1 Introduction
In the Metaphysics of Morals, Kant famously compares ‘the great moral forces’ of mutual love and respect that bind rational beings into a moral community to the physical forces of attraction and repulsion that hold together the physical world (cf. MM, 6: 449).Footnote 1 Kant’s treatment of our moral duties towards others – duties of love and respect – is central to his system of ethics. Yet, according to one persistent criticism raised in recent years (most prominently perhaps in the work of Darwall Reference Darwall2006, Reference Darwall2012, and Wallace Reference Wallace2007), Kant’s emphasis on self-legislation issuing from the first-personal deliberative standpoint alone makes him unable to account for the intrinsically second-personal, or ‘bipolar’, normativity that characterizes our duties and rights towards others. Bipolar obligations involve a relation between two agents – an obligee who is obligated and an obligator who imposes the obligation and to whom it is owed. These two ‘poles’ of a moral relation are bound in a way that gives an obligator a special standing to demand certain actions from an obligee as a matter of right.Footnote 2
Drawing on ideas from contractualist and other social theories of morality, critics worry that Kant’s emphasis on individual autonomy results in a theory that treats others merely as ‘beneficiaries’ of our dutiful actions, rather than as persons who have a distinctive authority to make claims on us.Footnote 3 The problem can be put this way. Standing in a bipolar relation of duty and right requires agents on both sides of this relation to stand (and take one another to stand) under the very same law, a law they must find mutually acceptable and binding. Yet, according to Kant, the authority and content of the particular moral law that I will is grounded in my autonomous practical commitment (my being bound by my own pure reason). It then appears that I simply owe it to myself, or to no one in particular, to act in a certain way, but I cannot genuinely say that I owe it to you (or any particular other person). Similarly, I cannot demand from you, as a matter of my right against you, that you act in a certain way. Call this the original version of the bipolarity objection.
Michael Thompson offers a novel neo-Aristotelian version of this objection (Thompson Reference Thompson2004). Contra Darwall and other second-person theorists, Thompson grants that Kant does have an answer to the original ‘bipolarity’ objection. He argues, however, that this answer comes at an unacceptably high cost – it is underwritten by untenable metaphysical commitments that are incompatible with an even ‘mildly naturalistic’ outlook, and, therefore, with ordinary practical thought. Central to his argument is the claim that Kant’s metaphysics of pure practical reason (as it abstracts from all natural-causal or socio-historical relations) renders ordinary bipolar judgements empty of determinate meaning. The novelty of this way of framing the bipolarity objection, and its apparent force, lies in the way it connects the concerns about Kant’s treatments of bipolar normativity with the older, more entrenched charge of ‘empty formalism’, now reframed as the problem of incompatibility with ordinary practical life.
I argue that none of these criticisms work against Kant: they fail to appreciate the inherently practical character of Kant’s approach to ethics – the extent to which his moral theory is based on the revolutionary idea that the nature of our moral practical relations can be properly understood only from the standpoint of the activity of embodied rational agency in practical thought and action. This activity unites nature and freedom, imbuing our shared natural world with normativity. In particular, its essentially inter-subjective form makes bipolar deontic relations possible. Recognizing this fact helps to shed new light on Kant’s conception of practical agency and its place in his system of morals.
I proceed as follows. In section 2 I begin by briefly sketching Thompson’s bipolarity objection. Proper assessment of this critique requires a closer look at Kant’s conception of a person. In section 3 I argue that central to Kant’s distinctively practical approach in practical philosophy is the claim that the categorical imperative applies to all rational beings because it determines, for us, what counts as a rational being, a person, in the first place. Moreover, the existence of particular moral relations of duty and right towards other persons depends essentially on the possibility of interaction with these persons, or others like them. Indeed, Kant’s conception of moral agency implies, rather than denies a place for, a shared natural-causal and socio-historical background in his system of morals, or so I shall argue. In section 4 I conclude that failure to appreciate Kant’s resources in the face of the bipolarity objection rests on a more general failure to appreciate Kant’s inherently practical outlook in moral philosophy.
2 A Neo-Aristotelian Argument, or ‘Sylvia and a Martian’
Thompson’s criticism provides an interesting lens through which we can consider Kant’s views on the concept of a person and the determination of bipolar moral content. He invites us to consider a human being – call her Sylvia – and to reflect on some generally unproblematic moral judgements about her, such as: ‘it would be wrong for you, me, or anyone else, to kill Sylvia for any prudential or charitable or any other aim’. He suggests that our intuitive agreement with these judgements points to the fact that you (I, or any other person) bear a special kind of practical relation to Sylvia – a bipolar relation of right, in which you and Sylvia are two opposing poles.Footnote 4 All practical thought-explanation having to do with your standing in this kind of relation to Sylvia expresses ‘the special sort of dent that Sylvia herself is making on your agency’ (Thompson Reference Thompson2004: 347).
Thompson suggests that wherever two agents are apt to be represented in true bipolar moral judgements, they stand together in a particular ‘order of right’. A person is an agent that falls under a particular order of right. This relational concept of a person is formally distinct from a ‘thin’ conception of an agent, such as ‘something that operates on the strength of practical reasons or thoughts or considerations’, a ‘realizer of concepts’, or the like. In order to imagine an interaction between two such ‘thin’ agents we need only suppose that each of them sees himself and the other as an agent and can attribute to himself and the other some aims. Such agents would treat each other as a kind of ‘raw material for wrong doing’, as ‘something with respect to which either agent might “mess up” instrumentally’. The normativity in play here is a merely monadic normativity of each agent’s own ends. On the other hand, the conception of a person who sees herself as standing in bipolar normative relations to others is a de-relativization of a prior relational concept X is a person in relation to Y (just as, for example, X is a friend is a de-relativization of X is a friend of Y). All bipolar judgements presuppose this relational conception of a person.
Thompson argues that for two subjects to stand in a bipolar normative relation, both must be the kind of persons who are typically capable of thinking that they stand in that relation, and their two thoughts must concern the very same relation, grounded in the same ‘order of right’, rather than one subject thinking of one relation and the other thinking of a different, twin relation. On the ‘mildly naturalistic’ view he favours, there must be a common natural cause of their two graspings, something (such as a common ancestry or history) that makes it no accident that two thinkers grasp the very same concept.
Now Thompson claims that Kant conceives of persons under the moral law abstractly – as rational beings, or rational animals who can act on principle, or Kantian ‘humanity’-havers. The key feature of an abstract concept of a person is that two agents falling under such a concept can stand ‘in complete natural-causal independence of one another, without any shared dependence on anything we can understand as a common source’ (Thompson Reference Thompson2004: 364). There is no naturalistically intelligible homogeneity among beings falling under abstract concepts of this kind. In this they contrast with such ‘concrete’ concepts as human, Martian or Norway rat. Thompson’s worry is that Kant cannot explain how individuals falling under his concept of a person can non-accidentally have the same bipolar moral thoughts and, therefore, share an order of right. To show why this presents a problem for Kant’s account of our duties towards other human beings, he offers the following thought experiment:
Sylvia and a Martian
1. A Kantian agent sees herself as bound not to kill Sylvia ‘in just the same way’ in which she is bound not to destroy any rational animal being, be it a Martian or a twin Earthling.
2. Even if these other rational beings have bipolar concepts, the Kantian agent cannot in principle deploy the same bipolar concepts as they do, since we can talk of deploying the very same concepts only in the context of a shared natural-causal or socio-historical background.
3. If the agent is required to think that she stands in the very same moral relation to a possible rational alien as that in which she stands to Sylvia, then she cannot form a meaningful thought about her relation to Sylvia as a relation of right either.Footnote 5 This kind of thought is, in effect, empty.Footnote 6
Kant can account for bipolar normative relations, Thompson argues, only by positing a single intelligible cause – pure practical reason – as the ground of possibility of the very same thoughts being thought by different agents; ‘a Platonic form’ that instantiates itself in the thoughts of various rational beings. This takes care of the original ‘bipolarity’ problem, but at an unacceptably high cost. Kant’s underlying metaphysical commitments (his ‘strange and wonderful’ metaphysics of pure practical reason’: Thompson Reference Thompson2004: 383) are incompatible with an even mildly naturalistic outlook, and, therefore, with ordinary practical thought. ‘I have nothing with which to oppose this orthodox Kantian conception of the matter’, Thompson writes, ‘apart from a mild naturalism. But that is enough, I think we should grant, to make a serious difficulty’ (Thompson Reference Thompson2004: 383).
If correct, Thompson’s critique does raise a serious difficulty for Kant’s account of moral obligation; a difficulty that has damaging implications for Kantian ethics in general. Indeed, a Kantian seems to face an unhappy dilemma: either accept Kant’s ‘strange and wonderful’ metaphysics of pure practical reason and stand in conflict with even mild naturalism, or jettison this metaphysics and lose the capacity to make sense of bipolar normativity.
I shall argue, however, that Thompson’s objection misses the mark because it considers Kant’s conception of moral agency from a standpoint orthogonal to the practical perspective inherent in this conception. In the next section, I explain how this practical perspective is reflected in Kant’s conception of a person and in his account of our duties towards others.
3 Kantian Persons – Who are they?
3.1 The Categorical Imperative and Kant’s Concept of a Person
Kant’s revolutionary shift from attempting to derive moral requirements from an antecedently defined concept of the good to grounding our conception and knowledge of the good in the fundamental principle of morality is often described as the Copernican transformation in ethics. The relation between the categorical imperative and Kant’s conception of a person is a crucial aspect of this transformation.
In the Groundwork of the Metaphysics of Morals, Kant describes the formulations of the categorical imperative as different ways of representing the moral law (G, 4: 436). In a sense, these formulations represent different fundamental aspects of the activity of pure practical reason that brings together rational and sensible elements of finite practical agency – the form of rational willing and the sensibly affected power of choice. In the third section of the Groundwork, Kant offers an illuminating description of the synthetic character of the categorical imperative:
[T]his categorical ought represents a synthetic proposition a priori, since to my will affected by sensible desires there is added the idea of the same will but belonging to the world of the understanding – a will pure and practical of itself, which contains the supreme condition … of the former will; this is roughly like the way in which concepts of the understanding, which by themselves signify nothing but lawful form in general, are added to intuitions of the world of sense and thereby make possible synthetic propositions a priori on which all cognition of a nature rests. (G, 4: 454; my emphasis)
The parallel between the role of the understanding in theoretical cognition (in making possible the transcendental laws of nature known as the analogies of experience, for example) and that of practical reason (in making possible moral laws) is significant. The analogies of experience form an a priori conceptual framework that makes experience possible, and first allows us to distinguish merely subjective associations of perceptions from objective theoretical judgements.Footnote 7 For Kant, this a priori framework is the basis on which our cognition of nature rests. The formulations of the categorical imperative play a formally analogous role in grounding practical cognition. They constitute the basic conceptual framework which gives determinate meaning to the distinction between subjective and objective practical judgements (judgements that are valid merely for the subject who holds them and judgements that are always valid for all). This framework first allows us to make sense of fundamental moral concepts such as duty, right, the good, and the very concept of moral agency.Footnote 8 The categorical imperative is the sole criterion for drawing the distinctions between what merely seems to be a duty, a right, a good and what actually is; and it tells us that a person is just the kind of being who is able, and is bound, to draw these distinctions and to apply basic moral concepts in action. A person, says Kant, is ‘a subject whose actions can be imputed to him’ (MM, 6: 223), ‘the subject of a morally practical reason’ (6: 434), ‘a being capable of rights, who can encounter wrong or can consciously do it, and who stands under the categorical imperative’ (OP, 22: 55). That is to say that the categorical imperative first specifies what it means to be a finite rational agent with a will, a person, and thus it constitutes the moral space within which actual mutual recognition, shared deliberation, and action are first rendered intelligible. Only after moral concepts are given determinate meaning through the categorical imperative, can we properly formulate questions about them and hope to provide definitive answers.
Now consider how the categorical imperative specifies moral agency. As an imperative, it is a representation of a command to do or to omit doing something because so acting would be good. It speaks to the practical necessity of acting in a certain way and is addressed to rational beings in the order of nature who act on reasons. In its Universal Law formulation, it represents a rational being (from a first-personal perspective) as an agent who recognizes a rational requirement to act on maxims that have the form of universal principles – principles that others at the same time should be able to will. The Formula of Humanity represents the moral law from the standpoint in which each agent recognizes herself and others as ends in themselves – objective ends and ‘limiting condition[s] on all merely relative and arbitrary ends’ (G, 4: 436). This introduces a second-personal perspective by representing a rational agent as someone who recognizes herself and others as sources of rational claims on her will, as persons in relation to each other. Moreover, it considers the moral law from the point of view of those persons who would be affected by the agent’s proposed action.Footnote 9 It addresses rational beings acting towards and with each other in the same order of nature. Finally, the Formula of Autonomy represents the will as essentially self-legislative and, at the same time, co-legislative with the wills of others and exhibits this character of the will as the very reason why an agent is subject to the moral law. The Formula of Autonomy leads to a representation of a person as a member of ‘a systematic union of rational beings through common objective laws’ (G, 4: 433; my emphasis), or the kingdom of ends. This moral order is a kingdom of ends, says Kant, because the objective practical laws to which all rational beings are subject require that every member treat oneself and every other as an end in itself.Footnote 10 To treat someone this way is to recognize her as a responsible being with distinct authority to place rational demands on every rational being in the world (cf. MM, 6: 435). Kant’s conception of a person as autonomous, and as a legislative member of the kingdom of ends, brings together two key aspects of rational agency emphasized in the formulas of the Universal Law and of Humanity – a view of a person as a source of claims on the will of every rational being and as subject to the objective moral laws, including the laws that govern what we owe to each other. It represents persons as members of a moral order that encompasses an ‘order of right’.
Hence, the concept of a person that grounds Kant’s doctrine of moral duty and right is a relational concept with a complex structure constituted by the categorical imperative.
3.2 What can we Judge about our Duties towards Others?
In part II of the Metaphysics of Morals, which sets out the duties of virtue, Kant explains to whom these duties apply:
As far as reason alone can judge, a human being has duties only to human beings (himself and others), since his duty to any subject is moral constraint by that subject’s will. Hence the constraining (binding) subject must, first, be a person; and this person must, secondly, be given as an object of experience, since the human being is to strive for the end of this person’s will and this can happen only in a relation to each other of two beings that exist (for a mere thought-entity cannot be the cause of any results in terms of ends). But from all our experience we know of no being other than a human being that would be capable of obligation … A human being can therefore have no duty to any beings other than human beings; and if he thinks he has such duties, it is because of an amphiboly in his concepts of reflection, and his supposed duties to other beings is only a duty to himself. (MM, 6: 442; my bold)Footnote 11
First, note that Kant’s explicit assertion that, given what we know, we must judge that a human being has duties only to human beings already indicates that something is amiss with the set-up of Thompson’s argument involving our duties to a hypothetical Martian.
Second, consider how Kant arrives at his conclusion. Kant’s conception of duty and right relates two rational beings with wills that are both sensibly conditioned and free; beings under the categorical imperative. The ‘amphiboly’ arises when a morally necessitating relation involving another being or thought-entity is taken for a relation of moral necessitation by (duty to) that being or thought-entity, and, moreover, when this is done without proper reflection on whether a supposed obligator is a person under the categorical imperative. For example, you may think (correctly) that a stray cat that wanders around your neighbourhood ought to be treated humanely, but then (incorrectly) take this requirement to be a duty to the cat. Or you may hold that morality prohibits suicide and think this requirement is your duty to God.
While the primary target of the charge of ‘amphiboly’ is the moral character of our relations to animals, inanimate nature and ‘superhuman beings’ (e.g. God and the angels), it has further implications. A duty of a human being to any subject involves ‘moral constraint’ by that subject’s will. This implies that to take yourself to be bound by duty to someone you must be able to recognize that someone as a being with a will, and, therefore, as a person – a being under the categorical imperative. You must be able to see her as a co-legislator; a person in relation to you who recognizes that you have the same moral standing in relation to her. Moreover, recognizing another’s moral standing involves recognizing a moral requirement to ‘strive for the end of this person’s will’. This striving must be the cause of ‘results in terms of [her] ends’ (MM, 6: 434) – the efficient cause that contributes to their realization. And you must be able to see her, in turn, as morally required to strive for your ends. Thus seeing another person as a co-legislator implies the ability of both agents to understand each other’s maxims and the possibility of mutual influence – acting in accordance with one another’s maxims, contributing to each other’s ends, etc. This kind of mutual recognition, understanding and influence is possible only between persons who, as Kant puts it, are given as objects of experience; beings with a will who actually exist (and take each other to exist) in a shared natural world.Footnote 12
Now Kant’s claim that, as far as reason can judge, a constraining subject must be given as an object of experience cannot mean that we must take ourselves to have duties only to persons whom we have actually encountered. Such a claim would be independently implausible – most would agree, for example, that we are morally obligated not to randomly kill, maim or belittle people whether we have ever met them or not. It would also be an implausible reading of Kant, since duties of virtue are universal laws – they are principles whose scope encompasses ‘the whole species’ of humanity’ (MM, 6: 451) and cannot be limited to those with whom we are acquainted. Kant’s meaning comes into focus more clearly when we consider that duties of virtue are metaphysical principles – they are concerned with a priori determination of a faculty of desire as a will, the concept of which, according to Kant, ‘must still be given empirically’ (CJ, 5: 182).Footnote 13 The rational requirement that the constraining subject be also given as an object of experience is then best understood as the claim that in order to see ourselves as constrained by another person’s will, we must see the constraining subject as a being with a certain kind of will – the kind of will, the concept of which is given empirically. That is, we must see the constraining subject not merely as a rational being, but also as a phenomenon – an empirical being acting in our shared natural world, subject to the natural laws of cause and effect.
Such beings exercise what Kant calls ‘practical freedom’ – a self-conscious capacity to make their decisions operative causes in the world, or ‘the ability of reason to be of itself practical’ (MM, 6: 214), to act for reasons. Kant thinks that we can cognize practical freedom through experience. We recognize that our power of choice is not determined by mere sensory impulses and desires, that our choices are determinable by reason.Footnote 14 Moreover, we necessarily ascribe practical freedom to all others with whom we are able to interact rationally.
3.3 Interacting Rationally with Empirical Beings
Kant tells us very little about how we are able to determine which empirical beings we interact with rationally. But we can see, in outline, how a Kantian story would go. We ascribe practical rationality and, therefore, practical freedom to all others with whom we can (or judge that we can) communicate by exchanging judgements; whose expressions and responses we can recognize as reasons, as opposed to seeing these expressions and responses as behavioural signals characteristic of mere animal interaction. The relevant kind of interaction results in at least some degree of mutual and reciprocal understanding, and mutually recognized coordination of conduct. It involves the ability to discern in the other’s conduct regularities or patterns characteristic of intentional activity – to interpret his doings in terms of beliefs, desires and, most importantly, rational rules or norms characteristic of action. These patterns of behaviour and behavioural dispositions express what Kant calls the ‘empirical character’ of an agent’s power of choice (A549/B577).
In the Critique of Pure Reason, Kant describes an agent’s empirical character as
nothing but a certain causality of his reason, insofar as in its effects in appearances this reason exhibits a rule, in accordance with which one could derive the rational grounds and the actions themselves according to their kind and degree, and estimate the subjective principles of his power of choice … (A549/B577; my emphasis)
We ascribe practical rationality to empirical beings (or kinds of empirical beings) in whose external movements we are able to discern a rule-governed activity of a certain kind – to interpret them as acting on the basis of ‘subjective principles of [their] power of choice’ or maxims; as acting on putative reasons. In particular cases, we are also able to estimate what these maxims are and to assess whether they amount to genuine reasons (considerations with normative force that extends across all persons). An agent’s empirical character, according to Kant, is an appearance of his power of choice. It ‘indicates’ (yet does not give us any knowledge of) the agent’s ‘intelligible character’, the fundamental principle of his rational will as the ground of his actions (cf. A546/B574, A551/B579). ‘To every rational being having a will,’ Kant declares in Groundwork III, ‘we must necessarily lend the idea of freedom also, under which alone he acts’ (G, 4: 448, my emphasis), and, for that reason, we must think of him as actually free in a practical respect – to affirm that ‘all laws that are inseparably bound up with freedom hold for him just as if his will had been validly pronounced free also in itself and in theoretical philosophy’ (G, 4: 448). To see someone as a rational being with a will is to see him as a person – a being under the moral law.
What makes it possible for us to discern the intentional patterns in human activity, and to see empirically given beings as practically rational beings, as persons? Kant’s answer seems to be that we are able to do so only because our practical reason projects its own active structure, its own basic concepts and norms, onto the world it is trying to understand and to shape. To say this is just to give a particular expression to Kant’s ‘Copernican’ thesis that we can cognize of things a priori only what our reason itself puts into them. Any attempt to interact with others (be it human beings or, perhaps, members of alien species) must issue within the framework of our own practical rationality – the very framework that constitutes our conception of what practical thought and practical judgement is. Sometimes we succeed, as we (typically) do when we interact with other humans. When this happens, we are able to recognize others as rational agents, as persons. That is, only by projecting our own framework of practical thought onto others can we represent or recognize them as persons at all.Footnote 15
Indeed, this account of how we represent and recognize other beings as persons (in a practical sense) parallels, and extends, Kant’s view on how we represent others as thinking beings in general. In the Paralogisms chapter of the Critique of Pure Reason, he explains it this way:
[R]ight at the start it must seem strange that the condition under which I think in general, and which is therefore merely a property of my subject, is at the same time to be valid for everything that thinks … But the cause of this lies in the fact that we must necessarily ascribe to things a priori all the properties that constitute the conditions under which alone we think them. Now I cannot have the least representation of a thinking being through an external experience, but only through self-consciousness. Thus such objects are nothing further than the transference of this consciousness of mine to other things, which can be represented as thinking beings only in this way… (A346–7/B404–5; my bold)
A Kantian thinking subject is able to represent others as thinking beings only by ‘transferring’ or ‘lending’ to them, as it were, her own self-consciousness. She can judge a being to be a thinking subject only if she is able to project onto him the active structure of her own ‘I’, with the categories of the understanding as modes of its activity. Similarly, a Kantian practical agent can judge someone to be a person, a practically rational being, only if she is able to ‘transfer’ or ‘lend’ to him her own practical self-consciousness; to project onto him the active structure of her own practical thought, with the categorical imperative as its fundamental principle and basic moral concepts as modes of its activity.
We recognize practical freedom of other rational beings through experience (or the possibility of experience) of rational interaction. All moral relations between such beings involve mutual limiting of one’s own will by the will of another (by their ends and rights). We can, therefore, judge that we stand in relations of duty and right only to those beings to whom we can ascribe a will and whose exercise of ‘practical freedom’ we are able to recognize in experience.
Moreover, Kant draws an important distinction between mere thinking (or speculating) and judging. Reasoning ‘is not the same as speculating, playing with mere experiments is using reason apart from the law of reason’ (Anth, 7: 228). When we deliberate about what to do or make judgements about our moral relations to others we are not speculating, and when we speculate about thought-entities or hypothetical rational inhabitants of Mars we apply the categorical imperative, to use Kant’s term, in a ‘transcendent’ way. This is fine as an imaginative play, but it would be a mistake to treat this activity as resulting in objective practical judgements, such as bipolar judgements of duty and right. While ‘I can think whatever I like, provided only that I do not contradict myself’ (Bxxvi, n.), in judging practically I am making a knowledge claim about certain actions being right and, therefore, objectively good. Ascriptions of a will to the agents in either the subject or the object position of a bipolar practical judgement must, therefore, also be based on practical knowledge and experience, rather than on mere speculation. Our present experience, however, limits all such ascriptions to human beings. Given that we have no knowledge even of the existence of rational aliens, let alone of the nature of their cognitive capacities, and no actual ability to interact with them, a thought about our duties to such beings is not only not a practical judgement, it also has no determinate meaning.
I do not want to suggest, implausibly, that thought experiments involving hypothetical situations have no value in practical life or that they cannot help us clarify and sharpen our judgements, especially about situations that we have not yet encountered, but could in principle encounter in the future. My claim is only that thinking about such situations and practical judgement are two different things.
We are now in a position to see where and why the bipolarity objections go wrong. In the rest of this article I will address these questions, while drawing further attention to the systematic structure of Kant’s moral theory and its focus on the activity of rational agency in thought and action. I will argue that both the original bipolarity objection raised from the perspective of a broadly contractualist, social theory of morality and Thompson’s neo-Aristotelian criticism fail for a common reason: they fail to appreciate the nature of Kant’s inherently practical outlook in moral philosophy – albeit in somewhat different ways.
4 Why the Bipolarity Challenge Fails
First, recall that the original bipolarity objection emphasizes the impossibility of vindicating second-personal or bipolar normativity from the first-personal deliberative standpoint. But we have seen that Kant’s theory of moral obligation is not based on the presuppositions of a first-personal point of view alone. Rather, it is based on a practical point of view that is essentially multi-faceted. Kant’s discussion of the formulations of the categorical imperative and of their relationship in the Groundwork shows that the practical point of view combines two complementary aspects of practical agency – the first-personal and second-personal deliberative standpoints. It is the point of view of a practical agent who sees himself as a co-legislative member of a moral-practical world. That is, the categorical imperative establishes a moral order which contains pairwise bipolar relations of every person to every person. It represents all rational beings as ends in themselves, as sources of reasons recognizable by others; it says that each of us is a person in relation to every other.
On this analysis of his views on practical agency, Kant has no problem accounting for the distinctive authority of another as a constraining (or binding) agent. In becoming aware of ourselves as persons under the moral law, we at the same time become aware of the essential interdependence of our own moral standing and that of all other persons.Footnote 16 In recognizing others as co-legislators, as persons in relation to ourselves, we recognize (at least tacitly) each person’s distinctive individual authority and corresponding individual responsibility in relation to ourselves. We see ourselves as standing to each of them in mutually binding relations of duty and right. When we actually interact and co-deliberate with others, we apply these general principles to the circumstances at hand: it is you who are a person in relation to me, whose authority I recognize, whose reasons I take as normative for me, and it is you who, I trust, takes the corresponding stance towards me in this interaction of ours. In this way we fix, as it were, the particular content of moral-practical relations between us: by acting in a way that creates a particular rights claim (as e.g. in behaving ungratefully in response to a benevolent action or stepping unbidden on another’s foot).
Now there is a further important point that our discussion of Kant’s conception of practical agency brings to light. This conception grounds the systematic doctrine of ethical duties Kant develops in the second part of the Metaphysics of Morals. Indeed, it is the complex, multi-faceted structure of the practical point of view that allows Kant to establish a systematic and non-reductive relation between duties to oneself (that are essentially first-personal) and duties towards others (that are essentially second-personal) by deriving both kinds of duties from one and the same fundamental principle of morality. This, I submit, gives Kant a clear advantage over social theorists of morality who, while criticizing his supposed inability to make sense of our duties towards others, tend to discount the normative importance of the first-personal deliberative stance and, as a result, have no clear way of accommodating self-regarding duties.Footnote 17
The analysis offered in section 3 also points to a number of problems specific to Thompson’s critique of Kant’s treatment of bipolar normativity. It shows that there is no sense, for Kant, in which I can judge that I stand in the very same relation to Sylvia as I stand to a Martian, since I cannot judge that I stand in any deontic relation towards him or towards any other being outside the scope of our practical cognition. And this is, in the first place, because I cannot recognize such a being as a person. I cannot recognize him as someone whose ends I can, in principle, promote or hinder, or whose will can constrain my own. Indeed, Kant would agree with Thompson that a thought about our bipolar deontic relations to a Martian is empty. But Thompson is wrong to think that the emptiness of this thought has problematic consequences for Kant’s ability to make sense of our ordinary moral relations to human beings.
From Kant’s point of view, the very set-up of Thompson’s criticism is untenable. The idea that an agent takes herself to have duties towards a Martian implies that she is also able to recognize that Martian as a rational being with a will, and so implies that she can, in principle and to some degree, understand his movements as actions. She must be able to subsume his movements under the ‘categories of freedom’ (CPrR, 5: 64–7) – basic practical concepts that allow us to see certain beings as agents and certain events as acts of their free choice. In order to see herself as having duties towards a Martian, a Kantian agent would have to think of herself not only as standing in a natural-causal relation to this being or others like him (physically sharing some corner of the world), but also as engaged in at least an incipient shared practice with him. Yet, ex hypothesi, the latter thought is not ascribable to her.
In imagining her Martian, our agent is imagining someone who is like a human being in all the relevant respects.Footnote 18 She imagines that she understands him, that she can rationally interact with him. But this merely means that she imagines him as a member of our own practice of exchanging reasons and this is why her bipolar thought (and Thompson’s thought experiment) is intelligible at all. Without taking him into this practice, as it were, she would not be able to think of him as a person and would not be able to imagine having duties towards him at all.
Were we actually to encounter Martians, we could not consider them as rational beings unless we could enter into at least minimal communicative interaction with them – perhaps a version of a Davidsonian triangulation in which we would correlate our respective responses to similar stimuli. But in doing so, we would be starting to construct a practice together, even if at first this practice was very thin. We could come to see such a being as a person if and only if we could take him into our form of life and subsume him (and others like him) under our concept of a rational being with a will – a concept that first receives determinate content through the formulations of the categorical imperative.
This means that Kant does not, as Thompson suggests, take every rational life-form to instantiate ‘a Platonic form’ of pure practical reason. Rather, we cannot represent (and recognize) anyone as a practical reasoner, as a person in relation to ourselves, unless we can see him as sharing our basic framework of practical thought, as falling under the principles of pure practical reason and fundamental concepts that follow from these principles.Footnote 19 Moreover, we cannot take anyone as falling under the principles and concepts of pure practical reason until we judge that we are able (at least in principle) to interact with him, which presupposes exchanging moral-practical reasons, agreement, disagreement and the like. We cannot recognize alien beings as practical reasoners prior to or independently of this kind of interaction. Until we share with other beings at least a thin social practice of exchanging reasons we cannot see them as persons, and as soon as we construct such a practice, seeing them as persons in relation to us becomes a default position. This is a key aspect of Kant’s practical perspective in moral philosophy and is reflective of Kant’s critical approach in general.
If this is correct, it is hard to see how Thompson’s critique can get traction. There is a fundamental difference between Kant’s ‘Copernican’ inherently practical outlook in moral philosophy and Thompson’s (in this context) largely theoretical perspective. The former refuses to consider the concept of practical agency as genuinely meaningful apart from the activity of practical reason governed by the categorical imperative which first determines the content of this concept. Thompson goes wrong when he assumes an independent moral-theoretical perspective from which we are supposed to make comparisons between our moral relations to unknown rational beings and our relations to fellow humans, and claim identity or difference between them.
To be a person under the categorical imperative is to be able to take the practical point of view. This point of view requires us to assess whether the principles of our actions are fit to be co-legislated by all persons with whom we share a natural social world. This presupposes mutual recognition of persons as persons under the categorical imperative and the real possibility of intentional mutual influence (the possibility of both hindering and promoting one another’s actions and ends). From this perspective alone, for Kant, can we make moral judgements about what we ought to do, including judgements about what we owe to each other. It follows that any comparisons between my (or any other human being’s) duties towards Sylvia and my duties towards another person, and any claims about agreement or implication between judgements about what I owe to various persons make sense only when made within this practical perspective. There is no standpoint outside of this perspective from which moral questions can be meaningfully considered.
The point can also be put this way. The categorical imperative is part of the language of our practical thought. It describes and governs how we go about rationally interacting with those with whom we are rationally interacting (or able to interact) in a common world. This is to say that making laws together is the form of rational interaction. By supposing that a Kantian agent can hold an unknown being to be rational and reasonable (a person) and can make practical judgements about her duties to such a being, Thompson sets up a scenario that is, in fact, incoherent from Kant’s point of view. The only way we can judge that another being is a person with whom we stand in moral relations of duty and right is through rational interaction under the categorical imperative, or in virtue of knowledge (based in experience) that such interaction is possible.
This should not be mistaken for denying the role of natural factors, such as shared ancestry and social practices, in shaping the content of particular moral requirements (or local systems of value). When we recognize other beings as persons, we need not, at least at first, share specific moral requirements. These, as Kant’s theory allows, could differ significantly even among different human social practices and historical stages.Footnote 20 Kantian moral philosophy begins with the idea of a moral law and general a priori moral principles that then gradually translate into a system of particular requirements through our practice of living together in the shared natural social world.
I conclude with a brief observation about the ‘abstractness’ of Kant’s concept of a person. Recall his claim that, given our present understanding of the world, we must judge (at least provisionally) that human beings stand in moral relations of duty and right only to human beings. Kant’s argument for this claim implies that, in our actual moral deliberation, the concept of a person under the moral law is, to use his term, ‘schematized’ as human – presented in concreto as cognizable in experience. Indeed, it is only by being ‘schematized’ in this way that it acquires objective content and plays a role in our practical judgements.
The abstract concept of a rational being with which Kant starts in the Groundwork, for example, is not, as such, intended to pick out an actual manifold of persons whose humanity comes into consideration in our practical reasoning. Rather, this concept is a part of the formal framework of Kant’s moral theory and has to be ‘schematized’ in order to have actual application in practical reasoning. This is what Kant does, in a sense, when in the Metaphysics of Morals he specifies metaphysical first principles of virtue by applying the moral law to human beings. In the Introduction to the book, he explains why it contains principles that apply the moral law to specifically human nature:
[J]ust as there must be principles in a metaphysics of nature for applying those highest universal principles of a nature in general to objects of experience, a metaphysics of morals cannot dispense with principles of application, and we shall often have to take as our object the particular nature of human beings, which is cognized only by experience, in order to show in it what can be inferred from universal moral principles. (MM, 6: 371; my bold)
Throughout the book, Kant appeals to specific, empirical aspects of human nature to derive and classify particular duties of virtue. He uses various ‘impulses of nature having to do with man’s animality’ as part of the basis for classifying duties to oneself (cf. MM, 6: 420), for example, and draws on a number of traits of human sensibility, such as appreciativeness, compassionate feeling, misanthropy, paternalistic attitude, etc., in specifying duties towards others. Thus in addition to general a priori duties (the obligatory ends of one’s own perfection and the happiness of others) that hold for all finite rational nature, Kant’s system of duties of virtue contains moral requirements that render these general duties more determinate by applying the categorical imperative a priori to the empirical concept of human nature.Footnote 21 If perchance we encounter a tribe of previously unknown humanoid species here on Earth, or aliens at the cross-roads of the Universe, we will have to extend and reconcile our system of moral requirements in a way that applies the categorical imperative to them too. And if we reflect on the history of our practices, is not that what we have been doing all along?Footnote 22
