Compassion (or ‘feeling with’) has attracted fervent admiration and vehement condemnation. Among the critics are some Stoics and early Christians, and early modern philosophers such as Kant and Spinoza. Characteristics that are regarded as ideal in humans tend to be regarded as divine perfections as well. Accordingly, traditional Christian and Jewish theology includes the belief that God is ‘impassible’, that is, both emotionless (in general) and incapable of suffering (in particular). Divine compassion is precluded on both grounds.

Central to the rejection of human and divine emotion (in general) and of compassion (in particular) is the idea that emotions and compassion are irrational, deceptive and misleading (see Nussbaum, 2001, 356–400). For example, compassion-critics argue that a non-emotional attitude of benevolence is a better response to others' misfortune than the emotional response of compassion. This is because benevolence is rational and impartial, whereas compassion is discriminatory and imbalanced. As a result, compassion gives us a false impression of who is valuable and leads us to behave ethically to some people and not to others. It is easier to be compassionate to someone who is like us, or whom we find attractive, than to someone who is completely different, or someone we feel disgust towards – for example, a fetid vagrant. Because compassion is seen to be irrational, misleading and deceptive in these ways, it is argued that emotions and compassion are incompatible with divine, and with perfect human, intelligence.

In this paper, against that view, I am going to point to ways in which compassion is intelligent. Moreover, I am going to suggest that compassion and related emotional responses provide a non-substitutable kind of knowledge – that is, a kind of knowledge that can't be gained by non-emotional means. If I am right, then compassion and other emotions are necessary to human intelligence rather than being things that detracts from it. Finally, I am going to conclude by putting forward some implications for epistemology (that is, our understanding of what knowledge is). In particular, I am going to argue that compassion and other emotions should not only be regarded as sources of knowledge, but actually as potential forms of knowledge. Central to that argument is the idea that emotions are mental states such as beliefs and judgements, rather than being something other than mental states, rooted either in the body, or in some other part of us, such as the metaphorical ‘heart’.

Before I start, we need to be clear about what I am talking about when I use the word ‘compassion’. I have already said that compassion is ‘feeling with’. In this, it is like empathy (with also means ‘feeling with’, though ‘empathy’ derives from Greek rather than Latin). Empathy is often described as the imaginative reconstruction of another person's feelings (in other words, in empathy, we put ourselves into another's shoes). At the same time, empathy is only ever approximate, because we can never experience someone else's feelings exactly as they experience them. Empathy is part of compassion, but empathy and compassion are not the same things. That is because a torturer or bully might be empathetic, and use his or her empathy to contrive new ways of hurting his or her victim. While we might want to say that the torturer or bully is empathetic, we would not want to say that they are compassionate. While a torturer or bully would experience pleasure at another's pain and perhaps pain at another's pleasure, a compassionate person would experience pleasure at another's pleasure, and pain at another's pain. Therefore, compassion should be seen as benevolent empathy. In other words, compassion is empathy that seeks the good of the other.

A distinction also needs to be made between compassion and pity. Compassion and pity are not the same, since to feel with someone is different from feeling sorry for them. The philosopher Aaron Ben-Ze'ev notes several differences between compassion and pity that are helpful here (Ben-Ze'ev, Reference Ben-Ze'ev2000). First, compassion is typical of situations in which the subject (the person feeling compassion) is close to the object (the person for whom compassion is felt). In contrast, pity is typical of situations in which the subject does not have a close relationship with the object. We are more likely to feel compassion for a friend, but pity for people we see on the News. Underlying this distinction is the idea that compassion is rooted in the subject's personal involvement with the object whereas, in pity, the subject remains not entirely, but rather more, personally detached. Second, ‘compassion involves a willingness to become personally involved, while pity usually does not’ (Ben-Ze'ev, Reference Ben-Ze'ev2000, 328). Therefore, compassion differs from pity in being self-giving. Compassion incorporates real help in which personal resources are expended. Third (and relatedly), in compassion the subject is actively discontent with the situation, while pity ‘typically includes a kind of acceptance of the present situation’ (Ben-Ze'ev, Reference Ben-Ze'ev2000, 330).

Of course, you are welcome to disagree with my definitions of compassion, empathy, and pity, but that is how I am using them here.

Now I will put forward five ways in which I think compassion contributes to our intelligence.

One way in which compassion contributes to intelligence is it makes us realise that people, and the things that happen to them, matter deeply. In other words, compassion tells us that people are intrinsically valuable rather than valuable for any utility they have, or as a means to an end. Because of this, compassion makes us understand why selflessly benevolent behaviour towards others is important. For this reason, a rule-based morality devoid of the imaginative resources essential to compassion can easily become confused with submissiveness to cultural rules or norms (see Nussbaum, 2001). This means that in addition to showing us that people matter deeply, compassion also makes morality meaningful. As Schopenhauer put it (though within a rather different framework) compassion is the basis of all morality.

Furthermore, compassion and other emotional responses have the potential to break through existing moral judgements and to cause the subject to depart from them. This is particularly the case when exposure to, engagement with, and eventual attachment to, a particular person or persons causes the subject to re-evaluate a negative judgement about that ‘type’ of person. An example of this is explored in Huckleberry Finn. The story focuses on the encounter and developing relationship between two runaways: Huck, a white boy escaping his violent father, and Jim, a black slave escaping his owner. Initially, Huck experiences conflict over whether or not to report Jim, but, as they travel together, Huck begins to have compassion for Jim when he hears of his difficult life. Huck's personal and emotionally engaged encounter challenges Huck's beliefs about black people and about slavery. His beliefs are gradually altered and develop throughout the remainder of the novel. In his lecture notes, Twain writes that ‘a sound heart is a surer guide than an ill-gained conscience’ and describes the novel as ‘a book of mine where a sound heart and a deformed conscience come into collision and conscience suffers defeat’. (Twain, cited in Hutchinson, Reference Hutchinson1993, 193). Compassion, then, is intelligent because it can break through our inherited moral norms and provide us with insights about the value of people that we would not be capable of having without it.

As Huckleberry Finn suggests, humans learn universal compassion through their attachments to particular humans and non-human animals from childhood onwards. In Martha Nussbaum's words, ‘The good of others means nothing to us in the abstract or antecedentally’ (Nussbaum, 2001, 388). We begin to learn compassion from an early age in the context of our own friendships and familial relationships, but (if we accept that compassion is a virtue it is important to cultivate) we need to learn to extend it to people we don't know, and even whom we don't like.

Compassion can also function to make otherwise undiscerning justice systems intelligent (Ben-Ze'ev, Reference Ben-Ze'ev2000). In Christian theological texts about redemption or Atonement, as well as in texts such as Shakespeare's The Merchant of Venice, compassion (or sometimes mercy) and justice are often portrayed as opposites and frequently as antithetical. However, I think that compassion is better seen as an instrument by which a more nuanced form of justice can be realised than is possible through rules and laws alone. This is because there may be differences between people (for example, upbringing or education) that become morally relevant when we come to judge responsibility for wrongdoing, but it is difficult or impossible to legislate for all such morally relevant differences (see Ben-Ze'ev, Reference Ben-Ze'ev2000, 349). In a perfectly just society, every situation would be treated as different, but since legislating for every combination of every possible difference is, in practice, impossible, the legal system goes for the second best option: treating every situation as the same. The deficiencies of using ‘sameness’ as a means to ‘fairness’ are to some extent compensated for by the mediation of a human (and, therefore, hopefully, compassionate) judge. A non-compassionate form of justice is likely to be less intelligent because it would be unable to ‘see beyond’ the rules and laws to the underlying personal circumstances that may make a particular penalty unjust. For this reason, a (non-emotional) computer that was programmed to apply laws to specific cases in order to give a sentence would be less able to be entirely just (despite its flawless rationality and objectivity) than would a human judge capable of compassion and able to see the situation with ‘compassion's eye’.

Compassion also contributes to our intelligence because it affects the way in which we relate to others. Simply put, in compassion, we relate to people as people who are ‘like us’, whereas in a non-empathetic response such as pity we relate to people as inferior to us, and (ultimately) not as people at all. Charitable actions motivated by pity lead to condescending behaviour which means that the subject fails to relate to the object, frequently behaves in a debilitating way towards her, and fails to understand her actions and intentions. In contrast, compassion can help us to perceive why people are behaving as they are, and how best to respond to them. This insight is highlighted by the theme of the wounded healer, according to which (as the Carthaginian Queen, Dido, puts it) someone has learned through their own suffering how to bring aid to the wretched (Virgil, Book 1, line 630). The sympathy of the wounded healer is compassion rather than pity because the healer's own woundedness causes her to be empathetic. The wounded healer is more able to relate intelligently to the person who suffers, precisely because her sympathy is compassion and not pity.

So far, I have made five claims about the intelligence of compassion. First, I have suggested that compassion reveals that people matter and, second, that it therefore shows us the reason for morality. Third, I have argued that compassion may allow us to transcend existing moral norms where these are limited, by revealing the value of people society does not regard as valuable. Fourth, I have argued that compassion can transform undiscerning justice into intelligent justice. Fifth, I have claimed that compassion informs how we relate to people. At the beginning of this paper, I suggested that compassion and related emotional states (for example, empathy and love) provide a non-substitutable form of knowledge. I will now try to support this claim.

Discussions of knowledge and intelligence, at least in philosophy of religion (my own field) have tended to prioritise propositional knowledge. In other words, if we are knowledgeable then we can state facts (propositions) about things. Knowledge is defined as ‘justified true belief’, and therefore an all-knowing being such as God would have a perfect and justified belief in all true propositions. For example, Anthony Kenny argues that everything we know about our sensations is communicable by propositions, and that anything that we are unable to communicate is not worthy of the name ‘knowledge’ (Kenny, Reference Kenny1979, 31). The prioritisation of propositional knowledge reflects the modern dichotomy between the head and the heart, where the head is seen as the seat of the intellect, and the heart is seen as concerned with sentiments, feelings, emotions, and the arational. This is shown in our everyday conversations. When someone has a difficult decision to make, they will often say that they need to think with their head (intellect) rather than their heart (emotions). Other people say they should follow their heart (their emotions). Either way, there is an apparent dichotomy between the head and the heart: they are seen as entirely separate, and often at war with one another.

However, maybe this is the wrong way to look at things. The idea that all knowledge is propositional overlooks the distinction between knowledge that something is the case (for example, knowledge that someone is in pain) and knowledge of how that something (in this case, pain) feels (see Sarot, Reference Sarot1992). Moreover, it is difficult to see how we could have knowledge of how something feels without having felt it, or without having felt something like it. If my friend tells me he has toothache, and I have only ever had backache, I could probably imagine approximately what his toothache is like. But I could not imagine what his toothache felt like unless I had experienced some sort of pain. I would know that he had toothache, but I would not know what it is like to have toothache. This suggests a distinction between propositional knowledge (that my friend has toothache) and experiential knowledge – that toothache feels like THIS:

Furthermore, our experiential knowledge is emotional because all of our experience is emotional. It may not obviously be ‘about’ specific emotions, but we experience it emotionally because it involves sadness and happiness, and because it affects the quality or ‘colour’ of our lives. As the novelist Robert Musil puts it: ‘things swim in emotions the way water lilies consist not only of leaves and flowers and white and green but also of “gently lying there”’ (Musil, Reference Musil, Wilkins and Pike1995, 1561). Therefore, the separation between our emotions and our intellects is deeply flawed.

If what I have been arguing is correct, then emotions can not only be intelligent, but can also provide us with a non-substitutable form of knowledge. In other words, compassion and related emotions not only reveal (for example) that people matter. Even more crucially, we couldn't know why people matter without experiencing compassion, love, empathy, and so on. To assent to the proposition that a person is valuable without experiencing them as valuable is, contra Kenny, to have a limited kind of knowledge. Or, as Oscar Wilde said, to look at a thing without seeing its beauty is not really to see it at all. This is where I get to the implications for epistemology. All of this not only suggests that compassion can contribute to our intelligence and knowledge, and can contribute to it in a non-substitutable way. It also suggests that compassion and other emotions should not merely be regarded as sources of knowledge, but should be regarded as potential forms of knowledge in and of themselves. In other words, the experience of compassion and of other emotions is a potential form of knowledge, just as much as the belief that x (where x is a proposition, such as ‘it is raining’) is.

Of course, this still leaves open the possibility that emotions can be misleading, even if they are not always so. We might still be misled by compassion into treating someone we find attractive or someone who is like us much better than we treat a fetid vagrant. However, in this, compassion is like non-emotional beliefs (such as the belief that it is raining). That too might be true or untrue, and we may or may not be justified in believing it to be true. Therefore, emotions are like non-emotional beliefs (in so far as non-emotional beliefs exist) in that both are potential forms of knowledge, but, equally, both require a process of discernment or wise judgement in order for us to work out whether or not we should trust them.

On the face of it, this seems to contradict the definition of knowledge as ‘justified true belief’, since an emotion is not usually thought of as a belief. But what if emotions are beliefs? Some contemporary and ancient philosophers of emotion (such as Nussbaum and the Stoics respectively) have argued, persuasively I think, that emotions are beliefs or judgements (albeit ones that aren't necessarily propositional or even conscious). That idea resonates with my suggestion that we should reject the dichotomy between the heart and the head (or the emotions and the intellect). If I, Nussbaum, and others are right, then emotions are mental states just as much as non-emotional beliefs are, rather than being something else – for example, non-mental bodily feelings, or something housed in another part of us; perhaps our metaphorical ‘heart’. This is important for what I am arguing, because if emotions are beliefs, then viewing emotions as a potential form of knowledge is compatible with the traditional definition of knowledge as ‘justified true belief’ after all.

In this paper, I have argued that compassion and other emotions contribute to our intelligence, and that they contribute to it in a unique or non-substitutable way. I have also argued that emotions should not only be regarded as sources of knowledge, but also as potential forms of knowledge, in much the same way as a non-emotional belief is a potential form of knowledge if it is both justified and true. This has implications for our everyday life such as how we assess our choices and decisions. If emotional states are potential forms of knowledge, then, while subjecting them to the same criteria as non-emotional beliefs (since both emotions and non-emotional beliefs can be misfounded or mistaken) we will nevertheless want to take them very seriously and to consider them of primary rather than secondary importance. As potential forms of knowledge, we will want to listen more to what our emotions are telling us. For those of us who are monotheists (for example, Christians, Jews or Muslims) we may also want to break away from the long-held ‘impassibilism’ of our traditions, and attribute emotion to God as an essential part of God's wisdom, intelligence and even knowledge.