Mere Maxims

This is not the place to discuss Kant’s theory of practical universalization, but in order to determine the quantity of this judgement we need to remind ourselves of why, for Kant, this maxim cannot be universalized. For this is what the categories of quantity are about: they determine whether the maxim is valid for one, for some or for all rational beings (ibid.). Kant’s main idea is that the end of this maxim (borrowing money) could not be achieved if the maxim became a law, where ‘law’ for Kant implies strict universality. There would be no such a thing as borrowing money anymore if it became a universal law that everybody in need of money falsely promised to pay it back. There is nothing problematic about a world without promises as such. But every rational agent must agree that it is impossible to bring about the end to borrow money if this maxim were to become a universal law. So from a universal practical standpoint, the standpoint of morality, this maxim can neither be willed nor even be thought without contradiction.

Now a maxim is supposed to function as a principle in a practical syllogism (CPR 5: 27). For this principle to be justified it should hold for any rational being, i.e. all other rational agents should be able to agree with it. But if the maxim cannot even be thought without contradiction, it cannot function as a sound principle of a practical syllogism, i.e. the conclusion lacks practical necessity. Thus the maxim is what Kant calls a ‘mere maxim’ (CPR 5: 19, 26) which does not qualify as a practical law. Hence, in Kant’s modal terms, it is morally ‘impermissible’ and ‘contrary to duty’. So we can already see that, similar to the table of the categories of nature, there is a dependence relation between the modality and the quantity of a practical judgement.

We will come back to this point later. For now I want to confine myself to the category of quantity. We already know that the maxim cannot become a universal law because it cannot even be thought without a contradiction. If this is true of a maxim, the question does not even arise whether the maxim is a precept or not. For all maxims that cannot be willed or thought without contradiction cannot possibly become any kind of imperative whatsoever. Otherwise it would be possible to have two valid but contradictory imperatives. The agent would, for example, be obliged both to keep and to break his promise, which is obviously absurd. In other words, hypothetical imperatives or precepts must be in line with the demands of morality on pain of losing their normative force.Footnote ⁴ The maxim in this case is therefore neither a precept nor a law but a mere maxim, or, as Kant puts it, a volitional opinion (Willensmeinung) only. The agent is making an exception for herself to the demands of morality. Thus this mere maxim has merely private validity (CPR 5: 66–7).

Calling the non-universalizable maxim a ‘mere maxim’ is supposed to convey the idea that maxims are not necessarily immoral. There are maxims which can at the same time be precepts or laws. But to determine that something is a maxim is not yet to determine whether it is good or evil; it could still be both. In this sense the first category of quantity is still ‘morally undetermined’.

It has been suggested that the categories of quantity quantify over ‘cases’ in which the agent commits or omits the intended action. To determine the quantity of a maxim means to determine whether the agent wants to Φ in one case, in many cases, or in all cases. Do we only want to tell the truth in one particular case, in many cases or in all cases (Haas Reference Haas1997: 54–5; Zimmermann Reference Zimmermann2011: 220ff.)? However, the categories of freedom are supposed to be constitutive of the objects of practical cognition, i.e. good and evil. If the case reading of the categories of quantity were correct, they would fail to determine moral value. For in Kant’s ethics it is essential for determining the moral value of a maxim whether it can be willed by only one individual or by all rational beings. The case reading, by contrast, leads to absurdity; for both the maxim ‘I want to increase my wealth in all cases’ and the maxim ‘I don’t want to increase my wealth in all cases’ would need to be classified as laws. However, it belongs to the most fundamental principles of Kant’s ethics that only universalizable maxims qualify as laws. The first maxim is not a law. Thus the case reading is not a satisfactory account of Kant’s view.

Since the case reading fails to determine moral value, it is not surprising that it leads to the view that the categories of quantity do not relate to the concepts of good and evil but to the morally neutral concepts of good and bad (Benton Reference Benton1977: 81; Beck Reference Beck1960: 136–7; Zimmermann Reference Zimmermann2011: 49–50, 72). But if, as I am suggesting here, the categories of quantity quantify not over cases but over agents, then there need not be any ambiguity in reference; all categories of quantity are rules for the employment of the concepts of good and evil. Moreover, those who think that there is in fact an ambiguity in reference have to solve the following puzzle: the categories of freedom are supposed to be constitutive of practical cognition. For Kant, there is no genuine practical cognition of the agreeable or disagreeable simply as such. The only practical cognition is moral cognition. How then could the categories refer to the concepts of the agreeable and disagreeable?

Turning now to quality, in general logic the ‘quality’ of a judgement signifies whether the subject term S is thought under the sphere of a predicate P (affirmative judgement), outside of this sphere of P (negative judgement) or whether the subject is posited in the sphere of another concept non-P that lies outside the sphere of P (infinite judgement). Derivatively, to determine the quality of a practical judgement is to determine whether the subject S (the agent) is to be thought under the sphere of a certain action Φ (commission), outside of this same sphere of Φ (omission) or under the sphere of another action non-Φ (exception).

In our example the maxim is a rule of commission; it expresses what kind of action is to be committed if the agent is short of money. If the maxim were ‘when I am in need of money, I will refrain from borrowing money when I know that I cannot pay it back’, then the maxim would be a rule of omission. This change regarding the quality, however, also changes the moral value, the modality, of the maxim. The change of moral value can of course only be determined if the maxim makes reference to the problematic conception of a lawful maxim. In other words, we need to ask ourselves whether the maxim has a lawful form, i.e. whether it has practical necessity. I want to claim that it is precisely for this reason that Kant takes the categories of quality and the categories of relation to be ‘morally undetermined’. This is important for it will be crucial in order to disambiguate the ambiguity in transition. Whether a maxim is a duty or contrary to duty is a question of its modality. Those categories are morally determined in the sense that their moral value is fixed. This does not hold for the categories of quality.

The categories of commission, omission or exception do not have a determinate moral value. In determining a maxim as contrary to duty it can only be evil. But a maxim that is determined to be a commission can still be both – good or evil. As we will see, this does not mean that these categories are irrelevant in determining the moral value. They do play a role, but morally undetermined categories leave it open whether the maxim is good or evil.

Now, as mentioned in the first section, this indeterminacy claim has led some interpreters to maintain that the heading of the table of the categories of freedom is too narrow (Beck Reference Beck1960: 136–7; Zimmermann Reference Zimmermann2011: 49–50, 72). According to them, at least the categories of quality and relation do not refer to the concepts of good and evil. However, as we have just seen, the moral value of our practical judgements can crucially depend on their quality. Whether I borrow money even though I know that I won’t be able to pay it back or I don’t borrow money since I know that I won’t be able to pay it back is a decisive qualitative and moral difference. Thus the move from ‘morally undetermined’ to ‘morally neutral’ turns out to be unwarranted (e.g. Zimmermann Reference Zimmermann2011). Categories are morally undetermined only if the category by itself does not fix the moral value of the maxim. To put it more concretely, a commission, omission or exception is not by itself morally good or evil. The moral value of the categories of quantity, by contrast, is fixed. It is impossible for a mere maxim to be morally good, as it is impossible for a law to be evil. And as I will argue shortly, even a valid precept must be morally permissible if it is impossible for it to be evil.

This reading, according to which the whole group of categories of quality is morally undetermined, differs fundamentally from those interpretations which resolve the transition ambiguity by trying to show that the transition from the morally undetermined to the morally determined categories takes place from the second to the third category within each group. According to the latter reading, rules of exception are morally determined and the rules of commission and omission are morally undetermined. According to Bader, for instance, ‘the former depend on the faculty of desire, telling us to seek out that which is agreeable and promotes our well-being. The latter depend on the faculty of aversion and command us to avoid what is disagreeable and leads to ill-being. Good and evil, on the contrary, pertain to practical rules of exceptions’ (Bader Reference Bader2009: 11). If I understand the argument correctly, only moral rules can be rules of exception because only they can trump or ‘dismiss’ any hypothetical imperative. ‘The imperative not to lie amounts to the practical rule of exception that says that one should not lie even though there is a hypothetical imperative telling one to lie if it is in one’s interest’ (Bader Reference Bader2009: 12). However, this view gives rise to the following question: why can’t the rule be reformulated as a rule of omission which obliges us not to lie? To call the moral rule a rule of exception seems literally absurd, for it requires the assumption that we are in principle obliged to lie, and must make an exception to this rule. However, there is no hypothetical imperative which obliges us to lie in the service of our own interests. Hypothetical imperatives are only imperatives at all if their ends do not conflict with what we are morally obliged to do. So the rules of exception do not tell us ‘to do x even though the hypothetical imperative told us not to do x’ (ibid.). Instead, they express the general volition that under certain circumstances we want or ought to Φ, even though under all other circumstances we do not want or ought to Φ. A rule of exception can be morally good or evil; it can be a duty or contrary to duty. A rule of exception is contrary to duty if, for example, I only want to tell the truth when someone threatens to kill me. A rule of exception is a duty if, for example, I only want to pursue my own happiness when it is compatible with the pursuit of other people’s happiness.

I believe that Bader’s reading is an expression of what I called in the introduction a mixed or hybrid conception of practical reason. According to this reading, there are different rules which pertain to different faculties. Each faculty issues its own valid imperatives, which later come into conflict. However, if we think of the formation of our volitions as a single, unified act, this conflict between different kinds of imperatives cannot arise. The hypothetical imperative is not hypothetical because the obligation is only an obligation under the condition that it does not conflict with the moral law. Instead, hypothetical imperatives are conditional because they oblige us only if we do in fact have a certain end. That this end is not in conflict with morality is presupposed for the hypothetical imperative to obtain any normative force in the first place. There is no imperative, not even a hypothetical imperative, which obliges us to lie or, to use one of Kant’s famous examples, not to return deposits.

So none of the three categories of quality is morally determined. This does not mean, however, that these categories do not refer to the concepts of good and evil. All our moral judgements are based on these fundamental categories. It is through them that we bring the ‘manifold of our desires to the unity of our consciousness’ and we thereby subject it to the moral law (CPR 5: 65). In other words, without those categories it would be impossible to know what we ought to do, or to know whether a maxim is good or evil. Every maxim as a principle of action is either a commission, omission or rule of exception. As I have argued above, changing the maxim in our example from a commission into an omission changes its moral value: ‘When I am in need of money, I will refrain from borrowing money when I know that I cannot pay it back’ can function as a practical law and is morally good while ‘When I am in need of money, I will borrow money and promise to pay it back even though I know that this will never happen’ cannot function as a practical law and is morally evil.

To say that some of the categories do not have any bearing on the moral value of the judgement and are morally neutral is like saying that some of the categories of nature do not have any bearing on the truth (or objective validity) of theoretical judgement. In both cases, judgement is one single unified act. If we did not determine the quality of a maxim we would not be able to judge morally at all. And if we did not judge the relation between the underlying subject and the condition of the subject there also would not be any moral judgement.

Let us now return to our example in order to acquire a better understanding of the category of relation. The maxim expresses (as all maxims do) an agent’s end from the perspective of the first person: ‘When I am in need of money, I will …’. This articulates a relation between the underlying subject I and the intended end which the subject has freely adopted. Only because we have the capacity to freely adopt and act according to universalizable maxims are we morally accountable beings. This capacity is part of the essence of our moral agency, which Kant also calls ‘personhood’ (RGV 6: 26). All our maxims are necessarily related to this capacity. This explains why the underlying relatum in the first category under the heading of ‘relation’ is ‘personhood’ (cf. Bader Reference Bader2009: 13–14).

The remaining two categories within the class of relations are concerned with the effects the agent’s action has on the condition of herself or on the condition of other persons. Either an agent’s maxim expresses a one-way relation to the state of a person, or the relation is reciprocal and the maxim expresses a two-way relation between the state of the agent and the state of another person. In our example the maxim expresses a reciprocal relation. The end is not only directed at the state of the agent as a dishonest debtor but also at the state of other people as possibly betrayed creditors.

As the categories of quantity are morally undetermined, so are the categories of relation. In other words, in determining the relation of a maxim we do not thereby fix the moral quality of that maxim. All maxims, whether good or evil, necessarily express a relation of the agent to the intended end. But whether the agent’s end is a duty or contrary to duty is a question of the modality of her maxim; it is not determined by the categories of relation. The fact that Kant, in employing the term ‘personality’, employs a moral term does not mean that the category is morally determined. For moral determination implies that through the moral law the moral value of the category is fixed as either morally good or evil. The same is true for the other two categories of relation. But even though the categories of relation are morally undetermined, they are not beyond morality or morally neutral. For all three categories do not express a relation to inanimate objects which, according to Kant, do not possess any moral rights. Instead they are limited to those entities which have moral rights, i.e. persons. The relation is between the subject of the maxim as the one who is morally accountable for adopting or rejecting a maxim and those persons (including herself) who are the object of her ends (duties towards oneself, duties towards others). Only under the condition that the categories are an expression of the relation between persons, as morally accountable beings, is Kant entitled to say that they are developed with respect to the concepts of good and evil. In other words, practical cognition as the cognition of the good necessarily involves cognition of a particular kind of relation: the relation between the subject of moral judgement (the moral agent) and another equally morally accountable and respect-worthy object.

As for the modality of the maxim, since it cannot be willed or even thought without contradiction as a universal law, it is practically impossible. That is, an agent cannot will to give a false promise when in need of money and at the same time will that her maxim (the practice) is a universal law. In instrumental reasoning, the irrationality arises when we take means that we know cannot bring about the desired (willed, wanted) end. Only if we act fully rationally do we always take the necessary means to our ends. But as finite rational beings we can adopt an end, thereby knowing that it commits us to certain means, and still fail to take those means. This is what instrumental irrationality consists in; it is a deficient act of willing because reason does not have ‘decisive influence on our action’ (GMM 4: 417).

The necessity in instrumental reasoning is attached to adopting the means; it is dependent on the agent’s willing the end. The obligation is therefore, as always in instrumental reasoning, conditional or, as Kant puts it, hypothetical. However, for Kant the concept of duty implies unconditional obligation (GMM 4: 389; CPR 5: 80). Kant’s main idea is that moral obligation cannot be the kind of obligation that depends on the agent’s discretion (about whether or not she is obliged to Φ). The concept of duty, as Kant understands it, therefore implies unconditional practical necessity.

Now returning to our example, it should be clear that, since this maxim involves a contradiction and therefore cannot be agreed to by all other rational agents, it must be contrary to duty. To be sure, the maxim does not involve a lack of instrumental rationality. The agent does not adopt the wrong means in order to get her money. In following this maxim she might even get her money. The more morally intact the world is, the greater the chances are that she will be able to achieve her end. But thinking about the maxim from a general standpoint, from a moral point of view, makes the agent aware that it could not become a universal law. The maxim therefore lacks the unconditional practical necessity that is needed for a maxim to be a duty.

We have just explained the concept of ‘law’ in terms of ‘necessity’. In the categories of nature the category of (strict) universality is also coextensive with ‘necessity’. Everything in nature that holds with strict necessity is also a universal law. And everything in nature that is a strict universal law is also necessary. For instance, the strict universality of the judgement ‘every event has a cause’ can only be justified if it could not have been otherwise, i.e. if it is necessarily true of events in nature. As we will see below, the difference between precepts which are only valid for some rational beings and laws which are universally valid in the table of the categories of freedom has to be spelled out from a moral point of view as follows: the practical laws are practically necessary while the precepts are only practically possible.

In the table of the categories of nature, everything that is impossible must also be unreal. If we transfer this implication holding between the impossible and the unreal to the table of the categories of freedom, we obtain the implication that everything contrary to duty must also be impermissible. Prima facie this seems to be correct. If it is contrary to duty to give a false promise, it must also be morally impermissible. However, on one of the few occasions where Kant comments on the table of the categories, he explicitly says that the three pairs of modal categories are supposed to line up with the three kinds of imperatives – problematic, assertoric, apodictic imperatives. Since the first kind of imperative is merely ‘technical’ it seems, as many have argued, that Kant wants to reserve the first pair of the modal categories (permissible–impermissible) entirely for technical deliberation. If this were correct, then the prima facie plausible relation of implication between ‘duty’ and ‘permissible’ no longer obtains. This would be grist to the mill for those interpreters who think that the transition from the morally undetermined to the morally determined is to be located in the second or the third pair of categories in the category of modality (Schütz Reference Schütz1788: 542–3; Bader Reference Bader2009: 2; Beck Reference Beck1960: 152). However, in turning to the second example I will show why this is not the correct response to the transition ambiguity.

Precepts

This maxim is not a mere maxim but, as Kant puts it, an ‘important practical precept’ (CPR 5: 20). In contrast to mere maxims, this maxim can be willed and thought as a universal law without any contradiction. However, this is not sufficient to establish it as a law. Being willable and thinkable as a universal law are only necessary conditions for a maxim to become a universal law. In order for it to be a universal law, a maxim must also be unconditionally valid. The practical validity of a precept, by contrast, depends on two conditions. First, the person must in fact will the end. Second, the means need to be necessary in order to achieve the intended end. In the case of this second maxim, it is likely that the first condition is met. As human beings we necessarily strive for our own happiness and we therefore do not wish to starve. However, whether the second condition is also met depends essentially on the special circumstances the agent lives in.

Kant mentions three possible reasons why the particular precept ‘you ought to be thrifty and industrious in your youth’ does not have general validity, even if all agents desire the end of happiness. First, we ‘look forward to other resources than those of [our] own acquisition’. Second, we may ‘not hope to become very old’. Third, in case of ‘future dire need [we] will be able to make shift with little’ (ibid.). So this maxim does not become a universal law by being adopted by agents who live in the appropriate circumstances for the means to be necessary. Nor would the precept turn into a law if everyone who adopted it happened to agree with all others in the pursuits of their ends. Only if agreement and harmony are based on their cognition of the universalizability of the maxim and are not a mere accident do their maxims qualify to be laws. For those individuals who do have this end and who are in the appropriate circumstances the maxim may indeed be a valid practical principle. Judged by its quantity the maxim is, therefore, not a mere maxim. It is not merely a ‘volitional opinion’ (Willensmeinung) with private validity. Nor is it a universal law which is valid for all rational beings, but a precept which for some agents is obligatory or has conditional practical necessary. ‘All imperatives express obligation’ (GMM 4: 413). But the obligation or practical necessity of precepts cannot be unconditional, for otherwise they would be obligatory for all rational beings. So when Kant says precepts ‘lack practical necessity’, he only means to indicate that they lack the unconditional practical necessity of a law. And again, all imperatives are ‘objective’ or express ‘objective necessity’. But hypothetical imperatives or precepts are only objectively necessary under subjective conditions. Once these conditions are met, it is no longer up to the agent whether or not to take the means.

The quality and the relation of this precept are straightforward. The precept is a rule that requires the agent who is in the appropriate circumstances to do something, namely to be thrifty and industrious. With respect to its quality the precept is therefore a commission. As to its relation, the precept expresses, as all maxims do, a relation between the underlying subject I and the end which the subject intends to adopt. By acting on a certain end the subject changes her own state. The maxim therefore also places the agent in a relation to her own state. Since this particular maxim is only directed towards the state of the agent, it only expresses a one-way relation. It does not express a reciprocal two-way relation between the state of the agent and the state of another person.

The main reason for analysing this maxim was that we wanted to get a better grip on the first pair of the categories of modality. The maxim is a precept; as such it is an imperative. All imperatives imply, as Kant explicitly says, necessity (CPR 5: 20). This does not make the precept a duty, for ‘duty’ implies unconditional necessity (GMM 4: 389; CPR 5: 80). In contrast to hypothetical imperatives, duties do not depend for their obligatory force on contingent facts about the agent. The necessity of a precept is, in contrast to the necessity of a law, as Kant puts it, subjectively dependent. A precept does not imply unconditional practical necessity; thus it cannot be a duty. Only laws are ‘independent of accidentally attached conditions’. Only laws are, therefore, unconditionally binding and imply ‘objective necessity’ (CPR 5: 21–2; my emphasis).Footnote ⁵ So the reason why Kant claims that unconditional necessity only refers to moral obligation is that only in moral obligation is the prescription independent of some given desire. In the case of mere precepts, the means to the realization of the willed end also are not at the agent’s discretion. However, the willing of the particular end is entirely at her discretion. Precepts or hypothetical imperatives are therefore not unconditionally practically necessary. In short: precepts are not duties. And as indicated above, the unity of practical reason requires that precepts can never be contrary to duty either, or otherwise we would face the absurd consequence that we are both obliged to Φ and not obliged to Φ.

Now if our exemplary maxim is neither a duty nor contrary to duty, it cannot fall under the second or third pair of the modal categories, but instead must fall under the first pair; it must either be permissible or impermissible. However, to classify the maxim as permissible or impermissible gives rise to the following problem. Calling this maxim merely permissible seems prima facie incorrect, for even though we are not categorically obliged to follow this maxim, it is (under the right conditions) still imperative to follow it. It seems, therefore, too weak to classify the maxim or its opposite as merely permissible (erlaubt). I think that the solution to this problem will help us gain a better understanding of what it means to say that the table of the categories is developed ‘with respect to good and evil’. I want to suggest that ‘permissible’ and ‘impermissible’, as they occur in the table of the categories of freedom, are notions that are used with an eye to moral obligation. If something is judged as permissible or impermissible it must be morally permissible or impermissible. This means that the maxim is qualified as merely permissible from a moral point of view. So it is not too weak to classify a precept as merely permissible, since the obligation is only valid under subjective conditions.

There is nothing practically impossible in willing to be thrifty and industrious in one’s youth or not willing to be thrifty and industrious in one’s youth; both ends as such can be willed and thought without contradiction. Our volition is only morally impossible if the volition is formed in such a way that it cannot be willed as a universal law. The permissible or impermissible maxims are morally permissible and not contrary to duty. They are, however, not duties because the opposite of these maxims can also be willed without a contradiction. So even when a maxim is judged permissible or impermissible, the moral law and the concepts of good and evil are at work, even though they do not play a role in determining the content of the precept itself. Cognition of the content of instrumental principles involves theoretical cognition (mostly a posteriori). But in order for the maxim to become a precept, it has to prove itself to be practically possible, which is to say that it is morally permissible.

This interpretation is also in line with Kant’s explicit claim in the preface to the second Critique, mentioned above, where he says that there is a correspondence between the three modal categories and three kinds of grounds of determination: problematic, assertoric and apodictic grounds. For a judgement or a concept to be problematic it is necessary that it is not contradictory. However, if the practical judgement already involves a contradiction in willing or thinking, it is not even a problematic judgement. So Kant’s correspondence claim in the preface confirms the main point that even the first pair of modal categories is developed with an eye to the moral law. Only if the presupposed end in a precept can be willed or thought without contradiction does it have normative force. In this sense even the first pair of the modal categories is, to use Kant’s phrase, ‘morally determined’ (Zimmermann Reference Zimmermann2011; contrast Bobzien Reference Bobzien1988; Bader Reference Bader2009). This does not mean that the moral law determines the end which a precept prescribes, but that the end which the precept presupposes must at least be morally permissible. This does not mean that a precept itself is morally determined, for whether something is a valid practical precept cannot be determined by pure practical reason without considering its content. Theoretical cognition is required in order to find out what the necessary and sufficient means for the realization of our ends are.

So even if our practical judgement is neither a duty nor contrary to duty, it is not thereby morally indifferent or neutral. For a maxim to be permissible or impermissible, it is required that we have already determined the maxim as morally permissible. This determination is an act of pure practical cognition and is only possible through the moral law via the concepts of good and evil. In other words, if a maxim turns out to be morally permissible and the end of this maxim is what the agent desires, then the action qualifies as ‘good’; it has been carried out only under the condition of its universalizability. So there is good reason to doubt that the merely permissible or impermissible maxims, as it has been suggested (Schütz Reference Schütz1788: 542–3), are morally neutral. For as Kant says in Religion, ‘[A] morally indifferent action (adiaphoron morale) would be one resulting merely from natural laws, and hence standing in no relation whatsoever to the moral law, which is the law of freedom; for such action is not a factum and regarding it neither command, nor prohibition, nor permission (legal privilege) occurs or is necessary’ (RGV 6: 22).

In short, all three modal categories are morally determined; it is impossible to employ any of these categories without thereby fixing the moral value. This distinguishes them, as we will see, from almost all other categories. Thus the question whether they refer to the concepts of good and evil does not even arise. None of the modal categories is morally neutral. Each of them is necessary in order to articulate the full scope of practical cognition. Practical cognition, for Kant, is cognition of good and evil. Any attempt to derive the categories of freedom as constitutive of non-moral practical judgements (Zimmermann Reference Zimmermann2011: 267) faces the problem that moral cognition, for Kant, is the only kind of practical cognition possible. There is simply no practical cognition to be had with regard to the agreeable and disagreeable or to any morally neutral kind of agency. The categories would therefore cease to be constitutive of practical cognition in these cases; they would no longer be categories. This resolves not only the ambiguity in reference but also the ambiguity in transition. None of these categories is morally neutral; they all refer to the concepts of good and evil. There is no transition from the morally undetermined to the morally determined categories within the group of the modal categories. However, that the transition to the morally determined categories occurs first and only with the modal categories faces a pressing problem that emerges when we turn to our final example.

Laws

This maxim is the opposite of our first example. In both cases, the material end is the acquisition of money. But in contrast to the first example, the action with which I want to bring about my material end here is an honest promise. Thus in this maxim the material end is only pursued under the condition of its universalizability, i.e. that every rational agent could agree to it. Thus the formal object of this maxim is the moral good. The maxim does not depend on some private interest, but on its practical universality. Hence it is not a mere maxim, but a law. However, as I have argued above, even precepts need to be morally permissible, i.e. their end must be in principle compatible with everyone else’s ends. In the case of a precept the opposite of the intended action does not necessarily involve a contradiction, so it is not merely morally permissible but obligatory. Whether I ought to be ‘thrifty and industrious’ in order not to starve in old age depends partly on some contingent facts about my life. But in the case of a truthful promise, Kant argues, its opposite involves a contradiction which can neither be willed nor even be thought. Thus a truthful promise is morally obligatory and not merely permissible. Hence keeping one’s promise is not merely obligatory for some agents who happen to be disposed in the appropriate way (precept), but for all rational agents, i.e. the maxim is not merely a subjective principle but at the same time a law. It is important to note that in employing the concept of a maxim we are not thereby committed to the view that it is a mere maxim. A principle can be subjective in the sense that it expresses an intention from the first personal perspective, but at the same time a law, i.e. a valid principle for all rational agents. This is what the categorical imperative demands from us. Again, this is not the place to assess the merits of Kant’s so-called ‘testing procedure’. My goal is only to explain how the categories are employed in our moral judgements.

Now Kant explicitly says that the modal categories first ‘introduce the transition to the morally determined categories’ (CPR 5: 67). This is what I call ‘Kant’s transition claim’. Interpreters who attempt to argue that the transition to moral determinacy takes place in the third category of each group cannot account for this claim (Bader Reference Bader2009). Others have attempted to interpret the category of law in a broader, non-moral sense (Bobzien Reference Bobzien1988: 211, following Benton Reference Benton1980: 187). However, it is difficult to make this compatible with Kant’s claim that only categorical imperatives can be rightly called practical laws (GMM 4: 416). So ‘counsels of prudence’ and ‘rules of skill’ are not possible candidates. Laws of nature also do not qualify for the employment of the categories of freedom. The reading I am developing here has the disadvantage that it cannot fully explain Kant’s transition claim, but it has the advantage that it does not have to hold that practical laws are morally undetermined. On my reading, not only are all modal categories morally determined, as Kant says, but also the third category of quantity is morally determined (cf. Beck Reference Beck1960: 146; Bader Reference Bader2009: 9–10). In other words, if a maxim is determined as a law it is no longer an open question whether it is good or evil. This reading is more compatible with Kant’s explicit transition claim if we keep in mind that there is a dependence relation between the modality and the quantity of a judgement. As we have seen, the concept of a law is coextensive with Kant’s concept of unconditional practical necessity. Everything that holds with strict necessity is also a law (strict universality). And everything that is a law (strict universality) must also hold with strict necessity.

Since this maxim is identical with respect to its quality and relation to the first maxim, we don’t need to add anything here to what was said in relation to it. It might suffice as a reminder that both quality and relation are essential to our moral judgements. In other words, if we did not determine the quality of a maxim, we would not be able to make moral judgements at all. And if we did not judge the relation between the underlying subject and the condition of the subject there also would not be any moral judgement. But even though the moral value depends upon those categories they do not fix the moral value. This maxim and its opposite are both commissions but have diverging moral values. The same cannot be said about the modality of a practical judgement. It is impossible for a maxim to be a duty and still to be morally evil. In this section I want to make clear how the modality of practical judgement is determined.

We cannot infer in a non-trivial way from the fact that a maxim can be willed without contradiction to its being also a duty. We can think of many maxims that can be willed without contradiction without thinking that they are also duties. It is not an accident that Kant, both in the Groundwork and in the second Critique, chooses to evaluate maxims that are contrary to duty whenever he examines the moral value of a maxim. So what one needs to know, as I have already pointed out, in order to know that a certain maxim is also a duty is that its opposite involves a practical contradiction. Duty as unconditional obligation implies that it is not up to the agent’s discretion to act on a maxim or its opposite. In other words, if both the maxim and its opposite do not involve a practical contradiction, then they are both merely morally permissible, and we would not be unconditionally obliged to follow only one of them. If someone acts contrary to duty, she does not will her maxim to become a universal law. Instead, she desires, as Kant explicitly says, that its opposite ‘remains a universal law’ and that her action may count merely as an ‘exception’ to it (GMM 4: 424).

Since we already know from our discussion of the first example that the opposite of this maxim cannot be thought without contradiction, we thereby also know that we have an obligation to adopt it. Or more concretely, if I know that I ought not to commit suicide, I also know that I ought to keep myself alive. If I know that I ought not to give false promises I thereby also know that I either ought not to make any promises or I ought to keep the promises I make. If I know that I ought not to waste my talents I thereby also know that I ought to cultivate them. And finally, if I know that I ought not to be indifferent to people in need, I thereby also know that I ought to help them.

So by now we have established that the maxim in question is a duty. This will help us to get a better grip on the transition ambiguity. If one maintains, as some interpreters do, that only the third category of each group of the categories is morally determined, one is thereby also committed to the claim that the second pair of the modal categories ‘duty–contrary to duty’ is morally undetermined (Bader Reference Bader2009: 16, following Mellin Reference Mellin1802: 534; Benton Reference Benton1980: 186). This would mean that our maxim in the present section as well as its opposite discussed above would not be determined through the second pair. This reading is supported by the fact that the contradictory opposite of ‘duty’ is not ‘contrary to duty’ (as Kant suggests in the table of the categories) but ‘not-duty’ (Beck Reference Beck1960: 150, following Schütz Reference Schütz1788: 541–2). So the proposal is that, contrary to Kant’s common usage, the term ‘contrary to duty’ as it occurs in the table has to be understood as ‘not-duty’, which is supposed to include all actions that are not duties and not only those that violate duties.

However, if the morally neutral reading of the pair ‘duty–contrary to duty’ were correct it would be impossible to modally determine the first maxim altogether. For the notion ‘contrary to duty’ is interpreted by those accounts as morally undetermined. And while the third pair of the modal categories (‘imperfect duty–perfect duty’) can account for the morally good maxim, there would be nothing in the present section corresponding to the modality of its morally evil opposite above.

Now one might think that a broad understanding of ‘duty–contrary to duty’ can at least also accommodate actions that are moral duties and actions that are morally contrary to duty. The domain of actions that are not-duties includes those actions that are simply contrary to duty, morally permissible and morally neutral. However, this reading is not very plausible, because not only does Kant explicitly reserve the concept of duty for unconditional obligation (unconditional practical necessity: GMM 4: 389; CPR 5: 80), which is fully identical to moral obligation, but the table of the categories is also developed with respect to good and evil. Thus there is no room for a non-moral notion of duty within the table of the categories of freedom. In short, there is no good reason to believe that the second pair of modal categories is morally undetermined and not (or not exclusively) developed with reference to good and evil in a moral sense. Thus there is no reason to believe that a transition from the morally undetermined to the morally determined categories takes place from the second to the third pair of the modal categories.

So far the maxim has been determined as a duty. This does not fully determine it with respect to its modality. Already in the Groundwork Kant distinguishes between two kinds of duties: perfect duties and imperfect ones. In the table of the categories of freedom he takes up this distinction in the third pair of modal categories. If a maxim merely cannot be willed without contradiction, it is a violation of an imperfect duty. The maxim itself is perfectly conceivable (or thinkable) as a universal law, but it does not cohere with other things a finite rational being necessarily wants (GMM 4: 424). These maxims therefore do not rule out a particular action as irrational or unthinkable. Their opposites therefore leave room for the agent to judge and choose for herself which action and how many actions are sufficient for acting in accordance with this maxim. It would be contrary to duty not to help others in need, but what particular action the agent ought to perform cannot be determined by pure reason and needs to be determined by the agent’s own power of judgement. The maxim by itself does not determine a particular action as contrary to duty but only determines the general principle on which the agent acts (MM 6: 446). If, on the other hand, a maxim can neither be willed nor be thought without contradiction, it is a violation of a perfect duty (ibid.). Perfect duties determine particular actions as practically necessary. This is the case in our example; it involves a false promise and cannot even be thought without contradiction. It is not up to the agent’s own choice to subsume a possible action under the general principle, since the maxim by itself already implies the particular action.