Practical Identity and Practical Principles: Kant contra Korsgaard
What is the relationship between practical reason and the self? Through a close analysis and critique of the work of one of the most influential Neo-Kantian philosophers, Christine Korsgaard, this paper begins sketching a response to that question. For Korsgaard, the self is fundamentally a normative posit, a construction made out of the “pragmatic necessity” to unify one’s desires and act as an agent. Although we may, in this picture, find utterly no self when we look at ourselves from the theoretical standpoint, we will need to act as if we have one when we practically reason. I will charge that this is a basically incoherent view of the self’s relationship to ethical theory, both because it sunders theoretical and practical reason, and because it ignores the multifarious ways in which the validity of our ethical principles might be dependent upon the actual existence of an enduring, rather than merely constructed self. In the end of the paper, I turn to the work of Kant to illustrate how acting from the practical standpoint involves making theoretical commitments, and how those commitments are themselves important to the validity of the practical standpoint.
Keywords: Kant, Korsgaard, Meta-ethics, Ethics, Theoretical Reason, Practical Reason.
In recent years, there has been a resurgence in Kantian Ethics catalyzed by the work of John Rawls, Christine Korsgaard, Barbara Herman, Allen Wood, and Onora O’Neill (Rawls 1999, Korsgaard 2014, Wood 1999, Herman 1996, O’Neill 2014). These thinkers have not only given historical exegeses of Kantian texts, but also made Kant speak to contemporary normative ethics. Of course, there exist tensions between appropriationist readings of past philosophy that focus on reconstructing arguments and more contextualist histories that aim to understand texts for their own sake. Generally speaking, though, these scholars have attempted to tread the line between antiquarian contextualism and hyper-modern appropriationism. Wood writes, for example, that the picture of Kant’s ethical thought he constructs is both “a more accurate portrait” and a “deeper and more engaging ethical view than that with which Kant is usually credited” (Wood 1999: Xiii). Korsgaard, sounding a similar note, writes that once the differences between Kant and his opponents are “correctly understood”, his views will emerge as “the more compelling” (Korsgaard 1996: xiii). Writing good Kantian Ethics, then, is a matter of both remaining faithful to Kant and mining his texts for insights in present debates in value theory.
Korsgaard’s work in particular has attempted to demonstrate the tenability of a Kantian perspective in contemporary debates, and hence much of her writing is in more or less explicit dialogue with other contemporary philosophers (Korsgaard 2014, 1996: 275-310, 365-331). In a piece in Philosophy and Public Affairs that has now garnered multiple responses (Anomaly 2008, Shoemaker 1996) Korsgaard addresses a problem that had previously flown under the radar in much of Kantian Ethics: the relevance of the self for ethical theory (Korsgaard 1996: 365-391). I will argue here that Korsgaard’s account of the self fails as a work of Kantian Ethics, both because it does not offer a compelling account of the relationship between the self, practical principles, and practical reason, and because it is not faithful to Kant. Although containing some valuable insights, Korsgaard’s account fails to reckon with the various ways in which practical and theoretical reason are, although separate modes of cognizing human action, entangled with one another. This leads Korsgaard to offer an impoverished account of the self’s relevance to ethical theory, which one-sidedly emphasizes the practically-motivated reasons for belief in a self, while ignoring the importance of the actual existence of the self for the validity of ethical principles.
Korsgaard’s article emerged as a response to Derek Parfit, a utilitarian who critiqued the Cartesian thesis that there exists some simple, distinct and singular self that exists through time (Parfit 1987). Parfit’s criticisms were part of a project to revive utilitarian ethics by undermining the “separateness of persons” charge that originated in the work of Rawls. In A Theory of Justice, Rawls argued that utilitarianism assumes that people are not separate agents but a single metasubject (Rawls 1999: 22-27). Parfit’s intuition is that by eroding our sense that we are distinct selves he could make it more justifiable to treat people as such a meta-subject. This is where Korsgaard steps in, trying to rescue Rawls (and by extension, Kantian ethics writ large) with a novel understanding of practical identity.
Curiously, Korsgaard’s article is discussed in Allen Wood’s 1999 article on Fichte and the ‘I’ (Wood 1999). Korsgaard’s inclusion in this article is not a matter of coincidence: both Fichte and Korsgaard see the self as a matter of taking up a certain kind of normative stance. By “normative stance”, I mean believing in or acting as if a proposition is true because of practical rather than epistemic reasons. Fichte speaks of the self as a practical “positing”, while Korsgaard writes that the “reasons” we have for regarding ourselves as “rational agent(s)” are “not metaphysical, but practical” (Korsgaard 1996: 369). She elaborates her view on pages 369-380. Her first argument for the “pragmatic” necessity of the self involves conflicting desires. In order to resolve them and act, we must treat ourselves as a “unified person” (Korsgaard 1996: 370). In-so-far as acting involves deliberation, we must regard ourselves as possessing some self over and above our various desires, which in turn chooses between them. Otherwise, we would be guided solely by our desires, and the entire deliberative, practical standpoint would drop out. In order to view ourselves as agents, then, we must regard ourselves as possessing some self that responds to reasons and thereby “regulate(s) your choices among your desires” (Korsgaard 1996: 370). The “need for identification with some unifying principle or way of choosing is imposed on us by the necessity of making deliberative choices, not by the metaphysical questions”, and therefore “It is practical reason that requires me to construct an identity for myself” (Korsgaard 1996: 370-371).
The conclusion is that the self is a matter of taking a certain kind of normative stance, and nothing more. It is a construction in response to pragmatic necessity. If I am to make decisions as an agent, I must regard myself as possessing a self at that moment that responds to reasons and unifies my activities such that I can take a singular action. I do this not because I know I have some self that is actually singular, but because I must if I am to take any action. In a similar vein, Korsgaard writes that in order to “carry out a rational plan of life”, I need to conceive of myself as a single, temporally identical person over time (Korsgaard 1996: 372). The self is thus, at least in-so-far as it is relevant for ethical theory (Korsgaard 1996: 380-388) a matter of a subjective response to the requirements of agency, rather than any sort of actually existent item in the world.
Korsgaard takes her view to be a “Kantian” one, since Kant believed that rational beings view themselves from both the practical and theoretical standpoints (Korsgaard 1996: 377-378). She thus believes herself to be merely explicating the notion of a practical standpoint. The self is something we discover as necessary to our status as agents, and it thus belongs essentially to this “practical standpoint”. There is a sense in which Korsgaard’s view is correct. Acting according to practical principles clearly requires us to regard ourselves as possessing a unified self. Consider the instrumental principle:
Whenever you have some end, you have a reason to do the means to that end.
There are three elements here: ends, means, and reasons. If we didn’t imagine them as each belonging to the same entity, this principle would have no validity for us. I would have no reason to believe that I have reason to do the means to some end, since the entity that has the end might be completely different from the entity that has reasons to perform the means to that end! Though we could perhaps still say this different entity has a reason to perform the means to that end, it could hardly be an instrumental reason if it was not a reason to perform the means to its own end. Hence, we need to take the normative stance that we have a singular self if we are to regard ourselves as bound by the instrumental principle. It is in this sense that the reasons we have for possessing our beliefs about the nature of the self derive from practical and not theoretical considerations. For if I am to regard myself as bound by the instrumental principle, I have practical reasons to believe in the singular self, at least in-so-far as I am to legitimately view the principle as having normative validity. That is Korsgaard’s insight.
Critique of Korsgaard
Although a normative stance is practically motivated, it might contain an epistemic component. A normative stance has an epistemic component if it requires the agent in the stance to hold propositions to be true. Presumably, if Korsgaard’s normative stance possesses an epistemic component when applied to the instrumental principle, it would be something like belief in the proposition that a single, unified self exists. If it does not contain an epistemic component, then it requires us to merely pretend as if we had a self, but entails no beliefs. One merely needs to act as if a given belief is true without subscribing to that belief. However, this latter stance clearly is not sufficient for the instrumental principle to be regarded as valid. If we don’t have a singular, unified self to which ends, reasons, and means all belong, then, as we have already seen, we would have no reason to believe that the instrumental principle actually generates reasons we ought to consider in action.
Indeed, if an agent acts as if they have a self, they may not be truly regarding themselves as having one. One could simultaneously act as if one had a self while holding the proposition that there is no such thing as a self. Under this picture, the instrumental principle will require a sort of self-delusion: we would need to temporarily suspend our beliefs about the self in order to will the instrumental principle. That cannot possibly be what Korsgaard is advocating. More importantly, this would strip the instrumental principle of its normative force. If we knew that willing it depended on deluding ourselves, then it is hard to see how it could be a normatively valid principle. I must sincerely believe that I am, in fact, acting as a unified self if I am to regard the instrumental principle as valid. If I did not, then I would not have reason to believe that the instrumental principle necessarily provides me with actual reasons, since I would not have any reason to believe that the X that possesses ends is the same as the X that responds to reasons. Hence the normative stance requires an epistemic component if the subject is to regard a practical principle as valid.
At this point, Korsgaard could respond that even if we need to actually believe in the existence of the self, such a belief is still motivated by practical, rather than theoretical reasons, and hence the self is still largely something that belongs to the practical standpoint. Yet notice that this epistemic stance must be at least potentially true. If it was evident that there clearly was not any singular, distinct self, then it would make no sense to take up a normative stance with an epistemic component that included belief in the existence of this type of self. Indeed, it would involve a manifest contradiction: it would be to believe both P and not P, that there exists a self and that there does not exist a self. Or it would be to delude oneself, and, as I’ve already noted, that doesn’t seem to be a stable ground for how someone could regard a certain practical principle as valid. This all suggests that theoretical and practical reason are not quite as independent as Korsgaard believes. If theoretical reason rules out certain possibilities, it prevents us from consistently taking certain kinds of normative stances, especially if they have epistemic components.
Consider the whole problem of the identity of the self across time. Korsgaard writes that even if we knew that there existed a “succession of rational agents” over time in one body, there would be a necessity of finding some way to “cooperate with them” by cognizing ourselves as one single self (Korsgaard 2008: 372). We would need to consider ourselves as unified over time because it is practically necessary for us if we want to be agents. Yet it is difficult to understand this as more than a delusion. Suppose we know A) that our status as agents requires us to believe that there is a continuously existing self and B) that no such self exists. In this case, we might indeed have some sort of pragmatic necessity to believe we are unified selves if we want to consider ourselves as agents, but we could not possibly regard this move as justified if we don’t actually have a self that exists through time. To do so would require believing B and not B simultaneously, and hence the “unification” she speaks of would involve contradiction, pragmatic necessity notwithstanding.
There’s another problem for Korsgaard, besides the mere coherency of our beliefs. It’s this: if it is important for us to posit a self to understand ourselves as bound by practical principles, then the actual existence of that self is likely to figure into whether those same principles are valid or not. If such a self is needed for us to believe that a principle is normatively valid, then it’s likely that it will also be needed for a given principle to be actually valid. The example of the instrumental principle is instructive. For us to regard it as being valid, we need to consider ourselves as singular selves. Simultaneously, if we were in fact inhabited by a number of different selves that each had their own means, ends, and reasons, it wouldn’t actually be valid. This just means that the practical positing of the “self” won’t be enough for a certain principle’s validity: that posit will also need to correspond to something in reality. A response to Parfit will thus need to do more than show that we find it pragmatically necessary to think of ourselves as having a single, unified self: it would also need to demonstrate that it is possible that such a self actually exists. It is not, in other words, just the ‘practical posit’ of the self that is relevant for ethical theory. It is also whether such a posit actually is veridical.
The Kantian Alternative
The roots of Korsgaard’s problems are simple: in her work, theoretical and practical reason are treated as two completely different domains, “incomparable” and “separate” from one another (Korsgaard 2008: 378). But in fact, they are mutually entangled. When we posit the existence of a self out of pragmatic necessity, we are simultaneously committing ourselves on a theoretical level to the existence of such a self, at least in-so-far as we wish to regard the various principles we bind ourselves by as valid. In other words, even though the reasons for these various forms of commitment are practical, the commitments themselves are still theoretical in the sense that we are not merely implicitly asserting that there “ought” to be a self, but also acting as if there actually is a self.
Consulting the Kantian texts reveals that practical and theoretical reason do not legislate over completely separate domains. Consider the following remark from The First Critique:
“Now suppose that morality necessarily presupposes freedom as a property of our will… which would be absolutely impossible without the presupposition of freedom, yet that speculative thought had proved that freedom cannot be thought at all, then that presupposition, namely the moral one, would necessarily have to yield to the other one, whose opposite contains an obvious contradiction; consequently freedom and with it morality…would have to give way to the mechanism of nature” (Kant 1998: 116).
To paraphrase: the presuppositions of morality may come into conflict with the findings of theoretical reason, and if they do, we will have to dismiss them because we have overriding epistemic reasons to abandon the conditions for our moral principles. However, notice that this conflict is itself only possible if the presuppositions of morality are not merely practical, but have a theoretical content that asserts how things in the world arei . This is why Kant writes in the Groundwork that when practical reason posits freedom, it “entangles” itself in theoretical matters (Kant 2017: 48). It is also why he is so careful in The First Critique to carve out a space for transcendental freedom and the existence of a ‘simple soul’ that has identity across time (Kant 1997: 29). These are, in other words, more than posits we make out of a “pragmatic necessity” to view ourselves as agents: they are the transcendental conditions for the validity of the moral law, and they must have the possibility of actually obtaining if we are to believe in them, and in turn, in morality.
Unconditional practical laws, Kant writes in The Second Critique, require transcendental freedom if they are to exist at all (Kant 1997: 26). If we adopted Korsgaard’s reading, this would mean that agents are required to view themselves as if they were the “authors” of their action in some practical standpoint. But Kant clearly does not mean this: he is asserting that there must actually be transcendental freedom, because practical laws are only binding if such freedom exists. Practical laws require various kinds of theoretical conditions to obtain if they are to be valid, which means that when we take up the normative stance that we have a self, part of what we are doing is committing ourselves to a theoretical proposition – even if our reasons for adopting it are practical.
Practical and theoretical reason, although two different “standpoints” from which we view ourselves, are thus not isolable. The findings of theoretical reason bear on practical deliberation, and Kant was concerned to delimit the claims of theoretical reason precisely so our practical faculty could have space to make various kinds of posits that, though practically motivated, are theoretical in their content (if only in the sense that they assert of the world that various states of affairs actually obtain). And, vitally, these posits are in turn necessary for the validity of practical laws. Practical deliberation and theoretical reason are hence deeply entangled in Kant’s writing, which means that any properly ‘Kantian’ response to Parfit will need to show that an enduring self is not only practically necessary, but also theoretically possible. In failing to accomplish this, Korsgaard’s work fails to live up to its Kantian heritage.
Anomaly, Jonny. 2008. “Personal Identity and Practical Reason: The Failure of Kantian Replies to Parfit.” Dialogue 331-350.
Herman, Barbara. 1996. The Practice of Moral Judgment. Cambridge: Harvard University Press.
Kant, Immanuel. 1997. Critique of Practical Reason. Cambridge: Cambridge University Press.
Kant, Immanuel, Allen Wood, and Paul Guyer. 2007. Critique of Pure reason. Palgrave McMillian.
Kant, Immanuel, and Johnathan Bennett. 2017. Groundwork for the Metaphysics of Morals.
Korsgaard, Christine. 1996. Creating the Kingdom of Ends. Cambridge: Cambridge University Press. —. 2014. The Sources of Normativity. Cambridge University Press. O’neill, Onora. 2014. Acting on Principle: An Essay in Kantian Ethics. Cambridge: Cambridge University Press.
Parfit, Derek. 1987. Reasons and Persons. Clarendon Press.
Rawls, John. 1999. A Theory of Justice. Belknap Press.
Shoemaker, David. 1996. “Theoretical Persons and Practical Agents.” Philosophy and Public Affairs 318.
Wood, Allen. 1999. Kant’s Ethical Thought. Cambridge: Cambridge University Press.
Wood, Allen. 1999. “The ‘I’ as Principle of Practical Philosophy.” In The Reception of Kant’s Critical Philosophy: Fichte, Schelling and hegel, by Sedgwick. Cambridge University Press.
i If morality did not require theoretical propositions to ground it, there would be no need to vouchsafe a space for freedom in the first place, since it wouldn’t stand at risk of being disallowed by theoretical cognition.