1 Duties of Respect Towards Moral Non-agents?

So far, I have suggested a way to understand the Categorical Imperative so that animals are not inherently excluded from moral concern (Chap. 4), and I have proposed an alternative to Kant’s view of directionality, according to which there can be duties towards animals (Chap. 5). It would be convenient if we could now simply apply Kant’s taxonomy of duties from the Doctrine of Virtue to the domain of animals as moral patients. We would effectively have transformed Kant’s oeuvre on ethics into an oeuvre on animal ethics. Unfortunately, Kant’s tacit (and sometimes not-so-tacit) assumption that all moral patients are finite rational beings leads to two additional difficulties. First, Kant assumes that all moral agents and patients are subjects of pure practical reason, who are equal in moral potential. This gives rise to a special class of duties—duties of respect—which demand that we do not exalt ourselves above others (MM 6:449.32). More specifically, duties of respect rule out not only being arrogant or contemptuous towards others, but also being overbearingly beneficent. Thus, duties of respect put a crucial limit on beneficence and put an important qualifier to the demands of duties of love. This gives rise to the characteristically Kantian idea that we should seek the right balance or reconciliation between beneficence and non-exaltation.

Secondly, as we saw in Chap. 2, Kant’s conception of practical love revolves around the idea that others pursue their own happiness as a matter of instrumental rationality. Because others set their ends with an eye to happiness, our best way of promoting their happiness is to support their self-chosen ends. But animals, at least as Kant sees them, are incapable of consciously pursuing their happiness as an end. Indeed, they pursue no ‘ends’ at all, being entirely steered by blind instinct (CPJ 5:172.10). Thus, Kant’s discussion of duties towards others rests on the assumption that others are subjects of both pure practical reason and instrumental practical reason. In this chapter, I will argue that with some modifications to the taxonomy, we can accommodate the ideas of duties of respect and of love even towards animals. Considering these duties in more detail also helps to see what our duties towards animals demand according to Kantianism for Animals.

Let me begin with the first difficulty, which concerns duties of respect. As we have already seen in Chap. 2, Kant describes these duties as a force of “repulsion” (MM 6:470.05), in that they ask us not to encroach upon others. They act as an important counterweight to duties of love, which on their own would simply demand that we benefit others without any limit. The reason why we ought not to encroach upon others is that, in doing so, we would be ‘exalting’ ourselves above them (see MM 6:449.32). In Kant’s view, generosity burdens the beneficiary with the expectation of politeness and flattery (Collins 27:341f.), putting her on an unequal footing with the benefactor. The very existence of duties of respect is thus predicated, first, on the assumption of moral equality, and secondly on moral agency and social expectations. But none of this appears to apply to animals, who are neither moral agents nor our moral equals.

As mentioned in Chap. 2, duties of respect present a taxonomical puzzle (Fahmy 2013). Kant categorises them as duties towards others, yet they do not appear to derive from the obligatory end of the happiness of others. Indeed, an important point about these duties is that they limit the extent to which we should be beneficent towards others. Kant also categorises duties of respect as duties of virtue, yet they do not appear to prescribe an end we should adopt, only the negative condition (MM 6:449.32; Baron 2002, 399) that we should not exalt ourselves above others in various ways. Whatever the solution to this taxonomical puzzle is, Kant clearly takes the recognition of another’s humanity—more importantly, their equal humanity!—to give rise to a duty not to exalt ourselves above them.

At first sight, it is hard to see how something like a Kantian duty of respect towards others could be justified along the lines of Kantianism for Animals. In this framework, a duty counts as a duty ‘towards’ others if, and only if, it derives from the obligatory end of the happiness of others (see Sect. 5.2 above). Of course, Kantianism for Animals can ask that we do not go so blatantly overboard in our attempts at beneficence that they backfire and make others less happy. But this is less a counterweight to the demands of duties of love than a built-in restriction of beneficence itself. What is more, animals are not our moral equals. They do not stand under a moral law, and they do not share our moral predicament. There simply seems to be no equality to be acknowledged here. So how could there arise a duty to acknowledge equality and not exalt ourselves above them? Do we have to abandon entirely the idea of a reconciliation of love and non-exaltation, of beneficence and recognition of equality?

The answer, I want to suggest, is no. We can preserve a part of the idea that we should not exalt ourselves above others even if those others are animals. We can even preserve the idea that this duty of non-exaltation is connected to the recognition of something like equality. But we need to depart from Kant one more time in order to do so in a way that considers animals. The first step to the solution is to put the duty of non-exaltation in another part of Kant’s taxonomy. In Kantianism for Animals, we should understand this duty not as a duty towards others, but as a duty to self, merely regarding others.Footnote 1 This represents another departure from Kant (though one concerning a point on which, as we have seen, his taxonomy stands in need of clarification anyway). Of course, this move is not without a certain irony, given that Kant made the same move the other way around, declaring our most important duties regarding animals to be duties to self. Conversely, in Kantianism for Animals, it is precisely those duties that would appear to hold only towards human beings, in virtue of their equal humanity, that are truly duties to self. If they still appear like duties towards others, that is an amphiboly in moral concepts of reflection.

Duties to self, on the conception outlined in Chap. 5, are those which derive from the obligatory end of our own moral perfection. This ‘perfection’ consists, on the one hand, in keeping ourselves in good moral shape as duty-observers (e.g. by not increasing our inclinations by overindulging in food, drugs, or sexuality, MM 6:427.01–428.26; MM 6:424.09–425.36). On the other hand, it consists in recognising, and acting in accordance with, proper moral self-esteem. We should not lower ourselves below others with whom we are truly “on a footing of equality” (MM 6:434.04), for instance by being servile (MM 6:434.20–436.13). But it seems perfectly plausible that the end of moral perfection, so understood, also requires that we do not exalt ourselves above our equals. In being arrogant, we “expect other human beings to esteem themselves but little in comparison with us” (MM 6:465.12–13). This expectation, just like the converse expectation that others should not respect us as potential observers of the moral law, is not appropriate to our equal status as moral agents. So the duty of non-exaltation fits the description of a duty to self in Kant’s framework very straightforwardly.

Recategorising the duty of non-exaltation as a duty to self makes sense on another count too: It emphasises how this duty is truly about a form of humility concerning ourselves, as much as it is about other-regarding stances. The duties Kant lists as specifications—not to be arrogant, not to backbite, not to ridicule others—are fundamentally about the recognition of a relation between equals. Another upside of the recategorisation I advocate is that the duty of non-exaltation can still play the role of a counterweight or limiting condition to the demands of duties of love, since in general duties to self and duties to others must be reconciled (see Vogt 2009, 238).

However, it must be said that conceiving of duties of respect as duties to self also presents some apparent problems. First, ordinary moral experience would suggest that we owe it to others not to be arrogant towards them or exalt ourselves above them in other ways. Is it not strange to claim that we truly only wrong ourselves by being arrogant? Consider, however, that Kantianism for Animals does not have to claim that acting from arrogance in a way that harms others only wrongs ourselves. As soon as the happiness of others is affected, our duties to others give us a foothold for moral criticism. What is at issue here is the duty we violate merely by having an expectation that others should view themselves as below us, the inner exaltation by which we see ourselves as inherently more worthy than others. That this type of exaltation should violate only a duty to self does not seem so strange.

Secondly, there is the further problem that Kant views only duties of respect as owed duties, while duties of love are meritorious (see Chap. 2). If we recategorise duties of respect as duties to self, only meritorious duties to others remain. Does that mean that we never owe others anything? This would again appear to conflict with ordinary moral experience, where we do seem to owe a certain treatment to others (including animals). However, recall that when Kant calls a duty ‘meritorious’, he does not mean to say it is not a duty. Others can morally object to our failure to observe meritorious duties, and those most affected by our failings are often in the best position to point them out. So, importantly, the meritorious is not the supererogatory. Kant draws the owed-meritorious distinction mainly by appeal to whether observance of a duty puts the other under a reciprocal obligation (say, to be grateful). But we should not expect this to be a very important distinction in an ethical system designed to capture duties towards animals, who can never acquire obligations anyway.

When it comes to capturing the ordinary experience of feeling like we ‘owe’ another some treatment, we are better served by a perfect-imperfect distinction than Kant’s owed-meritorious distinction. This perfect-imperfect distinction applies to our duties of love towards others as much as to our duties to self. What we ‘owe’ others is what anyone could expect us to do, a duty which we can have no good reason not to observe. That is a feature characteristic of perfect duties, which we can observe in every single instance. Imperfect duties, by contrast, are those that require a choice concerning when and how to observe them. Although anyone could expect us to observe our imperfect duties at some point in some way, there can be good reasons in terms of other (perfect or imperfect) duties that keep us from observing them in any particular instance. Now, since duties of virtue according to Kant prescribe the adoption of a certain end, they are always ‘wide’ and can get in each other’s way. However, to every duty of virtue there correspond ‘vices’—stances such as arrogance or contempt, which run counter to the observance of our duties. While the positive duty to adopt a certain end is never truly perfect, the duty not to adopt and act on a contravening vice is perfect. Even if we cannot always, without restriction, adopt an end of beneficence or sympathetic participation, we can always not take pleasure in another’s misfortune. To capture the feeling of owing others something, we should appeal to such a perfect-imperfect distinction as it applies to our duties of love and their corresponding vices.

Assuming that what Kant calls ‘duties of respect’ can be recategorised as duties to self, there remains the problem that these duties hinge on the recognition of moral equality. The reason why we should not go overboard in our generosity is that generosity puts the beneficiary under duties and expectations of gratitude, lowering their status vis à vis the benefactor. Neither are animals our moral equals, nor is there an expectation that animals should be grateful for any benefits they receive from human beings.

However, even though expectations of gratitude play no role in our moral relations to animals, there can still be a duty on our part not to exalt ourselves over them, neither in the sense of taking an arrogant or contemptuous stance towards them, nor in the sense of going overboard in our beneficence out of delusions of grandeur. Granted, we can have no Kantian duty to recognise, and act in accordance with, the moral equality of beings who are not our moral equals. But why should there not instead be a duty to recognise, and act in accordance with, the fundamental moral incomparability of a being capable, and one incapable, of morality? The moral inequality between human beings and animals, as Cholbi has pointed out (2014, 348), lies not just in the fact that animals lack the potential for moral goodness, but also in that they lack the propensity to evil. Not standing under a moral law, they can be neither good nor bad when they go about their actions. Hence, they are neither morally better nor morally worse than human beings. But we fail to act in accordance with this fundamental moral inequality when we view animals as morally worse or less deserving of their happiness or less important qua moral patients than human beings. So we have a Kantian duty not to exalt ourselves above animals, at least not in any sense that would have us view ourselves as ‘more worthy’ of our happiness than they are of theirs (for, in contrast to human beings, they do not need to be worthy of their happiness). Kant’s egalitarianism—the injunction not to exalt ourselves above others and hence not to be overbearingly beneficent upon them—can be drawn from the fundamental inequality between moral agents and non-agents, as much as from the fundamental equality between moral agents.

To be sure, the duty of non-exaltation will make different demands on us depending on whether we regard human beings or animals. To stay as close as possible to Kant, we can endorse his consideration that other human beings feel constrained or humiliated by overbearing beneficence (MM 6:448.24–25) and that we should strive not to embarrass others or otherwise lower their moral self-esteem. This is of no concern regarding animals who neither have nor need moral self-esteem in the first place. Still, the duty of non-exaltation asks us to be beneficent towards animals only from the right stance. We should act with a kind of humility which acknowledges that we are fallible, finite creatures, capable of doing good but always liable to doing wrong. We should not regard ourselves as would-be demigods of whose gracious assistance animals always need more and never less. The recognition of our own limitations puts a certain restriction on the way in which we should benefit animals, and this might impact the extent to which we try to benefit them as well. However, this restriction is less tight than that on our beneficence towards other human beings.

One might wonder why the move I suggest should apply only to our treatment of animals, and not to the whole rest of nature. Rocks and trees too are neither morally better nor morally worse than human beings, since they do not stand under the moral law. However, rocks and trees are not subjects of happiness, hence no moral patients according to Kantianism for Animals. In this moral framework, all specific duties towards others stem from the obligatory end of promoting their happiness (that is what makes them duties ‘towards others’). More specific duties are essentially specifications as to how we should promote this obligatory end. Here, the moral inequality between human beings and animals gives rise to a specific qualification that we ought not to exalt ourselves above animals in our promotion of their happiness. But that does not imply that we have duties of respect to things that are incapable of happiness.

2 Adopting Another’s Ends as Our Own

Kant’s system is built on the assumption that the ‘others’ to whom we have duties are rational beings in yet another way still. Consider that Kant asks us to ‘adopt another’s end as our own’ (see MM 6:388.05–08; G 4:430.24–27; MM 6:340.03–05; Sect. 2.5). This formulation is not incidental. Kant purposely lays an emphasis on what others want for themselves, not on what we happen to think they should want. The reason for this emphasis is not, mind you, that instrumental practical reason has any kind of moral value which we must honour by furthering its ends and means. Kant makes it very clear that instrumental practical reason alone does not elevate the human being above other things in the world (CPrR 5:061.32–062.01). The reason why we ought to promote the ends of others is more banal: Our chief duty towards others is to promote their hedonic happiness, and the (non-moral) ends others set by means of instrumental practical reason serve as means to their hedonic happiness as well. So the most straightforward way to promote the happiness of others is usually to help along their own endeavours. But of course, this is only the case because others set their ends in an instrumental-rational way, as a means to their happiness. So despite the fact that instrumental practical reason has or produces no moral value in itself, it has a mediatory role to play in moral patienthood according to Kant’s conception.

The idea has some intuitive pull that our chief duty towards animals, too, is to help along their own endeavours. It would be attractive if this feature of Kant’s framework could be preserved in Kantianism for Animals. However, difficulties soon arise: Kant would deny that animals have any ‘ends’ whatsoever which we could adopt. So it is quite simply impossible to give them the Kantian treatment without amending Kant’s picture to some extent. This becomes clear once we consider Kant’s various remarks on animal behaviour (Sect. 6.3). However, I am going to argue that we can grant, within a Kantian framework, that animals have necessitating states that are non-conceptual, or ‘conceptual’ in another sense than Kant’s, or which qualify as ‘obscure’, even if they are not proper ‘ends’ in Kant’s technical sense (Sect. 6.4). So there is more than enough in terms of ‘ends’ we could adopt.

Furthermore, even if we grant that animals have such necessitating states, they do not stand in an instrumental-rational relation to their happiness. So even if we could help along their endeavours, the question remains why we should. In response, I am going to propose another line of reasoning in favour of concern for the necessitating states of animals. What matters most for Kant’s ethics, I argue, is that another’s ends serve an indicative function for their happiness. Though the motivations of animals may not serve this indicative function due to instrumental practical reason, they serve the same purpose based on biological functionality. In other words, animals may not plan to promote their own happiness by means of their actions, but many of their behaviours serve a biological function for their health and flourishing, which in turn stand in close relation to their happiness. The upshot of this argument is the view that the promotion of animal happiness is primarily a matter of helping along animals’ own endeavours. This again gets us close to Kant’s original picture. Before concluding, I want to take a moment to explore how much of Kant’s specific list of ethical duties towards others we can transfer upon animals, based on the arguments in this chapter.

3 Kant’s Denial of End-Directed Animal Agency

Although he never dedicated a separate piece of writing to the topic of animal behaviour, Kant has a somewhat detailed and coherent account of the differences between human and animal agency. The first and major difference, of course, is that human beings are transcendentally free, animals are not (A534/B562). Human beings can at least consider acting against their strongest natural impulse by acting from duty (CPrR 5:030.22–35). Acting from duty is however only possible through practical autonomy—the capacity to act on a self-imposed law independent of all natural laws. Animals, since they are not autonomous, lack this capacity. Their will is necessitated by sensibility, not merely affected by it (A534/B562). Along similar lines runs a claim recorded in the Mongrovius lecture notes: “Animals have no free choice, their actions being necessarily determined by their sensory impulses” (Mongrovius 29:611). Kant also calls this mode of agency the “animal power of choice (arbitrium brutum)” (A534/B562, see Mongrovius 29.611), as opposed to the human power of choice (“arbitrium sensitivum, yet not brutum but liberum”, ibid.).

However, this view would still permit Kant to grant that animals have a limited form of instrumental practical reason. This would amount to the view that animals are ultimately bound to act on their sensory impulses, but are able to strategise to some extent, resisting momentary impulses for the sake of greater satisfaction later on. In fact, Kant acknowledges that there could conceivably be creatures who have all the capacities of instrumental practical reason, but who are still unfree (CPrR 5:449.04–07).

But in Kant’s view, animals are not such creatures. This is clear in a passage from the Collins lecture notes: “Animals are necessitated per stimulos, so that a dog must eat if he is hungry and has something in front of him; but the human being, in the same situation, can restrain himself” (Collins 27:267). In Kant’s own words from Syllogismen, a dog always acts according to “the natural connection which exists between its drives and its representations” (Syllogismen 2:060.08). That is to say, there is a hardwired behaviour for any sensation (which is all “representation” refers to at this point). And so, while human beings act on a complex system of ends and considerations, animals simply act on the next best impulse. Reason, even purely instrumental practical reason, does not enter the picture.

The absence of reason in animal behaviour has deeper implications: Not only can animals not strategise in their choice of means to preconceived ends, but they cannot pursue any ends at all. Acting for the sake of ends is itself an exercise of practical reason (MM 6:385.01–04). And this connection between reason and end-directed agency is not established by Kant arbitrarily. Reason is required for end-directed agency because acting on ends is essentially a conceptual capacity for Kant. Ends are always conceptually structured (see Guyer 2006, 349; Graband 2015, 16). This, in turn, is because ends are the objects of an intentional state, and they receive their object-shape from concepts. As much is clear from Kant’s explicit definition: “An end is an object of free choice, the representation of which determines it to an action (by which the object is brought about)” (MM 6:384.33–34). The kind of object Kant has in mind here is a logical object of an intentional state, not necessarily a physical object in the natural world. To be sure, it could be a physical object, such as the cup of coffee I intend to produce by using the coffee maker. But the happiness of others, which we are supposed to make our end, is certainly not a physical object. What is important is rather that ends are clearly represented somethings picked out by means of concepts. They could be physical objects, substances, processes, states, properties, or really any referent of a clear representation.

However, Kant assumes that animals lack concepts (as McLear points out, McLear 2011, 4). Of course, this implies that animal actions cannot be guided by clear representations, even though awareness of perceptions may still influence their behaviour. Whatever animals do, they do it unknowingly, without a goal in view. As Kant puts it in the Anthropology, animals “manage provisionally” (Anth 7:196.26).

It appears, then, that Kant is committed to the following inference:

Argument against end-directed agency in animals

  1. (1)

    Ends are objects.

  2. (2)

    Representing objects requires concepts.

  3. (3)

    Animals lack concepts.

  4. (4)

    Therefore, animals cannot represent ends.

But if Kant accepts (4), how does he think animals do what they do? In a word, by instinct. Kant gives us a coherent picture of instinctive animal behaviour, even if it has attracted little attention in the literature.Footnote 2 It is a picture of animal behaviour that does entirely without end-directed agency or any other conceptual capacities. For the rest of this section, let me flesh out this picture before discussing alternative Kantian views in the next section.

In Kant’s view, what sets the causes of animal actions apart from ends is their non-conceptual form. As he puts it in the third Critique: “The will, as the faculty of desire, is one of the many kinds of natural causes in the world, namely that which operates in accordance with concepts” (CPJ 5:172.04–06, emphasis added). So what is distinctive about the will is precisely that it operates with concepts. Kant then goes on to contrast the will with an example of a cause which operates without concepts and names animal instinct:

everything that is represented as possible (or necessary) through a will is called practically possible (or necessary), in distinction from the physical possibility or necessity of an effect to which the cause is not determined to causality through concepts (but rather, as in the case of lifeless matter, through mechanism, or, in the case of animals, through instinct). (CPJ 5:172.06–11, emphasis added)

As this passage also reveals, Kant does not strictly equate instinct with mechanism. There is still something special about what Kant calls ‘pathological necessitation’ or ‘psychological causality’. In psychological causality, not just physical bodies can be causal relata, but also representations (CPrR 5:069.30–31). The notion of ‘representation’ at work here is very broad: Mere sensations count too. Kant intends to reject what he takes to be Descartes’s view, namely, that animals are mere bodily machines devoid of sensation (CPJ 5:464FN).Footnote 3 So even though animal action is caused by non-conceptual causes, they have psychological causes.

Kant’s views about the causes of animal actions have certain implications concerning how those actions work. As we have already seen, end-directed human agency has some object clearly ‘in view’. Agents can only be clearly aware of what they strive for if they represent it by relying on a concept. Without the concept of a cup of coffee, one cannot strive, with awareness, to obtain a cup of coffee. So instinct, as Kant also puts it, is “blind” (MM 6:376.22). Kant makes instinct’s ‘blindness’ particularly salient in a passage from the Anthropology. Here he posits that instinct is an in-between category between mere propensity and inclination. Propensity is the “subjective possibility of the emergence of a certain desire, which precedes the representation of its object” (Anth 7:265.21–22; see Rel 6:028FN). In other words, propensities are dispositions to desire that require no awareness of what we may come to desire. Kant’s own example is the propensity to drink: We can be disposed to desire alcohol long before we try our first sip (Rel 6:028FN). Now, instinct is like a mere propensity in that it requires no clear representation of an object. The difference is merely that instinct already involves a feeling of desire: Instinct is “the inner necessitation of the faculty of desire to take possession of this object before one even knows it” (Anth 7:265.23–24, emphasis removed). One of Kant’s examples is the instinct to procreate (Anth 7:265.25): An animal in heat feels something that causes it to act a certain way. But it does not have to think about any particular mate, or about mating, or about the young produced as a result. Desire is hardwired to activity.

To sum up, Kant does not believe that animals are things in nature like any other, but that they are subject to a specifically psychological form of causality. What distinguishes them from human beings, however, is their lack of concepts. Although animals may act on ‘instincts’, they cannot orient their behaviour on clear representations of objects in the way human beings do. This presents a challenge for those of us who want to include animals in Kant’s account of moral concern, which is tied to the idea of ‘adopting another’s ends as our own’.

4 Animal ‘Ends’: Conceptual, Non-conceptual, ‘Obscure’

It is not uncommon in Kantian animal ethics to accept a more generous picture of animal agency than Kant’s. Korsgaard (2018), for instance, speaks of animals’ ‘ends’ throughout her book. She even asserts that “an animal just is a being that takes its own functional good as the end of action” (Korsgaard 2018, 146). On another occasion, she asserts that in what she calls an ‘instinctive’ action, there always is the “animal’s own purpose”, such as “avoid the lion’s attention” for an antelope who then chooses to duck in the grass (Korsgaard 2018, 42). A ‘purpose’, at this point, is not merely a biological function, in the way protecting the cornea is the biological function of an unintentional eye closure reflex. The animal’s own purpose is something the animal itself represents as the goal of its action. The difference to the ends of a rational being is merely that rational beings are capable of choosing ends for themselves, while animals merely represent theirs.

How substantial a divergence from Kant does it represent if we assume that animals have ‘ends’? It depends on what exactly we mean. As we have just seen in Sect. 6.3, Kant ties end-directed agency to conceptual capacities. When Kant speaks of ‘ends’, what he has in mind are object-shaped, conceptually structured representations that determine action. If we want to be able to claim that the Kantian injunction to adopt the ends of others applies to animals, it appears we have to disagree with Kant either on whether animals have concepts or on whether end-directed agency requires concepts. However, there is a third option. We can claim that animals have ‘ends’ in another sense than Kant’s, but that these ‘ends’ can still play the role that matters for Kantian interpersonal ethics. In fact, we can make this move in several different ways. We can claim that animals have non-conceptual necessitating states or necessitating states that are ‘conceptual’ in another sense than Kant’s, or so-called obscure ends. They do have a perspective on their own actions with which we can and should be morally concerned.

First, however, what would be the trouble with simply ascribing Kantian concepts to animals? The problem is that we would be attributing too much for it to be plausible. In order to have an end ‘Kantian-style’ to avoid a lion, antelopes would need to represent the lion by means of an ‘empirical’ concept, which is a concept whose content is an object given in experience. One important role of empirical concepts lies in allowing the subsumption of specific representations by the power of judgement (Heinz 2015). It allows a rational being to lend unity to what is initially a manifold of intuition. The other important role of empirical concepts is to be the building blocks of judgements, such as ‘lions are dangerous’. But in order to play either role—for subsumption or for judgement—it is crucial that empirical concepts be general. That is, their application should be largely independent of context. Whoever can subsume a certain manifold of intuition under lion in one context should also be able to subsume the same manifold under lion in other contexts. And whoever has the concept of lion must be able to form and understand various judgements about lions—not just ‘lions are dangerous’, but also ‘lions are to be avoided’, or ‘lions avoid fire’, provided the concepts are known. In the language of analytic philosophy, having Kantian concepts is a capacity with a very strong Generality Constraint (McLear 2016, 182; see Evans 1982). That is to say, to have a Kantian concept is to be able to combine it with other concepts, or with other manifolds of intuition, with only few limitations.Footnote 4

Ascribing a capacity with such a strong Generality Constraint to animals like antelopes is implausible. If antelopes do have a concept of lion, a concept of danger, and presumably some other concepts too, it seems that they only use their conceptual vocabulary in tightly restricted ways. They seem to keep judging that lions are dangerous, but they do not combine this judgement with other judgements in a way that would enable them to develop more elaborate strategies of avoiding lions than merely by ducking in the grass. If antelopes could flexibly recombine concepts, we should expect much more productivity and systematicity from their thought (see Beck 2012, 222ff.). The more plausible view, by far, is that antelope thoughts and actions are not structured in a way Kant would accept as conceptual. They may be capable of ‘conceptual’ thought in another sense of the term, but it is uncontroversial that they lack concepts in Kant’s sense. And since Kantian ‘ends’ are structured by Kantian concepts, it is implausible to ascribe Kantian ‘ends’ to animals.

But of course, Kantian ‘ends’ are merely a species of a larger class of action-determining or necessitating states. So even if we deny that animals pursue ‘ends’ in Kant’s technical sense, we can ascribe to animals ‘ends’ in the sense of some other, non-conceptual species of necessitating state. Of course, the instincts Kant ascribes to animals already fit this description, since they are non-conceptual necessitating states. However, the vocabulary of instinct can obscure the true complexity of animal behaviour. Within the broad class of ‘instincts’, we might find necessitating states which consist of discrete parts that can be systematically recombined to a certain extent, or they might be based on some capacity to stably discriminate between Xs and non-Xs (Beck 2012, 222). Certain animals may even be capable of drawing inferences and thinking in a means-ends-rational way within certain contexts (so-called islands of rationality, Hurley 2003, 2006). In short, animals’ necessitating states may have a lot more to them than Kant’s account of instinct would have it, all without Kantian concepts.

Another way to expand Kant’s view of animal agency is to grant that animals are capable of a non-conceptual mode of objective perceptual awareness (see McLear 2011, 2016, 2020). For philosophers interested in Kant’s views on perception, to claim that object perception is possible without concepts may seem problematic. After all, Kant gives great weight to the claim that concepts, and ultimately the categories, are a condition of the possibility of object perception (Kitcher 1984; Naragon 1990). However, we need to be clear about what kind of perceptual object awareness we are concerned with. As McLear points out, Kant himself at times asserts that animals are in some ways acquainted with objects (McLear 2011, 5). This acquaintance comes in the form of what Kant calls ‘obscure’ or ‘unconscious’ representations (Jäsche 9:33.25–26; see McLear 2011, 6). Such representations provide some sensory awareness, but do not enable us to individuate the object. For example, we might be aware of a violin’s sound in an orchestra, but still be unable to pick it out individually (McLear 2011, 6). Similarly, an antelope may be aware of a lion in its environment without however being able to individuate the lion in a way that would allow, say, reidentification of the lion in another environment. But if animals can have such ‘obscure’ representations, we may as well grant them ‘obscure ends’ structured by just these representations. For instance, an antelope may want to get away from lion-environments, though not from individuated lions. So even within the restrictions of a Kantian philosophy of perception, we can accommodate a more sophisticated view of animal behaviour than Kant himself advances.

To sum up, I have suggested that animals do not plausibly pursue ‘ends’ in Kant’s narrow, technical sense, but they can still have necessitating states that are non-conceptual, or ‘conceptual’ in another sense than Kant’s, or ‘obscure’. There is enough here in the way of ‘ends’ that we could adopt if a Kantian framework demanded it.

A second issue remains, however: Even if animals have ‘ends’ of some kind, why should these ‘ends’ concern us? If animals’ ends are to be morally relevant in the same way as the ends of Kantian rational beings, they need to be relevantly similar. Here, the crucial question is what makes our ends ‘valuable’ in the sense that others have a duty to adopt them. Korsgaard, with her influential interpretation of Kantian ethics, begins this explanation from the perspective of the moral patient. As we saw in Sect. 4.3, Korsgaard argues that we confer goodness upon our ends by rationally choosing and pursuing them. Her argument then revolves around showing that we must grant a version of this goodness-conferring capacity (‘end-in-itselfhood’ in Korsgaard’s sense) to all animals, since animals pursue their natural ends in much the same way as we do (Korsgaard 2004, 102–6, 2018, 143f.).

By contrast, the reading of Kant’s ethics I have suggested focuses steadfastly on the moral agent: We have a duty to adopt the ends of others, not because of any morally important feature of those ends themselves, but because our duty is to promote another’s happiness. The ends of human beings should concern us only because we can presume that these ends point towards the subject’s happiness in an instrumentally rational way. We set most of our ends merely as means to further ends, and the ultimate (non-moral) end is our own happiness. Thus, if we are looking for means to promote the happiness of others, we are well advised to use their ends as indicators.

Kant emphasises that we can be beneficent to others “only according to the concepts of him whom I would like to render a benefit” (MM 6:454.20–21). Fittingly, however, he at the very same time puts a restriction on his own demand that we let others judge for themselves what makes them happy:

I cannot be beneficent to anyone according to my concepts of happiness (except for children during their minority or the mentally disturbed), but only according to the concepts of him whom I would like to render a benefit by urging a gift upon him. (MM 6:454.18–21, emphasis removed and added)

Kant’s point, evidently, is that only certain adult human beings are fully competent judges of their own happiness. Only with regard to these specific human beings should we understand our duty of beneficence to demand that we help along the realisation of their own ends. Of course, we can and should disagree with Kant about the true capacities of those of us who are young or have disabilities, and about the extent to which they should get to determine the course of their own lives. For animals, however, the case seems fairly clear: It is implausible that animals set their ends (in whatever sense of the term) in an instrumentally rational way across the board, so that they all serve the ultimate end of happiness. At best, certain animals are capable of instrumental reasoning within certain contexts (see again Hurley 2003, 2006).

According to this understanding of Kant’s ethics, however, it is not truly instrumental rationality that does the most work. The instrumentally rational relation that connects our ends to our happiness itself matters only because it makes our ends indicators of our happiness. The question is hence not whether animals’ ends stand in an instrumentally rational relation to their happiness, but whether they indicate what makes animals happy. And they clearly do, at least to a significant extent.

There is actually a way of making this point in a distinctively Kantian way. It consists of two steps: first, in distinguishing between two types of functionality, practical and natural. Though animal behaviours may not ‘aim’ at happiness in the practical sense (so that happiness is the highest end in an instrumentally rational hierarchy of ends), they can still ‘aim’ at happiness, or at least at some of its prerequisites, as a matter of natural or biological functionality. Secondly, the solution consists in acknowledging that animal behaviours actually do serve biological purposes closely related to the animal’s hedonic happiness. As I will argue in the remainder of this section, both steps are readily possible within Kant’s framework. The upshot is that we should think of beneficence towards animals largely as the promotion of their own ‘ends’, not because animals choose these ends as means to their happiness, but because they serve the same indicative function for biological reasons.Footnote 5

Kant himself devotes much attention to the question whether there is purposiveness in nature in the second part of the Critique of the Power of Judgment, called the Critique of the Teleological Power of Judgment. He discusses ‘relative purposiveness’, which we might nowadays call ecological functionality, as well as ‘internal purposiveness’, which translates more closely to biological functionality (CPrR 5:366f.). Kant subsumes both types of functionality under the general label ‘end of nature’ (Naturzweck). Though Kant rejects realism about natural ends (CPJ 5:394.13), he vindicates a teleological perspective on nature on the grounds that “because of the peculiar constitution of my cognitive faculties I cannot judge about the possibility of those things and their generation except by thinking of a cause for these that acts in accordance with intentions” (CPrR 5:397.34–398.02). Hence, though Kant does not believe that ecological functionality is a matter of anyone’s intentions, let alone God’s, he considers teleological judgement to be indispensable for human beings.

To be sure, the perspective on animals we need in order to include them in Kantian moral concern has little to do with their status as a ‘relative means’ for other things in nature. What counts instead are the functional relations between animals’ ends on the one hand, and their happiness on the other. Even if there may not be an instrumental-rational hierarchy of practical ends that leads from momentary actions all the way to the ultimate end of happiness, there can be a string of biological functions of the same extent. In this sense, animals do ‘pursue’ their happiness, even if they cannot represent their own happiness in general as a practical end.

In more contemporary language, we can say that many animal behaviours and the necessitating states that cause them serve a biological function for prerequisites of the animal’s happiness. For example, food approach and wanting-food are biologically functional for nutrition, and nutrition of course contributes to hunger satisfaction and health, both of which typically promote the happiness of the animal. So even though a certain practical-functional order is absent in what many animals want, there is more than enough of a biological-functional order to make animals’ necessitating states important moral guideposts for us.

When it comes to the claim that animal behaviours actually do serve biological functions closely related to their happiness, Kant would not have to disagree. In fact, in the Religion he claims that human beings have an entirely pre-rational mode of self-love, which he terms ‘physical and merely mechanical self-love’ (Rel 6:026.13, see Rinne 2018, 22). This mode of self-love promotes unwitting behaviours that are biologically conducive to self-preservation, species propagation, and community with other human beings (Rel 6:026.14–18). Nothing stops Kant from endorsing the same view with regard to animals.

What is more, to some extent Kant already associates instinct with happiness, as a well-known passage from the Groundwork reveals:

Now in a being that has reason and a will, if the proper end of nature were its preservation, its welfare, in a word its happiness, then nature would have hit upon a very bad arrangement in selecting the reason of the creature to carry out its purpose. For all the actions that the creature has to perform for this purpose, and the whole rule of its conduct, would be marked out for it far more accurately by instinct. (G 4:395.04–16)

To be precise, Kant here only claims that nature could have created some happiness-conducive instinct in human beings, not that the actual instincts of animals produce their happiness. His main point is that if the purpose of reason lied in the natural world, that purpose would have been better served by some hardwired natural mechanism. Still, Kant’s view would easily allow him to claim that animal actions typically do serve to produce their happiness, or at least prerequisites of their happiness such as self-preservation and species-typical community.

To add a caveat, note that the argument in this section does not demand that we promote just any animal impulse, just as Kant himself does not demand that we help along the next best end others happen to have. In certain cases, the actions of animals and human beings might fail to indicate what makes them happy overall. For instance, an obese animal may pursue food intake too vigorously. This is a perfectly functional type of behaviour, but in this instance, it is a threat to health. A threat to health is usually a threat to happiness. So we ought not to help along this behaviour. Within these reasonable restrictions, however, we should take an animal’s own ends as guideposts for our beneficence.

Before moving on, let me address a potential worry. The move of the present section is to liken the biological functionality of certain animal behaviours to the practical functionality of human actions as Kant thinks of them. This might be a red flag to some Kantians, and addressing it may well be useful for animal ethicists too. As Altman warns:

If natural purposiveness is what is morally relevant, then all living organisms are directly morally considerable because things may frustrate or promote their teleological development. Although most environmental ethicists accept this, it takes us a long way from Kant. (Altman 2011, 25f.)

The reason it takes us a long way from Kant is that for Kant, moral concern hinges on the mutual sharing of the moral law. Altman is right to caution against acting as though an internal biological-functional structure could simply replace the moral law in Kant’s picture of interpersonal moral obligation. So I should emphasise that this is not the move I am advocating. What I have proposed (in Chap. 5) is that we can replace Kant’s second-personal view of interpersonal moral obligation with a thoroughly first-personal one. According to this view, what gives rise to our ethical duties towards others is our own autonomy, paired with the fact that others are subjects of hedonic happiness. To this line of reasoning, the present section adds an argument why a good Kantian agent should pay attention to animal behaviours and necessitating states. Though these may not be part of an elaborate, instrumental-rational project conceived to promote the subject’s happiness (as Kant thinks human actions are), they still serve biological purposes that relate to the animal’s happiness. So this chapter’s arguments do not imply that the line of moral concern is to be drawn at biological-functional structures. Nor does it assume that any kind of inherent value or value-conferring power attaches to biological functionality. Rather, they imply that in our treatment of beings with a happiness at stake, the biological functionality of their behaviours gives us reason to adopt their ‘ends’ as our own, at least within certain reasonable restrictions.

With these considerations ends the core argument for Kantianism for Animals. I have argued that the formulations of the Categorical Imperative do not present a serious obstacle to the inclusion of animals, since they do not settle the issue of who deserves moral concern. We need not tamper with this central part of Kant’s ethical system. What we must indeed change is Kant’s understanding of the directionality of ethical duties. I have argued that we can think of duties ‘towards’ others simply as duties that derive from the obligatory end of the happiness of others. This end is obligatory because it is a part of the highest good in the world, one that we can actually bring about by means of our actions. But animals’ happiness belongs to the highest good in the world, since it is a happiness that is not tainted by vice. Therefore, on this amended Kantian view, we have duties towards animals. The final step was then to remove the last anthropocentrisms from Kant’s system, particularly concerning the notion of duties of respect and regard for the self-chosen ends of the individual. These features do not need to be purged from Kant’s system, however, but can be accommodated in an animal-friendly way.

Along the way, I hope to have conveyed a glimpse into the kind of ethical outlook that Kantianism for Animals can bring to the table in animal ethics. It provides an account of what it means for ethical duties to be directed ‘towards’ an animal, and of what we ought to do, expressed at different levels of generality (adopting specific ends of others, adopting practical-emotional stances, promoting happiness). Having developed the Kantianism for Animals framework in general, let me try to situate it in the theoretical landscape of animal ethics and highlight what makes the framework distinctive. That is the purpose of the next chapter.