Can Animals Be Moral? 10 Years On, Part 2 Response to Streiffer

Can Animals Be Moral? 10 Years On, Part 2 Response to Streiffer Robert Streiffer (Mind, Vol.125, No. 498, April 2016, pp. 619–623) levels several interesting objections to the account I developed in CAM. I don’t know if its perversity on my part, but I seem inclined to deal with them in reverse order, since, to me at least, the most serious objections seem to come last (i.e., first in the order I propose to address them). I don’t know if Streiffer would accept this judgment of severity, of course. But here they are, my responses from last to first. In his last objection, Streiffer is unconvinced by my account of moral sensitivity – sensitivity to the good- and bad-making features of situations – that, I argue, can supply a sufficient condition of moral motivation. My thesis was that animals can possess morally laden emotions – emotions with moral content. An emotion has moral content if it stands in a certain relation to a moral proposition. In explaining this relation, I first introduced the notion of an emotion being misguided. An emotion is misguided if it presupposes a moral entitlement that does not really exist (as opposed to being misplaced if it is based on an erroneous assumption of fact). Smith’s indignation that Jones snubbed him, for example, is misguided if Smith has no right to expect otherwise from Jones, given Smith’s unpleasant behavior at their last meeting. Then, the concept of a morally laden emotion can be defined as follows: an emotion, E, is morally laden if (a) there is a particular moral proposition, M, and (b) the non-misguided status of E guarantees the truth of M. (It is the non-misguided status of the emotion that guarantees the truth of moral proposition, not the emotion itself). When an animal acts on the basis of a morally laden emotion, I argued, the emotion provides a moral motivation – a moral reason – for its action: acting on the basis of such emotions is a sufficient condition for moral motivation. I took, and continue to take, no stand on 1 whether it is a necessary condition – largely because I was not, and continue not to be, interested in that sort of thing. My project in CAM was to establish that animals can act for moral reasons by providing a sufficient condition for such action and arguing that animals meet it (or that there is no reason to suppose they can’t). Identification of necessary reasons for moral action was not my goal. I emphasize this because I think Streiffer’s final objection is best viewed in this context. He asks us to consider a device that is sensitive to the good- and bad-making features of a situation: “But consider an automated postoperative pain-management system. The system detects the patient’s pain and responds by administering appropriate doses of morphine. The system’s sensitivity can be normatively assessed because it can do a better or worse job of tracking some of the good- or bad-making features of the situation and we can suppose that its sensitivity and responses are grounded in a reliable module. Rowlands’ view implies the false conclusion that this system is a moral subject.” In fact, my view implies no such thing. Streiffer’s argument here seems to be based on assuming my position is expressed by way of a universal quantifier, when, in fact it is expressed in terms of an existential quantifier. There are various ways in which one might be sensitive to good- and bad-making features of situations. Animals exhibit one of these: they can be, I argued, emotionally sensitive to such features and this ultimately underwrites their possession of moral motivations and, so their status as moral subjects. I was very clear that, “This is intended as a sufficient condition for an individual to qualify as a moral subject … It is not intended as a necessary condition.” (p. 231). Further, “Moral subjects are ones sensitive to the good- and bad-making features of situations in the sense 2 that they entertain intentional content emotionally.” (p. 231). A postoperative pain-management system does not entertain intentional content emotionally and so is not a moral subject in this sense. I am also willing to allow that there are other forms of sensitivity that can underwrite an individual’s status as a moral subject. I am, of course, skeptical of the traditional picture of moral subjecthood as underwritten by critical scrutiny of one’s motivations, largely because I don’t think this model yields what it is supposed to yield – namely, control over one’s motivations. But, sice I don’t think moral subjecthood has much to do with control anyway, I am happy to allow that if there were person who did not display emotional sensitivity to morally salient feature of situations but who did habitually scrutinize their desires, examining them in the light of moral principles they antecedently accept, then this person would also be a moral subject – a status they have achieved by another route. This route we might dub, for want of a better word, rational sensitivity. Rational sensitivity, of the sort described, would be another sufficient condition of moral motivation. There might be additional forms of sensitivity. Perhaps, God, if He existed, is aware of the morally salient features of situations in a way that is neither rational (in the sense described above – i.e., He doesn’t have to work it out) nor emotional – a form of divine sensitivity, if you will. So, I am clearly committed to this: there is some x (indeed, more than one x) such that x is a form of sensitivity to morally salient features of situations and x underwrites moral subjecthood. However, it does not follow from this that for all x, if x is a form of sensitivity to morally salient features of situations then x underwrites moral subjecthood. One would only think this does follow if one has confused existential and universal quantifiers. The postoperative pain-management mechanism displays neither emotional nor rational (nor, for that matter, divine) sensitivity to morally salient features of situations. There is no reason, on my account, why it should be thought of as a moral subject. The only reason one would think that my account entails this is if one 3 mistakenly believes I am committed to the universally quantified formulation, for all x, where x is a form of sensitivity to good- and bad-making features – x is moral sensitivity. In providing only a sufficient condition for moral motivation I am emphatically not committed to this. I am not sure why Streiffer thinks I am committed to the universally quantified – any form of sensitivity will do – formulation. But he does cite my remark that I intend my condition to be, ‘neutral with respect to the type or mode of sensitivity involved’. This remark, however, is clearly not an endorsement of the universally quantified, ‘any form of sensitivity will do’, idea. Rather, it is an expression of the idea that I neither want to rule in nor rule out the possibility of other forms of moral sensitivity. There is some x (probably more than one x), such that x is a form of sensitivity to morally salient features and x underwrites moral subjecthood. It does not follow from this that any form of sensitivity to morally salient features will do. If my goal had been to provide a necessary condition of moral sensitivity, then I would be able to say which forms qualify as moral sensitivity and which do not. But my concerns in CAM lay elsewhere. Streiffer’s expresses his penultimate argument as follows: “Third, there is a related concern that Rowlands sufficient condition fails to correctly classify certain individuals as moral subjects. Suppose that Stephen is a conscientious desire-satisfaction theorist who strives to regulate his behaviour in light of his judgments about what is moral and so is reliably sensitive to desire-satisfaction. Suppose, though, that the correct theory is a Scanlonian view according to which desire-satisfaction is not a good-making feature. Stephen does not satisfy Rowlands sufficient condition but, regardless, he is a moral subject.” Yes, I agree that Stephen is a moral subject. The question, however, is why is he a moral subject? That is, what makes him a moral subject? There is more than one way of being a moral subject (this is a point that featured in my earlier response to Fitzpatrick). Stephen clearly 4 qualifies as a moral subject in a traditional Kantian/Aristotelian sense. When he is motivated to do something, he scrutinizes his motivation through his preferred desire-satisfaction moral framework. This is enough to make him a moral subject via the traditional critical scrutiny model. I have grave reservations about this model in that I don’t think it yields what it was designed to yield – control over first-order motivations. But since I think control is irrelevant to normativity anyway, I am happy to allow that if there were an individual who operated in this way, they would qualify as a moral subject. However, what, emphatically, does not make Stephen a moral subject is his emotional sensitivity to morally salient features of the situation – for, ex hypothesi (namely, Scanlon is right and the desire-satisfaction theory wrong), he has no such sensitivity. Happily for Stephen, he has found another, more traditional, route to moral subjecthood: he, therefore, is a moral subject, and there is nothing in my account that entails he is not. Streiffer’s second argument: “Second, an argument against the moral subjecthood of animals could appeal to their lack of moral concepts rather than their lack of control: motivation counts as moral motivation only if it arises from a mental state with explicitly moral content. It isn’t enough that the appropriateness of one’s emotions presupposes the truth of a moral proposition; rather, one is a moral subject only to the extent that one has moral judgments that play a motivational role in one’s psychology.” It is not as if I didn’t consider this objection. In fact, this kind of objection is the target of the long, and admittedly rather dense, Chapter 2, ‘Attributing emotions to animals’. The concern of this chapter is how to understand the mental states of animals when the concepts we must use (viz., our concepts) are not ones that animals can plausibly be thought to deploy. The framework I developed there was based on the notion of tracking. This applied an idea familiar from 5 situated semantics that the contents of animal mental states can be understood as functions from contexts to contents. When a dog barks at a squirrel that has evaded it by running up a tree, the dog is in a state with a (narrow) content such that were this state inserted into our context – a psychological milieu made up of a network of concepts and associated judgments – it would have the (broad) content that the squirrel is in the tree. In such a case, relative to a context, the (narrow) content of my dog’s mental state guarantees the truth of the proposition used to attribute the (broad) content to my dog. The result is an account of animal behavior that is sensitive to, and explainable in terms of, our concepts but is not driven by our concepts. Morally laden emotions, I argued, should be understood by way of a specific instance of this general strategy: the non-misguided status of an emotional state that forms an animal’s motivation guarantees the truth of a given moral proposition. This truth-guaranteeing role is what makes the motivational state a morally laden emotion and allows the animal to act on the basis of this emotion even if it does not understand the moral proposition and the concepts employed. The result of a view of animal behavior that is sensitive to moral concepts without being driven by moral concepts. One may hate this situated account, of course. But since it was developed with a view to heading off the lack of moral concepts objection that Streiffer levels here, it is strange that he doesn’t mention it. Streiffer’s first objection: “… while I share Rowlands view that regress-related concerns about metacognition are fatal for higher-order thought theories of consciousness, the jury is very much out on whether they are fatal for theories of free will. So one might deny, along compatibilist lines, that a lack of control over higher-order mental states implies a lack of control over lower-order mental states.” This objection is seems to be another – mutated – version of what I called the miracle-of-the-meta. One specific version of this is the idea that normative 6 control over a first-order states can be introduced via appeal to a higher-order state – when the same issue of control arises for that higher-order state. This, I argued, leads either to a regress or the assumption that something miraculous happens at the meta-level. Streiffer appeals to compatibilism to undermine this regress. This is (a) curious and (b) simply another version of the miracle-of-the-meta, and the two are connected. Streiffer’s appeal to compatibilism is curious because the appeal to a meta-level to explain some feature of a first-order state or states is typically an admission that compatibilism has failed at the first-order level. Streiffer mentions the free will debate, which is not quite what I was talking about, but let’s work with this. Suppose I am tempted by a traditional, first-order version of compatibilism: a desire, decision or choice is free, I claim, when it is caused in the right way (whatever that is). This is a first-order account, it appeals only to first-order states (decisions, choices, desires) and a relation between them (causation, of the right sort). If am happy with this, there is no need to appeal to anything meta. If I do appeal to a meta-level, this will only be because I am unhappy with the first-order account. As such, this appeal is simply another version of the miracle-of-the-meta. What I am saying, in effect, is this: Okay, I accept that compatibilism has failed here, at the first-order level. But – look! – it works here, at the meta-level. In other words, compatibilism – miraculously – works at the meta-level when it did not work at the first-order level. The miraculous nature of the meta-level has been transferred from states to philosophical views, but it is, nonetheless, still just another version of the miracleof-the-meta. 7