2. ‘Maxims’?

In order to be able to evaluate the question of whether, as Kant sees things, we always act on maxims, we need to know what Kant means by maxims. And for this discussion not to be too abstract or too general, it is useful to have a set of concrete examples of maxims that we can refer back to and test our interpretation against. Here therefore is a set of examples of maxims. I have sorted these into five examples of maxims Kant says are unfit to be universal laws and five examples of maxims that Kant says are fit to be universal laws:

These examples might seem like a rather motley crew.Footnote ⁵ I think, however, that a look at Kant’s various statements about what maxims are makes it possible to explain both what these examples of maxims have in common and what Kant means by maxims in general. To anticipate, I argue that Kant’s idea of a maxim has the following key features: a maxim is (i) a basic principle in the form of (ii) a self-imposed rule that is (iii) intended for repeat-use, and which (iv) either states an end or says how to pursue certain types of ends.

3. Problems with the Descriptive Reading

The descriptive reading, to repeat, holds that ‘everyone always acts on maxims’. If this were so, the following would hold. Everyone always acts on the basis of self-imposed (i.e. reflectively endorsed) fundamental principles that state certain basic ends and that are intended for repeat-use. There are, I think, two main problems with this view. First, it is rather implausible to think that people always act in such ways, which makes it uncharitable to attribute this view to Kant. Second, Kant makes various claims that suggest that he himself does not hold this view in the first place.

It is quite plausible to interpret human beings as differing in how reflective and principled they are. Some people are very conscientious and reflective. Others are the opposite. Some people dislike the thought of having too many rules and principles, like Lord Henry in Oscar Wilde’s The Picture of Dorian Gray, in whose mouth Wilde puts the line ‘I like people better than principles, and I like people with no principles better than anything in the world’ (Reference Wilde2006: 12). Even among those who are very conscientious and reflective, it is hard to swallow the idea that there are people who consistently and robustly govern and direct their own actions on the basis of the sort of self-imposed rules that Kant described.

Let Kant himself serve as our example of somebody who tries to stick very firmly to a regime of self-imposed principles and rules. Kant was convinced that a daily walk was a good way of maintaining his health. He was known in Königsberg for the regularity with which he adhered to his routine of taking a walk at a certain hour every day. According to one famous anecdote, however, Kant broke this routine on one occasion when he was so fascinated by one of Rousseau’s books that he could not put it down. He subsequently kept reading rather than stopping for his daily walk. It seems clear that when Kant was reading with great interest, thereby offending against his self-imposed rule of taking his strictly scheduled walk, he (a) was performing an action (namely, reading), even though (b) he did not seem to be acting on the basis of a firm self-imposed rule, but rather (c) against a strict self-imposed rule of his. Whether or not this anecdote is true, it does seem to illustrate that our common conceptions of agency very clearly allow for the possibility of actions (such as Kant’s reading of Rousseau’s book) that are not regulated or directed by self-imposed principles or rules intended for repeat-use.

Consider next the case of Ethan Couch, the Texas teenager whose claim to fame (or infamy) is that his lawyer successfully was able to lighten his sentence by means of the ‘Affluenza defence’. On a fateful day, Couch had borrowed his father’s car without permission. He had been caught on surveillance camera stealing beer in a convenience store. And he had gotten himself into a car accident through reckless driving, all while under the influence of Valium. Couch’s actions resulted in the deaths of four people, and serious injuries to an additional two people. His lawyer managed to lighten his sentence to rehab and ten years of probation, by arguing that Couch was ‘the product of wealthy, privileged parents, who never set limits for the boy’.Footnote ⁸ Whatever the facts of the case may have been, it is quite plausible to think that Couch might be the sort of person who spends very little time reflecting on what principles and rules to set for himself and equally little time trying to conform his actions to such self-imposed principles and rules. Does this mean that Couch is not an agent? Does it mean that he does not perform any actions? Intuitively, we want to say no. And the criticism with which the ‘Affluenza defence’ was met suggests that the general public would not accept any such conclusion either.

Cases such as Kant’s and Ethan Couch’s suggest that it is not the case that people always act on maxims. They suggest that there is variation in how prone people are to act on maxims, and also that even those who are very prone to do so (such as Kant) do not always act on maxims.Footnote ⁹ This means that, if Kant held the view that we always act on maxims, his view would offend very strongly against our common-sense ideas about human action. It is therefore uncharitable to interpret Kant in the way that defenders of the descriptive view do.

This problem – that it seems clear that we do not always act on maxims – is not something that has been lost on defenders of the descriptive reading. Various attempts to explain how it might be that we always act on maxims without realizing this have, accordingly, been made. Here are three examples. Onora O’Neill has suggested that, whereas we always act on maxims, we do not always know what our maxims are, because the notion of a maxim is a mere ‘idea of reason’ (O’Neill Reference O’Neill1989).Footnote ¹⁰ Second, Dieter Schönecker and Allen Wood hold that while we are often aware of our maxims, it is nevertheless the case that we sometimes follow maxims in a non-conscious way, just as we tend to follow rules of grammar in non-conscious ways (Schönecker and Wood Reference Schönecker and Wood2008: 102–3). Third, Timmermann claims, as noted in the Introduction, that whereas there are certain kinds of maxims that we do not always act on, it is nevertheless the case that there is one kind of maxim, ‘first-order maxims’, that we always act on (Timmermann Reference Timmermann2000).

I think that these kinds of responses are problematic. Take, for instance, O’Neill’s claim. It is hard to square this idea – according to which we often do not know what our maxims are – with the idea that we are under a categorical imperative of reason to act on the basis of maxims we can at the same time will as universal laws. In order for the categorical imperative to be an imperative we can intentionally follow, it must be that we are able to know and decide what maxims we act on (Raedler Reference Raedler2015). But this is not the only problem with O’Neill’s suggestion. We have also seen that Kant thinks of maxims as principles or rules that we impose upon ourselves. This clearly implies that we are conscious and aware of what our maxims are, if and when we adopt maxims. To be sure, circumstances and situational pressures can lead us to behave in accordance with social norms and various other patterns without our realizing it. But to unknowingly behave in ways that conform to certain patterns is very different from deciding to adopt certain basic principles or rules (namely, maxims) for ourselves.

This also suggests that Schönecker and Wood’s suggestion is problematic. It might be true that we sometimes follow rules we have reflectively endorsed and adopted for ourselves without putting too much thought into it. But Kant seems to think of the paradigmatic case of acting on maxims as being a case in which we deliberate about how to act and do so with the help of some principle or rule that we endorse. He seems to hold the view that the more we are inclined to act in certain ways without reflecting on what rules or principles we are following, the less we can be said to be acting on maxims of our own choosing.Footnote ¹¹ As I noted above, following maxims of action seems to be quite different from unreflectively following rules of grammar. Whereas the former involves reflection and personal choice, the latter is done without reflection and without much or any choice.

Consider next Timmermann’s suggestion that we always act on certain ‘first-order maxims’. This seems to be an idea that is best understood as the thesis that we always act on the basis of certain intentions. But as I noted above, to act on an intention is not sufficient to qualify as acting on a maxim we have adopted for ourselves. We can form intentions about how to act on specific occasions, without making any kinds of commitments about how to act in other cases. Again, we might decide to go to St Petersburg on a holiday just this one time. Or to use a nastier example, we might decide to do something immoral, such as cheating on some test, ‘just this one time’ – without making any commitments about how to act on other occasions. And so even if it is plausible to say that we only act if we act on certain intentions (which we can here assume for the sake of the argument), that does not mean that we always act on maxims. Maxims aren’t identical to intentions. The conclusion that we always act on maxims does not follow from the premise that we always act on intentions.Footnote ¹²

More generally, we face the following choice. Either we follow Kant in understanding maxims in the way specified above, in which case it becomes highly implausible to say that we always act on maxims. Or we insist that we always act on maxims, in which case we need to understand the idea of maxims in some way that does not conform to what Kant himself takes maxims to be. It seems that writers who insist that we always act on maxims go down the second path. This way of thinking about maxims appears to use a trivialized or watered down understanding of the idea of having certain maxims. It is better to reject the idea that we always act on maxims.

The considerations raised in the foregoing paragraphs prompt the following question: does Kant himself explicitly or implicitly comment on whether we always act on maxims? Kant does not directly address the question of whether or not we always act on maxims in his most famous and most widely read works on practical philosophy, like the Groundwork and the second Critique. But he speaks to this issue in other works, both earlier and later. In addition to remarks that speak directly to whether people always govern themselves on the basis of principles and maxims, we will here also focus on two main kinds of remarks: first, expressions of approval towards people (actual or imagined) who Kant thinks of as acting on maxims. Second, remarks about how it is by adopting and acting on maxims that people can build character and become virtuous.

Recall first that one of the defining features of maxims, as Kant described them, is that maxims are practical Grundsätze, i.e. basic principles. It is a recurring theme in Kant’s writings that it is admirable and respect-inspiring to have certain basic principles that we act on. It is also a recurring theme – which runs from earlier works to later works – that we do not always act on basic principles. Consider first Kant’s distinction between what is beautiful and what is sublime in human nature – or in the proper functioning of the human agent – in his early work Observations on the Feeling of the Beautiful and the Sublime. There Kant writes that ‘Those among human beings who act on basic principles (Grundsätzen) are but few’ (2: 227).

This, Kant comments, is a not necessarily always a bad thing. Some people who act on basic principles have bad principles. It can sometimes be better if people act on good instincts (ibid.). However, what is sublime and dignified in human nature, Kant nevertheless adds, always has to do with our ability to act on principles, whereas our capacity to act on, say, benevolent impulses or sympathetic feelings is a beautiful aspect of our nature.Footnote ¹³

At the same time, Kant does express a certain amount of admiration even for those who act on bad maxims, or bad basic principles. Consider the following two claims, the first one taken from the 1798 Anthropology from a Pragmatic Point of View, the second from Kant’s lectures on pedagogy:

If it were the case that Kant thought that people always acted on firm and steadfast maxims, it would be very odd of him to claim that it is pleasing and admirable for people to act on firm and steadfast maxims, even if these are bad and violent in their contents. Suppose, in contrast, that Kant thinks (as he did when he wrote the Observations in 1764) that it is a rare thing – and a sort of achievement – for people to act on the basis of firm and steadfast maxims. Then it suddenly makes more sense that he would express a kind of qualified approval even of those who act on evil or bad maxims.

Now as these claims just quoted suggest, Kant thinks that adopting certain maxims and being firm in our adherence to these maxims is a way of building a character. This is something that Kant thinks takes effort, and that not all of us succeed in doing. Consider the following crucial passage from the Anthropology:

For anybody who wishes to claim that Kant thinks that we always act on maxims – and who also wishes to have ‘maxim’ refer to the same thing as Kant takes it to refer to, i.e. self-adopted basic principles, etc. – this passage is a pretty bitter pill to swallow. Kant is saying that it is a rare thing to have a firm character, where this consists in having certain firm principles. He states that it is admirable and valuable when people do have such principles. Again, if maxims are defined as self-imposed basic principles, this directly implies that having maxims that we follow is something rare – rather than a necessary aspect of all actions. It implies that it is an achievement to have and act on maxims.

The theme of having maxims as a way of building character and being virtuous is also one that we find in the part of the Metaphysics of Morals that Kant calls the ‘doctrine of virtue’. There Kant at one point writes that ‘virtue is the strength of people’s maxims in their following their duty’ (MM, 6: 394) This suggests that being virtuous, as Kant sees things, requires first having character (i.e. having certain basic principles that we follow). It also suggests that this character needs to be in accordance with our duty. It suggests that virtue has two parts: it requires having certain maxims, and it requires that those maxims be fit to serve as universal laws. And Kant does not think that all people are virtuous. Instead, he thinks that it is quite possible to make practical decisions that undermine our character, i.e. that undermine our ability to act on the basis of self-adopted principles. He writes, for example, that:

Suppose it were the case that we always acted on maxims, and this were not only a normative ideal. It would then not make sense to claim, as Kant does here, that it is a duty and prerogative to act on principles and to avoid making ourselves into the plaything of mere inclinations. Something that we always do whenever we act cannot be a basic duty and special prerogative. It cannot be a distinctive sign of character and virtue. So we should conclude that the idea of acting on maxims is a normative ideal, not a defining feature of all human action.

4. The Normative Reading of the Idea of Acting on Maxims

To take stock: in the last section, I argued that (a) the idea that we always act on self-imposed principles and rules clashes starkly with our common convictions and assumptions about human action. We also noted that (b) it is the case that Kant thinks of acting on self-imposed rules and principles as something rare, as an admirable achievement, and as a way of building character and being virtuous. These are strong reasons to substitute a normative reading of the idea of acting on maxims for the widely held descriptive reading. That is to say, it is not the case that everyone always acts on maxims. Rather, as Kant sees things, it is a normative ideal that people should always try to act on maxims. In other words, it is a distinctively Kantian normative ideal to regulate, direct and govern our specific actions and overall agency on the basis of a set of fundamental principles that we reflectively endorse and adopt for ourselves. This is what the normative reading of the idea of acting on maxims holds.

If this indeed is Kant’s view, then how can we fit it into Kant’s overall practical philosophy? I will discuss this question further in the next two sections. But I first wish to make the following observation about what it means for the idea of acting on maxims to be a normative ideal within a Kantian theory of ethics. The idea is actually rather simple, having only two main parts. The first part is that what we ought, or have ethical reason, to do is determined by ‘our will’ as it would be if our ‘reason were practical without obstacles’ (G, 4: 449). The second part is that Kant thinks of reason as a ‘faculty of principles’, i.e. as a law- and order-imposing, or rule- and principle-imposing, faculty (A299–300/B356–7; MM, 6: 214). Reason is, Kant writes, ‘the capacity to judge and (with respect to practice) to act on the basis of foundational principles’ (Anth, 7: 199). So suppose that, following Kant, we understand reason as an essentially ‘active’ faculty whose distinctive activity is to impose order and principles upon objects of its consideration (G, 4: 452). Suppose also that we think of what we ought to do as being what we would do if our reason had complete control over our decision-making. Then it immediately follows that one of the things we ought to do is to govern ourselves on the basis of a set of basic principles. It follows, in other words, that acting on maxims (= basic practical principles) is one of the things we ought to do.Footnote ¹⁴ This is one of the things we would do if our reason had full control over our practical lives as agents.

To make this a little more concrete suppose that, like Kant and most people, you have a general interest in maintaining your health. The way to make this interest conform to the ideal of acting on maxims is to adopt, say, a general principle to promote your own health, and perhaps also certain more specific rules for what to do in order to maintain good health. Suppose next, by way of comparison, that you indeed have a general desire to be healthy, but that you lack any principled commitments regarding your health. Perhaps you tend to experience unpleasant feelings at the thought or experience of ill health. And you might instinctively avoid what you spontaneously perceive as clear threats to your health. If so, you still do not yet fully use your reason (as Kant thinks of it) in your concern to stay healthy. In contrast, if you start formulating general goals and principles with respect to your health, and you decide to try to firmly stick to these goals, then only at this point can you be said to be adopting maxims relating to your health. Only then are you fully using your reason with respect to your concern to be healthy. Only then, in other words, are you following the normative ideal of acting on the basis of maxims in the particular domain of health.

Return now to the issue of how Kant’s view is related to common-sense ideas about human action. I think that it is clear that the normative account of acting on maxims fits better with our ordinary ideas than the descriptive reading does. It is surely not the case that our actions always make full use of our ability to adopt principles and formulate rules. We certainly do not do this with respect to all different aspects and domains of our lives. It is much more plausible to think that we might do this only with respect to certain aspects of our lives, and that we might then do this in more or less elaborate ways.

How we think about our own health seems to be a case in point. People do sometimes reflect on whether or not they are living in a healthy way. They do sometimes adopt principles and rules for themselves regarding how to maintain good health. These include choices pertaining to diet and exercise and the like. So this seems to be a domain in which many people at least partly tend to conform to the Kantian ideal of having and acting on maxims. But it is at the same time hardly a domain within which people conform to this ideal in a very perfect way.

We might think, for example, that we should abstain from certain kinds of foods for health reasons, or that we should do a certain amount of exercise each week, without perfectly conforming our actions to these maxims of health. Whether or not we agree with Kant that it is an admirable and valuable goal to have and act on maxims, it does seem very plausible to think that this is an ideal, rather than a description of human action. And insofar as we do conform to this ideal within some area of life (such as health), this will usually be a partial conformity. As Kant says in the quotes above, it does seem to be a rare thing to find people who have a very strong character based on a set of basic principles or maxims that they firmly and steadfastly adhere to.

We should make a distinction here between (i) whether people have reflectively endorsed, or imposed upon themselves, certain guiding principles and rules, on the one hand, and (ii) the robustness or firmness with which people stick to their principles or rules, on the other. Reflectively endorsing certain rules and principles and also firmly sticking to these is a more demanding ideal – and therefore something more rare! – than reflectively endorsing certain rules and principles and then not consistently sticking to these. Think, for example, of New Year’s resolutions, and the firmness (or lack thereof) with which people stick to these. Or think of many dieters’ ability (or lack thereof) to stick to their diet in a consistent and firm way over time.Footnote ¹⁵ When Kant says that it is a rare thing to find somebody with a true character, what he means is most likely in part that it is rare to have people who are able to firmly and robustly stick to principles and rules that they impose upon their own agency. This all implies that the normative ideal of acting on maxims actually has two different parts: first, to lay down certain basic maxims for ourselves and, second, to remain firm and constant in our adherence to these maxims. Choosing a set of basic principles for ourselves only amounts to a partial realization of this ideal. Full realization also requires sticking to these maxims in a firm and constant way.Footnote ¹⁶

5. Interlude: Good and Evil

I have just argued that the idea of acting on maxims is an ideal, not a description of human action. Before considering two possible objections to this view, I wish to now briefly comment on how the normative reading of maxims relates to Kant’s doctrine of radical evil in Section One of Religion within the Boundaries of Reason Alone. Many interpret this doctrine as implying that we always have a ‘highest maxim’ that determines whether we are good or evil. This would mean that there is always at least one maxim on which we are acting. I think, however, that we do not need to read Kant’s discussion of radical evil in that way.

The issue that Kant is dealing with in Section One of the Religion is whether human nature is innately good or evil. And when he discusses this question, Kant does not say that we all actually have a highest maxim that makes us, as individual human beings, good or evil. What he instead does is to argue that there are certain predispositions and propensities within our shared human nature. These include a predisposition towards moral goodness and a propensity towards evil. Specifically, Kant argues that there are predispositions in our nature that incline us to adopt both maxims of a moral nature and maxims related to ‘self-love’ (i.e. maxims with whose help we can promote our own happiness). These can both, Kant thinks, be seen as predispositions towards goodness (Rel, 6: 26–8). But there is also, Kant argues, a ‘propensity towards evil’ within our human species. This is a propensity to subordinate morality to self-love. It is a propensity to make it our principle to, in cases of conflict, favour self-love over morality. This propensity is evil not only because it subordinates morality to self-love. The worst thing about it is that it is a propensity that inclines us to subordinate morality to self-love as a matter of principle (Rel, 6: 34–5).

Notice that this doctrine does not state that all human beings actually do adopt maxims according to the just-described pattern. It instead states that there are predispositions and propensities within our human nature that incline us towards this pattern. And just as, say, a genetic predisposition to contract some disease does not necessarily result in our contracting that disease, nor do predispositions and propensities to adopt maxims according to certain patterns necessarily result in our adopting maxims according to those patterns (or, for that matter, in necessarily adopting any maxims at all). So we cannot use Kant’s doctrine of the radical evil in human nature – a theory about what predispositions and propensities towards maxim-adoption we have in our nature – to argue that all people do, as a matter of fact, always act on maxims whenever they act.

Some writers, however, think that a remark that Kant makes in the course of his discussion of radical evil nevertheless suggests that we always act on maxims. This is the remark that ‘free choice has the characteristic, entirely peculiar to it, that it cannot be moved by any incentives expect insofar as it incorporates these into maxims’ (Rel, 6: 23–4). On the basis of the assumption that all human choices are free in this particular sense, this is sometimes taken to imply that Kant holds the view that all our choices involve the incorporation of certain incentives into certain maxims (cf. Allison Reference Allison1990). But this interpretation ignores the overall context in which the just-cited remark occurs. When we zoom out a little and look at the context, I think we can see that Kant does not mean to suggest that all choices are based on incentives incorporated into maxims.

In the context where he makes this remark, Kant is talking about the conditions under which it is possible to regard exercises of agency as morally good or morally evil. He is arguing that it is only if we do something as a matter of principle (if we make something into a maxim of ours) that it is possible for our exercise of agency to qualify as an instance of either moral goodness or moral evil. This allows him to consistently take many exercises of agency to be neither morally good nor morally evil. And he clearly thinks that there are instances of human agency that are not best judged in those terms. In the second Critique, for instance, Kant writes that it is ‘beautiful’ to benefit others out of sympathy, but that we do not yet exemplify moral goodness if we are moved by such motives (CPrR, 5: 82). The idea is that genuine moral goodness, as opposed to beautiful agency, requires the presence of a principle or maxim on which we act. Similarly, if we lash out at somebody in anger, this would not necessarily qualify as an instance of evil, even though it might be very bad in some other sense. For an instance of violence to be an instance of evil, on this view, it must be based on a personal principle or maxim of ours.

It should also be noted here that, in the passages discussed above, Kant himself observes that common experience suggests that people are neither wholly good nor wholly evil (Rel, 6: 22). And he does not say that this is a mistaken impression. What he instead says is that in ‘ethics’ (by which I think he means ethical theorizingFootnote ¹⁷ ), it is important to ‘as far as possible’ try to avoid anything ‘morally intermediate’ (ibid.). Kant then describes a position he calls ‘rigourism’. This is a theory that does away with the ambiguities and messiness we encounter in real life. It portrays everyone as having a basic mind-set (Gesinnung) whereby they either adopt the moral law as a highest maxim (= good) or choose not to do so (= evil) (6: 24). But this is presented as a piece of theoretical idealization useful for ethical theorizing, not as a descriptive theory of actual human agency. Hence Kant’s discussion of ‘rigourism’ cannot be used as an argument for the descriptive reading of maxim-adoption or an objection to the normative reading. What is more, Kant does not pledge allegiance to rigourism. The most he does is to say that it is a form of ‘praise’ to call it ‘rigourism’ (6: 22).

Kant’s view of good and evil, then, does not imply that people always act on maxims. On the one hand, he posits certain predispositions and propensities towards good and evil in human nature – which does not mean that he thinks that people necessarily always act on maxims. On the other hand, he thinks that only those exercises of agency that involve maxims can be understood as instances of moral goodness or evil – which also does not entail that people always act on maxims. What this means is that neither of the two main theses that Kant defends regarding goodness and evil in human nature and agency implies that people always act on maxims. Kant can consistently propound those two theses about moral goodness and evil while also taking it to be a normative ideal for us human beings to always govern ourselves on the basis of certain maxims.

6. Concluding Remarks

Having articulated and motivated the normative reading of the idea of having maxims, I now wish to end by briefly discussing two possible worries that might arise. Recall first that I said at the beginning that one reason that people endorse the descriptive reading is that they think of Kantian moral reasoning as beginning by attributing a certain maxim to agents and then asking whether that maxim is also fit to be laid down as a universal law. If it is the case that people do not always act on maxims, and this instead is an ideal, does this not mean, it might be worried, that Kantian moral reasoning becomes impossible? Does it not mean that it becomes impossible to reason about whether actions are right or wrong using the formula of universal law?

I think that this worry is not as serious as it might seem. The reason is that it overlooks the distinction that Kant makes in his major ethical works between the ‘legality’ and the ‘morality’ of our actions (e.g. CPrR, 5: 71; MM, 6: 219). This is sometimes also expressed, as it is at two places in the second Critique, as a distinction between whether we are following merely the ‘letter’ or also the ‘spirit’ of the moral law (CPrR, 5: 83, 152). To put it in yet another way, this is the distinction between acting in accordance with duty and acting from duty (G, 4: 397). It is always possible to decide whether actions have the property of ‘legality’ by following the ‘letter’ of the law and being ‘in accordance with duty’. We decide this by asking whether a person who was governed by a maxim that could at the same time be a universal law could perform the given act, on that basis, in the circumstances (cf. G, 4: 439). It is possible to do the right thing, even if we do not do it for the right reason. And to reason about whether somebody is doing the right thing, what we can do is precisely to attribute a maxim to her – whether or not this is actually a maxim she is acting on, if she is acting on any maxim at all – and then ask whether that maxim could at the same time be a universal law.

We cannot directly know what maxims others are acting on anyway. And so any reasoning about others’ actions requires thinking in terms of whether somebody who acted on a maxim fit to be a universal law could perform their actions in the circumstances. When it comes to our own actions, if we know ourselves not to be acting on any self-imposed rule or principle in some situation, we can ask what rule or principle might lead anybody to act as we do, and then ask whether that would be a principle or rule that at the same time could be laid down as a universal law. This is how to reflect on whether given actions under evaluation would have the property of ‘legality’ by being in conformity with what the law requires.Footnote ¹⁸

When it comes to the ‘morality’ of our actions, or whether we are acting from duty and thereby exhibiting true moral worth, this does depend on whether we are actually acting on a self-imposed maxim and whether that maxim is also fit to be laid down as a universal law for all. And so it can be that many actions that have the property of legality, or of being in accordance with the law, lack the property of morality (Moralität) precisely because the agent lacks any maxim that is guiding her decision-making in the relevant way.

As this helps to illustrate, and as is also indicated by the fact that the universal law formula is supposed to be the principle of a ‘good will’, the universal law formula actually contains two separate precepts: first, to govern ourselves on the basis of maxims (i.e. self-imposed basic principles, etc.) and, second, to choose those maxims on the basis of their fitness to also be laid down as universal laws. It is possible to act without acting on maxims. It is also possible to act on maxims without thereby acting on maxims that are fit to be universal laws. And so a formula such as the formula of universal law requires of us both that we adopt certain maxims and that these maxims have the further property of being ones that could also be universal laws for all (Nyholm Reference Nyholm2015b; cf. Bittner Reference Bittner1974; Höffe Reference Höffe1977).

What, lastly, about passages that can make it seem as if Kant does think that we always act on maxims? What, for example, about the definition of a maxim in the Groundwork according to which a maxim is ‘the principle on which the subject acts?’ Does not this and other similar passages suggest that people do always act on principles? By way of conclusion I wish to suggest that such passages can – and should – be read as elliptical. More fully spelt out, what Kant is saying in such passages, I believe, is the following. Insofar as people do act on certain self-adopted principles, then those principles are called maxims. Kant never explicitly claims that people always act on maxims. As we saw above, he claims that having a steadfast character – where this means having certain firm maxims that we govern ourselves on the basis of – is a rare and admirable achievement. So it is quite plausible to suppose that passages where Kant says that maxims are the principles on which the subject acts are to be read as being about the principles on which subjects act if and insofar as they do act on certain self-imposed rules and principles.

The normative reading of the idea of acting on maxims just presented – on which this is an ideal, rather than a description of human action – might strike some as overly unorthodox. So in conclusion I once more wish to make salient the choice faced by readers of Kant’s texts on practical philosophy. Either we understand Kant as saying that we always act on maxims. If so, we need to use a watered-down conception of what it is to act on maxims, and we also need to ignore passages in which Kant talks about acting on maxims as a kind of character-building achievement. Or we reject the commonly held view that Kant thinks that people always act on maxims. If so, we can understand the idea of a maxim in the complex and demanding way that Kant himself seems to have in mind. And our view need not offend against our common-sense ideas about how people clearly differ in how principled they are. I submit that, on balance, the latter is the better of the two options.Footnote ¹⁹