Do the concepts of dignity and value play a significant role in Kant’s ethics? Or is it rather that for Kant these concepts are dispensable, or at least dispensable when it comes to the foundations of his own ethical thought? In this paper, we shall provide answers to these questions with regard to Kant’s Groundwork of the Metaphysics of Morals (GMS). We will argue that for Kant dignity and value are crucial, indispensable concepts that are part of what we call Kant’s ground-thesis, i.e., the claim that rational nature exists as an end in itself and is the ground of the categorical imperative (CI).

Our analysis was sparked by Oliver Sensen’s truly challenging book Kant on Human Dignity. Sensen raises a very interesting question: Do humans beings have dignity because or inasmuch as we must respect them, or do they have dignity on the basis of a worth they already possess? His basic answer is this: “It is not because others have a value that we should respect them, but it is because one should respect them that they have an importance and a dignity”.Footnote 1 On Sensen’s and other revisionist readings, to have dignity simply means to be elevated, and therefore “‘Has value’ is merely another way of saying ‘should value’”.Footnote 2 For these revisionist readings, both the concept of absolute value and the concept of dignity are dispensable. Revisionists also argue that according to Kant value (dignity) is not a “metaphysical property” that exists “‘out there’”.Footnote 3 Furthermore, some defenders of the traditional reading understand Kant’s theory of value in terms of moral realism, others have a rather constructivist approach, but still hold on to the concept of value and dignity.Footnote 4 For our purposes, we shall ignore this important difference. We would like to point out that at least within Kant’s own philosophy and terminology, there can be no doubt at all that freedom and everything based upon it are metaphysical properties. Freedom is “a particular kind of causality” (460,16)Footnote 5 that is both non-natural and real (where these qualities – ‘non-natural’, ‘real’ – can be understood both from a contemporary and from Kant’s point of view and terminology).

At first glance, at least, Kant’s GMS contains a number of famous passages which strongly support the standard reading.Footnote 6 According to this line of interpretation, Kant claims that human beings as autonomous beings are ends in themselves that possess dignity and value, and that this moral status to be an end in itself is the ground of the CI. Hence, dignity as well as value are concepts that are essential to Kant’s ethics; although dignity is strongly related and, in some sense to be specified, even based on the concept of law and lawgiving, it cannot be replaced by these terms. We shall argue that the burden of proof is on those who advocate the revisionist reading; because of the strong prima facie evidence that speaks in favor of the standard reading, it is up to them to show we have reason to think that this reading is wrong nonetheless.

The underlying question – is value or dignity in Kant’s ethics the most fundamental moral property or is this property subordinate to the formal law itself – should be of interest not only to Kant scholars. After all, the difference between the right and the good has been central to many ethical debates, and it will be helpful to see that an alleged proponent of ethical formalism is much closer to a deep value theory than is often thought. Put in a nutshell, Kant’s argument is that the very idea of a categorically (morally) commanding norm presupposes the existence of something in the world that has absolute value. This, in turn, is the basis of what one could call (Kant’s) moral realism.Footnote 7

We shall first (1) provide an outline of Kant’s ethical foundations as we read it, in order to prepare our analysis of the ground-thesis. We will then have a closer look at the parallel Kant draws between hypothetical imperatives and their ground and categorical imperatives and their ground. The purpose of this part is not to offer a detailed interpretation (let alone a systematic defense) of Kant’s ground-thesis and the literature on it; rather, we shall outline such an interpretation arguing that this will suffice to demonstrate that the burden of proof is on the revisionist reading and that it cannot account for Kant’s ground-thesis (2).Footnote 8 There is one sentence in GMS II, however, that does seem to support the revisionist reading. This sentence is: “For nothing has a value except that which the law determines for it” (436,1); let us dub it the law-sentence. Our close reading shall show that the law-sentence does not support the revisionist’s claim; rather, it supports the standard reading that dignity itself is the ‘ground’ of the CI and hence the ‘ground’ for the respect we have to pay to all persons.

1 An outline of Kant’s ethical foundations

Before we provide an interpretation and defense of Kant’s ground-thesis against the revisionist reading, let us look at some key terms related to it. They will help us to better understand the ground-thesis. Our claim is not that everyone who defends the standard reading will agree with our interpretation of those key terms; but this interpretation certainly implies at least one version of the standard reading.

1.1 The good will and autonomy

We have to differentiate three different aspects of Kant’s concept of a good will: The noumenally-good will, the practically-good will, and the holy will. The noumenally-good will is the autonomous will that as such wills the good. As a moral faculty, it gives the law (the CI) for imperfect beings and, by means of moral feelings, it is also a motivating force. Every human being has such a will, even if he or she acts morally bad. The importance of this idea of the noumenally-good will is well established by a considerable number of passages.Footnote 9

The noumenally-good will is the basis both for the practically-good will and the holy will. The practically-good will is the will that finite beings have when their volition is indeed moral; it is the noumenally-good will considered as a will that manifests itself successfully in a finite being against the influence of inclinations and desires. For imperfect beings, to act morally (to act with a practically-good will) means to act from duty. The noumenally-good will that is manifest in a person without (active) sensual hindrances is what Kant calls the ‘holy will’; it only belongs to God and other holy (infinite) beings. These beings have no inclinations and desires contrary to the good; the “will whose maxims necessarily harmonize with the laws of autonomy is a holy, absolutely good will” (439,28). The noumenally-good will as such (regardless of its being incorporated in a finite being) cannot be differentiated from the holy will (regardless of its being incorporated in an infinite being).

At first sight, autonomy seems fairly easy to understand: It is “the property of the will through which it is a law to itself” (440,16). However, Kant’s distinction between the world of understanding and the world of sense as well as, accordingly, his distinction between a noumenally-good will and a practically-good will, complicate this ‘property of the will’ considerably. For Kant uses the term ‘autonomy’ not only for the human being and its capacity for a practically-good will, but also for the property of the noumenally-good will, just considered as the intelligible will. Recall how Kant understands the will itself as a noumenal causality: “The rational being counts himself as intelligence in the world of understanding, and merely as an efficient cause belonging to this world does it call its causality a will” (453,17, first emphasis ours). This will is then identified with the will that is autonomous, i.e., with autonomy itself: “if we think of ourselves as free, then we transport ourselves as members into the world of understanding and cognize the autonomy of the will, together with its consequence, morality” (453,11, o.e.). Note how Kant continues: “… but if we think of ourselves as obligated by duty, then we consider ourselves as belonging to the world of sense and yet at the same time to the world of understanding” (453,14, o.e.). Thus the free will is the noumenal will, and autonomy is its property; and in some contexts, this will is considered not as the will of a human being that is also part of the sensible world, but as a noumenal will only: as “a mere member of the world of understanding, all my actions would be perfectly in accord with the principle of the autonomy of the pure will” (453,25, o.e.). It is important to keep in mind that the noumenally-good will as such is not only a mere capacity to act morally; for this will as such wills the good. Nonetheless, it is the noumenally-good will that enables the human being to act morally; thus, for the human being – who is a member both of the noumenal and the sensible world – the noumenally-good will is indeed a capacity. And unless autonomy and having a practically-good will are not the same, a scoundrel would not be autonomous – which he actually is insofar as even he, to some extent, wants to be morally good, i.e., insofar he has a noumenally-good will. We will return to this once again.

1.2 End in itself and rational nature

The concept of ‘autonomy’ is strongly related to the concepts of ‘end in itself’ and ‘rational nature’, and yet it in GMS II Kant says very little about what exactly an ‘end in itself’ is and what ‘rational nature’ means.Footnote 10 To be sure, “rational beings […] are called persons, because their nature already marks them out as ends in themselves” (428,21). Yet this does not tell us what a ‘person’ or precisely a ‘rational being’ is. In the Religion, Kant famously distinguishes between animality, humanity, and personality, and it has been a recurring misinterpretation to ascribe to Kant the position that humanity as the ability to set ends is what deserves respect.Footnote 11 But clearly it is not.Footnote 12 It is personality, even though Kant often speaks of ‘humanity’ when he means ‘personality’, i.e., freedom as the ability to give the moral law and obey it. In his so-called lectures Naturrecht Feyerabend (1784) it becomes clear that it is freedom of the will which makes human beings ends in themselves: “The freedom of the human being is the condition under which the human being himself can be an end” (Fey: 1320). Or again: “I must presuppose the freedom of this being if it is to be an end in its own eyes” (Fey: 1322). It is also quite obvious from this lecture that a being is not an end in itself simply because it is rational and capable of setting ends: “If rational beings alone can be ends in themselves, they cannot be so because they have reason, but because they have freedom. Reason is merely a means. – Through reason the human being could produce in accordance with universal laws of nature, without freedom, what the animal produces through instinct” (Fey: 1321–1322). Only if a rational being is free in the positive sense that this freedom is “a law for itself” (Fey: 1322) is such a being an end in itself and possess value: “The inner worth of the human being rests upon his freedom, upon the fact that he has a will of his own” (Fey: 1319). This ‘inner worth’ is also called here “dignity” (Fey: 1319).

Thus, when it comes to be an end in itself, the relevant rationality is moral rationality and thus autonomy. These terms, in turn, are related to the term ‘dignity’, because autonomous beings are ends in themselves; and since ends in themselves have dignity, autonomous beings have dignity. But is it really the noumenal (‘autonomous’) will that grounds dignity? The following places seem unambiguous: “Autonomy is thus the ground of the dignity of the human and of every rational nature” (436,6); and: “the dignity of humanity consists precisely in this capacity for universal legislation, although with the proviso that it is at the same time itself subject to this legislation” (440,10); or: “the will of one rational being must always at the same time be regarded as universally legislative, because otherwise the rational being could not think of the other rational beings as ends in themselves” (434,23).

Yet there are contexts in which Kant seems to say that what is an end in itself is not the human being as an autonomous being (with its capacity to give the moral law and obey it) but the human being who has a practically-good will.Footnote 13 One passage, of course, is the very first sentence of GMS I on the good will being the only thing that is ‘good without limitation’. Another is found in Kant’s theory of the realm of ends. To have a practically-good will seems to be a necessary condition in order to be a member in the realm of ends: “Such a realm of ends would actually be brought about through maxims, the rule of which is prescribed by the categorical imperatives of all rational beings, if they were universally followed” (438,29). We shall now have a closer look at the concept of dignity and then see that it cannot be the practically-good will that grounds absolute value, but autonomy in a human being as a capacity for morality.

1.3 On dignity and value

Kant uses the term Würde (dignity) in numerous places in various ways. Sensen quite rightly points out that ‘dignity’ in Kant’s writings is not limited to a moral meaning.Footnote 14 One can have dignity as a king, a teacher, a mathematician, and so on. We do not dispute this, of course; we do not mean to say that in all possible contexts ‘dignity’ means the same as ‘absolute inner value’. Here, in our context, when we speak of dignity we mean the dignity of a being it has due to its autonomy, and we have seen already that it is strongly related to other terms such as ‘end in itself’, ‘nouemenal will’, ‘autonomy’, and so on. To prepare our interpretation of Kant’s ground-thesis, let us now look at dignity more specifically.

We suggest that Kant has a two-level-model of being an end in itself. According to this model, every human being is an end in itself and has intrinsic value because every human being has a noumenally-good will and therefore the capacity to be moral. And since every human being (even a scoundrel) has this noumenally-good will, every human being is an end in itself and has to be respected. On a second level (in another respect), however, one has to take into consideration whether someone actually intends and acts morally, i.e., whether she also has a practically-good will. And if she does, she is an end in itself in an elevated sense and therefore has value in an elevated sense.

How is this two-level-model related to the concept of dignity? Kant’s theory of a realm of ends seems to leave no other choice than to interpret him as saying that having a practically-good will (being a virtuous person) is not only a sufficient, but also a necessary condition for being a member in the realm of ends.Footnote 15 It is in this context of the realm of ends – in which all members have a practically-good will – that the term ‘dignity’ appears for the first time (434). Therefore, and because of the thesis associated with the very first sentence in GMS I, it seems conceivable at first glance that Kant wants to ascribe dignity only to the second level, i.e., to beings with a practically-good will. There are, however, not only passages in which Kant seems to say a practically-good will is necessary for a being to be an end in itself; there are also other passages, already quoted (see 436,6; 440,10), in which he says it is just the capacity. So let us assume that there is textual evidence for both interpretations. In this case, there are still two systematic reasons that make it highly implausible that Kant would hold that a being with a practically-good will alone is an end in itself: First, Kant says that what is an end in itself is the ‘ground’ of the CI. If what makes a being an end in itself were it having a practically-good will, then the practically-good will would be the ‘ground’ of the CI – which makes no sense since to have a practically-good will already presupposes to be determined by the CI. Second, there are human beings who do not have a practically-good will. Could it really be Kant’s position that they are not ends in themselves, i.e., that they are not persons and hence have no dignity? After all, rational beings are called persons, says Kant, “because their nature already marks them out as ends in themselves” (428,22, o.e.). But to have a practically-good will is not a matter of ‘nature’; it is a matter of acquisition. (And to say that bad persons are no persons seems also strongly at odds with what Kant says about imputability.) Thus, there are two quite convincing arguments suggesting that it is rational, autonomous nature which is an end in itself and thus rational, autonomous nature that has dignity. Indeed, Kant makes use of this expression (‘dignity’) even when he talks about an end in itself in terms of a being which is only capable of being moral: “Thus morality and humanity, insofar as it is capable of morality, is that alone which has dignity” (435,6, o.e.), and as already quoted: “the dignity of humanity consists precisely in this capacity for universal legislation” (440,11, o.e.). Hence we are justified in holding that Kant has a two-aspect-model of dignity as well: In one respect, “every rational nature” (436,7) has dignity, inasmuch as rational beings are autonomous (and with regard to human beings it is the “dignity of humanity as rational nature”, 439,4). In another respect, Kant also speaks of the dignity of the “way of thinking” (Denkungsart, 435,25), i.e., of the “morally good disposition or virtue” (435,29), or of the “dignity in a person who fulfills all his duties” (440,1). Thus every rational being in its capacity to give the moral law and to obey it has dignity; and every finite rational being has its own level of dignity, depending on how virtuous it is.

It is very important to see that Kant also speaks quite often of the end in itself as having an ‘absolute value’ or an ‘inner value’.Footnote 16 This ‘absolute’ or ‘inner value’ in turn is identified in both ways with dignity: “…inner value, i.e., dignity” (435,4), or, the other way round, “…dignity, i.e., an unconditioned, incomparable value” (436,3). Indeed, in ethical contexts ‘dignity’ is just a catchword for ‘absolute inner value’. To say that x has dignity is tantamount to saying that x has absolute inner value. Since an autonomous being as an end in itself has absolute value, and since ‘absolute value’ is tantamount to ‘dignity’, Kant relates an end in itself to (autonomy and) dignity.

Now it is true, of course, that one aspect of possessing dignity (morally understood) is that whoever does possess dignity is elevated.Footnote 17 But this does not at all imply that “‘dignity’ means nothing but ‘raised above’”.Footnote 18 A person who is taller than another person certainly is taller in relation to that other person. Unless there is another person, and a person that is smaller than her, she is not taller; still she would not be taller in relation to that other person if she had not a certain height as such in the first place. By the same token, Kant often says that human beings are elevated above animals. But this cannot simply mean that animals do not possess a certain quality (freedom) which human beings do possess. Probably all animals possess qualities that human beings do not possess; many animals run much faster than we do, but that does not imply that they are raised above human beings, because the ability to run fast is irrelevant to the question of what species is superior. As living beings we are elevated in toto above animals, and this requires that we have a quality with a certain absolute value.

Furthermore, that which raises human beings above animals is not simply the descriptive fact that they possess freedom as a moral capacity. If this were all, then all we could say would be that human beings possess freedom and animals do not. But to say that human beings are ‘above’ animals only makes sense if there is a standard, i.e., a value, in regard to which we can say that something has a higher value, that is, is raised above someone else (more on this later). It is the term ‘value’ by which we can state that to be a moral subject is a “prerogative” (TL: 420,17).

2 Kant’s ground-thesis

Now let us take a closer look at Kant’s ground-thesis, first, at the ground-thesis itself (2.1), and then at the law-sentence that seems to undermine it (2.2).

2.1 The end in itself as the ground for the categorical imperative

In GMS 426, Kant makes a step forth into the “metaphysics of morals” (426,30). His leading question is how “reason for itself alone determines conduct” (427,15). The context for everything that follows is Kant’s claim that “which serves the will as the objective ground of its self-determination is the end” (427,22). Kant distinguishes between “subjective ends” (427,28; 428,25) and “objective ends” (427,29; 428,27). Objective ends, unlike subjective ends, are ends that are “valid for all rational beings” (427,25; 427,29, o.e.); after all, it is one of the questions to be answered by the ‘metaphysics of morals’ whether the moral law is a “necessary law for all rational beings” (426,22). It is beyond any doubt that Kant parallels the “ground of hypothetical imperatives” (428,2) with the “the ground of a possible categorical imperative, i.e. of a practical law” (428,5). Again, an end ‘serves the will as the objective ground of its self-determination’. To say that such ends are ‘subjective’ is to say that they only have a “relative value” (428,20), and this is to say that ‘subjective ends’ “can provide no necessary principles valid universally for all rational beings and hence valid for every volition, i.e.,, practical laws” (427,36, o.e.). From this Kant concludes: “Hence all these relative ends are only the ground of hypothetical imperatives” (428,1). Kant then continues:

“But suppose there were something whose existence in itself had an absolute worth, something that, as end in itself, could be a ground of certain laws; then in it and only in it alone would lie the ground of a possible categorical imperative, i.e., of a practical law. [new paragraph] Now I say that the human being and in general every rational being exists as an end in itself, not merely as means to the discretionary use of this or that will, but in all its actions, those directed toward itself as well as those directed toward other rational beings, it must always at the same time be considered as an end […] The ground of this principle [i.e., the CI] is: Rational nature exists as end in itself”. (428,3–429,3)

Subjective ends can only ‘ground’ hypothetical imperatives. But if there are to be ‘necessary principles valid universally for all rational beings and hence valid for every volition, i.e., practical laws’, they must have their ‘ground’ in objective ends; for only these ends, and not ‘subjective ends’, have an “absolute value” (428,4; 428,30) and therefore ‘can provide’ principles valid for all rational beings. Only objective ends, i.e., ends in themselves that are of absolute value, can be a ‘ground’ for the CI. Every rational beings exists as an end in itself; hence human beings (as rational beings) exist as ends in themselves. Therefore, human beings as ends in themselves can be a ground for the CI. This is the basic idea of Kant’s ground-thesis.

It is of the utmost importance to understand the parallel between the ‘ground’ of hypothetical and the ‘ground’ of categorical imperatives. Kant draws this parallel in the first two paragraphs which mark the inception of the ‘metaphysics of morals’; let us call this passage (the two paragraphs, 427,19–428,6) the parallel-passage. Now there is no doubt that Kant claims that relative (subjective) ends are the ‘ground’ of hypothetical imperatives, for this is plainly what he says by asserting that relative ends are only the ‘ground of hypothetical imperatives’. There is, secondly, also no doubt that, parallel to this, Kant claims that rational nature (i.e., autonomy and its absolute value) is the ground of the CI; for again, this is plainly what he says in that crucial passage cited above, to wit, that rational nature as an ‘end in itself’ has an ‘absolute worth’ and that ‘only’ such an end can be the ‘ground of a possible categorical imperative’. This – Kant’s ground-thesis – is what he says ad litteram; in whatever way one interprets this parallel, it needs to be taken into account seriously.

So let us take a serious look. The immediate context of the ground-thesis is Kant’s claim in the very first sentence of the parallel-passage that the “will is thought as a faculty of determining itself to action in accord with the representation of certain laws” (427,19). These laws are hypothetical and categorical imperatives. The term ‘ground’ then appears (in the parallel-passage) for the first time when Kant says that the end “serves the will as the objective ground of its self-determination” (427,21, o.e.). To say that ‘relative ends’ are the ‘ground’ of hypothetical imperatives is to say that relative ends make hypothetical imperatives possible. Yet hypothetical imperatives are reasons to act, they necessitate; hence relative ends make necessitation possible. And if a rational being follows these reasons and their necessitation, it ‘determines itself to action in accord with the representation of hypothetical imperatives’. Therefore, to say that relative ends are the ‘ground’ of hypothetical imperatives is to say that they are their ground of possibility: Without relative ends, there simply are no hypothetical imperatives. The necessitation through a hypothetical imperative is possible because subjects set and pursue relative ends which provide reasons to act in certain ways, i.e., to achieve means to the given end. Without relative ends, no corresponding (hypothetical) reasons would be given and therefore no hypothetical imperatives which necessitate the subject. And so the ‘end’ is a ‘ground’ inasmuch as the ‘end’ has a function: It gives rational beings reasons to act. (On the basis of Kant’s motivational internalism, this also leads the subject to have a motive to act in a certain way, but this is not the point here.)

By paralleling this relation between a relative end and the ‘ground of hypothetical imperatives’ with the relation between an end in itself and the ‘ground of categorical imperatives’, Kant claims that the ‘end in itself’ is the ‘ground of categorical imperatives’ in structurally the same way. Absolute ends are the ‘ground’ of the possibility of categorical imperatives: There simply would be no categorical imperatives unless there are absolute ends. Categorical imperatives are reasons to act; they necessitate; hence absolute ends make categorical necessitation possible. If a rational being with a will follows these categorical reasons and their necessitation, it ‘determines itself to action in accord with the representation of categorical imperatives’. So again, as regards ‘absolute ends’, they are a ‘ground’ inasmuch as they have a function: They give rational beings categorical reasons to act.

It is not the first time that Kant draws a parallel between the ‘ground of hypothetical imperatives’ and the ‘ground of a possible categorical imperative’; as a matter of fact, this parallel is the context and basis for the ‘step into the metaphysics of morals’. In GMS 417, Kant raises the following question: “How are all these imperatives possible?” which he identifies with the question of “how to think the necessitation of the will” (417,5). The answer is given with regard to hypothetical imperatives, and it is basically the same answer Kant gives in GMS 427–428: We easily understand, he says, the “ground of making the act of the will actual” (417,17, o.e.), since hypothetical imperatives necessitate by presupposing an end that is already willed such that the hypothetical imperative commands one to use the necessary means to achieve that end. Kant postpones the answer to the question of how the CI is possible, i.e., he postpones the answer to the question of “how to think the necessitation of the will that the imperative expresses” (417,5). He first develops the very idea of a CI and its formula (420–425), and he concludes once more with the remark that he is “still not ready” (425,7) to answer that question at this point of the argument. Kant next discusses the necessity of a ‘metaphysics of morals’, and the question is still whether the moral law is “a necessary law for all rational beings” (426,22). In order to answer this question, Kant argues, one must take that ‘step into the metaphysics of morals’. For it is a question of metaphysics to ask whether “reason for itself alone determines conduct” (427,17); the moral law only is a law for all rational beings if it is “bound up (fully a priori) with the concept of the will of a rational being in general” (426,25). This is the task of the ‘metaphysics of morals’ – which we enter into in the parallel-passage. Kant’s ground-thesis therefore is related to the question in GMS 417 of ‘how categorical imperatives are possible’, or how we can think of the categorical necessitation of the will of all rational beings. The answer is: because there is an objective end ‘for all rational beings’ which provides categorical reasons and therefore necessitates all rational beings.Footnote 19

One might think that the parallel between the grounding function of relative and absolute ends is not this parallel Kant has in mind here, but another. Given that autonomy is what makes a rational being an ‘end in itself’, one might think that Kant addresses this autonomy as the ‘ground’ of the CI inasmuch as this autonomy is what enables us to be obliged by the CI: Unless we are autonomous, we cannot be obliged by the CI; ought implies can. Autonomy would be the ‘ground’ of the CI in the way freedom is discussed in sec. 2 of GMS III, to wit, as a quality of the will that must be “presupposed” (447,27, o.e.); it is a quality that “we have to presuppose […] if we would think of a being as rational and as endowed with consciousness of its causality in regard to actions, i.e., with a will” (449,1, o.e.). So in this reading autonomy as the end in itself would be the ‘ground’ of the CI in terms of being the latter’s presupposition.

Now it is true, of course, that freedom (autonomy) in that well-known sense is the ‘ground’ (“ratio essendi”, KpV: 4) of the CI. Nevertheless this is not the topic of the parallel-passage, and it is not the topic of the parallel. To be sure, Kant does say at the beginning of the parallel-passage that the faculty of determining oneself to action in accord with the representation of ‘certain laws’ can “be encountered only in rational beings” (427,21). But clearly this presupposition is not what Kant is occupied with in the parallel-passage: Immediately after having made the very brief introductory remark that only rational beings can determine themselves in accord with the representation of imperatives, Kant introduces the concept of an ‘end’; and thus almost the entire parallel-passage is just about this concept of an ‘end’, not about that presupposition. Hence the parallel between the ‘ground’ of hypothetical imperatives and the ‘ground’ of categorical imperatives cannot be that the former presupposes rationality and the latter autonomy; this is true, but simply not the topic here. The topic in the parallel-passage is the concept of relative and objective ends, and their grounding function.

2.2 The law-sentence: ‘nothing has a value except that which the law determines for it’

On the revisionist account (such as Sensen’s), ‘end in itself’, ‘value’, ‘dignity’ are more or less expendable concepts and certainly have no ‘grounding’ function. All the more, it needs convincing evidence that the ground-thesis must be understood as a “rare exception[.]”Footnote 20 in Kant’s theory rather than a serious philosophical thesis; the burden of proof is on the revisionist reading.

In Sensen’s approach, refuge is sought in what we call the law-sentence: “For nothing has a value except that which the law determines for it” (436,1).Footnote 21 The entire passage (G) reads as follows:

[G1] And now, what is it that justifies the morally good disposition or virtue in making such high claims? [G2] [G2a] It is nothing less than the share that it procures for the rational being in the universal legislation, thereby making it suitable as a member in a possible realm of ends, [G2b] for which it by its own nature was already destined, as end in itself and precisely for this reason as legislative in the realm of ends, as free in regard to all natural laws, obeying only those that it gives itself and in accordance with which its maxims can belong to a universal legislation (to which it at the same time subjects itself). [G3] For nothing has a value except that which the law determines for it. [G4] [G4a] The legislation itself, however, which determines all value, must precisely for this reason have a dignity, i.e., an unconditioned, incomparable value; [G4b] for which the word respect alone yields a becoming expression for the estimation that a rational being must assign to it. [G5] Autonomy is thus the ground of the dignity of the human and of every rational nature”. (435–436)

Let us first turn our attention to the law-sentence ([G3]). In doing so, we will, for lack of space, abstract from [G1] and [G2], the preceding paragraph of G as well as from the parallel-passage and the other passages on the realm of ends, autonomy, and dignity. We will also presuppose the two-level-model, i.e., we will assume that it is the noumenally-good will (autonomy) that is an end in itself.

At first glance and without respecting the immediate context, it seems as if Kant were saying that whichever x has value it does so because of the moral law. It turns out, however, that by the ‘law’ Kant means ‘legislation’. If value is endowed by moral legislation, the source of this legislation itself must have value; value is not brought about by something that has no value, says Kant. And what has value is rational nature with its ability to legislate, and this is just what the ground-thesis says. Let’s take a detailed look.

Abstracting then from the relation of [G3] to [G2] (marked by ‘For’, Denn), we can reconstruct [G3] (the law-sentence) in a way that can hardly be disputed:

[G3]* Nothing has a value except the value which the moral law determines for it (i.e., for that very thing that has a value).

In [G3], Kant speaks of ‘the law’. However, this term (‘the law’) is not used again in [G4]. Rather, it is replaced by ‘legislation’ (Gesetzgebung). This yields:

[G3]** Nothing has a value except the value which moral legislation determines for it (i.e., for that very thing that has a value).

Three main arguments speak in favor of this claim (that ‘law’ in [G3] is identical with ‘legislation’ in [G4]):

  1. 1.

    The key term in [G3] is determination of value. The very same term is used again in [G4]: ‘The legislation itself, however, that determines all value…’. Unless one assumes that Kant has two different kinds of ‘determination of value’ in mind, we must assume that the thing that ‘determines all value’ is one and the same in each sentence; and since it is ‘legislation’ in [G4], it must be legislation in [G3] as well (not the other way round, however, since autonomy, i.e., legislation, is undeniably the ‘ground of the dignity of the human and of every rational nature’).

  2. 2.

    [G4] begins with ‘The legislation itself, however (aber),…’. This conjunction (grammatically speaking) connects [G3] with [G4a] by way of emphasizing that the expectation raised by [G3] – that strictly speaking ‘nothing has a value except the value which the moral law determines for it’ – is not adequate, or put another way, that there is a restriction to [G3]: ‘nothing has a value except the value which the moral law determines for it’, except the moral law itself; and rather than saying that the moral law itself has an ‘unconditioned, incomparable value’, Kant says it is the ‘legislation itself’ that has this value.

  3. 3.

    In [G5] Kant draws the conclusion (‘thus’ [Also]) that ‘autonomy’ (i.e., selflegislation) is ‘the ground of the dignity of the human and of every rational nature’. Once more, the key word is ‘legislation’.

Let us now turn to [G4]. Regardless of the elements that connect [G4a] with [G3], what [G4a] says is this:

[G4a]* The legislation itself that determines all value has a dignity, i.e., an unconditioned, incomparable value.

And this, of course, implies:

[G4a]** The legislation itself that determines all value has an unconditioned, incomparable value.

Sentence [G4b] is more complicated because it includes a pronoun (‘it’, sie) the reference of which is not clear: As a feminin pronoun, the German ‘sie’ can refer both to ‘legislation’ (Gesetzgebung) and to ‘dignity’ (Würde). Thus there are two possible readings, and they both make sense in a certain way. Yet we propose to refer ‘sie’ to ‘dignity’, for the following reason: The German “welchen” (‘for which’, 436,4) is masculine; its reference can only be ‘value’ (Wert). Thus the first part of [G4b] says: ‘The word respect alone yields a becoming expression for the estimation for the value (of the legislation)’. As one can see, Kant then continues this sentence by basically repeating his point: ‘…that a rational being must assign to it’. Since ‘respect’ refers to ‘value’, and ‘value’ is identified with ‘dignity’, ‘it’ (sie) is best understood as referring to ‘dignity’:

[G4b]* For the unconditioned, incomparable value of legislation the word respect alone yields a becoming expression for the estimation that a rational being must assign to the dignity of legislation.

In [G5] ‘legislation’ is identified as ‘autonomy’:

[G5]* Autonomy is the ground of the dignity of the human and of every rational nature.

Note that [G5] is connected with [G4] by a ‘thus’ (also). This is justified because ‘autonomy’ is nothing but legislation; it is self-legislation, to be sure, but this is evident from the context. Note also (and most importantly) that according to [G4] the ‘legislation itself, however, that determines all value, must precisely for this reason have a dignity, i.e., an unconditioned, incomparable value’. This is to say: because the law, or rather, because legislation determines the value of whatever has value, it must itself also have value; this is Kant’s main thought in [G3–G5]. So as it stands, [G5] sums up what Kant says in [G4]. Even if one translates the German also with ‘therefore’ indicating an argument proper, one has no difficulties in reconstructing it: 1. What determines all value has itself an unconditioned, incomparable value. 2. Legislation determines all value. Therefore, legislation has an unconditioned, incomparable value.

[G3] states that ‘nothing has a value except the value which moral legislation determines for it’. But what is ‘it’? Which things obtain value by strength of the moral law? Given the basic elements of his ethical theory, two things come to mind first: actions and dispositions (Gesinnungen). In some sense this is true, of course; but in [G5] Kant does identify what has dignity − it is the ‘human and every rational nature’. In [G5] Kant also identifies the ‘ground’ of this dignity − it is ‘autonomy’. Since ‘autonomy’ is ‘legislation’, we can now reconstruct [G3] as follows:

[G3]** Moral legislation determines the dignity of the human and of every rational nature.

At first sight, it might seem that [G5] says that only the ‘human and every rational nature’ has dignity but not autonomy because it says that autonomy is (merely) the ‘ground of the dignity’. But that would miss the crucial point: As Kant says in [G4], legislation ‘itself’ has dignity. Since to have a ‘human nature’ or, more generally speaking, a ‘rational nature’ is to possess a will that is autonomous (a noumenally-good will), it is this quality (autonomy) that endows dignity to the ‘human and every rational nature’.Footnote 22

Part of Sensen’s revisionist reading is the claim that both freedom and being an end in itself are not normative properties; rather, he says, they are descriptive.Footnote 23 For Sensen, a term is ‘normative’ if it is either a duty-term or a valuative term. But how should freedom as the property of a noumenally-good will, hence as a moral causality, not be understood as normative? And how could being an end in itself not be a normative fact, since such beings have absolute value (which is a valuative) fact? We can also see here that Sensen goes astray in claiming that defenders of the standard reading need to find passages “in which Kant does not merely use phrases like ‘has absolute inner value’, but in which he also specifies it as something prior to and independent of the moral law”.Footnote 24 But why so? The absolute inner value (dignity) of human beings stems from their autonomy, and autonomy certainly is not ‘prior to and independent of the moral law’.

According to the standard reading of the GMS, Kant understands autonomous beings as ends in themselves with dignity and value, and he claims that the existence of such beings is the ground of the CI. Hence, dignity and value are indispensable components in Kant’s GMS (and, as we’d argue, in his entire ethical theoryFootnote 25). The burden of proof is on those who propose the revisionist reading. But our analysis of the law-sentence – the basis for the revisionist reading – reveals that it does not only not contradict the ground-thesis; as a matter of fact, it corroborates it.