: 677). Putting ourselves in others’ places is necessary but not sufficient for prompting sympathetic feelings. We can ‘put ourselves in the position of another’ in a merely ‘logical’, ‘heuristic’ way, for example ‘a follower of Crusius’, to ‘get better at certain things’ (Eth-H, 27: 58) such as understanding the structure of another’s philosophical views. We can also put ourselves into a universal position: when following the second ‘maxim of the common human understanding’, one ‘think[s] in the position of everyone else’ in order to reflect on his ‘own judgement from a universal standpoint … which he can only determine by putting himself into the standpoint of others’ (CPJ, 5: 294–5; also see Anth, 7: 228). Neither of these necessarily prompts sympathy. But when we put ourselves in another’s place for the purpose of sympathetic participation, ‘we really feel ourselves to be in his place’ (Eth-H, 27: 58), and ‘[w]e are sensible of this sympathizing feeling in our entire soul’ (Anth-F, 25: 606).

It is also noteworthy that, in what may be Kant’s most detailed remark on the phenomenon, he says that it occurs with reference to both fictional and actual people, just as proposed above:

The contrast Kant makes here between the world’s foci and its centre is inevitably expressed in pre-critical language at this time (1775–6), but it is recognizable as a progenitor of the contrast between subjective and objective unities of experience.

Two passages about sympathy and social subordination in Friedländer’s Anthropology notes provide helpful detail about how voluntarily putting ourselves in others’ places moulds our sympathetic feelings. One is about expanding the range of our feelings to sympathize more accurately with how the other feels. Kant says that ‘a humble person can easily put himself in the position of the higher one and assume greater dispositions. However, the distinguished one cannot assume the state of the humble one, hence he also does not sympathize (sympathesirt) with his misfortune’ (Anth-F, 25: 607). ‘If the ills are natural, for example, famine, then the distinguished person sympathizes with the humble one just as well as the latter with him, but in the case of … ideal ills, the distinguished one does not sympathize (sympathesirt) with the humble one, but the latter does in fact sympathize (sympathesirt) with the former’ (25: 606–7). The distinguished one ‘thinks that the one who is thus not accustomed to the refined life is indeed just a humble man, hence he always gets on [in life], if he can just live’, and does ‘not become as aware’ of the ‘distance’ of the humble man’s ‘social standing from the civic one in general’ (25: 607). Kant says that while a commoner ‘has compassion (Mitleiden) for an unfortunate king’, the ‘unfortunate thing with kings’ is that they ‘have no inclination’ to ‘imagine the misfortune of their subjects’ (25: 607).

As we saw earlier, when the imagination roves involuntarily, it is guided by inclination, and (ironically, given the passage currently under consideration) we must exercise the power of ‘monarchy’ to voluntarily direct it. Kant’s implicit point in this discussion is that when the ‘distinguished’ sympathize naturally, their inclinations may dispose them to imagine what it is like for the ‘humble’ to be hungry or in pain, but not to imagine their ‘ideal’ misfortunes – in particular, they do not imagine that the ‘humble’ have ideas of happiness which include more than just living, and resent the way their social standing makes it hard to do more than just live – and that the ‘distinguished’ should resist their inclinations, and sympathize rationally, and put themselves in the others’ place more accurately, in a way that brings them a greater range of sympathetic feelings.

A second passage on this theme appears nearby, where Kant argues:

The frequency with which the oppressed lose the idea of the right of humanity is certainly debatable, and we must be cautious about assuming that social structures which people do not actually resent are really oppressive. But if it is clear that a society is oppressive, then we should sympathize with people who manage to live under that oppression without occurrent resentment by projecting ourselves into a version of their position inflected by the idea of right. So while our primary task in rational sympathy is to be accurate to the other’s actual feelings, we should in some cases sympathize with their actual feelings as well as the feelings they would have if their social circumstances were adequate to their rational nature, by adjusting our sympathies in light of ideas of reason. In this way we can access feelings on their behalf, which we may be able to help them experience if we can do it without paternalism.

Associating intuitive contents we imagine others to have in terms of their imagined space, time and similarity relations allows us to empirically knit together and occupy an imaginary first-person perspective for the other. But this only gives us a theory of how we produce the imagined form of the other’s perspective. It does not give us a theory of how we produce the imagined intuitive content which we are meant to have under voluntary control in rational sympathy, such as the additional imagined content that the kings discussed above ought to have. How would rationally sympathetic versions of Kant’s kings perform better? Kant’s answer, already indicated above, is to communicate better (Eth-V, 27: 677), and this requires us to draw on our ability to regulate our imagination by controlling our associations, both in communicating about how we feel and understanding others’ communications about their feelings.

5. Communicating Feeling in the Third Critique

We can discern Kant’s views on such communication by pursuing the connection we observed between sympathy and art. Both sympathy and art require us to communicate feelings, and both involve imaginatively putting ourselves in another person’s place, whether it is a real or fictional person. All Kant’s references to sympathy imply that we must put ourselves in others’ places. The Friedländer passage cited above states that we must do this ‘when we read something’ (Anth-F, 25: 476), which suggests that we must do this whenever we read literature. The third Critique makes no general claim that all art or all reading involves putting ourselves in characters’ places, but it does state that imaginative projection is a way to engage art:

Kant thinks that concepts communicated by the poet to the reader prompt the reader to put herself in the place of a character and imagine that concept applying to her, and that this sparks associations in the reader’s imagination which prompt feelings. If we assume that we can use the same powers of the imagination in sympathy that we use in the kind of case Kant mentions in this passage, then further details about how the imagination works in this kind of case are also applicable to the sympathetic imagination.

This passage is one of two examples Kant offers of ‘aesthetic ideas’. Aesthetic ideas are central in his account of communicating feeling through artistic language. An aesthetic idea is a certain sort of ‘representation of the imagination associated with a given concept’ (CPJ, 5: 316). It is not clear whether aesthetic ideas include feelings, or we respond to aesthetic ideas by having feelings. But it is clear that when one successfully expresses an aesthetic idea by means of expressing the associated concept in language, one is able to prompt another person to have the same feelings one has in connection with the aesthetic idea. The capacity to form aesthetic ideas is ‘genius’, and the capacity to express them is called ‘spirit’ (5: 317), and while the artists that we refer to as geniuses are ‘exemplary’ (5: 318) in their exercise of these capacities, these capacities are ‘really only a talent (of the imagination)’ (5: 314) which we all have in one degree or another. What Kant is describing in the ‘beautiful morning’ passage is the expression of an aesthetic idea by way of concepts, which the recipient receives by imaginatively putting herself in the place of someone to whom those concepts apply, in a way that allows the recipient to have feelings that she takes to be like the feelings the imaginary person has.

In explaining how this works, Kant contrasts two uses of imagination, one ‘for cognition’ (CPJ, 5: 316), and another ‘through which the subjective disposition of the mind … can be communicated to others’ (5: 317), including ‘inner feeling’ (5: 296), which Kant regards as the only purely subjective aspect of sensible content. This latter use is the one at issue in aesthetic ideas. When we use the imagination for cognition, ‘the imagination is under the constraint of the understanding and is subject to the limitation of being adequate to its concept’ (5: 316–17); concepts and the intuitions provided by imagination ‘flow together into a cognition’ (5: 296) as we attend to the aspects of our subjective unity that can be synthesized into the objective unity, which Kant thinks does not include how we feel about things.Footnote ¹² The aspect of the imagination which assembles intuitions for synthesis into objectivity is reproductive imagination, as discussed above.

When we use the imagination for communication of aesthetic ideas, on the other hand, the imagination and understanding are related differently: ‘the imagination is free to provide, beyond … concord with the concept’, a ‘manifold of … representations in the free use of the imagination’ which the ‘imagination … associates with’ that concept. These voluntarily associated representations ‘belong to the concept’ but ‘aesthetically enlarge … the concept itself in an unbounded way’, and this ‘arouses a multitude of sensations and supplementary representations’ (CPJ, 5: 316–17) which prompt feelings. In the ‘beautiful morning’ example, the poet’s goal is to convey the ‘multitude of sublime and calming feelings’ she has when she thinks about virtue, and she does this by carefully making voluntary associations to select concepts she thinks will prompt the reader to imagine things that will cause her to have those same sublime and calming feelings. The reader understands the concepts and imagines sensory content on the basis of them, and has feelings in response to them which she represents as conveyances of someone else’s feelings by imaginatively putting herself in that person’s place.

Given Kant’s focus on art here, he is especially interested in simile. He appears to think that it is straightforwardly true of virtue that tranquillity streams from it in a way that is like the way sunlight streams forth – this is what he means in saying that this representation ‘belongs’ to the concept. But it is not part of the meaning of the concept – it is not something that we must have in mind to correctly use the concept, or that another person must have in mind to understand what we mean in saying ‘virtue’. It is merely something that the poet associates with the concept in her own imagination, choosing it with the goal of helping the reader imagine sensory content which prompts the multitude of sublime and calming feelings.

Something very similar can happen in conversations that aim at sympathy. There are important differences, but they make our picture of sympathy simpler than the picture Kant gives us of aesthetic ideas. Kant appears to think of poems on the same model we saw him use above for novels, that is, as fictions with which we engage by identifying with fictional characters, so he says readers acquire feelings like those of the poet by imaginatively putting themselves in the place of a fictional person. In communications aimed at creating or refining sympathy, the only perspectives we need to include in our model are the perspective of the person communicating about her feelings, and the perspective of the person seeking to sympathize. We do sometimes use similes in such conversations – the ‘humble’ person may say to the ‘distinguished’ person that anxiety over his children’s next meal is like being in a fog, or having a weight pressing upon him – but much of the basic structure Kant describes in his remarks about aesthetic ideas applies in just the same way to conversations meant to convey feelings which use language that does not include literary devices. Because a concept is necessarily general and abstract, there are infinite specific, concrete intuitions which ‘belong’ to that concept in the sense of being appropriately subsumable under it, any of which can be coherently associated with it by the imagination, but none of which belong to the meaning of the concept, in the sense that none of these particular intuitions are such that the imagination must associate them with the concept for the concept to be understood. Concepts are all that we immediately convey when we communicate in language, and so we have the freedom to associate an infinite variety of intuitions with what language conveys to us (this is part of the ‘free play’ of the imagination). Given the finitude of experienced time and of the number of associations we can actually make, however, the variety of intuitions we associate with the concepts conveyed to us in language is also finite, and since different associated intuitions prompt different feelings, it makes a difference to our feelings that we associate some intuitions rather than others.

The passages we saw earlier about the continuous functioning of imagination imply that our imaginations always follow along in any conversation, in at least an involuntary way, and associate imagined intuitive content with concepts. If natural sympathy is part and parcel of that constant involuntary work, then natural sympathy can arise at any time as a by-product of those conversations. This can be valuable if it means that we happen upon morally useful insights, and it can be problematic if it surprises us with affect.

But we can also use our capacity to associate intuitive content with concepts voluntarily. The ability to do this is really just the productive correlate of reproductive imagination’s capacity (discussed above) to associate intuitions according to ‘similarity’ in the use or generation of empirical concepts – in this case, it is the ability to (as it were) inventively ‘backtrack’ from a concept to intuitive content that embodies that concept in a way that is effective for the conversation at hand. We can converse for the purpose of sympathizing, and in such conversations, responding with feeling to the imagined intuitive content associated with the concepts is the point of the conversation. The communicator can voluntarily select the concepts she thinks are most likely to prompt imagined intuitive content in the sympathizer which will prompt feelings like hers, and on the other end of the conversation, the sympathizer must exercise discipline to associate imagined intuitive content she thinks likely to facilitate accurate and careful development in her of feelings like those the communicator seeks to convey. (This makes a ‘lawful business’ of the imagination’s ‘free play’, to borrow an idea from Kant’s discussion of the sublime (CPJ, 5: 269).) If the ‘humble’ person says to the ‘distinguished’ one, ‘I would like just one day when people like you spoke with me as if my opinions mattered’, the ‘distinguished’ person can discipline her imagination to make associations with remembered sensible content from experiences in her own life when her views were not respected, and transform them in imagination into an imagined life story in which few people ever respect her views, and thereby prompt relevant sympathetic feelings.

An important disanalogy between literature and sympathy is that Kant thinks we can know a priori that the feelings of a poet who writes beautiful poems are universally communicable, but we have no such knowledge regarding many of the feelings we seek to communicate through sympathy.Footnote ¹³ Kant thinks that the transcendental analysis of judgements of beauty shows that the pleasure we take in beauty is the result of judging that the contents of the imagination, even in its free play, are nonetheless harmonious with the understanding, in a way that makes this ‘relation suited to cognition in general’ (CPJ, 5: 217–18). It is a pleasure that results from judging that it is possible to bring aspects of our subjective unity to objectivity. Kant thinks that because cognition is universally communicable, pleasure in the possibility of cognition must be universally communicable too. But as mentioned earlier, Kant thinks of feeling in general as the most subjective aspect of experience, and he holds that it cannot be universally communicated (5: 213, 224). This is of a piece with his view that happiness is an entirely empirically based idea of the imagination, and there is no way to reason a priori about what will make anybody happy (G, 4: 418; CPrR, 5: 25). This does not imply that feelings are not communicable between particular individuals. But it does imply that our knowledge that our sympathetic feelings correspond accurately must be a posteriori – the only route to such knowledge would appear to be ongoing communication or other kinds of informative interaction.

A noteworthy exception to this disanalogy is the adjustment of our sympathies in light of ideas of reason, as we saw Kant thinks we must do in cases of people oppressed for so long that they no longer actively resent the injuries to their dignity. It is reasonable to assume that there are universally communicable feelings about dignity in the same way that Kant implies there are about virtue, as he indicates in the ‘beautiful morning’ passage. But we only have reason to make such adjustments in cases where the other lacks the feelings at issue, and it is acquiring feelings like other’s actual feelings which I take to be the vital moral core of sympathy. Our epistemic access to these is a posteriori as just noted, and very likely falls short of perfect certainty as a result. However, as mentioned earlier, the duty of sympathy subserves the duty of beneficence according to the present interpretation, and the duty of beneficence is imperfect, so the lack of perfect knowledge here is not an obstacle to my claim that morality requires sympathy for the fulfilment of the duty of beneficence.