Changing How the World Thinks

An online magazine of big ideas


Sorell vs Rowlands: Morality Beyond Humanity - part 4

Anthropocentrism is an obsession of yesterday.

Sorell vs Rowlands 45 2

Read part 1: Tom Sorell on why human beings are the only animals who act morally.
Read part 2: Mark Rowlands argues that animals can act morally – we need to
look at the evidence.
Read part 3: Sorell warns against the dangers of anthropocentrism in moral philosophy.

Tom Sorell and I are in agreement on at least one thing. As he puts it: “What I think is necessary for moral behaviour is acting for certain kinds of reason, where the reasons operate to produce the relevant behaviour.” I agree. We differ, however, on this: I think at least some (non-human) animals can act on reasons of this sort. Sorell is sceptical of this. I’m really not sure why he is, and I’m reasonably confident he shouldn’t be.

First of all, we should be aware that the notion of a moral reason for action is neither clear nor univocal. Indeed, neither is that of a reason for action more generally – whether moral or not. As Philippa Foot once said: "I am sure that I do not understand the idea of a reason for acting, and I wonder whether anyone else does either." There are different conceptions of the reasons for acting – practical reasons as philosophers call them.

Here is one well-known model of practical reason – I don’t necessarily endorse it, but it’ll do for present purposes. According to this model, a reason to act is composed of two states: a pro-attitude towards a certain thing, and a belief about how that thing might be attained. Desire is the most obvious example of a pro-attitude: I desire to respond decisively to Tom Sorell’s objections (i.e. I have a pro-attitude toward this), and I believe – I expect Sorell will think mistakenly – that writing down these words will help satisfy this desire.

Explaining a person’s behaviour in these terms is part of what is known as folk psychology, which Tom mentions late on in his reply. Some of the things he says there suggest he thinks folk psychology is not applicable to animals, or is dubiously applicable at best: “The problem is that folk psychology is designed for people, and that it risks anthropomorphising the behaviour of non-human animals.” I’m really not sure whether folk psychology was designed for people – assuming by ‘people’ Sorell means ‘humans’. This is an empirical speculation. It may be true. Or it may be that folk psychology was at least in part developed to predict or explain the behaviour of animals that early humans depended on – most obviously, the animals they hunted. But, anyway, the origins of folk psychology are irrelevant. The utility of any device is not restricted to the reason for its original invention: the device may be co-opted for other use, and may be just as good or even better in this new role. If Sorell thinks that folk psychology is genuinely inapplicable to animals this, it seems, can only be because he doesn’t think that the theoretical posits it requires – beliefs, desires and similar states – are possessed by animals. Thus, Sorell uses the word ‘anthropomorphising’ – the attribution to animals of qualities only humans can possess. In this, I suspect, he is swimming against the current. At one time, anthropomorphism was something that everyone worried about. Now, not so much. Few cognitive scientists studying the behaviour of animals now doubt that they have beliefs, desires, and other mental states of this sort. Indeed, there is an entire branch of animal science – cognitive ethology – that wouldn’t make much sense without this assumption. Anthropomorphism, of this sort at least, is an obsession of yesterday. Of course, fashions change, but I am pretty certain of at least one thing: no convincing argument has ever been given against the claim that animals have states such as beliefs or desires. (Some very good philosophers have tried. Their arguments were poor).

Perhaps, however, Sorell’s scepticism pertains to the idea that an animal can possess specifically moral reasons for what it does – rather than reasons in general. If this is his position, then he is in much better company. But I still think he is mistaken.

There is a tendency that philosophical disputes have to degenerate into verbal disputes. This happens when there are disputed conceptions of a given phenomenon. The question of whether animals can act for moral reasons is in danger of becoming one of those if there are disputed conceptions of moral reasons (and believe me there are) for then it may be that animals can act for moral reasons in one sense of that expression but not in another of its senses. That is, I think, the case. So, to avoid the possibility of my disagreement with Sorell degenerating into a merely semantic or verbal one, let me explain clearly what I mean when I claim that animals can act for moral reasons.

First, at least some animals – many of the social mammals are the most obvious examples, but I wouldn’t want to restrict it to them – possess a mechanism that functions in a certain way. It provides animals with the ability to respond empathically to other animals. Mirror neurons are the most likely candidates for neural realisers of this mechanism. Mirror neurons mirror. That is their function. Thus, if an animal is confronted with, for example, distress behaviour exhibited by a conspecific, it will – if the mirror neuron system is engaged – feel distress also.

Suppose, second, that this sort of empathic response tracks – not perfectly by any means, but sufficiently – the features of situations that make them good or bad. The conspecific’s distress behaviour is the result of its suffering, and this is a bad-making feature of the situation (i.e. suffering, once it reaches a certain level of size and intensity, is a bad thing – not all the time, but usually). The empathic monkey’s distress, in this sense, tracks a bad-making feature of the situation. When the animal responds empathically to the distress of another, this is because distress is typically – not always but often enough – a bad thing.


"Do animals possess the relevant moral mechanisms? Do they possess mirror neurons? These are empirical, testable, questions."

Third, the empathic response is directed at the distress-behaviour of the other, and not merely caused by it: it is distressed at the other’s distress (or, if you prefer, at the other’s distress-behaviour – it makes little difference in my overall account). To be distressed at something is to want it to end. Thus, built into the animal’s empathic response is a desire to ameliorate the distress of the other.

Once these conditions are met – empathic emotional response, grounded in reliable mechanism, that tracks (albeit imperfectly) the good- or bad-making features of situations, we have a moral reason – or, at least, we have one sort of moral reason. It is the sort of moral reason that many humans act on, and when they do their resulting behaviour is moral behaviour (or, there is no compelling justification for supposing it is not).

With this brief account of a moral reason in hand – I go into more detail in my book Can Animals Be Moral? – let me return briefly to Sorell’s remaining objections.

First, he writes: “Mark Rowlands needs to explain how his account does not legislate away deflation by selfish gene explanations or other kids of explanation.” In fact, I legislate away nothing. The above account is empirically testable and, therefore, hostage to empirical fortune.

With regard to selfish-gene explanations, part of the reason that interest in the possibility of moral behaviour in animals exploded in recent years is because there is strong evidence of such behaviour in cases where there is clearly no genetic relation. Consider, for example, the various (it happens more than you would expect!) boy-falls-into-gorilla-enclosure stories, where the unconscious boy was met with a seemingly sympathetic response on the part of the gorillas. And when Carole Zahn-Waxler conducted her classic studies on sibling autism, one of the things that emerged was that the family dog would often respond to the feigned suffering of the mother in ways that seemed as empathetic as the child. These merely scratch the surface. The evidence of apparently moral behaviour in cases of non-genetically related individuals is large and growing. This precludes any general selfish-gene type of deflation.

There are other forms of deflation. When Binti Jua – a gorilla in Brookfield Zoo – rescued an unconscious boy who had fallen into her enclosure, carrying him to the access gate where she handed him over the a keeper, attempts at deflation were rife, and often almost endearingly desperate: perhaps, some said, she was just doing what she had been trained to do with her doll (Binti Jua was born in captivity and zoo workers taught her mothering skills using a doll). Perhaps, yes – but refusing to accept that she can distinguish between a boy and a doll (something that a dog can do with 100% accuracy) does carry with it the distinct aroma of desperation.

Deflationary explanations can be true in some cases but no reason whatsoever has been given for thinking they are true in all. The crucial point, however, is that this is an empirical matter. Given the account of moral reasons I have defended, the presence of moral reasons in animals is testable – even if this testing is not straightforward. Do animals possess the relevant mechanisms? Do they possess mirror neurons? These are empirical, testable, questions. Do animals exhibit evidence of empathic response to the distress of other animals? Again, this is an empirical question. Is this empathic response directed at the distress of another rather than merely caused by it. Again, testable – we simply need to look at the relevant counterfactuals. For example, rats, like Capuchins, refuse food in the presence of conspecific distress, but they also do so in the presence of white noise – this suggest that we should consider this behaviour in rats merely aversive stimulus behaviour rather than empathic response (although there are also counterarguments). None of this is beyond our ability to test, to garner evidence one way or another. The question of whether animals can act morally does require much conceptual ground clearing – we need to know what a moral reason is, for example. But once this has been done, the question is ultimately empirically testable. Legislation on my part is neither required nor desired.

Sorell’s final remarks concern my appropriation of sentimentalism. This is understandable. When he thinks of sentimentalism, he thinks (naturally enough) of David Hume. I tend to think of my colleague, Michael Slote, whose book, Moral Sentimentalism, I take to be a canonical statement of contemporary sentimentalism. So, on the view I prefer, sentimentalism is entirely compatible with the existence of moral reasons. Indeed, the sort of empathic emotional response to the distress of another that an animal (or human) might exhibit is precisely such a reason.

Sorell writes: “This requirement smuggles in a near relation of the view that emotions have to stand to reason or a dispassionate view, which may reintroduce moral reasons and, on Rowlands’s view, the slippery slope toward intellectualism.” I reject the implicit premise that my view is incompatible with moral reasons. On the contrary, it assumes there are such things and that animals can possess them. Emotions are what put us, and some other animals, in touch with such reasons, and they do so because they track (partially and imperfectly but sufficiently) the morally salient – good- and bad-making – features of situations.

Read part 1: Tom Sorell on why human beings are the only animals who act morally.
Read part 2: Mark Rowlands argues that animals can act morally – we need to
look at the evidence.
Read part 3: Sorell warns against the dangers of anthropocentrism in moral philosophy.

Image credit: Alexander Cahlenstein

Latest Releases
Join the conversation