Paul Guyer is rather fond of one particular passage from the Mrongovius and Collins Lectures on Ethics where Kant declares “if all creatures had a faculty of choice bound to sensuous desires, the world would have no value; the inner value of the world, the summum bonum, is the freedom to act in accordance with a faculty of choice that is not necessitated. Freedom is therefore the inner value of the world” (Mrong 27:1482). Guyer quotes this passage in at least eight different places.1 I am not even counting in that total the related passage he is also fond of quoting from the Naturrecht Feyerabend course lectures to the effect that the value of the human being is due to freedom rather than rationality as such: “If only rational beings can be an end in themselves, that is not because they have reason, but because they have freedom” (NF 27:1321). Guyer takes Kant to have a “morality of freedom” in which the highest value is the freedom of human beings (and other beings like us). Guyer presents this emphasis on freedom in contrast to what others have seen as possessing absolute value for Kant, such as rational nature itself.2
In this chapter I want to argue that Guyer’s use of the claim that “freedom is the inner value of the world” is misplaced. My concern is with the “inner value of the world” part of the claim, which raises the metaethical question regarding the meaning of the term “inner value.” I am not concerned here about the claim that freedom rather than some other thing is of value. To say that a value is “inner” implies an ontology in which value is understood as a property of objects independent of any perceiver or moral evaluator. I will argue that such a property conflicts with Kant’s own epistemology and ontology.
What kind of thing is “inner value”? Specifically, I want to determine where this absolute moral value lies in the debate over whether Kant is a moral realist or moral idealist (or anti-realist or constructivist, depending on whose terminology is in use). An inner value of the world identifies value as an intrinsic property of the world; in the case of freedom this value presumably lies in moral agents themselves as an intrinsic property. Properties are an ontological matter and so the question about value realism is properly viewed using a metaphysical conception of realism. The definition of moral realism that I favor is metaphysical:
Moral realism: the moral principles, properties, or objects of the world are independent of the transcendental or empirical moral agent.3
Correspondingly, moral idealism (a better term than “antirealism” with its negative connotations) sees principles, properties, and objects as all dependent on the moral agent.4 I understand dependence on the moral agent to mean dependence on the moral subject, the agent as active.5 For example, the value of particular ends that an agent chooses to pursue is uncontroversially dependent on that agent’s valuing those ends and so would be ideal rather than real. My assessment here is targeted at the ontological claim that there is an intrinsic value to something in the world that is not dependent upon the moral agent in any kind of similar way.
In my definition I distinguish between the “transcendental” and “empirical” as two levels of possible realism in Kant. This novel approach has important implications for my argument. The transcendental level – and here I am using transcendental in its methodological sense in relation to transcendental arguments – consists of the very conditions for moral agency as such, parameters that would apply universally to all possible moral beings, or, in modified Kantian terminology, the transcendental conditions for the possibility of moral experience. The empirical level consists of the specific characteristics of actual moral agents that do not have such transcendental justification. There would be only one description of the transcendental moral agent, just as there is only one description of the subject of the transcendental unity of apperception in theoretical philosophy, but there are many different empirical moral agents with different characteristics. These two different conceptions of moral agents do not imply that there are two different moral agents in each human being; rather, human beings as empirical agents can be exemplars of transcendental moral agency if they possess these universal features. At the very least Kant argues that as empirical moral agents, human beings must also view themselves as transcendental moral agents. It is also possible conceptually that the transcendental moral agent would require some non-empirical property or ontology, which would then have to be attributed to specific moral agents as a non-natural property or ontology. Whether this is so or not is to be determined by an examination of the transcendental conditions for moral experience that Kant actually advocates.
Most of Kant’s discussion of morality in the Groundwork and the Critique of Practical Reason are at the transcendental level, concerned with the practical nature of finite rational beings and with pure practical reason. There is some discussion of empirical moral agency, for example, in Groundwork III when Kant asks whether actual human moral agents may assume that they are transcendental moral agents, and in the second Critique when the fact of reason plays a similar role. Overall, I take the derivation and formulas of the categorical imperative, the distinction between autonomy and heteronomy, and the requirements for freedom to operate at the transcendental level.
Some of Guyer’s statements indicate that he takes Kant to be a moral realist about value. Many others call Guyer a value realist6 and he himself takes the critical Kant to base a claim of the absolute value of freedom on an argument that claims that we really are free, rational, noumenal selves.7 The picture that one has is that Kant’s critical moral theory requires the existence of moral agents as things in themselves independent of nature in space and time, freely able to determine their own actions, with that freedom possessing value of itself. Guyer accepts this as an interpretation of Kant but also argues that it is philosophically untenable because of the difficulties in, for example, explaining how this purely rational acting being should end up making some of the bad decisions it exhibits in nature in space and time. Guyer would prefer a more naturalistic Kant who avoids transcendental claims and instead understands the value of freedom as simply the fact that actual human beings really do desire freedom, a tendency that Guyer sees in the pre-critical Kant.
But there are other indications that Guyer rejects value realism as an interpretation of Kant.8 He often notes that Kant steps back from the strong metaphysical claim that human beings are really free, rational, beings in themselves independent of nature in space and time to the less realist position that human beings merely act under the idea that they are free, rational, beings in themselves independent of nature in space and time. After presenting Kant’s Groundwork III in strongly metaphysical terms, Guyer quickly notes that “Kant drops the argument we have just examined almost as soon as he has expounded it,” that is, Kant retreats from the position that human beings are rational beings in themselves to the less demanding position that they can think of themselves that way because humans might be other than the way they appear.9
There is thus inconclusive evidence from Guyer’s published writings to determine whether he interprets Kant as a value realist or not.10 Either way the issue remains whether it would be a correct interpretation of Kant to say he is or is not a value realist, and to understand the term “inner value” to refer to an intrinsic property of the object. This is a complex issue, and in this chapter I will present only a few key points in support of my claim. I will examine, in turn, the lack of a place in Kant’s ontology for any inner value property and the nature of value as a formal ordering of ends by reason.
In this section I argue that moral value cannot be understood as something existing as a property of objects independent of either transcendental or empirical moral agency. The textual evidence at first glance is itself ambiguous. There are places where Kant says that a rational being is something “the existence of which in itself has absolute value” (Groundwork 4:428) but also passages in which he just as emphatically states that “nothing has any value other than that which the law determines for it” (Groundwork 4:435).11 Instead of trying to settle the issue by throwing textual passages at one another, a better method is to see which fits better with Kant’s broader philosophy. I claim that there is no room in Kant’s ontology for any such property.
1 The first reason to reject any independent value is that the very idea of independent normative properties does not square with Kant’s division of philosophy into the theoretical – a study of what is, or ontology – and the practical – a study of what ought to be, or the normative. This division is not just a separation of one larger homogeneous topic into two sub-topics, as one might divide the study of life into various taxa such as bacteria, animals, or plants. The theoretical and practical have fundamentally different tasks covering fundamentally different domains (e.g., KrV A841/B869). The domain of theoretical philosophy is knowledge of what exists. The domain of practical philosophy is the determination of free acts, and practical reason provides the law for determining those acts (this is clearest at 27:243 in the Collins metaphysics lectures). The practical concerns behavior, not being. Prima facie the burden is on a value realist to show how the practical is supposed to provide an ontology.
One might object that surely this simple division does not prevent the practical from possessing an ontology of its own. Doesn’t Kant, in fact, discuss on ontology that practical reason apparently provides in the form of the postulates of practical reason: the immortality of the soul, the existence of God, and freedom of the will. But an appeal to the postulates does not help. One reason is that the postulates themselves have nothing to do with moral value. If the postulates do contain Kant’s practical ontology, then the fact that no moral value property is mentioned as a postulate would count as evidence against any claim that practical reason requires that ontology. Further, if the way that practical reason were to offer an ontology would be through the postulates, then it would not help in defending an ontology of moral value unless moral value were also a subject of theoretical reason. In the Critique of Pure Reason, Kant explains the relation of ontology to the practical when in the Canon he discusses the famous three questions of philosophy: what can I know, what ought I to do, and what may I hope? (KrV A805/B833). He says that the first question is speculative, the second practical, and the third “simultaneously practical and theoretical.” Kant raises the need for immortality and God only in answering the last question. The ontology related to the practical is not a product of the practical alone but of the practical in conjunction with the theoretical. The postulates are fully consistent with Kant’s general division of philosophy into practical and theoretical, with only the latter concerning ontology.
2 The second reason that value cannot be an independent property of things is that there is no place for it in Kant’s conception of nature, which is to say it could not be empirically real. Nature is constituted by concepts applied to intuitions and is extended past actual intuitions only through laws of nature.12 Oliver Sensen makes this argument by specifying the manner in which intuitions would give human beings access to independent things or properties.13 Any value property would have to come from sensation (outer intuition) or feeling (inner intuition). No value comes through sensation, and any value derived through feeling would be contingent and not the necessary value of humanity and would also reflect the agent and not merely be a property of the object. These considerations show that, absent any other way of determining the ontology of nature, there could be no empirically real intrinsic value properties in nature.
3 For there to be an inner value to the world or to humanity at the transcendental level of moral agency, it would have to be a property of things in themselves that would be independent of the transcendental moral agent as subject (and thus also independent of the empirical moral agent). Presumably this would be a property of moral agents as objects as beings in themselves independent of their appearance in space and time. But there is a clear difficulty with this position: Transcendental moral agents as subjects would have no access to this property of the transcendental moral agent as object.
The obvious candidate for another mode of access to value properties is practical reason itself. And the third reason that Kant cannot be claiming that there are independent value properties is that reason could not know them directly. Practical reason cannot function in this way. Suppose that practical reason were to be the mode of access for an independent value property. It would have to access it either actively or passively. It is reason’s nature to be active, but by hypothesis reason cannot create this value property because it is independent of the moral agent’s pure practical reason. Reason would have to be able to intuit the independent value of humanity as an end in itself in other beings in themselves, in which case it would have to be passive relative to the value property and the latter would have to actively affect reason. But there is no mechanism by which one being in itself can affect another being in itself except through intuition as appearance, which brings back the point that there is no place in nature for any independent value property.
4 But perhaps reason has no direct access to the independent value property of moral agents in themselves, either actively or passively, but reason can know or prove that there must be such a thing. The fourth point against independent value is that such a claim would be restricted by the nature of reason itself. This option would require that there be some proof by reason that there must be such an independent value. The first Critique’s main goal is to show that a priori arguments by reason cannot be trusted. The best that reason can do is to provide us with ideas that we can use to enhance other claims regarding experience, as the idea of God can be used to see the world teleologically. Reason can make no ontological claim that objects or properties it requires for its own systematic purposes actually exist. One might object here that Kant’s restriction applies to theoretical but not practical reason, since practical reason seems to make existence claims in the postulates. A few paragraphs earlier, I provided reason to think that any ontological claim in the postulates would have to conform to theoretical reason. I also argue elsewhere that the status of the postulates is more like the theoretical ideas of reason than like an assertion of reality (Rauscher 2007). A defender of an a priori argument by practical reason would have to provide some argument to the effect that practical reason, despite as reason being generally untrustworthy in determining existence claims, and despite centering on normative and not ontological claims, is better positioned than theoretical reason to make a priori arguments with ontological conclusions.
5 Perhaps one would claim that the argument is not a priori but depends upon some a posteriori premise. The premise would have to be based on something in moral experience rather than nature. Given the lack of accessibility to any external normative properties as I have argued, the most plausible candidate is the fact of reason, which presents us with the categorical imperative. Assuming that we can take our awareness of the categorical imperative to be a solid enough experience to be the basis of a further argument with ontological conclusions, the issue now comes to precisely what Kant means when he talks about the categorical imperative’s need for a necessary end, that is, humanity as an end in itself.
The specific argument that value realist interpretations use to claim that humanity is intrinsically valuable occurs in Kant’s discussion of the formula of humanity. What Kant claims there is that there must be an end in itself, something the existence of which is an end, and he appears to equate this property of being an end in itself with absolute value (GMS 4:428). Now, what is an end? An end is not a property of anything considered in isolation, as an independent property would have to be, but is conceivable only in relation to some active means-ends reasoning by a moral agent. Even a rational being in isolation could be an end as moral object only in relation to herself as active moral subject. To say that a being is an end in itself is just shorthand for saying that it must be treated as an end at all times. So when Kant does employ reason in an argument that rational beings have absolute value, the conclusion is not that an independent value property exists in them but that they must be treated in a certain way.
6 But perhaps the reason that they must be treated that way is because they have an independent value property that is not “being an end in itself” but is something else. This brings up the sixth reason that value is not an intrinsic property. Kant’s language does not seem to support any other value as being the basis for the status of end in itself; rather, the basis is the non-normative property of humanity. Kant says, “rational beings are called ‘persons’ because their nature already marks them out [auszeichnet] as ends in themselves” (GMS 4:428, my emphasis). The word “auszeichnet” has the connotation not of a revelation of something intrinsic but of the assignment of a value to something – one receiving an honor or a good receiving a price. The term then suggests that there is a property of rational beings, their humanity, whether that be their freedom or their rationality or their capacity to choose to follow the moral law, that makes them something that reason would elevate in the order of ends above all others. So, to say that rational beings in themselves have value is not to say that the value is in the rational beings, but that there is a non-normative property of the rational beings that pure practical reason recognizes. This is the same thing as saying that the moral agent, through the categorical imperative given by pure practical reason, is commanded to treat certain beings as ends in themselves because they possess some non-normative property that reason itself holds is deserving of that treatment. The existence of rational beings and even their humanity are intrinsic properties of the moral object independent of the moral subject, but the value of that humanity is dependent upon practical reason in the moral subject.
These reasons show that value cannot be an intrinsic property of objects independent of the transcendental or empirical moral agent. This is not to say that rational beings are not ends-in-themselves but only to say that what is meant by “end-in-itself” is not a property that rational beings have as objects independent of moral agency. There is no independent value property, there are only non-normative independent properties that reason uses as a basis for its value claims. This relation is best explained by relating value to the order of reason, the topic of my next section.
The value of humanity is not an “inner” or intrinsic property of rational beings. Rather it is merely a part of the formal order of practical reason.
Kant talks about the formal order of reason in several places, using terms such as “intelligible order” and “moral world.” In the solution to the third antinomy, during a discussion of the causality of reason in moral decisions, Kant says that “reason does not give in to those grounds which are empirically given, and it does not follow the order of things as they are presented in intuition, but with complete spontaneity it makes its own order according to ideas to which it fits the empirical conditions” (KrV A548/B576). In the Canon he denies that the “moral world” has any ontological status except to the extent that rational beings, through their actions, impose that order on the world (A808/B837). In the second Critique he refers to an “intelligible order” in at least three places (5:42, 5:86–7, and 5:106). He is clear that the direct application of the moral order to the sensible world provides it with the form of an intelligible world.
This law is to furnish the sensible world, as a sensible nature (in what concerns rational beings), with the form of an intelligible world, that is, of a supersensible nature, though without infringing upon the mechanism of the former… supersensible nature, in so far as we can make a concept of it, is nothing else than a nature under the autonomy of practical reason.
The intelligible order of things is the formal ordering that practical reason provides to nature through the categorical imperative. The intelligible order of things is not an order of intelligible things but an order of things in our experience, particularly in relation to rational beings. Any value has to be understood in relation to this formal ordering by reason.
The best way to understand the formal ordering by reason is to compare it to ordinary empirical ordering of values. An empirical agent is faced with many possible ends, the pursuit of each contingent upon the choice of the empirical agent. No empirical agent can pursue all ends simultaneously and only a finite number successively. Pragmatic empirical agents select from among all possible ends those to pursue and by so doing place a value upon them. As Kant recognizes, this value is entirely dependent upon the subject’s discretion (GMS 4:427) and is not an intrinsic property of the ends or objects related to those ends. Further, an empirical agent values some ends more than others and so creates a hierarchy of ends. For an empirical end to be valuable at all is for it to have a place in some empirical agent’s set of chosen ends, for an end to be more valuable is for it to have a higher rank in such a hierarchy.
The intelligible order of reason is analogous to this empirical order of contingent ends. At the transcendental level, the moral agent is considered only abstractly and does not have access to particular ends. But the moral agent is aware that there are two possible kinds of ends: the contingent ends from which she will be able to select her particular ends as an empirical agent, and necessary ends that would not be optional at the empirical level. Pure practical reason, as a faculty of the transcendental moral agent, identifies humanity (understood by Guyer as freedom) as this necessary end. Pure practical reason’s selection of humanity as an end is what gives humanity value, and the ranking it provides in setting this necessary end above all possible contingent ends is what gives humanity absolute value. Reason provides its own order at the transcendental level that parallels the empirical level in that in both levels the value of the ends, whether contingent or necessary, is dependent upon the moral agent. Rather than being an independent property of rational beings independent of reason’s imposition of order through the categorical imperative, absolute value is only the place of some ends in the formal ordering of ends by pure practical reason.
The order of reason is a reflection of the formal character of the categorical imperative. Practical reason operates through form alone. Through the categorical imperative, pure practical reason imposes its own intelligible order on nature without adding any content to nature except the actions that reason itself causes. It does this by insisting on a certain order of ends in which humanity is given priority over all other ends. The demand of reason is for empirical moral agents to use humanity as a limiting condition when deliberating about subjectively chosen ends, and the means used to reach those ends. Those empirical moral agents are restricted in their choices by the demands of transcendental moral agency.
The specific formulas of the categorical imperative that impose this necessary order of reason in relation to ends are the formula of humanity and, derivatively, the formula of the kingdom of ends. Here is the place to show how I can account for Kant calling the value of humanity the “matter” of the categorical imperative even though the ordering of reason is merely formal. He also says, as I quoted already, that a rational being is something “the existence of which in itself has absolute value” (GMS 4:428, my emphasis).
Absolute value as ranking in the merely formal order of ends and not an intrinsic property of rational agents is still compatible with these claims. Consider the claim that the categorical imperative has the value of humanity as necessary ends as its “matter.” This need not mean that the “matter” is an independent existent that reason recognizes; it can just as easily mean that reason determines a priori that certain ends are of value and will then constitute the “matter” at issue. There are two possible scopes for the necessity of the end: It can be a necessary end for maxims or a necessary end of the moral law itself. Kant’s summary of the formulas of the categorical imperative identifies the necessary end to be an end for maxims, not for the categorical imperative as such.
all maxims have, namely … a form, which consists in universality, … a matter, namely an end, and in this respect the formula says that a rational being, as an end by its nature and hence as an end in itself, must in every maxim serve as the limiting condition of all merely relative and arbitrary ends.
The necessity is for maxims to conform to humanity as an end in itself, not for the categorical imperative to conform to humanity as an end in itself. When Kant presents the formula of humanity he says that rational being as an end in itself “is at the same time an objective principle from which, as a supreme practical ground, it must be possible to derive all laws of the will” (GMS 4:429, my emphasis). Kant here treats ends in themselves not as the ground of the categorical imperative but as the ground of laws that we would derive using the categorical imperative. The categorical imperative can, as the command of reason to order ends in a certain way, itself provide the necessity to the ends that maxims must take into account. The categorical imperative as the means by which pure practical reason imposes its formal order onto nature is still purely formal in the broad sense.
A typical realist objection to this claim is to ask for a basis on which reason would assign value. Is there a ground for reason’s assignment of value to humanity and not something else? Surely, the objection goes, that ground would itself have to be a preexisting value, for otherwise it would appear that reason is creating value arbitrarily. This is a broad topic about which I will only make a suggestion. Many understand “humanity” in Kant as referring to the ability to rationally deliberate and decide upon courses of action. We have value in virtue of that practical rational capacity. If this is true, then there is a ground for practical reason to assign value to humanity because in doing so practical reason is assigning value to its own free exercise. Reason valuing itself should be no surprise. Reason’s selecting of rational activity as an end is not arbitrary at all. Further, there is good reason to think that if humanity is understood this way but the value of humanity is understood in realist terms as an intrinsic property not stemming from reason, then it would appear that practical reason simply lucked out that it is able to command as a moral imperative the maximization of itself.
In this section I have argued that the best way to understand Kant’s claims about the value of humanity is in terms of the ordering of ends imposed on moral agents by pure practical reason. No independent value property is needed to explain the role of absolute value in Kant.
The answer to the broad question of whether Kant is a value realist is “no.” There is no property of objects “absolutely valuable” that is independent of either the transcendental moral agent or the empirical moral agent. Kant’s ontology of nature does not allow for such a property, the nature of practical reason precludes ontological claims, and when examined closely the alleged independent value, that of “end in itself,” is at best relational because of its dependence on active beings who pursue ends. Instead I have shown that Kant’s theory of absolute value is both transcendentally and empirically idealist because it is constituted by the formal order imposed by reason.
So why would Kant say, as so frequently quoted by Guyer, “freedom is the inner value of the world”? Kant never received any awards for consistency, but I speculate that when Kant uses language about inner value, particularly to the young students in the lectures on ethics from which this quotation comes, he is simply trying to stress that we human beings are subject to dictates of reason that command absolutely and that, in this particular case, freedom is the focal point of reason’s normative ordering of our experience. I have no problem with occasional imprecision for the sake of rhetorical flourish, but we must read his claims in their broader context.14
1 He quotes the passage at different lengths, and using either the Mrongovious (Mrong 27:1482) or the Collins Lectures (Collins 27:344) versions, in Guyer 2000d; Guyer 2000a; Guyer 1996a; Guyer 1996c; Guyer 2005b; his survey book Kant (Guyer 2006b:178); at least one book review (Guyer 1996d); and at least one paper in Portuguese (Guyer 1995).
2 Guyer particularly argues that freedom rather than reason must be of moral value in a review essay on Christine Korsgaard’s book Creating the Kingdom of Ends (Guyer 1998), but lays out his argument for the unconditional value of freedom in the two papers just cited (Guyer 2000a and 2000d).
4 This kind of realism differs from one focused on the existence of moral facts or the objective validity of moral claims. A widely used definition from Geofrey Sayre-McCord centers on truth: He takes realism to consist of two theses: “(1) the claims in question, when literally construed, are literally true or false (cognitivism) and (2) some are literally true” (Sayre-McCord 1988:5). I take this kind of definition to be lacking when applied to Kant, for surely it is wrong to equate “true” and “real” when discussing transcendental idealism. Kant certainly claims that all a priori knowledge is true even when it is understood as ideal rather than real in a transcendental sense. To equate “true” and “real” is wrongly to place the dividing line between realists and non-realists, who make different kinds of metaphysical claims, precisely atop the dividing line between those who accept and those who reject morality, who make much broader claims about the validity of moral practice as a whole. Certainly in Kant interpretation the interesting question that divides realists and anti-realists is not whether the moral law is valid but what makes it valid. This kind of definition is particularly unhelpful with regard to moral value, since no one disputes that Kant thinks that there is absolute value. The disagreement centers on what kind of thing that value is.
5 My emphasis on the agent as moral subject as opposed to the agent as moral object precludes any quick and trivial argument to idealism through a claim that freedom (or humanity or dignity) are properties of moral agents as objects (items of concern in morality) and so dependent on the very existence of rational agents. If it were to turn out that moral value were an intrinsic property of rational agents as objects independent of whether any moral agents as subjects considered them to be so, then moral value would be real, not ideal.
7 In Guyer 2007b:452, he argues that the pre-critical Kant possessed a naturalistic view of morality that the critical Kant superseded with his transcendental, a priori account.
8 But for an argument that Guyer does not take value to be an independent property, see Sensen 2011:79–87. Sensen claims that “Guyer does not present the value of freedom as a distinct metaphysical property” and a footnote to this sentence claims that Guyer confirmed this fact in a personal conversation (Sensen 2011:84).
9 Guyer notes this reversal in a few places: e.g., (Guyer 2007a:167) and (Guyer 2007b:461) but is more strongly metaphysical in Guyer 2009b.
10 In conversations he has stepped back from the realist ontological implications of the term “inner value,” including conversations stemming from the version of this chapter read at the “Nature and Freedom in Kant” conference in 2013.
11 Two recent studies evaluate these and other passages in detail. Robert Stern argues that they support realism (Stern 2011:26–40); Oliver Sensen that they do not (Sensen 2011:21–3, 39–51).
12 Kant makes the point that ontology of the actual is a product of intuitions and laws of nature when discussing the possibility of magnetic matter at KrV A225–6/B272–4.
13 Sensen 2011:19–20, where his main point is that even if there were an independent value property, it would be inaccessible to moral agents and could play no role in morality.
14 I would like to thank my commentator Rob Hoffman at the Nature and Freedom in Kant conference, as well as that audience and the audience at Western Michigan University for comments on earlier versions of this chapter.