Joshua Glasgow. Journal of the History of Philosophy. Volume 45, Issue 2. April 2007.
Contemporary Kant scholarship generally takes Kant’s conception of humanity in his ethical writings to refer to beings with rational capacities. According to this interpretation, when Kant tells us in the Categorical Imperative’s Formula of Humanity (FH) to “act so that you use humanity … always at the same time as an end, never merely as a means,” we are to treat anyone with rational capacities this way. However, Richard Dean has recently revived an alternative interpretation that he traces to H. J. Paton. According to this interpretation, by ‘humanity’ Kant really means the good will, and, furthermore, Dean takes this interpretation to be the more defensible view within Kant’s ethical system.
In addition, while Dean appears to hold a minority view among Kant scholars, reconciling Kant’s separate claims that the good will is the only unqualified good and that humanity is unconditionally valuable has been a tricky subject for several others. Samuel Kerstein has wondered how to render these claims consistent, and in his review of Thomas Hill’s Human Welfare and Moral Worth, Kerstein criticizes Hill for distinguishing between the good will and humanity without also adequately resolving that tension. While not explicitly identifying the good will with rational capacities, Paul Guyer seems to collapse the two when he writes that, in the transition from section I of Kant’s Groundwork for the Metaphysics of Morals to section II, “all that is happening is that the intuitive conception of the good will is being replaced with the more abstract notion of rational being as an end in itself.” Finally, while in many places Christine Korsgaard holds that humanity and the good will are two different things, she also writes that, “[o]n Kant’s view there is only one thing that has what he calls unconditional value … and that is the power of rational choice (when the choices are made in a fully rational way, which is what characterizes the good will).” The parenthetical remark seems to commit her to the view that rational capacity has unconditional value only when it is actualized in the form of the good will.
Thus, despite the appearance of near unanimity that humanity is different from the good will in Kant’s ethics, in Kant scholarship, there is actually something of a problem on this front. For not only is it the case that Korsgaard, Guyer, Kerstein, and-if Kerstein’s criticism is on target-Hill have trouble distinguishing humanity from the good will, but we also now have a full-fledged attack on that distinction from Dean. Here, I want to defend the view that humanity is rational capacity and the attendant distinction between the good will and humanity. In section 1, I briefly sketch a rough interpretation of what the good will is. In section 2, I critically examine Dean’s general argument for his interpretation. In this section, I also aim to tackle Kerstein’s challenge of explaining away the apparent tension in Kant’s statements that the good will and humanity are each of singular, unqualified, or unconditional value; I do so by identifying different kinds of value at work in Kant’s ethics. If successful, the arguments in this section will establish the propriety of distinguishing between the good will and humanity, contra the claims of Guyer and Korsgaard. Section 3 addresses the question of which interpretation better coheres with Kant’s overall ethical system. I conclude, in section 4, with a discussion of how Kant’s texts should be read with respect to this question.
1. The Good Will
Kant’s understanding of the good will can be examined through the two main senses of ‘will’ in his writing: Wille and Willkür. He writes that,
[i]nsofar as [the faculty of desire in accordance with concepts] is joined with one’s consciousness of the ability to bring about its object by one’s action it is called Willkür … The faculty of desire whose inner determining ground, hence even what pleases it, lies within the subject’s reason is called Wille. (MM, 213)
On this account, Wille is one’s practical reason: it is what gives rational commands. Willkür, by contrast, is one’s faculty of choice: it is the capacity to choose with or against the commands of Wille.
In drawing this distinction, Kant achieves two aims. First, in the idea of Wille, Kant has secured the internality or autonomy of moral commands: they proceed from the agent’s own will, or reason, and so the agent herself provides the authority for morality’s commands. This ensures that morality is self-legislated and that a reason to be moral is always present. Second, however, Wille cannot causally determine our choices. For, if an agent could will only via her Wille, then any irrational—and so any immoral—“willings” would not truly be willings (that is, all true willing would, by definition, be practically rational), in which case immoral actions could not be imputed to the agent. Thus, the power of choice, Willkür, is separated from Wille, such that one may freely and imputably choose to act in accordance with or against the commands of reason (Wille). To privilege one’s inclinations against what reason, or Wille, dictates is to have a bad Willkür. Thus, when Kant says at the beginning of the Groundwork that the only unconditional good is the good will, we can take this to mean that the only unconditional good is a good Willkür. Accordingly, one has a good will when (1) one’s faculty of choice, or Willkür, chooses in accordance with the commands of Wille, and (2) does so from duty, and not because Wille’s commands contingently happen to converge with one’s inclinations.
Furthermore, we can agree with Dean that an agent has a good will when her Willkür is committed to her Wille’s commands, even if she occasionally acts immorally. The point is that having a good will is a fundamental, enduring feature of a person’s character. It expresses a bottom-line commitment to morality, as opposed to the satisfaction of one’s inclinations, in cases where the two may come into conflict. One’s particular willings are rather secondary, though they may indicate the more fundamental character of the will. Kant indicates that the good will mainly consists in a fundamental character trait in the Religion: if a person is to become morally good,
i.e., virtuous according to the intelligible character [of virtue] [virtus noumenon] and thus in need of no other incentive to recognize a duty except the representation of duty itself-that, so long as the foundation of the maxims of the human being remains impure, cannot be effected through gradual reform but must rather be effected through a revolution in the disposition of the human being. (R, 47)
The post-revolution disposition Kant has in mind here is one in which one subordinates inclination to the moral law. If one is disposed to such a subordination as a sort of fundamental maxim or baseline character trait, then one can act from the representation of duty itself and be morally good.
2. Dean’s General Argument
Dean holds that for Kant, FH “orders us to treat as an end in itself the will of an agent who is committed to governing her Willkür with the moral principles legislated by Wille.” For convenience, call this the ‘good will interpretation’ of Kant’s conception of humanity. The more common contemporary interpretation of Kant’s position holds that by ‘humanity,’ Kant means to refer to those beings with practical reason: they have a Wille and a Willkür, and they are autonomous in Kant’s technical sense that they can act on principle rather than on inclination. Call this the ‘rational capacity interpretation.’
Dean’s argument for the good will interpretation is as follows:
(1) If X has value independent of inclinations and X has dignity, then X must be an end in itself.
(2) A good will has value independent of inclinations and has dignity.
(3) Thus, a good will is an end in itself.
(4) Since only humanity is an end in itself, the good will must be humanity.
With Dean, we can stipulate that both humanity and the good will have value independent of inclinations and dignity: it is consistent with the rational capacity interpretation that the good will has this value, for if every rational will has it, then the good will does too. But we can stipulate these points-represented in (1)-(3)-and still reject (4): the claim that only humanity is an end in itself, as stated in (4), and (3)’s claim that the good will is an end in itself do not jointly entail (4)’s consequent, that the good will is identical to humanity. These points are also consistent with the principle that since the good will is one kind of humanity, it is an end in itself, but only one end in itself among other kinds of humanity (i.e., the not-necessarily-good kinds) that are also ends in themselves. So, to bar the possibility that the good will merely constitutes a subclass of humanity, Dean offers the following complementary argument, which seems to be the kind of argument that motivates Kerstein’s concern.
(5) If X is valuable in all possible circumstances, and X has dignity, then X must be good without qualification.
(6) Humanity is valuable in all possible circumstances and has dignity.
(7) Thus, humanity is good without qualification.
(8) Since only a good will is good without qualification, humanity must be a good will.
If (8) is true, then it is no longer possible that the good will is merely a subclass of humanity; now humanity just is the good will.
One might try to disrupt the link between the good will and humanity by stressing that, while the good will has unqualified value, humanity has unconditional value, and holding that the two are distinct. I am not confident whether (and if so, how) we should think that there is a substantive difference between qualifiedness and conditionality for Kant, so I will not pursue that strategy here, and I will largely treat ‘unqualified’ and ‘unconditional’ as roughly equivalent. (Doing so does not seem to threaten the substantive claims I will make.)
Instead, I want to respond to Dean’s argument by suggesting that Kant is working with two different kinds of value here. If this suggestion is right, then we can resolve the tension that motivates Kerstein’s concerns. We will also have reason to dispute Guyer’s claim that the “intuitive” good will can simply be replaced with the more “abstract” humanity, and we will not need to hold with Korsgaard that humanity has its value only when it takes the form of the good will.
Kant has several kinds of value at work in his overall ethical system. The distinction he most clearly draws is that between das Wohl and das Gute (CPrR, 59-62). Unlike the latin ‘bonum’ (and, we might add, the English ‘good’), German has two expressions, one for ‘good’ in the moral sense (das Gute) and one for ‘good’ in the sense of well-being (das Wohl). When Kant argues that the good will is good without qualification, he means that since it has goodness in the sense of moral excellence, it has unqualified goodness simpliciter (though this is not Kant’s term) (G, 393-94). And, while the good will is the only unqualified good simpliciter, this does not (to use Kant’s own example) rebut the commonsense idea that the scoundrel’s coolness is good for the scoundrel’s well-being (if we exclude, for the sake of drawing the distinction, moral perfection from being a constituent in one’s well-being). It only means that his coolness is not unqualifiedly good simpliciter. So while coolness is neither morally good nor good simpliciter unless accompanied by a good will, it may be prudentially valuable. This is possible because well-being (or prudential value), moral goodness, and goodness simpliciter are conceptually distinct.
One might object here that the good will’s unconditional goodness is not mere moral goodness, since when Kant holds that it-as opposed to, say, courage or wit-is unqualifiedly good, he does not use ‘moral’ to modify ‘unqualifiedly good.’ Rather, the good will’s goodness is goodness simpliciter. Yet, while the good will does have goodness simpliciter, I think it would be a mistake to suggest that the goodness referred to in ‘good will’ is anything other than moral goodness, for three reasons. First, Kant clearly does distinguish between das Wohl and das Gute. Thus, there is a kind of goodness (in English) or bonum that is related to prudential well-being, and another, moral kind that is independent of well-being. The latter, das Gute, the goodness of a good will, cannot, therefore, be some sort of all-encompassing goodness, since its scope does not include one kind of value, namely prudential value (das Wohl). Relatedly, Kant’s highest good is a state of affairs in which virtue is rewarded with happiness-a state of affairs that is contingent in part because the good will cannot conceptually encompass happiness. Second, in the context of the distinction between das Gute and das Wohl, Kant holds that the opposite of goodness is evil, and whether one’s will is good or evil is a matter of the will’s maxim. Given the overall conceptual scheme of Kant’s ethics, according to which a will’s having adopted the proper maxim from duty makes it morally worthy, this also indicates that the good will’s goodness is specifically moral goodness. Finally, sections I and II of the Groundwork proceed from the concept of the good will to the concept of duty, and from there to the categorical imperative, or the moral law. For Kant, these moves are supposed to be analytic, which suggests that he is analyzing moral concepts from the beginning, i.e., from the point at which Kant declares that the only unqualified good is the good will. Indeed, Kant’s concern in the second Critique’s distinction between das Gute and das Wohl is that, because of the ambiguity in the latin bonum, one might be seduced into thinking that goodness is what is desired, in which case morality would end up heteronomous. But while it is true that prudential goodness is dependent on our desires, what is morally good must be determined independently of desire, so equivocating between the two kinds of goodness can lead to abandoning the autonomous nature of morality.
The objection under consideration now, however, rightly indicates that, for Kant, it is also the case that only the morally good will can be called, simply and unqualifiedly, ‘good’. Kant’s argument here is that the good will is the only unqualified good simpliciter, because anything else can always be used directly in the service of evil. The distinction that renders this point consistent with the good will’s goodness being specifically moral goodness is the distinction between general goodness, or goodness simpliciter, and specifically moral goodness. Given this distinction, Kant’s claim is that unqualified moral goodness is the sole, sufficient and necessary condition for something’s being appropriately called unqualifiedly ‘good’ simpliciter. Call this the Principle of unqualified Goodness. Given that the distinction between moral goodness and goodness simpliciter is required for the Principle of unqualified Goodness to be possible, the most coherent interpretation of Kant seems to be that the good will’s goodness is moral goodness and that this moral goodness makes the good will alone unqualifiedly good simpliciter. It is, therefore, safe to say that the good will’s goodness is neither a prudential good nor an all-encompassing good that covers both morality and well-being; rather, to have a good will is to have moral goodness, which then allows the good will to also be called simply good without qualification.
If the foregoing is correct, then the second half of Dean’s general argument will need to be revised. In particular, (5) should be revised to read as follows:
If X is good simpliciter in all possible circumstances and X has dignity, then X must be good simpliciter without qualification.
But if that is the case, then (6) also must be modified:
Humanity is good simpliciter in all possible circumstances and has dignity.
Only then can Dean’s argument go through. What we need to determine, then, is whether (6) is true. For Kant, does humanity have goodness simpliciter in all possible circumstances? If, as I will argue, the answer is ‘no’, then it will not follow that (7) humanity is good (simpliciter) without qualification, which will block this putative point of identity between humanity and the good will.
By way of working towards a negative answer to the present question, I now want to argue that the end in itself that has unconditional value is distinct from both moral and prudential value. For, according to the Principle of unqualified Goodness, something can have unqualified goodness simpliciter only if it has unqualified moral goodness. So, if it is true that humanity, as an end in itself, has a value that is distinct from moral goodness, then it will be the case that humanity, qua humanity, also will not have unqualified goodness simpliciter, which will undermine (6) and (7).
Ends in themselves have unconditional value, in the sense that they have a certain moral status that is unconditioned by circumstance or our inclinations. I will develop the idea of moral status in more detail shortly, but for the moment let me briefly characterize moral status as the value that, for Kant, generates moral obligations. The reason we must, for example, further the ends of others and not make false promises to them is because those others have value. This obligation-generating value is what I am referring to as ‘moral’ or ‘right-making’ status. Pebbles do not have moral status in the sense that they have no value that gives rise to (direct) obligations to them. If the rational capacity interpretation is correct, then we also have no (direct) obligations to the good will qua good will; rather, we only have obligations to the good will insofar as the good will is a being with rational capacities. Its rational nature is what creates our obligations to it, in which case all beings with rational capacities will have moral status.
This value of moral status is distinct from moral goodness, prudential goodness, and goodness simpliciter. As I will defend the distinction, then, there are at least four different kinds of value at work in Kant’s overall value theory:
Value 1: Goodness simpliciter. This is a general category, encompassing whatever is common to all specific kinds of goodness.
Value 2: Prudential goodness (das Wohl), the first specific kind of goodness. This is what constitutes the agent’s well-being.
Value 3: Moral goodness (das Gute), the second specific kind of goodness. This is the morally good character trait possessed by the good will. Moral goodness gives it unqualified goodness simpliciter, and it qualifies the goodness simpliciter of every other character trait and gift of fortune. To say that the goodness of, say, wit (a character trait) and happiness (a gift of fortune) are qualifiedly good is to say that their goodness is contingent upon their not being used for evil purposes or gotten by evil gain. The good will’s goodness is not, in this sense, contingent on anything else (after all, it cannot, by definition, be directly used for evil purposes); it is unqualified.
Value 4: Moral status. This is the value that generates the obligatoriness of morally obligatory actions. When it is unconditional, such as with humanity, it conditions the moral status of all other things, such as happiness. To say that the value of happiness is conditional in this sense is to say that we have a moral duty to produce happiness only insofar as an end-in-itself is being made happy, so happiness might be an end, but not an end in itself, not an end with unconditional status. Humanity’s status is not, in this sense, derivative from something else with moral status; it is, unconditionally, an end in itself.
If this distinction is appropriate, then humanity has unconditional value in the sense that it has moral status independently of anything else having right-making status. But a good will has unqualified value in the sense that it is the character trait that qualifies the goodness (simpliciter) of everything else, without its goodness (simpliciter) being qualified by anything. The good will has unqualified goodness simpliciter because, according to the Principle of unqualified Goodness, moral goodness is the sole, sufficient and necessary condition for something’s being unqualifiedly good simpliciter. But now note that, according to this Principle, humanity’s value of moral status does not by itself entail that humanity has unqualified goodness simpliciter. For humanity to have unqualified goodness simpliciter, humanity would have to have not only moral status, but also moral goodness.
But then (7) is without justification; that is, humanity does not (by virtue of its moral status) have unqualified goodness simpliciter. This is what makes room for the rational capacity interpretation, but at the same time it might seem objectionable. after all, Kant is clear that humanity does have unconditional value. So how is humanity not unqualifiedly good simpliciter?
Recall here that goodness simpliciter is simple “goodness-in-general,” i.e., whatever is common to all specific forms of goodness. In that case, it should not seem objectionable that humanity does not have unqualified goodness simpliciter. For humanity would not have the element of goodness-in-general unless humanity also had some specific kind of goodness. But whether humanity qua humanity has any particular kind of goodness-especially moral goodness-is a contingent matter, or so I am arguing. On the interpretation advanced here, humanity qua humanity only has moral status; it does not have any kind of goodness (as characterized in Values 1-3).
Strictly speaking, it seems that, for Kant (as for some forms of ordinary discourse), “good” is a subset of value; it is reserved for those things that we want to pursue, promote, or otherwise bring about. Moral goodness, for Kant, is a property we are to aim for in our wills, and other kinds of goods, such as prudential goodness, are also goods that are (ceteris paribus) to be effected. Now, while the good will need not be effected in the phenomenal world to be good, non-moral goods must be; in either case, however, both have in common that they are ends to be produced (whether in the will itself or in the phenomenal world). But, for Kant, humanity is not a value that is to be brought about; it is not a “good” at all in this sense, qua humanity. For he is clear that humanity is an “independently existing end,” rather than an end to be produced (G, 437). So it makes sense that, while Kant’s value-talk includes the statements that humanity is an end in itself, and has “absolute value” and “dignity,” he does not call humanity ‘good’ in any sense-prudential, moral, or simpliciter-in and of itself. For that matter, as we will confirm in the next two sections, it seems that Kant allows that one could have humanity and not be morally good.
Given this parsing, it should not seem objectionable that humanity is not unqualifiedly good simpliciter. Again, it is not a good in any sense. Rather, it is an independently existing value, in Kant’s terminology. However, even though it is not a good simpliciter, it is still unconditionally valuable in the sense of having unconditional moral status. Thus, the view that humanity has no unqualified goodness simpliciter remains faithful to Kant’s statements about the unconditional value of humanity.
If, for Kant, humanity only has moral status, and not moral goodness, then humanity will not have unqualified goodness simpliciter, and (6) and (7) will be false, or at least unjustified, in which case Dean’s general argument will not go through. To further defend this claim, I want to, first, sharpen the distinction between moral status and moral goodness, and, second, mine for evidence that Kant distinguishes these two kinds of value in the way that I am suggesting.
Perhaps the clearest way of demonstrating that there is a sharp distinction between moral status and moral goodness is to leave Kant altogether and to focus on another moral theory, Mill’s utilitarianism. Mill argued that what makes right acts right is that they promote happiness. In what is now more commonplace jargon, happiness-conduciveness is Mill’s “moral criterion”; it is the right-making feature of right acts. Within this criterion, Mill assigns fundamental moral ‘status’, as I am using the term here, to happiness. It has fundamental moral status in that it has the value that at the most fundamental normative level generates our moral obligations.
But Mill assigned what I am calling moral ‘goodness’ a different place in his theory. He stated the following, in response to the objection that “utilitarianism renders men cold and unsympathizing”:
[C]ertainly no known ethical standard decides an action to be good or bad because it is done by a good or bad man … These considerations are relevant, not to the estimation of actions, but of persons; and there is nothing in the utilitarian theory inconsistent with the fact that there are other things which interest us in persons besides the rightness and wrongness of their actions.
What Mill contended, then, was that there is one kind of value that provides the criterion for when an act is right and when it is wrong (the value of happiness), and another kind of value, moral goodness, that attaches to our characters (if we have the virtuous disposition to act in utility-maximizing ways). Accordingly, a person might not have a morally good character in saving a life-perhaps he is doing it to impress a third party and not because he is disposed to maximize utility. But that person’s action still has the property that makes it right. And the reason it is right is because it produces what is fundamentally valuable with respect to right action, viz., happiness. This separation of the value of the agent’s character and the value that makes his action right is what I am identifying as the difference between moral goodness and moral status; this conceptual distinction seems coherent, regardless of the merits of Mill’s substantive views about moral goodness and moral status.
Now, some theories of rightness might not include a property with moral status, because moral criteria that include moral status are (by definition) always valuebased, and not all moral criteria are value-based. So, by way of contrast, consider a purely right-based theory of rightness, which holds that for any right action a, one should a merely because a-ing is right-and not because a-ing produces happiness, or treats humanity as an end in itself, or any other principle to which another moral theory might subscribe. This kind of theory’s moral criterion, we might say, is brute. There is no further reason, other than the nature of a itself, why a-ing is right. So this theory’s moral criterion is not value-based; this contrasts with Mill’s moral criterion, which is a’s production of the fundamental morally relevant value, namely, happiness. So, while Mill has a property to which he assigns moral status, our purely right-based view includes no such moral status.
To summarize, then, there is a difference between a theory’s account of what makes right acts right (the moral criterion) and its account of when an agent is a morally good agent. Furthermore, among different theories’ moral criteria (i.e., among their theories of rightness), some will be value-based and some will not. Those that are value-based will assign some property of actions moral, or right-making, status. So there is a real difference between moral status and moral goodness. Moral status is the value that, for value-based theories of rightness, attaches to the property that fundamentally makes right acts right; moral goodness is the value that attaches to persons whose characters are morally admirable.
The remaining question is whether Kant’s texts really encourage an interpretation according to which his theory distinguishes between moral status and moral goodness in the way that I have suggested. Kant does not clearly and explicitly tell us he has such a distinction in mind, whereas he does clearly and explicitly tell us to distinguish between moral and prudential value. However, when we look at the roles played by the good will and humanity in Kant’s texts, it does seem appropriate to apply the distinction to his view. For, when he talks about the good will, he is referring to a morally good character trait, as we have seen. And when he discusses humanity, it is in the context of fixing the rightness of right action and the wrongness of wrong action: you are to “act so that you use humanity … always at the same time as an end, never merely as a means.” This is a claim about what makes right acts right; i.e., it presents Kant’s moral criterion, which is conceptually distinct from whatever he holds about the nature of a morally good character (as captured in his discussions of the good will).
Finally, since Kant is clear that his moral criterion-humanity-has the value of being an end in itself, it makes sense, within our conceptual scheme, to say that Kant assigns humanity moral status. Kant, in this respect, has a value-based theory of rightness. On this point, it is worth recalling that, for Kant, humanity is an “independently existing end,” rather than a good to be produced. In this sense, while both Kant and Mill have value-based theories of rightness, only Mill’s is “good-based”; of the two, only he holds that the object with moral status is a good-to be-produced, for Kant explicitly holds that the object with moral status, humanity, is not a good to be produced. It is an already existing end in itself. Thus, it makes sense that, while Mill characterizes happiness as a good, Kant never says the same about humanity, though they both share the language of ends and values.
So humanity and the good will, while both valuable, possess different kinds of (unqualified or unconditional) value: humanity possesses moral status (it is what makes right acts right) and the good will possesses moral goodness (it is the distinctive valuable property that is predicated of morally admirable people), which generates its unqualified goodness simpliciter. The distinction between moral status, moral goodness, and goodness simpliciter, recall, required us to revise (5) to (5), which required us to revise (6) to (6). And, since humanity does not necessarily have goodness simpliciter, let alone unqualified goodness simpliciter, (6) and (7) appear to be false, which undermines Dean’s general argument. This distinction also alleviates the tension that worries Kerstein, viz., it allows us to understand how humanity and the good will each can have unique unqualified or unconditional value: humanity has unconditional moral status and the good will has unqualified goodness simpliciter. Finally, given the difference between the good will and humanity, we should reject Guyer and Korsgaard’s overly close associations of the two. In short, humanity and the good will appear to be distinct.
3. The More Defensible View within Kant’s Ethical System
If the distinction so far defended-that the unqualified values of the good will and humanity are not identical sorts of value-is correct, then we have good reason to doubt that Kant intends ‘humanity’ in FH to refer only to those among us who have good wills. But there are two other ways in which Dean defends the good will interpretation. One way is to appeal to textual evidence. The other, which I will consider first, is to show that the good will interpretation better comports with some of Kant’s other ethical doctrines than does the rational capacity interpretation.
The first way in which the good will interpretation might seem more defensible in this regard is that only it can support the view “that an agent who chooses to act immorally is sacrificing her own most valuable possession,” which itself is a reason to act morally. Since it is impossible to sacrifice our rationality, this does not hold true for the rational capacity interpretation. Yet this argument leaves open two rejoinders. First, it is questionable whether FH is designed to provide a reason to act morally, for Kant offers two wholly independent arguments that morality really does obligate us. Kant’s synthetic argument that we are obligated by the moral law in section III of the Groundwork tells us why we should be moral: since we are free, we are obligated. And, after Kant’s so-called ‘Great Reversal’ in the Critique of Practical Reason, he held that freedom instead is derived from morality, and the authority of morality is simply given as a “fact of reason” (CPrR, 29-33). Furthermore, the Formulations of autonomy and Kingdom of Ends provide another rationale for us to act morally: we legislate the moral law, so we have an immediate reason to follow it. Second, if the good will is our “most valuable possession” (even in the restricted senses of value provided above), then sacrificing this by acting immorally will be bad for us, regardless of whether it is what ‘humanity’ means in FH.
Another way that the good will reading might do more doctrinal work than the rational capacity reading is that FH would then share “stronger connections” with the universal law and Kingdom of Ends formulations of the Categorical Imperative. Dean’s first point here is that, if we take the rational capacity reading, humanity’s “central subject matter is not the making and following of universal law.” But, while on this reading humanity would not necessarily will in universalizable ways, nevertheless humanity will necessarily have the “central subject matter” of “making and following universal law.” For humanity, on the rational capacity interpretation, refers in part to the possession of Wille, and Wille-practical rationality-will always (for Kant) legislate universal law and prescribe that the agent follow it. Dean acknowledges this reply to his first point, and he argues that it still leaves out a second element found in the Kingdom of Ends formulation of the Categorical Imperative, namely the idea that in that formulation right action is determined by a realm of agents who not only legislate the law, but also comply with it.
Defenders of the rational capacity interpretation seem forced to concede that their view leaves the idea of ideal compliance out of FH, but three points are in order here. First, to the extent that each of the different formulations can generate the same duties—such as duties against false promising, non-beneficence, etc.—they are extensionally equivalent, and leaving ideal compliance out of FH does not seem to threaten that equivalency (i.e., the same duties seem to follow from FH on the rational capacity interpretation). Second, the different formulations of the Categorical Imperative feature different ideas, and it is not evident that ideal compliance is not one of those features that is unique to the Kingdom of Ends formulation. Given these first two points, it is not obvious that leaving ideal compliance out of FH threatens Kant’s overall view of how strongly the different formulations are connected. But, third, even if this were an interpretive problem for the rational capacity interpretation, this would be a bullet I would be willing to bite, for the rest of this section and the next are designed to show that all other interpretive questions either come out in favor of the rational capacity reading, or are neutral between the two interpretations.
Dean’s third argument for the superiority of the good will reading is that, on the rational capacity reading, there is no “principled reason” for prohibiting agents from promoting others’ immoral ends. This would not be a problem for the good will reading, since, for it, FH requires us to further a good will, and as such we are limited to furthering only the ends of others that comport with the choices made by the good will. It is allegedly a problem for the rational capacity reading because, according to it, we are not primarily concerned with furthering a good will. However, treating a rational being as an end in herself does not entail promoting all of her ends. Since she has end-in-herself status in virtue of her rationality, what we are to do is further her rational ends, not all of her ends. Thus, we can have the rational capacity reading and still hold that, in some cases, we must limit our furtherance of others’ ends to those ends that are consistent with good, or rational, willing. Although, because FH is founded on rational agency, according to this interpretation, this may mean that sometimes we cannot impede her free choice, even if it seems irrational. This, indeed, is an advantage of the rational capacity interpretation, rather than the good will interpretation: forcing morality upon others is at best secondary to preserving, promoting, and respecting rational agency.
Finally, Dean responds to a predictable objection to his view: something is morally wrong with not treating all persons-regardless of how good their wills are-as ends in themselves. His response is to note that, even on the good will interpretation, we are still required to treat all people, rather than good wills alone, with benevolence and respect. Dean holds that three justifications for this requirement are available to the good will interpretation. (A) Even someone with a good will can occasionally act immorally; a good will is about one’s “fundamental commitment” to morality. (B) wills are opaque; we usually cannot know if they are good or bad-and we must give agents the benefit of the doubt. (C) Even if a person’s bad will is obvious, we still must respect that person because to not do so would “lessen our respect for all humans” and would wrongly remove an incentive for the bad person to reform.
When taken in conjunction, (A) and (B) obligate us to assume that everyone besides the most obviously evil people of the world have good wills (even if they actually do not). This might seem questionable, but I want to grant it and focus instead on Dean’s strategy for dealing with the obviously evil will, as presented in (C). As much as we might like to, there seems to be something intuitively wrong with slowly torturing an evil person to death (say, by cutting off one square inch of his skin per day for a year). Yet it is not obvious that doing so would lessen our respect for all humans; it might even bolster it insofar as such an action might be seen as redemption for crimes against humanity and a deterrent against future violations of respect for persons. And, on similar grounds, it seems our last concern should be the evil person’s chance to reform if torture would increase our respect for all humans, given that Dean is right that such a person is not an end in himself. Thus it seems that (C) is not so plausible; even if the good will interpretation does not recommend “rampant mistreatment” of others, it is at least arguable that it allows some of what we intuitively consider mistreatment.
Finally, consider how Dean’s view meshes with Kant’s distinction between (nonhuman) animals and humans. Putting aside the reformation worries, the good will interpretation amounts to treating humans with not-so-good wills as animals. That is, Dean’s argument is that we have what Kant calls an ‘indirect’ duty to the not-so-good person: we have a direct duty to maintain respect for all good humans, and one indirect way to achieve this is to not disrespect the not-so-good humans. Kant gives the same argument for the humane treatment of animals: even though animals have no moral standing on their own, we must treat them well because this will prevent a scenario in which animal abuse “weakens and gradually uproots a natural predisposition that is very serviceable to morality in one’s relations with other men” (MM, 443). This is significant because Kant explicitly distinguishes humans as ends in themselves from animals and the rest of nature by virtue of our being “the subject of a morally practical reason” (MM, 434, emphasis added). It thus seems, contrary to Dean’s claims, that insofar as an evil will is even subject to Wille’s commands, we have direct duties to that will, rather than indirect duties, which are restricted to duties to animals, in which case all rational wills, good and evil, are ends in themselves.
4. Kant’s Texts
Kant’s distinction between duties to humans and duties to animals is the sort of claim that lends textual credibility to the rational capacity interpretation. For another famous example, Kant defines ‘humanity’ as “the capacity to set oneself an end-any end whatsoever.” Being able to set ends, of course, includes the ability to set immoral, as well as moral, ends, and so ‘humanity’ here seems to refer to all rational creatures, good and bad. There are several other passages that reinforce this rational capacity interpretation. After marshalling some of these in support of the rational capacity interpretation, I will address the texts that Dean cites in support of the good will interpretation.
Kant writes that “[e]very human being has a legitimate claim to respect from his fellow human beings and is in turn bound to respect every other” (MM, 462). By putting this in terms of “every human being,” Kant seems to indicate that he has in mind all members of the species, not just the good ones. So, if humanity is the object of our duty of respect, then all members of the species are to be ascribed the property of humanity.
Furthermore, Kant holds that because morality cannot be based on any incentive, it must not be derived from anything particular about human nature. Rather, morality “must therefore hold for all rational beings … and only because of this be also a law for all human wills” (G, 425). Kant’s point here is that, instead of deriving morality from human nature in particular, it must be “connected (completely a priori) with the concept of the will of a rational being as such” (G, 426). This shows us that rationality-rather than any quality of humans in particular, or, worse yet, particular kinds of humans (such as those with good wills)—is that to which morality “connects.” and, in a claim that should be considered in conjunction with his claim about indirect duties to animals, Kant famously contrasts persons with things, because persons’ “nature already marks them out as an end in itself” while things are “beings without reason” (G, 428). Again, Kant clearly grounds our humanity on our rational capacities here. Finally, consider the passage where Kant writes that we can have contempt for immoral people, but not for the humanity in them (MM, 441). This implies that immoral people cannot sacrifice their humanity by virtue of being immoral.
I now want to turn to the passages that Dean cites that, in his words, “seem clearly to identify humanity as good will or obedience to morality.” On balance, I do not think that these passages clearly support the good will interpretation. For instance, one text cited by Dean in support of the good will interpretation is where Kant writes that a distressed person who sticks to morality from duty “has maintained humanity in its proper dignity in his own person and honored it” (CPrR, 88). But this is, at most, one of many texts that is ambiguous and hence incapable of supporting either of the rival interpretations. In the comment on “honoring” one’s dignity, Kant seems to be discussing how one pays one’s dignity its due; this is different than identifying the property in virtue of which one has one’s dignity in the first place. And when Kant makes the comment about “maintaining” one’s proper dignity, he can again be taken to mean that acting from duty is the way in which one admirably puts one’s dignity into action. So, we can agree that acting in accordance with duty from duty does honor one’s dignity without thinking that one has dignity because of one’s good will.
Dean also has us consider the following two passages. The first is where Kant writes that “it is just this fitness of his maxims for giving universal law that marks him out as an end in itself” (G, 438). That is, a being that wills maxims that are universalizable, and hence rightful, is “marked out” as an end in itself. But for Kant, having rightful maxims is not equivalent to having a good will, for he is clear that there is a distinction between acting “from duty” and acting “in conformity with duty” (G, 397; CPrR, 71). The second text Dean cites is this:
… we thereby represent a certain sublimity and dignity in the person who fulfills all his duties. For there is indeed no sublimity in him insofar as he is subject to the moral law, but there certainly is insofar as he is at the same time lawgiving with respect to it and only for that reason subordinated to it. (G, 439-40)
Again, though, merely fulfilling one’s duties is not to have a good will: one might act in accordance with duty simply because the actions required by duty coincide with the actions that would satisfy one’s inclinations, and not because one has privileged morality over inclination, like the grocer who gives correct change merely to preserve a reputation for honesty.
What I have said about these two passages still leaves open the possibility that only those who will rightly-whether or not their wills are also good-are ends in themselves. This would run against both the rational capacities and the good will interpretations, for it would mean that rightful wills, not just any rational will and not just good wills, have the property of humanity. But this possible interpretation seems misguided. The dignity Kant speaks of in the passages now under consideration originates in the fact that one is both the author of the moral law and-by virtue of that authorship-subject to it. And this fact is as true of the not-so-good will as it is of the good will and the merely rightful will, for the lawgiver is Wille, in which case one’s dignity comes from one’s Wille; it does not come from the particular leanings of one’s Willkür. This explains why Kant writes that “a human being cannot renounce his personality as long as he is a subject of duty, hence as long as he lives.” Kant’s point is that any rational being that is subject to duty (i.e., everybody) has her lawgiving capacity and hence personality no matter how evil her Willkür.
Finally, Dean claims that the following statement by Kant should remove “any doubt” that Kant means that humanity is the good will: “morality is the condition under which alone a rational being can be an end in itself.” But Kant follows this by stating that “morality, and humanity insofar as it is capable of morality, is that which alone has dignity.” It seems plausible to suggest that what Kant is stating here is not that morally good agents are the only dignified agents, but that moral capacity is the sole, necessary and sufficient condition for having the dignity of an end in itself. Since all humans have this capacity, all humans are ends in themselves. It is this ‘capacity’ qualifier that allows for a different reading of the passages cited in defense of the good will interpretation. That is, when Kant says that “giving universal law” is the mark of being an end in itself, he means that beings capable of giving such law-i.e., moral legislators or beings with Wille-are accorded the status of an end in itself. On balance, then, it seems that what gives us our dignity is our being autonomous authors of-and consequently subjects to-the law, whether or not we then comply with it, let alone whether we comply with it because we have a fundamental commitment to acting morally.
Consider two passages that should entrench the interpretation that our autonomous authoring of the law, rather than our abiding by it from duty, is what gives us our dignity: “[a]utonomy is therefore the ground of the dignity of human nature and of every rational nature” (G, 436); and “[a]utonomy of the will is the property of the will by which it is a law to itself (independently of any property of the objects of volition)” (G, 440). In this definition of ‘autonomy,’ Kant is asserting both its negative definition (the ability to act independently of one’s inclinations) and its positive definition (the ability to act on a law given to oneself). If these abilities neatly contain Kant’s view of what it means to be a rational being, and if autonomy is the ground of one’s dignity, then, for Kant, all rational beings are dignified.
In short, it seems that the rational capacity interpretation captures Kant’s texts better than the good will interpretation does. Passages that seem to support the good will reading are, at best, ambiguous. Indeed, those passages seem to be more naturally read as countenancing the view that one’s autonomy and concomitant moral subjectivity are the grounds of being an end in oneself-that is, they support the rational capacity interpretation. Furthermore, we have seen that the rational capacity interpretation seems more defensible in light of both Kant’s other moral doctrines (particularly regarding the distinction between animals and humans) and our considered moral intuitions about the mistreatment of evil wills.
Finally, Dean’s general argument in support of the good will interpretation trades on a hidden ambiguity in the uses of ‘valuable’ contained in that argument. once the various kinds of value are separated into goodness simpliciter, moral goodness, and moral status, any putative conceptual identification of the good will and humanity dissipates. And it is this separation that allows us to resolve the tension that concerns Kerstein, that Kant’s claims that the good will and humanity, while distinct, each have unique unqualified or unconditional value. Finally, if this tension is resolved by distinguishing between the three senses of value in Kant’s overall value theory, then we also should not follow Guyer and Korsgaard’s respective claims that seem to collapse the good will and humanity.