2.1 Empathy, sympathy, compassion

In order to better situate what I mean by compassion, I propose in this section to compare and contrast it to empathy and sympathy. I acknowledge the large literatures on all these terms – for present purposes, what is important is to offer stipulative definitions of them that will allow me to carve out what I take to be distinctive about compassion.

By empathy, then, I propose to mean the ability and process of understanding what another person might be experiencing and the situation in which they are experiencing it. Defined this way, empathy is cognitive – crucially, it need not involve experiencing any emotion, of actual feeling. Of course, the (initial and continual) development of empathy as an ability surely requires the experience of emotion, but it is not necessarily the case that any one exercise of empathy is (on this definition) an affective state or process. Further, empathy does not necessarily involve – as I argue, below, sympathy does – evaluation. It is, rather, the bare process of understanding – making inferences – as to what another person may be experiencing.

Sympathy is related to empathy but, in the sense I use it here, goes one step further – that is, having attempted (as in empathy) some understanding of the situation of the other and their experience of it, sympathy involves an evaluation of that experience. Sympathy, then, is cognitively evaluative (empathy being only cognitive). This sense of sympathy is arguably close to the way Adam Smith uses the concept in his Theory of Moral Sentiments (Reference Smith and Haakonssen2002 [1759]), where the impartial spectator approves or disapproves of the emotions of the person being observed – asking, for instance, ‘would I, as an impartial spectator, also feel or have felt resentment in this situation?’, with consequences for one's moral judgment. Evaluation here clearly depends on cognition – one must first be capable of understanding the situation of the other person before one can evaluate how they experience it. Crucially, as with empathy, we can exercise sympathy without feeling any emotion (though, again, our ability to sympathise develops as a result of emotional experiences and reflections on them – as indeed Smith is at pains to show). Thus, both empathy and sympathy are cognitive processes, with sympathy also being evaluative, and neither necessarily involve feeling any emotion.

We come, finally, to compassion. I acknowledge that compassion is necessarily cognitive (like empathy) and evaluative (like sympathy). With respect to the evaluative dimension, however, one has to be careful: I will be suggesting that compassion is necessarily thinly evaluative, meaning that the other person is salient to us and that we have some concern for them. This thin evaluation is necessary, but thick evaluation is only contingent. By thick evaluation, I mean approving or disapproving of their feelings – for example, whether they are proportionate or deserving. But, despite the overlap with empathy and sympathy in these ways, compassion is also different from them both in that it necessarily involves feeling. In short, the crucial point is that there is something it feels like to experience compassion – whereas this is not necessarily so for empathy and sympathy.

It is notoriously difficult to articulate what is meant by ‘feeling’. I also acknowledge the debate in the literature on emotions as to whether emotions require feelings. Later, when I speak of the process of compassion, I refer to feeling some of the suffering one takes another to be in, and then feeling sadness for that suffering. These states may be experienced in different ways, but it seems to me arguable that they include: (1) some bodily changes (which are likely to differ across people), such as a certain heaviness in one's stomach when feeling some of the suffering one takes another to be in; and (2) a more immediate focus or concentration upon the object of one's emotion, which may include a greater readiness or willingness to perform some action with respect to that object. It is important for my argument that this action may refer to imagination – to (as I will articulate it later) the mental activity of constructing images of possibilities and alternatives – for I shall be arguing that, by feeling, we imagine more richly. Compassion, then, includes experiencing these states and, in this respect, compassion is unlike empathy and sympathy. The latter are more distant from the object and one's actions, less focused, less self-involving – and, again, crucially for my argument, less likely to stimulate the imagination.

2.2 Cognitively evaluative relational feelings

Emotions, for present purposes, are then cognitively evaluative feelings. By thinking of emotions in this way, I express solidarity with those theories that treat emotions as intelligent, as ‘concerned with receiving and processing information’ (Nussbaum, Reference Nussbaum2001, p. 23), and to some extent as judgments or appraisals of value. Compassion is an emotion – and it is thus also a cognitively evaluative feeling. However, it is a particular kind of emotion. To help see this, let me draw an analytical map of some of the different dimensions of emotions.

For a start, emotions can have three objects. Our emotions can be: (1) world-directed (as when we fear something in the environment, such as a tidal wave, a storm, etc.); (2) self-directed (as when, for instance, we experience self-pity or are scared of ourselves)Footnote ¹ ; and other-directed emotions (what I call ‘relational emotions’, of which compassion is one). The point of making such distinctions is that there may well be different cognitive and evaluative abilities that are exercised, or exercised in different ways, depending on the objects our emotions are directed to.

Relational (i.e. other-directed) emotions can be further broken down in various ways. First, one can distinguish between three different modes of interaction: first, one can experience relational emotions in a directly interactive (second-personal) way, such as in face-to-face interaction; second, one can experience them communally, as when, for instance, one feels safety in numbersFootnote ² ; and third, one can experience them spectatorially. Once again, there may be different abilities at stake in these different modes – for example, one must surely be quicker, make faster guesses, more fragile and tentative inferences in a direct interactive encounter than one can (usually) afford in the spectatorial mode. The spectatorial mode itself, however, can also be further broken down: arguably, it makes a difference whether one is actually observing someone and their situation (e.g. witnessing an accident on the street) – the ‘observational’ submode – or reading a report/description of it (as in a novel or a witness statement) – the ‘testimonial’ submode. For the purposes of this paper, I am interested in the spectatorial mode of relational emotions, and especially those in the testimonial submode, namely those that are experienced in the course of reading a report or hearing a description of someone else's experience. Thus, I acknowledge that compassion can be experienced in various modes of interaction, but I focus here on it as experienced in the spectatorial (testimonial) mode.

Further, one can also break down relational emotions into non-mediated and mediated ones. Thus, one can speak of the emotions we experience vis-à-vis others in a way that is not mediated by another emotion; for example, one can feel envious of another. But one can also experience a relational emotion that is first mediated by another one – for instance, in compassion, one arguably needs to experience some of the sufferingFootnote ³ one takes (knows or infers) the other to be in before one feels sadness for them. One can think of other examples too: schadenfreude or emotions closely related to compassion (such as pity). For the purposes of this paper, I am interested in the mediated relational emotions.

In sum, then, for the purposes of this paper, I focus on compassion as a mediated cognitively evaluative relational feeling, experienced in the spectatorial (testimonial) mode.

2.3 The process of compassion

Situated in the way I have above, it is important now to outline the process of compassion. The first point to make is precisely that it is not a state or episode, but a process – a temporally extended series of linked (mental and bodily) states. There is a tendency to sometimes speak of emotions as immediate, transitory and sudden – and of persons being swept up by them. This way of thinking about emotions tends to be connected to thinking of them as opposite to reason, as non-cognitive or non-evaluative, or as non-intelligent. I acknowledge that it is important not to lose sight of some emotional experiences being (somewhat) like this – but, on the whole, and in keeping with the above cognitively evaluative positioning of emotions, I am proposing here we think of emotional experiences as complex, multilayered, multistage processes, in which a mixture of abilities (cognitive, evaluative and at various levels of development and dexterity) is exercised. It is also important to emphasise that reflection can be part of an emotion, and that it arguably necessarily is part of some particularly complex emotions (especially those, as above, that are mediated).

Compassion, as a process then, can be thought of as – at the very least – containing the following five stages:

• 1 A thinly cognitively evaluative state of both awareness of another person (as being an other) one is interacting with, observing or imagining, accompanied by – and this is really inseparable from that awareness – an interest in or concern for that other person. Another person, then, must be both cognitively and evaluatively salient for one – they must be in the sphere of one's phenomenal (whether sensory or imaginative) reach and interest. This salience of and interest in the other can either be voluntary or involuntary (e.g. our attention and interest can be prompted by some loud or sudden noise, or we can make ourselves focus on some other without being prompted to do so).

• 2 Stimulated by this initial awareness and concern/interest, one then needs to exercise further cognitive effort, imagining the other's situation and perspective. In this second stage, however, one need not yet feel any emotion (aside from that initial interest in/concern for the person, as per stage 1). It is also part of this stage that one begins to take the other person to be suffering – as a result of imagining something of their situation and perspective, one cognises that they may be suffering.

• 3 Having imagined something of the other's situation and their perspective of it, and inferred that they may be suffering, in the third stage, one then feels something of the suffering that one has imagined the other person to be in. It is important to emphasise that one cannot assume that one is feeling the same as the other person is feeling – one is merely feeling something of what one imagines or takes the other to be suffering. One can be entirely mistaken about this – one may, for instance, have misread the expressions of another person, possibly as a result of cultural differences. It is, thus, an important part of the process of compassion that one is aware of the possibility that one has misinterpreted what state the other person is in. In that sense, self-reflexive humility is an important part of the process of compassion. Nevertheless, I take it to be key to the process of compassion that one actually feels some of the suffering that one has taken or imagined the other to be in.

• 4 Having felt something of the suffering one imagines or takes the other person to experience, one is then likely to return to imagining the situation of the other and their perspective on it. Indeed, there is likely to be a multiple back-and-forth between feeling some of what one takes to be the suffering of the other and imagining (increasingly more richly) their situation and perspective. The two processes are naturally difficult to disentangle (I am disentangling them somewhat artificially for analytical purposes). As I will go on elaborate, feeling some of the suffering one takes the other to be in stimulates one to imagine the situation from their perspective more richly, and imagining it more richly, may well stimulate one to feel more of the suffering one imagines them to be in.

• 5 Finally, having both imagined and felt some of the suffering of the other person, one then feels sadness for their suffering.Footnote ⁴ This is also a crucial step, without which there is no compassion. One has to be capable of stepping outside the perspective of the other, and experience the feeling (that one has oneself, as oneself) for another's suffering. The feeling of sadness for another's suffering may also result in one returning to imagine the situation and perspective of another more richly, and to better understand the suffering they may be in, thereby also intensifying the feeling of sadness for another.

These are the minimal bones of what is a multilayered, multi-stage and also back-and-forth dynamic process of the emotion of compassion.

To this basic structure, one may add several extra elements. Space will not permit me to discuss these in detail, so let me simply flag two of them. For some, most notably Nussbaum (Reference Nussbaum2001, pp. 311–315), there is necessarily a stronger evaluative element in compassion, namely the evaluation of the other's suffering as undeserved. I suggest that it is possible for me to experience sadness for the suffering I take another to be in without having evaluated them as undeserving. Yet more strongly, though also more rarely, it is also possible to experience that sadness even after having evaluated someone as deserving of suffering. I acknowledge that this requires effort and may be a question of skill – one may need to train oneself to be able to experience sadness for the suffering of those whom one does evaluate as undeserving of it, or that one does evaluate the other as deserving of that suffering. Indeed, this may be a crucial skill (let us call it the skill of overcoming strongly evaluative compassionate resistance) in many circumstances in which one is given the task of evaluating the conduct of a wide variety of persons (as judges in legal institutions are). For these reasons, I would consider this stronger evaluative element a contingent one. I would also regard as contingent Nussbaum's other condition, namely that the suffering for the other be serious rather than trivial (see Nussbaum, Reference Nussbaum2001, pp. 306–311). Treating the other's suffering as serious may be part of some experiences of compassion, but I also think it is possible to experience compassion for what may initially strike one as trivial, but which one comes to revise (precisely as a result of the process of compassion) as serious for that particular person.

3.1 The perspectival and multi-perspectival imagination

The perspectival and multi-perspectival imagination is a subset of imagination in general. The philosophical debate over the bounds of the imagination has intensified in the past few decades, and it has become common to speak of the variety of exercises of the imagination, all needing to be carefully distinguished.Footnote ⁵ While acknowledging that there is little in these debates that is not controversial, I offer the following stipulative definition of the imagination for present purposes: it is the process of deliberatelyFootnote ⁶ and effortfully constructing mental imagesFootnote ⁷ of what it might be like for someone else to experience a certain situation.

This definition already – in its final feature – references the idea of perspective. Indeed, for a number of philosophers, perspective is the key feature that differentiates thinner, typically propositional imagining (e.g. imagining that, or also supposing or conceiving) from thicker, non-propositional imagining. Thus, for instance, Richard Moran, in a classic paper, distinguishes between the ‘hypothetical and dramatic imagination’ (Reference Moran1994, p. 104), arguing that the latter involves adopting a perspective, and as a result requires a different kind of effort, and with different things at stake for the imaginer.Footnote ⁸ More recently, Magdalena Balcerak (Reference Balcerak, Kind and Kung2016) has also argued for a distinction between supposing, conceiving and imagining, on the basis that it is only in imagining that we adopt ‘experiential perspective-taking’ (p. 49). Like Moran, Balcerak stresses the self-involvement of this perspective-taking, saying that it involves putting ‘myself in the position of somebody who actually perceptually experiences [e.g.] an apple’ (Reference Balcerak, Kind and Kung2016, p. 49).

I think there is an important ambiguity in this way of thinking about perspective – namely, as something self-involving. I suggest it is crucial to keep in mind two kinds of imagining of the perspectives of others: first, one can project oneself into the situation and experience of another, asking ‘what would I feel if I was in that situation?’, or, second, one can instead consider all that one knows about that person and their situation and ask ‘what might that other person be experiencing in that situation?’. The second is a more robust way of exercising the perspectival imagination; it arguably requires greater sustained, imaginative effort to try to understand how another person might be experiencing a certain situation.Footnote ⁹

For these reasons, I prefer not to focus on this distinguishing feature of perspective.Footnote ¹⁰ Instead, I want to draw on a different kind of distinction – argued for by Joshua Landy (Reference Landy2004) in his work on Proust – namely, the distinction between perspective and viewpoint (or point of view). In brief, a viewpoint is something that we might share in common with others – it is literally the angle from which we might look upon a situation. Thus, we can think of standing and facing in a certain direction, and from that viewpoint seeing a series of church steeples. A viewpoint, then, is not something that uniquely belongs to that other person. A perspective, however, goes considerably further than something we can all see from a certain viewpoint. According to Landy:

In the case of Marcel's perspective – Marcel being the character that Proust explores in his great novel – we must therefore pay attention to the associations and connections that inform his own construction of the situation. Perspective is an active, affectively rich, meaning-constructing process. It takes time and effort to learn anything – even a small fraction – of someone else's perspective (for Proust, a master, it takes many years and over 4,000 pages). We learn something about Marcel's experience of the steeples from the following quotes (in a piece of writing by Marcel that we are made privy to by Proust):

Birds, flowers and maidens – the steeples are constructed to have a meaning by Marcel; those associations and connections refract his interests, needs and values (see Landy, Reference Landy2004, p. 60). Proust here is giving us access to Marcel's perspective – and not just the viewpoint from which he sees the steeples.

For Landy, there is similarity between the way Proust articulates and explores Marcel's perspective and the perspectivism of Friedrich NietzscheFootnote ¹¹ – and drawing this out will allow me to introduce the idea of the multi-perspectival imagination. According to R. Lanier Anderson, whom Landy draws on, ‘a perspective [is] a way of organising our experience’, and this is not because of some Kantian transcendental concepts that make experience possible, but because of our ‘contingent’ and ‘variable’ set of ‘needs, interests and values’ (Reference Anderson1998, p. 3). What Proust adds here, arguably, is that he shows very concretely just how such needs, interests and values manifest themselves in how persons (especially Marcel) experience the world and others. More explicitly, Proust shows us that we cannot rest content with the abstract thesis that experience is perspectival in the sense Nietzsche means that (as above). Rather, in order to imagine another's perspective, we must try to visualise – in imagination – those needs, interests and values in the concrete experiences of persons in their relations with the world and others.

But Nietzsche also adds something – which perhaps Proust does not do enough of in the novel (given the focus on Marcel). At stake is the usual problem with any more general (philosophical) celebration of perspectivism – namely, relativism and epistemological/normative anarchy. Nietzsche deals with that challenge – ‘What now for objectivity? Are there any standards?’ – by saying that the idea of objectivity as a form of ‘contemplation without interest’ (neutrality, the view from nowhere, or non-perspectivism) is ‘a nonsensical absurdity’, adding more positively:

Perspectivism, then, does not collapse into relativism, but neither does it mean we must adopt non-perspectival objectivism. Instead, we are enjoined to improve our knowledge – for example of some situation or some issue – by seeing how differently it is constructed by multiple perspectives (how different needs, interests and values are refracted concretely in particular experiences of that situation). Indeed, our knowledge is further improved by not only constructing what that situation might be like from multiple other perspectives, but in switching back and forth between them.Footnote ¹² The multi-perspectival imagination, then, is valuable because by deploying multiple perspectives we can construct a more complete (but never entirely complete) picture of the constellation of needs, interests and values at stake; and this value is enabled by the task of perspective-switching, including the experience of friction (axiological, affective) between different perspectives.Footnote ¹³

Thus, to summarise on the topic of imagination: by the imagination, in general, I mean the process of deliberately and effortfully constructing mental images of what it might be like for someone to experience a certain situation; by the perspectival imagination, I mean more than seeing a situation from a certain angle (viewpoint) – I mean the active construction (in mental imagery) of the needs, interests and values of another person as they are refracted in the concrete and particular experiences of persons in their relations with the world and others (e.g. via the associations and connections they make between things in their experience); and by the multi-perspectival imagination, I mean the process of constructing a variety of such perspectives, and switching between them, thereby accumulating greater knowledge of a situation. Imagining the perspectives of others requires (1) resisting my own background knowledge and attempting to characterise the other as richly as one can; (2) recognising that this process is a difficult one, which requires training and development; and (3) exercising humility, in particular by recognising the necessary incompleteness of any one attempt to imagine the perspective of another.

3.2 Emotion and imagination

Having articulated what I mean by relational emotions and the perspectival and multi-perspectival imagination, it is now time to turn to the relationship between them – a relationship that, as I will indicate in the final part of the paper, is important for the quality of legal reasoning. In doing so, I will be drawing mainly on Adam Morton's Emotion and Imagination (Reference Morton2013). Morton makes some bold (though also interesting) claims as to that relationship – for instance, asserting that ‘all emotion involves imagination’ (Reference Morton2013, p. 3); that imagination ‘can be seen as a special case of emotion’ (p. 21); and that ‘[p]erhaps emotion made thought [including imagination] possible’ (p. 21). I am not here concerned with whether one is the subset of the other, or whether one makes the other possible. Indeed, for present purposes, I can concede that we can experience emotion without the imagination, and that we can imagine without experiencing emotion. What is most helpful and important, however, is Morton's idea that ‘emotion has an object and involves imagining scenarios around it’ (Reference Morton2013, p. 30) and, more elaborately, that ‘[h]aving an emotion typically involves a complex of imagined possibilities, imagined reactions to them, and imagined sensations, and these imaginations are much of what the fear [or some other emotion] feels like’ (p. 33). I have already spoken of emotions as complex, multilayered processes. With Morton's help, I now suggest that: (1) part of what makes emotions complex is that they often (though not necessarily) involve some imagination – indeed, sometimes a great deal of imagination; and (2) that part of what makes at least the perspectival imagination distinctive (from other kinds of imagination) is that it is typically accompanied by emotion – and, again, sometimes by a great deal of emotion – such that typically there is something it feels like to imagine how someone else might be experiencing a certain situation.

Further, and this is also critical for present purposes, Morton notices that, in many cases, emotion (and here I would include compassion) increases the richness and vivacity of imagining, which might further increase the intensity of the experience of an emotion. The relationship, then, is not only one of overlap at the same time, but also mutual stimulation over time. Here is one example he gives:

The observer here first begins to imagine more richly on the back of an increase in feeling (of greater concern, having recognised the skier as her own child). Imagining more richly, in turn, intensifies her emotion (of concern for, fear of, her child) and this intensification of emotion further spurs her imagination, not only more richly, but even quite extravagantly (e.g. imagining wolves). We can also surely relate to this example in various ways: we can think of other circumstances in which we have felt fear, where part of what it is like to feel fear is to imagine, indeed for the imagination to almost take over (e.g. childhood fears of darkness and imagining what was under the bed). But we can extend the example considerably beyond fear – consider jealousy, for instance, or resentment. Part of what it is to feel jealous (and what it feels like to experience jealousy) is to imagine what a beloved is up to with someone else – and to imagine such scenes is to intensify the feeling of jealousy.

The same dynamics as to the mutually reinforcing relationship between emotion and imagination are at play in the case of compassion. Part, then, of the process of compassion is to imagine (concretely, with mental images) someone's suffering in a situation – understood from their perspective. The initial salience of and interest in another stimulates an initial act of imagining the situation from their perspective. This initial act of imagination, in turn, stimulates feeling (in compassion, feeling some of the suffering one imagines the other to be experiencing). This feeling, then, stimulates the perspectival imagination further. The more one feels, the more one imagines – and vice versa. In other words, the more one feels, the more richly one constructs the concrete associations and connections one takes to be constituting the other's perspective. What is critical here is to see the relation (in time and over time) between the perspectival imagination (again, in its concreteness – and in concreteness from a perspective, that is as filtered by that person's needs, interests and values) and the feelings one experiences as part of the process of compassion.

Morton's example, of course, also illustrates various dangers – the imagination, stimulated by emotion, can construct wildly implausible possibilities; and the fact that we imagine more richly may be caused by bias, that is by whom or what we are already predisposed to care about. These dangers are real, but I do not think they warrant the conclusion that we should thus dismiss the value of feeling imaginatively or imagining by feeling – that would be to give up on the considerable insights that engaging in such imaginative acts can provide (precisely into the perspectives of others). Thus, to the contrary, I would argue (along with Nietzsche) that we should instead (at least in contexts of judgment) exercise the multi-perspectival imagination. Aiming to understand better a certain situation, we do not withdraw into allegedly non-perspectival abstractions, but instead enter into multiple concrete perspectives – and, in order to increase the quality of our imagining those perspectives, we do so with feeling. It is, to repeat, only by feeling that we can get closer to (by imagining, of course always imperfectly) the concrete and particular experiences of someone else's perspective.

There is one final and important aspect to this affective dimension of imagination that I wish to mention, before going on in the next part to briefly illustrate the argument in the adjudicative context. That is the role of emotion in what is referred to as ‘imaginative resistance’. As Balcerak explains it, ‘imaginative resistance occurs when a subject is asked to imagine a particular situation, but is either unable or unwilling to do so’ (Reference Balcerak, Kind and Kung2016, p. 47). She goes on to claim that ‘most convincing examples’ of it ‘involve requests to imagine situations where morally highly deviant behaviour and attitudes are endorsed’; for example, confronted with the case of Alice, who ‘took her new-born baby, put it into a cotton bag, closed the bag tightly with a rope, and threw it into the lake’, there may be resistance to imagining the scene, let alone Alice's perspective (Balcerak, Reference Balcerak, Kind and Kung2016, p. 47). What Balcerak does not address here is the role of emotion – and yet that role must be considerable. Faced with such an example, there may be not only resistance to imagining – but also anger (moral anger, or even outrage) against Alice, as well as stomach-churning sickness at any attempt to imagine the scene (which leads to further resistance). Thus, just as imagining more richly is tied to emotion (in the ways I have discussed above), so is refusing to imagine or resisting imagining. Such resistances may be tolerable if reading a novel or watching a play or film – but they are different if the reader is tasked with evaluating Alice. What one needs to develop, then, is an ability to deliberately feel – in this case, some of Alice's suffering and then sadness for her – in order to not only better imagine her perspective, but also to challenge (and move somewhat beyond) resistance to imagining it. This is necessary if Alice is to be given the respect she deserves, not as someone who has murdered their baby, but as a subject of others’ judgment.