Most advocates of moral realism defend their view primarily by addressing objections and by pointing out the shortcomings of rival metaethical positions. Realists often assume that their view is the default position in metaethics and, therefore, do not undertake a full-fledged defense of moral realism by putting forward positive arguments in favor of it. Colin Marshall’s Compassionate Moral Realism is an ambitious enterprise in that it aims to provide a positive case for moral realism.

Marshall’s core argument is twofold. First, he argues that there are certain irreplaceable epistemic goods that are possessed only by compassionate agents, i.e. agents who perceive other sentient beings’ pains, pleasures and desires. Second, he argues that the fact that compassion allows us to get reality right in a unique way supports a naturalistic, reductive but non-eliminative version of moral realism.

Given the vast amount of philosophical and empirical literature that focuses on a number of related phenomena described by various authors by terms such as “compassion”, “empathy”, “sympathy” and “fellow-feeling”, let me briefly clarify how Marshall’s notion of compassion differs from other related notions. The phenomenon Marshall calls “compassion” can be characterized as follows: (1) it is an affective state; (2) it is perceptual, non-inferential in nature; (3) the term “compassion” can properly be applied to it if and only if a certain condition is satisfied (i.e. if and only if the affective response matches or resembles the state at which it is directed); (4) it provides an epistemic good called “being in touch”, which involves perceiving or experiencing a given object in a way that reveals a property that the object really has, thus perceiving or experiencing it as it is in itself (e.g. to be in touch with another creature’s pain means to perceive or to experience the suffering creature in a way that reveals the property of pain). 

Marshall asks us to imagine two people who see an injured wombat that struggles to release his foot from a trap. They both know that the wombat is in pain, but only one of them is “pained by the wombat’s pain” and is moved to help, while the other one is “not pained at all by the sight” and watches it with cool curiosity (p. 40). In a certain sense, both of them understand what is going on. Yet Marshall argues that, despite having test-passing propositional moral knowledge about the situation, the observer who is indifferent to the wombat’s pain is missing something epistemically. What she fails to perceive is the unpleasantness of the wombat’s pain. Another way of putting this is to say, following Locke, that the compassionate observer’s experience resembles the wombat’s pain, whereas the indifferent observer’s experience doesn’t. In virtue of that resemblance, the experience of the compassionate observer lets her know what pain is like, whereas the property of pain is not revealed in the experience of the indifferent observer. This suggests that there are certain irreplaceable epistemic goods that are possessed only by compassionate agents. Over the course of several chapters, Marshall sets himself the task of “clarifying [this claim], defending it against various objections, and applying it to more complicated cases” (p. 41). His discussion touches on a wide range of topics, including John Locke’s theory of world-matching ideas, Arthur Schopenhauer’s phenomenological analysis of compassion as providing access to others’ mental states, the implications of Frank Jackson’s famous “Mary” example for the existence of an irreplaceable epistemic good of a perceptual nature, as well as neuroscientific findings about mirror neurons.

The book consists of three parts. The first two parts defend the claim that there are certain aspects of reality that can be grasped only by compassionate agents. While the first part focuses on short-range compassion, i.e. compassion directed at “a single present mental state of some creature” (p. 13), the second part extends the argument to long-range compassion, showing how we can be in touch with past, present but spatially distant, and future pain states. Furthermore, it is argued that long-range compassion applies also to cases involving other creatures’ pleasures and desires. Showing how long-range compassion is possible helps to dispel familiar worries about grounding morality in our capacity for compassion. Some of the most problematic cases, such as those involving compassionate reactions to sadism or masochism, cease to be problematic once we turn our attention from short-range compassion to the possibility of long-range compassion. Marshall argues that, in cases of sadism, long-range compassion “leads through” to compassion for the creature tormented by the sadist, whereas in cases of masochism, long-range compassion “leads through” to potential future pain states that could result from the masochist’s actions, and therefore, his view does not have the implausible implication that compassionate agents must be pleased at the sadist’s or the masochist’s pleasures. By defending the epistemic asymmetry between compassionate and non-compassionate agents, the first two parts of the book provide an answer to the question “why be moral?”, a question that is rarely regarded as central in contemporary metaethics. By contrast, Marshall argues that answering this question in a non-trivial way should be viewed as an integral part of defending moral realism.         

The third part of book relies on the conclusions of the first two parts to develop a novel form of moral realism, called Compassionate Moral Realism (hereafter CMR). After identifying three criteria that are deemed sufficient for a metaethical view to count as a realist, Marshall moves on to argue that CMR satisfies these criteria. First, he argues that his conclusions about being in touch with some kinds of pain together with a conceptual truth called “Bad Enough”, according to which “[s]omething is objectively bad if anyone who was in touch with it would feel would feel adverse to it” (p. 196), imply that it is literally true that some pains are objectively bad – hence, CMR satisfies the semantic criterion. Second, Marshall claims that CMR also satisfies the metaphysical criterion given that, on this view, there is at least one stance-independent moral fact, i.e. the fact that some pains are objectively bad. Third, he points out that, given the epistemic asymmetry between compassionate and non-compassionate agents, CMR has no trouble satisfying the epistemic criterion, which requires that paradigmatically bad agents are missing out on an irreplaceable epistemic good. Furthermore, Marshall argues that both its naturalism and the fact it can make sense of an internal link between moral representation and motivation count in favor of the plausibility of CMR.

I found Marshall’s Compassionate Moral Realism insightful and thought-provoking. The book is highly recommended to those interested in exploring new resources that can advance the debate over moral realism. Marshal’s discussion of the phenomenon of compassion is wide-ranging and displays a vivid philosophical imagination. However, his attempt to defend moral realism by appealing to our capacity for compassion is not free from difficulties. In what follows, I will outline two worries about his view.

First, it is not entirely clear what it means for compassionate reactions to “sufficiently resemble” the states to which they respond. On the one hand, Marshall claims that “[t]he main point of resemblance between compassionate experience and pain is their affective aspect, their unpleasantness” (p. 72) and that what matters for his argument is that “we can have certain motivational states that match other motivational states” (p. 14). Yet if compassion is understood merely as being in a motivating negative affective state, then the criterion for sufficient resemblance is too weak. Once we allow different types of psychological pain into the picture, it is hard to see how Marshall’s test for whether a compassionate experience reveals a certain property (pp. 57, 68) could be met. More concretely, if compassion consisted merely in being in a negative affective state irrespective of the specific negative affective state of the target (e.g. sadness, grief or anxiety), then a subject who had no other representational encounter with a property (such as sadness, grief or anxiety) wouldn’t know what that property is like. On the other hand, if compassionate states are viewed as isomorphic with the affective states at which they are directed, we end up with a criterion for similarity that is too strong. Marshall is well aware of the challenges implicit in defending such an epistemologically ambitious view and seems to opt for a middle way between the two aforementioned extremes, suggesting that that both resemblance and revelation of properties “come in degrees” (p. 72).

The question is whether Marshall’s account of compassion can steer a middle way between the above-mentioned extremes while retaining both its initial appeal and its plausibility. What allows Marshall to hold that there is a sense in which only compassionate agents can fully grasp reality (p. 161) and that compassion is an epistemic achievement, is precisely the claim that compassionate states match the states to which they respond. Absent such a requirement, his view would ultimately collapse into a less attractive view of our affective responses to other creatures’ pain. Note that understanding compassion merely in terms of being in an affective state would make Marshall’s view undistinguishable from other views centered on sympathetic moral concern. Note further that, even if compassion is understood as being in a negative affective state, adding this further requirement does not suffice to exclude certain cases that cannot properly count as cases of compassion. Consider an example in which A reacts to B‘s suffering by being angry at B‘s being treating badly. In such a case, A‘s reaction does not qualify as compassion even though A is in a negative affective state. A‘s anger is connected to B‘s situation, but not to B‘s state of mind, so perhaps such cases can be excluded by introducing a further requirement concerning the object of compassionate reactions. Interestingly, however, some authors view anger in such situations as an indispensable vehicle of knowledge. (See, for instance, Roberts and Wood (2007: 53)). It is also worth stressing that, although Marshall agrees with such authors about the preoccupation with propositional knowledge being too narrow and about the importance of emotional acquaintance, appealing to an emotion such as anger is not well suited for the purpose of defending realism, since anger can help us attain epistemic goods, but it can also lead us astray. Nor does it serve this purpose to appeal to sympathetic concern. The question is whether an appeal to compassion understood simply as being in a negative affective state (directed at a proper object) can do the required work.

Now, compare a compassionate reaction to the grief experienced by someone who has lost a loved one to a compassionate reaction to the sadness of a child who cannot take part in a skiing competition because of a sprained ankle. There is something odd about viewing compassionate responses to such cases as uniform in intensity and affective quality while claiming that compassion allows us to fully grasp certain aspects of reality. A further complication arises in connection to the suggestion that there is a neat separation between propositional knowledge and the distinctive form of knowledge derived from perceiving others’ affective states. The more we insist on this separation, the less plausible it becomes that compassionate responses vary in intensity and affective quality depending on the states to which they respond. If our affective responses to grief or sadness are not informed by how we understand the details of the case (e.g. by inferences about the cause and severity of psychological suffering), then the view that our responses can match these states in intensity or affective quality turns out to be too demanding and controversial. (Our capacity to experience others’ affective states would indeed have to be similar to “flash sonar”, i.e. a technique discussed by Marshall which allows the visually impaired to experience spatial properties by echolocation. The reliability of this capacity would be nothing short of remarkable given its reliance on such an indirect mechanism, when on a more natural picture of what happens in cases involving psychological pain, our affective responses cannot be so easily teased apart from the process of inferring or imagining others’ affective states).

I turn now to a second, related worry. Marshall’s argument in defense of the thesis that compassion provides a distinct epistemic good is developed along several lines. The idea behind his strategy is that, for the main argument to go through, it is not necessary that all lines of argument developed be successful. He characterizes the relevant epistemic good in Lockean terms, relying also on phenomenological and neurological considerations to support his view. Those unconvinced by one line of argument might find another one compelling. Yet, if successful, each line of argument adds strength to the conclusion. This strategy is problematic because we are left with the impression that Marshall can always retreat to a weaker, more modest position, when, in fact, his position is quite a bold one. His discussion of Lockean resemblance-based epistemic goods encourages optimism about our capacity to perceive others’ mental states, going so far as to suggest that compassion “may be our only capacity for perceiving external objects as they are in themselves” (p. 164). It is doubtful that phenomenological considerations can help vindicate this account of compassion. The difficulty stems not only from the fact that Marshall uses a success-requiring notion of compassion, but also from the fact that he views compassion as a perceptual, non-inferential process, and, as the literature on perceptual knowledge makes clear, introspection is not a reliable guide to whether we are engaged in inferential processes. Are considerations about mirror neurons meant to provide decisive support for Marshall’s view? No argument is provided to support the idea that, in cases involving psychological suffering, mirroring enables us to perceive others’ mental states in the rich sense of “perceiving” that is required for compassionate experiences to reveal mental properties (i.e. a subject who had no other representational encounter with a certain property would know, in virtue of such an experience, what that property is like). Do considerations about mirroring in cases of physical pain, together with “Bad Enough”, go all the way to establishing that CMR satisfies the semantic and metaphysical criteria for realism? Is the mere possibility of mirroring in cases of physical pain enough to render plausible a response-dependent version of moral realism? One would expect a more thorough discussion of these issues. Furthermore, it is far from clear that CMR can meet the epistemic criterion for moral realism if the proponent of CMR retreats to a modest position according to which compassionate responses to others’ pain consists merely in being in a negative affective state. While some would undoubtedly find seductive the idea that compassionate agents are epistemically better off because they perceive other suffering creatures as they really are, as long as compassion is understood as being in just any (motivating) negative affective state, it is difficult to resist the suspicion that this account of the epistemic good provided by compassion is tailor-made for the realist’s purposes.

To conclude, Marshall’s book makes a valuable contribution both to the literature on compassion and metaethics. Many of the arguments presented deserve further discussion and readers will benefit regardless of whether Marshall’s attempt to develop a novel form of moral realism by appeal to our capacity for compassion is successful.

 

References

Roberts, Robert C., and Wood, W. Jay. (2007). Intellectual Virtues: An Essay in Regulative Epistemology. Oxford: Oxford University Press.

 

© 2019 Laura-Janett Vasile

 

Laura-Janett Vasile, PhD in philosophy from the Central European University in Budapest