Recent Work in the Philosophy of Empathy


Professor Derek Matravers


Derek Matravers, Philosophy Department, The Open University

July 2017 – It is an exciting time in Anglo-American philosophy in the study of empathy. In addition to a host of articles, there have recently been two significant collections of essays published (Coplan & Goldie, 2011; Maibom, 2014). There have been two books – one a serious monograph, and the other an overview (Stueber 2006, Matravers, 2017). Finally, a true sign of the debate having arrived, a handbook has just appeared, which contains no fewer than 33 papers (Maibom, 2017). However, it would be a mistake to imagine that there is a consensus, or, indeed, even an agreement on the nature (or existence) of the phenomenon under discussion. In this contribution, I will briefly outline the recent history of empathy within philosophy, and then say something about what I take to be the most interesting question currently engaging the field.

The absence of agreement is not much of a surprise, given the history of the concept. Aspects of the contemporary discussion bear a marked similarity to what certain Enlightenment thinkers, most particularly Adam Smith and David Hume, called ‘sympathy’. Unfortunately, one such similarity is a failure precisely to demarcate the phenomenon; to distinguish it (or not) from such phenomena as emotional contagion or sympathy (in the modern sense). The term (‘empathy’) has a much more recent history. As is well known, it was coined by Edward Titchener in 1909, as a translation of the German, ‘Einfühlung’. The debate from which this German neologism sprang was also confused; the core seemed to be that beauty is a matter of our ‘animating’ parts of the world by projecting our inner states into them. This notion of ‘projection’ was subject to excoriating criticism by Vernon Lee in 1910 (Lee & Anstruther-Thomson, 1912), in 1914 the theory was magisterially dismissed by E.F. Carritt: ‘we have here nothing but an attempt to explain in figurative language an unconscious process by which some beautiful objects may have become so’ (Carritt, 1962, pp. 191). Discussion then largely disappears until around the 1980s, when two separate areas of philosophy become hotly debated. The first was the debate as to how we know the contents of another’s mind – I shall refer to this as the ‘mind reading’ debate. The second was the debate over the nature of the emotions. I shall start by examining this distinction, and then move do discussing empathy and the emotions – as that will be of most interest to readers of this newsletter.

The mindreading debate sprung out of a dissatisfaction with the consensus view – functionalism. Functionalism holds that person A works out what person B is thinking by relying on some general theory that connects perceptual inputs and behavioural outputs with our mental states. That is, A is supposed to rely on some tacitly known theory which is full of theorems of the sort ‘In circumstance C, if X desires Y, and believes that Z-ing is the best way to get Y, then X will Z’. Such a view faced a number of problems both in principle an in practice. Independently, Jane Heal (in the UK) and Robert Gordon (in the US) proposed an alternative (Gordon, 1986; Heal, 1986). Instead of using a theory to discover what someone else is thinking, we (broadly) imagine that we in that person’s situation, note what we would think in such circumstances, and attribute that thought to the person. This notion of taking someone else’s perspective drew on thinking on empathy. Indeed, for a while the view was known as ‘the empathy view’, before settling on ‘simulation theory’ (Davies & Stone, 1995, pp. 1).

Possibly stemming from the same basic dissatisfaction with an excessive focus on functionalism, philosophical work on the emotions also become popular in the 1980s. For a while empathy merely had a walk-on part, with few if any mentions. This changed when Peter Goldie, in his 2000 monograph, The Emotions, discussed it more extensively (Goldie, 2000). Such discussions linked with the discussion of simulation theory, and the touch-paper was lit.

However, it is not obvious the two debates have much in common. The first, mindreading debate, was basically about working out what other people are thinking and the second, emotion debate, about feeling what other people are feeling.

Those engaged in the mindreading debate did not deny that some people used the term ‘empathy’ in a way that linked it to affective states. Their view was, rather, that this was not where the concept could do most work; a view they (rightly) linked to the tradition out of which the term had emerged – the discussion of Einfühlung. Hence, Kirsten Stueber defines ‘empathy’ as ‘a form of inner or mental imitation for the purpose of gaining knowledge of other minds’ (Stueber, 2006, pp. 28). Indeed the gulf between the mindreading debate and the emotions debate soon got wider. Working out what people will think – that is, working out the inferential connections they will make – has greater prospects of success than the rather messy business of working out what people will feel. Hence, the project of ‘gaining knowledge of other minds’ has sometimes left the emotions behind altogether, and been restricted to purely cognitive states (Heal, 1988). Across the divide, work on empathy as an affective state has side-lined the cognitive. In the emotion debate, this definition of ‘empathy’ by Heidi Maibom would be fairly standard: ‘S empathizes with O’s experience of emotion E in C if S feels E for O as a result of: believing or perceiving that O feels E, or imagining being in C’ (Maibom, 2014, pp. 3).

Having laid out the contrast, I will, as indicated above, focus specifically on a problem with empathetic emotions. Philosophers are interested in understanding emotional empathy both because it is an interesting phenomenon in itself and because of the role such emotions play in other areas of enquiry. Particularly fruitful is the link to morality. Hume and Smith, mentioned above, both made empathy (or, as they called it ‘sympathy’) foundational for their moral theories (see in particular Hume (1739-40) and Smith (2002)). However, as contemporary debate shows, understanding the role empathy plays in our moral thinking and conduct is particularly difficult. This debate is not purely (or even principally) conducted by philosophers. Among the psychologists, Martin Hoffman holds views ‘empathy as the bedrock of morality’ (Hoffman 2011, pp. 96), while Daniel C. Batson holds that ‘empathy-induced altruism is neither moral nor immoral; it is amoral’ (Batson, 2014, pp. 47). A similar division can be found in philosophy. Michael Slote thinks empathy is the key to morality (Slote, 2007) while Jesse Prinz thinks that ‘empathy is not necessary for the capacities that make up basic moral competence’ and ‘can interfere with the ends of morality’ (Prinz, 2011, pp.  211-213). The philosophical view is that much of this debate (which, through Paul Bloom has become part of popular culture (Bloom, 2016)) is greatly helped by the kind of distinction-drawing, conceptual clarity, and interpretation of evidence, in which the discipline excels.

On, then, to the empathetic emotions. Let us restrict ourselves to the usual case where someone imagines themselves occupying the perspective of another (I shall ignore various problems with this, some of which I will return to below). I shall call the person who is being empathetic ‘the empathiser’ and the person for whom empathy is felt ‘the target’. Consider an example in which the target feels sad because their dog, Fido, has died; that is, the emotion (E) is the target’s sadness at Fido. The question is: what does the empathiser feel?

According to Maibom’s definition, the empathiser feels E. That, however, is a complicated claim that needs some unpacking. One issue is how finely we want to individuate E. Let’s be old-fashioned cognitivist about this, and individuate emotions by their cognitive content: that is, E is ‘sadness at the death of Fido’. Could this be what the empathiser feels?

One might think not for at least two reasons. First, the relation between the empathiser and the death of Fido is unlikely to be the same as the relation between the target and the death of Fido (Matravers, 2011). This is, the latter is likely to have ramifications throughout the target’s mental economy; a sense of loss, starting every time a dog barks or a door is pushed open, weeping in the pet-food isle, and so on. Imagining being in the target’s situation, as we are assuming the empathiser does, is unlikely to bring about the same effects. One could reply that the empathiser feels what the target feels but not to the same degree. However, it is not clear that is coherent. Consider a parallel example. The target is in love with Susan; he cannot concentrate, his mind returns to her constantly, he is intoxicated by the sound of her voice, the smell of her hair. Clearly, the empathiser (who, let us assume, is not in love with Susan) does not feel all (or any) of that. It is not obvious it makes sense to say that he feels it, but just not to the same degree.

One solution is to say that the cause and object of the empathiser’s emotion is not the death of Fido, but rather some object that plays an equivalent role in the empathiser’s life (say, Rover – a dog of his that died). That is, in imagining being in C, the empathiser takes on the target’s perspective, but substitutes Rover for Fido and thus generates the kind of ramifications in her (the empathiser’s) mental economy as the death of Fido does in the target’s mental economy. Something like this might be the case, but it needs a bit of finessing. At first pass it does not look to be empathy at all; it looks, rather, as if the target’s situation has triggered the empathiser into their own emotion – feeling sad for themselves at the death of their own dog.  That might, however, not be the best way of specifying the cause and object of the empathiser’s emotion. Instead we might say this: essentially, the cause and object of the empathiser’s emotion is something of a sort that would be the appropriate cause and object of the target’s emotion. That is, what the empathiser tries to do is to find something (anything) in her (the empathiser’s) life that matches the role the death of Fido has in the target’s life. So it is not really right to say that the empathiser is feeling sad about the death of Rover; they are feeling sad about some state of affairs whose nature is dictated by the object of the target’s emotion. As a matter of fact, that thing is the death of Rover, but the most informative way of describing what the empathiser is feeling sad about is not ‘sad-at-the-death-of-Rover’ but ‘sad-at-something-that-matches-what-the-target-feels-sad-about’.

Let us let that worry lurk in the background as I shall return to it below. For the second reason for holding that E is not ‘sadness at the death of Fido’ is the more interesting: that it cannot be the emotion the empathiser has in mind as she could have that emotion without implicating the target; but the target clearly is implicated.

The obvious way to sort this out is to have the target, rather than Fido, as the object of the empathiser’s emotion. This is the literal reading of Maibom’s definition above: S feels E for O. Prima facie, this is not an attractive solution. If the object of the empathiser’s emotion was the target, rather than the dog, that would be sympathy (feeling sad for someone who is sad) rather than empathy (feeling the sadness of someone who is sad).

Maibom herself is aware of the problem and has this to say:

‘Feeling for’ is to be understood broadly so as to include cases where I am angry with a person because that person wronged you, where, in a sense, you are neither the object not the subject of my emotion. What makes it a case of empathic anger is that I am feeling it not directly as a sort of objective moral anger, but rather I feel it on your behalf. (Maibom, 2014, pp. 5; see also Maibom, 2017)

What is it to feel an emotion on someone’s behalf? Stephen Darwall discusses this under the name ‘proto-sympathetic empathy’.

A person grieves the loss of his child, and in sharing his grief projectively my focus is on the child who was lost, not the person whose grief I share. When, however, I turn my attention to what it must be like to live with this loss, I focus on the person himself and the ways his grief pervades and affects his life. Before my thought was: What a terrible thing – a precious child is lost. Now my thought is: What a terrible thing for him – he has lost his precious child. (Darwall, 1997, pp. 271)

Darwall does not go into detail about how this might be done. Here is a possible scenario. The empathiser imaginatively identifies with the target. In doing so, he imagines that his child (that is, the target’s child) has died. However, the empathiser also has a perspective on how, within the imaginative project, he (the empathiser) feels. Assuming that the attempt at empathy has been successful, this will also be how the target feels. Hence, the empathiser is able to note, and track, how the target feels. From this second-order perspective, the empathiser can (a) note that they are feeling ‘on the target’s behalf’ and (b) form a judgment, on the basis of what they are feeling, about what the target is feeling. This would seem to implicate the target in the right kind of manner.

The field is split as to whether or not this scenario is plausible. Peter Goldie has been particularly sceptical. His scepticism is rooted in the possibility that the empathiser can replicate, in any reliable way, the flow of thoughts and emotions around the target’s mental economy. In particular, the psychological role of emotional dispositions is not the kind of thing that can be duplicated in imagination (Goldie, 2011). This is slightly different to the problem I considered earlier: that the same object (the death of Fido) might have different effects on different people because of different people will have different relations to that object. Here the issue is more that the structure of the mind (in particular, the relation between our emotional dispositions and our emotional experiences) cannot be replicated in imagination. I shall not dwell on this here; suffice to say, it might be an advantage if we are able to implicate the target without relying on this controversial model.

Let us focus on (b): what is the judgement? One option would be for the empathiser to decide on the sort of emotion they feel – and given what Darwell has said, that might be complicated – and judge that the target feels an emotion of that sort. However, they might instead use the emotion they are feeling as an exemplar; they simply make the judgement that the target feels like this; where ‘this’ refers to the emotion they (the empathiser) is feeling. This solution has recently been put forward by Kendall Walton (see also Joel Smith  (Smith, 2015)).[1] In Walton’s example, Emily is the empathiser and Oscar is the target.

Emily’s judgement or impression is not merely that ‘I am panicked, and so is Oscar,’ but rather, ‘Oscar is as I am, like this.’ She can appropriately say, ‘I know how it is with him’ or ‘I know how he feels,’ where ‘know’ carries a connotation of intimacy, acquaintance… Notice that the content of what she knows is in propositional form: She know that Oscar feels like this. But this is propositional knowledge of a special kind, with the sample taking the place of a linguistic predicate in the formulation of what she knows. (Walton, 2015, pp. 9)

The key point to notice about Walton’s view is that, at least with respect to (b), we do not have to rely on the controversial scenario detailed above. All that needs to be the case is that the empathiser is experiencing an emotion they believe is the one felt by the target. There is no reason to restrict the method by which they generated this emotion to imaginatively identifying with the target. They could, for example, simply recall what it was like for them when they were in the target’s circumstances, and, having dredged up that emotion from memory, claim that the target feels like this (where ‘this’ is the emotion they are feeling). Walton’s definition of empathy postulates no necessary link to the imagination: ‘I propose to define “empathy” as, simply, using some aspect of one’s current mental state as a sample to understand another person, in the way I have described, i.e., judging or experiencing the target person to be feeling “like this”’ (Walton, 2015, pp. 9-10).

Walton’s account not only has the advantage of not relying on the controversial scenario, it also avoids the earlier problem that, as the empathiser stands in a different relation to the object of the target’s emotion as does the target, the effect on the empathiser will be different. My suggested solution to that problem was that the empathiser focusses on ‘some object that plays an equivalent role in the empathiser’s life’. This fits perfectly into Walton’s proposal. Empathising with you on the death of your dog involves me recalling or imagining the death of my dog, thus creating an emotional experience, and judging that you feel like this.

Thus, if we focus just on (b), there is good reason to favour Walton’s proposal. What, however, about (a): that we are feeling ‘on the target’s behalf’? Well, this is partly a matter of judgement. On some views, empathy is the key to moral and spiritual progress – the process of identifying with another, taking on their issues, and thus being motivated to help them. Walton’s view certainly seems ‘thinner’ – perhaps too thin to give us everything a pre-theoretical understanding of empathy (if there is such a thing) promises. This would not bother Walton; he is a ‘theory builder’ in philosophy; rather than someone who analyses concepts as he finds them, he is happy to take them away, tidy them up, and return them to the linguistic wild. If we follow Walton, as I have suggested we should, it reveals that the link between empathy and the imagination might not be as tight as we might have assumed. The pay-off is a neat solution as to how we can feel our own emotion, implicate the target in that emotion, and yet not have to explain this in ways that some have found incredible.

Empathy is a good example of a topic where philosophy can make progress by working together with other disciplines including psychology (in its clinical, developmental, and social forms), phenomenology and hermeneutics, and cognitive neuroscience. Indeed, it is an area in which progress will only be secure once certain philosophical issues are bashed out: what exactly is it to ‘take someone else’s perspective’? What is it to feel an emotion ‘on behalf of’ another? What is egoism, altruism, morality, and how are they related? What are the differences between emotional contagion, empathy, and sympathy, and when are they important? These, and other issues, leave plenty of work to do.


Batson, C. D. (2014). Empathy-induced altruism and morality: No necessary connection. In H. L. Maibom (Ed.), Empathy and Morality. (pp. 41-58). Oxford, UK: Oxford University Press.

Bloom, P. (2016). Against Empathy: The Case for Rational Compassion. New York, NY: Ecco.

Carritt, E. F. (1962). The Theory of Beauty. London, UK: Methuen.

Coplan, A. & Goldie, P. (2011). Empathy: Philosophical and Psychological Perspectives. Oxford, UK: Oxford University Press.

Darwall, S. (1997). Empathy, sympathy, care. Philosophical Studies 89, 261-282.

Davies, M. & Stone, T. (1995). Introduction. In M. Davies & T. Stone (Eds.), Folk Psychology: The Theory of Mind Debate (pp. 1-44). Oxford, UK: Blackwell.

Goldie, P. (2000). The Emotions. Oxford, UK: Clarendon Press.

Goldie, P. (2011). Anti-empathy. In A. Coplan & P. Goldie (Eds.), Empathy: Philosophical and Psychological Perspectives (pp. 302-317). Oxford, UK: Oxford University Press.

Gordon, R. M. (1986). Folk psychology as simulation. In M. Davies & T. Stone (Eds.), Folk Psychology: The Theory of Mind Debate (pp. 60-73). Oxford, UK: Blackwell.

Heal, J. (1986). Replication and functionalism. In Mind, Reason, and Imagination (pp. 11-27). Cambridge, UK: Cambridge University Press.

Heal, J. (1988). Co-cognition and off-line simulation: Two ways of understanding the simulation approach. In Mind, Reason and Imagination (pp. 91-114). Cambridge, UK: Cambridge University Press.

Hoffman, M. L. (2011). Empathy, justice, and the law. In A. Coplan and P. Goldie (Eds.), Empathy: Philosophical and Psychological Perspectives (pp. 230-254). Oxford, UK: Oxford University Press.

Hume, D. (1739-40). A Treatise of Human Nature. London, UK: Thomas Longman.

Lee, V. & Anstruther-Thomson, C. (1912). Beauty and Ugliness and Other Studies in Psychological Aesthetics. London, UK: John Lane.

Maibom, H. L. (2014). Empathy and Morality. Oxford, UK: Oxford University Press.

Maibom, H. L. (2014). Introduction. In H. L. Maibom (Ed.), Empathy and Morality (pp. 1-40). Oxford, UK: Oxford University Press.

Maibom, H. L. (2017). Affective empathy. In H. L. Maibom (Ed.), The Routledge Handbook of Philosophy of Empathy (pp. 22-32). London, UK: Routledge.

Maibom, H. L., Ed. (2017). The Routledge Handbook of Philosophy of Empathy. London, UK: Routledge.

Matravers, D. (2011). Empathy as a Route to Knowledge. In A. Coplan and P. Goldie (Eds.), Empathy: Philosophical and Psychological Perspectives (pp. 19-30). Oxford, UK: Oxford University Press.

Matravers, D. (2017). Empathy. Cambridge, UK: Polity.

Prinz, J. (2011). Is empathy necessary for morality? In A. Coplan and P. Goldie (Eds.), Empathy: Philosophical and Psychological Perspectives (pp. 211-229). Oxford, UK: Oxford University Press.

Slote, M. (2007). The Ethics of Care and Empathy. London, UK: Routledge.

Smith, A. (2002). The Theory of Moral Sentiments. Cambridge, UK: Cambridge University Press.

Smith, J. (2015). What Is Empathy For? Synthese, 194, 709-722.

Stueber, K. R. (2006). Rediscovering Empathy: Agency, Folk Psychology, and the Human Sciences. Cambridge, Mass., MIT Press.

Walton, K. (2015). Empathy, Imagination, and Phenomenal Concepts. In In Other Shoes: Music, Metaphor, Empathy, Existence (pp. 1-16). Oxford, UK: Oxford University Press.

[1] Strictly, Walton does not think it is necessary for the empathiser to judge; they can merely experience the target ‘as feeling “like this”’ (Walton, 2015, pp. 9). I shall ignore this complication.

Print Friendly, PDF & Email

Traffic Count

  • 215965Total reads:
  • 115Reads today:
  • 171754Total visitors:
  • 91Visitors today: