Skip to the main content

Original scientific paper

Morality without Categoricity

Elizabeth Ventham orcid id ; University of Salzburg, Austria

Full text: english pdf 302 Kb

page A4-23

downloads: 80


Download JATS file


This paper argues that an agent’s moral obligations are necessarily connected to her desires. In doing so I will demonstrate that such a view is less revisionary—and more in line with our common-sense views on morality—than philosophers have previously taken it to be. You can hold a desire-based view of moral normativity, I argue, without being (e.g.) a moral relativist or error theorist about morality. I’ll make this argument by showing how two important features of an objective morality are compatible with such a desire-based account: 1) morality’s authoritative nature, 2) our ability to condemn immoral agents.


meta-ethics; practical reasoning; hypothetical imperatives; desire; moral realism

Hrčak ID:



Publication date:


Article data in other languages: croatian

Visits: 223 *


It’s common to think of some of our normative reasons and obligations as being moral. Bren, for example, might have a moral reason to attend a protest in her city. We can suppose it’s an important cause, and one where anyone expressing their support will be doing something positive. We can also suppose that Bren is a good person, who cares about the cause. But what exactly gives Bren this reason? And if we subscribe to a theory that explains Bren’s (practical, normative) reasons as being necessarily contingent on her desires, does that mean that her reasons can’t really be moral reasons?

This paper argues that a desire-based theory of normativity need not lead us away from moral realism. Furthermore, it argues that even though this means people without the right moral desires won’t have moral obligations, this account is still less revisionary than it might seem. I begin by explaining the positive view (that of desire-based normativity) in section 1. In section 2 I will introduce two kinds of opponent: advocates for desire-based theories of reasons, who take this to be evidence that moral realism is in some sense a mistake, and those who are committed instead to the truth of morality, and who think that the desire-based view, therefore, must be wrong.

In section 3 I will make my argument against this false dichotomy by addressing two different features that seem to be important for morality to have, and that my opponents might believe are incompatible with a desire-based view of normativity. The first feature is morality’s inescapable and authoritative nature. After all, a key component of morality seems to be that it is in some sense objective and that it holds some power over us. I respond here in two ways: firstly, just because the normativity of morality is conditional on agents having certain desires, that doesn’t mean that we can escape the hold that morality has over us any more than we can ‘escape’ our own desires. I cannot stop myself wanting to be good, or wanting to make other people happy, for example, just because it would be convenient. Secondly, I make an important distinction about what it means to say that normativity is desire-based. It’s not to say that moral principles themselves are subjective, only that desires are informative of who moral principles and rules apply to. I give an argument here about the marginal cases—showing that it’s plausible to think of which creatures have moral obligations based on which ones have moral desires.

The second important feature of morality that I address is our ability to condemn people who don’t follow the moral law. When someone acts badly, it seems important to be able to criticise their actions for failing to meet that objective moral standard. Here I respond in three parts. I begin by arguing that most agents will still be susceptible to the kind of moral and rational criticism that we expect. Next, for when that isn’t the case, I argue firstly that we have better kinds of criticism to offer, and then that certain types of criticism in these cases would be inappropriate anyway.

1. Hypothetical Imperatives and the Desire-Based View of Normativity

Hypothetical imperatives are, as I will understand them, very simple. They’re (1) imperatives (statements that prescribe what to do) that are (2) hypothetical (conditional on something). The thing that they’re conditional on is the agent’s current set of desires. 1 They take this form:

If A desires X, then A ought to φ 2

Where A is an agent, X the state of affairs that they desire to come about and φ is an act that might 3 bring about that state of affairs. Take Bren as an example, and we have the following:

If Bren wants to help others, then Bren ought to go to the protest.

Imperatives of this form can be very wide-ranging. They can include agents’ better and worse desires, and a number of different actions that might bring about those ends. Not all imperatives of this kind will be very important, but the connection with desire is what gets the normativity going.

A desire-based view of practical normativity, then, is one where everything that an agent ought to do takes this same form. 4 There is some desire, either explicit or not, in the foreground or the background, that plays an essential role in the explanation of why she ought to perform the action.

In the contemporary literature, a lot of the discussion of desire-based normativity is framed specifically in terms of normative reasons. According to this view, an agent has a practical normative reason to act if there is an appropriate connection between that reason and their desires. It can be helpful to contrast this with ‘motivational’ or ‘explanatory’ reasons, which look to give explanations for why someone acted. ‘I fell over because I was clumsy’ gives an explanatory reason, ‘I fell over because I wanted sympathy’ gives a motivational one.

Normative reasons do more than just explain actions or explain an agent’s motivations—they justify those actions. Alvarez, who explains the distinction well, says, “it seems clear that reasons can have normative force. By that I mean that reasons can make something right—not necessarily morally right, but right in some respect” (Alvarez 2009, 182). That’s what makes a reason a normative one, rather than an explanatory or motivating one. It’s not just that we want to understand why I fell over mechanically or in terms of what mental states featured in my reasoning—normative reasons cover the why as well as the how.

There are several reasons why a desire-based view of reasons (or practical normativity more generally) is appealing. The source of the normativity’s justification is clear: it comes from the desire, we see in exactly what respects these reasons are making something ‘right’. We know why Bren ought to go to the protest, and it’s because of her desire to stand up for justice. It also shows us why the specific agent in question ought to act, and under which conditions the imperative stops applying. It’s not metaphysically weird. 5 It gives us a link between the action and the psychology of the agent, narrows the list of things she ought to do down to things that she is (in some sense) capable of doing (Williams 1981; Goldman 2009; Markovits 2014). It gives us a way to distinguish between normative behaviour like reasoning (and appealing to an agent’s desires) with non-normative behaviour like restraining (Williams 1981; Manne 2014). All of these arguments are spelled out more carefully elsewhere. My own task will be to demonstrate the account’s compatibility with important features of morality that are usually thought to be necessary sacrifices for such views. In doing so I will give the reader even more reason to find the desire-based view persuasive. Or at least, take away a reason they might’ve had to rule it out.

Before moving on to describe my opponents’ views, I will briefly pause for some clarifications on what I mean when I talk about desire.

1.1. On Desires

One key part of my understanding of hypothetical imperatives is the broad understanding of desire mentioned above. I take desires to be a certain kind of pro-attitude that can manifest in a wide variety of ways, but there’s plenty about desire that I want to remain neutral on. I will say that I don’t take ‘desires’ to only mean desires that an agent is conscious of at a given time, or desires that the agent feels particularly strongly. Even our ordinary idea of desire, after reflection, includes a far greater range of desires than that.

Foot gives the following examples when she discusses hypothetical imperatives:

Sometimes what a man should do depends on his passing inclinations, as when he wants his coffee hot and should warm the jug. Sometimes it depends on some long-term project, when the feelings and inclinations of the moment are irrelevant. If one wants to be a respectable philosopher one should get up in the mornings and do some work, though just at that moment when one should do it the thought of being a respectable philosopher leaves one cold. (Foot 1972, 306).

When I wake up in the morning and feel only the intense desire to stay in bed, that doesn’t mean that all of my other desires have gone away. I still have projects that I desire to continue, people I care about and desire to do well, etc. I still want to become (or continue to be) a respectable philosopher, and I still want to get up in the mornings and do some work, even when those desires aren’t reflected in my current phenomenology.

As well as the desires that I’m conscious of at a certain time, I also have what Pettit and Smith call ‘background desires’ (Pettit and Smith 1990), that is, desires that feature in the background of my thought rather than the foreground. I also want to include what Tim Schroeder calls “standing desires” (Schroeder 2017), desires that do not necessarily play a “role in one’s psyche” (ibid.) at a certain time. After all, we sometimes act to overcome the desires we feel most strongly. What moves us to do this are our other desires. 6

The broad understanding of desire used in this paper is also compatible with a number of different theories of desire itself: that is, theories about what desire actually is. Arpaly and Schroeder, for example, argue that to have an intrinsic desire for P “is to constitute P as a reward”, and to desire not-P is to constitute it as a punishment (Arpaly and Schroeder 2013, 127). There are also a variety of theories in which desire is a particular disposition, such as a disposition to act in a way to bring P about, a disposition to believe that P is good, or a disposition to feel attraction towards the prospect of P. 7 This paper argues that desire—whatever that turns out to be—can be a necessary feature of practical normativity, without that necessary relationship ruling out moral realism. 8 The only thing I require of a theory of desire is that it isn’t too narrow, and that includes the wide range of phenomena listed above: from the background desires and ‘passing inclinations’ to the full-blown projects and passions.

I also take it that, for most of us, at least some of our desires will be moral desires. Desires to change the world for the better, to help our friends, to feed the hungry. Desires that lead us to doing good actions. The details of what might make a desire a ‘moral’ one might rely, to an extent, on which moral theory turns out to be true. It might be the case, for example, that our moral desires are those altruistic ones which compel us to maximise good for others, or to honour and respect the people we do meet. It might be the case that they are those which compel us to exhibit the right virtues, to be kind, generous, and forgiving.

But this doesn’t mean that we would need to know the right moral theory for our desires to be moral ones. It might turn out that Marla’s desire to help a hungry woman on the street is a moral desire even though she doesn’t know the specifics of why it’s a good thing to do, she just knows that she wants to help the woman.

However it is that the distinction is cut, I take it that some of an agent’s desires can be moral ones, and others not. 9 And if an agent can have moral desires, then perhaps they can also have moral reasons and moral oughts that apply to them in virtue of those desires. This paper defends the idea that these moral oughts can apply to agents in virtue of the agent’s desires, and that such dependence on desire is compatible with those oughts being genuinely moral oughts.

Next, I will turn to understanding the rival accounts according to which such compatibility isn’t possible.

2. The Opponents

As I mentioned in the introduction, this paper has two main forms of opposition. Firstly, it will target those who agree that normativity must be connected to desire but go on to suppose this gives them reason to be sceptical of moral realism. Secondly, I will cover those who see that same dilemma, and, if forced to choose, would rather hold on to moral realism than be tempted by an account of normativity that makes imperatives contingent on desires. I hope to give people in both camps a reason to look again at whether they need to make that choice at all. 10

A common route to go down, for people who are persuaded by desire-based theories of normativity, is error theory about morality. Joyce (2001) makes one such argument, although similar ones are made by other error theorists such as Olson (2014) and Mackie (1997). Joyce’s argument can best be explained by looking at a distinction between two kinds of categorical imperative:

2.1. Weakly Categorical Imperatives

Firstly, some imperatives apply to agents in a ‘weakly’ categorical sense. This amounts to not much more than being a description of a set of rules, or perhaps a description of what the speaker would prefer for the agent to do. Joyce gives an example of the rules of gladiatorial combat (such as not to throw sand in your opponent’s eyes), and an unwilling gladiator called Celadus.

He says,

When we say that Celadus ought not throw sand in his opponent’s eyes, this is a weak categorical imperative. A Roman spectator—someone who heartily endorses gladiatorial combat and all its rules—will not retract her judgment “Celadus ought not throw sand” just because it is pointed out to her that Celadus wants to throw sand, and throwing sand is the best means of Celadus satisfying his own desires. (Joyce 2001, 36)

He also refers to these weakly categorical kinds of rules as the “non-evaporating” kind: the kind of rule that doesn’t evaporate just because the agent has none of the relevant desires.

It’s also the same kind of ‘categorical’ that Foot (1972) addressed when she makes an analogy between morality and rules of etiquette. Just because someone doesn’t care about the rules of etiquette, that doesn’t mean that the rules don’t apply to them. But these kinds of weakly categorical imperatives don’t—on their own—seem to come with any “practical oomph” (Joyce 2006, 63) of normativity. They seem to be descriptive more than prescriptive. This isn’t what my opponents are looking for.

2.2. Strongly Categorical Imperatives

The alternative for my opponent is to say that some imperatives apply to agents in a ‘strongly’ categorical sense. This is the sense that goes further than weakly categorical imperatives, further than just describing rules that apply to certain agents. Strongly categorical imperatives apply regardless of an agent’s desires (just like weakly categorical imperatives do), but they also come with the oomph of normativity that’s missing from weakly categorical imperatives. Joyce doesn’t say much more to explain what that extra step looks like, and for good reason—he doesn’t think it exists. Foot (1972, 314), too, talks about it as a feature that’s “missing”.

According to the error theorists, morality needs to consist of imperatives that apply in this strongly categorical sense. Morality needs something stronger than weakly categorical imperatives can supply, and that something is a “non-negotiable” (Joyce 2001, 8) part of our moral discourse. That is, without that necessary feature, our moral discourse turns out to be in error. Because the error theorists think that this extra ingredient does not exist, that we can only understand normativity as being related to an agent’s desires, they also think there can be no such thing as morality after all—our moral discourse is all predicated on a mistake. 11

Unfortunately, most moral realists tend to agree that morality must be strongly categorical in Joyce’s sense. In a recent paper on error theory, for example, Lofitis says “Pretty much everyone accepts the unconditionality of moral requirements (…). Crucially, both error theorists and moralists accept the reason-supplying force of morality” (Lofitis 2020, 40). Shafer-Landau argues against desire-based theories of normativity in ‘Moral Realism’. 12 He, like the error theorists, seems to think that such theories do not make room for some key concepts that are important to moral realism. It’s these key concepts I’ll turn to now, as I begin my defence of this hitherto unpopular view.

3. Two Features of Morality

3.1. Authority (and Inescapability)

The first feature of morality to discuss is its inescapable nature. 13 This is the idea that our moral imperatives have an authority over us: an authority that we can’t escape from or choose to ignore.

Before I begin, I’ll briefly say something about what I mean by imperatives a) having authority over us and b) being inescapable. I take these to be similar ideas, as they aim to explain a way in which the imperative is important, and something that we cannot opt out from. Imperatives have authority over us when they come from an (authoritative, important) source, and they are inescapable when they apply to us whether we like it or not. I aim to show that both of these ideas can be the case for hypothetical imperatives, and I’ll flesh out more about what I mean as I show why this is so.

First Mistake: Escapable Desires

Firstly, agents can no more ‘escape’ moral hypothetical imperatives than they can ‘escape’ their moral desires. I can’t stop wanting to be a good person when that want is inconvenient for me. That’s why we feel pangs of guilt when we do (or we’re tempted to do) something we know is wrong, and why we’ll often at least consider doing what’s right, even when the less morally good option is so appealing.

Suppose that Bren has to choose between going to a protest for a good cause and staying in to have a quiet day at home. Suppose as well that the right thing to do is to go to the protest, and that she knows it, but the better thing for her and her happiness would be to stay at home. Despite this, she still wants to help others, and she would best do this by going to the protest. This desire isn’t something that Bren can opt out of. It might be raining outside and her flat is warm and cosy. Her cat is asleep on her lap and there are games that she wants to play. The protest is a crowded bus journey away, and the weather is cold and uninviting. All of these facts might be running through her mind as she decides what to do, and the temptation to stay at home can be incredibly strong. But the reason that it feels like a difficult decision at all is because her desire to help others is still there, the moral imperative still hangs over her. And if she’s made her decision and she wants to stop feeling bad about it, she might have to trick herself into not thinking about it any longer, or kid herself that it wasn’t really the right thing to do after all. Perhaps, she thinks to herself, her presence wouldn’t have made a difference anyway. And it really wouldn’t be right to move the cat.

Not all of our decision-making will be like this, but the example’s familiarity is indicative of the persistence of some of our desires. In many more cases, our moral desires will still affect us but be less obvious to us. Think back to the breadth of desire—including those that the agent isn’t conscious of at the time. For many of us, the desire to be good is not a whim or a short-term desire that just pops into our heads in certain situations and then goes away again; it’s a long-term preference, a standing desire. We tend to want to be good people throughout our adult lives, and this can persist even when the alternative actions seem much more tempting for one reason or another. It’s something many of us want for ourselves overall, when we get the chance to sit back and think about what’s really important. We shouldn’t be surprised, then, that such a desire can sometimes fail to translate proportionately into a strong motivation, or even into a noticeable feeling. But when the desire isn’t manifesting in those ways, it doesn’t mean that the desire isn’t still there, or that the moral imperatives would cease to have a hold on us. 14 Bren still ought to go to the protest, and the tired philosopher still ought to get out of bed and do some work.

It’s important that a theory of normativity be able to explain what it’s like to feel conflicted about our obligations, and to account for how some of our strongest obligations can be the ones that seem to apply to us despite the fact that we really don’t want to fulfil them. But having an agent’s imperatives be contingent on her desires doesn’t mean that there’ll never be conflict and struggle, that there’ll never be tough situations, or that there’ll never be times that she is compelled by moral imperatives to act against what her strongest desires seem to be. With morality as a system of hypothetical imperatives, the struggle is (still) real.

Second Mistake: A Relative Morality

A related point is one about whether there can be objective moral standards, which exist independently of ourselves. The existence of such principles might be the kind of thing that concerns some people when they think about the ‘authority’ of morality—the fact that we must be held accountable to a standard beyond what we happen to desire. 15

The existence of strongly categorical imperatives isn’t necessary for an objective moral standard itself to still exist. Morality can still form an independent standard without everyone always having reason to follow it, without moral imperatives applying to every creature out there. Only people with the right moral desires will have moral imperatives apply to them, but there can still be an objective truth about which desires are moral. An analogy here might be with other weakly categorical rules, like the law. The law of a country (we can suppose) is a defined set of rules that exist independently of that country’s citizens and their desires. This doesn’t mean that everyone has a reason to follow all of the laws of that country. Some laws might be unjust or irrelevant (the former being something less likely to be the case in morality), or some people might have cares, projects or other reasons that prevent the law from being reason-giving. Perhaps most significantly, not all people are even going to be citizens of that country.

Morality is like the law in that way. It can be an objective standard and authority, one that exists whether or not there are any agents in the world to whom the moral obligations apply. For example, it might be the case that moral goodness consists in agents exemplifying the virtues of compassion and wisdom, or it might be when people work to improve the well-being of others. And if there are agents who are capable in at least some sense 16 of doing these things, who have the right desires, then those are the people who this objective law applies to, the people who have moral obligations. Fortunately, these are all real desires that people have. The normativity of morality is explained.

There’s one more important point I want to make before I move on. An account of hypothetical imperatives doesn’t imply that moral principles change with the desires of the agent, but rather the extent to which the objective principles apply to the agent can change with the agent’s desires. Now is a good time to argue why that’s plausible, and I’ll do so here making particular note of marginal cases.

In the legal analogy there are people for whom the laws don’t apply. The law doesn’t apply to babies, to animals, or to people who live outside of its range, for example. The same applies in the case of the moral law. There are lots of entities without moral obligations: definitely rocks and trees, for example. Animals probably don’t have moral obligations either, at least certainly not most animals. Babies also don’t seem to start out with any, but will usually get closer and closer to being moral agents the older they get and the more they develop (or you might think that they acquire them all at once at a certain point in their lives; either way they tend to start out with none and end up with some). 17 An account of moral ‘oughts’ that connects them necessarily to the agents’ desires gives us an explanation as to why this is: moral oughts apply to entities if/when those entities start to

have moral desires. 18 And that answer seems to track the marginal cases in the right way. 19 Here I’ll go through some of them.

Inanimate objects don’t have desires, and they don’t have moral obligations. That one’s fairly simple, but also not particularly informative. There are lots of other things, after all, that inanimate objects don’t share with humans and which might explain why they don’t have moral obligations. They also don’t act, for example. Animals, then, might have desires, but they’re unlikely to have the kinds of desires that would satisfy an ethical theory’s conditions for being moral desires. Animals don’t generally desire to be good in the way that we take to be morally relevant. But what if they did? If there are intelligent animals out there who form genuine friendships or loving relationships, who are able to understand some concept of goodness that’s recognisably moral and to then intrinsically desire to act in good ways or bring goodness about, then it seems like there would be a good case for those animals being subject to moral oughts. This all depends, of course, on our understanding of the psychology of animals and what our preferred ethical theories take to be the right kinds of moral desires. But so far, having moral desires is a plausible criterion for moral agency.

The case might be easier to imagine with children, who regularly do turn into functioning moral agents. Again, according to my account, they do so not when they want to tell the truth or resist drawing on the walls only in order to avoid getting into trouble (because that isn’t likely to qualify as a moral desire), but when they start wanting to do so just because they want to do what’s right, or they understand the harms involved and want to avoid them, or because they see truth as having value (depending, of course, on the details of what you think a moral desire looks like).

Until people have these desires (if they ever do acquire them), when we tell them that they ought to do what’s right, and talk to them about their moral reasons, it doesn’t seem like we’re using the terms in a seriously morally normative sense, as much as it seems like we’re training them. The imperatives seem only to be weakly categorical. We’re showing them the kinds of things that we want them to desire, that we want them to take to be important. We’re describing the moral law as we see it and we keep doing it until they can see it for themselves. Until then, we’ll make do with the fact that they still ought to avoid drawing on the walls because it will get them into trouble when they’re caught.

To complete the picture, we should think about what happens when humans really never do acquire the right kinds of moral desires. Firstly, it’s worth saying that these sorts of people are very rare. They’re who we might find ourselves describing as amoralists. They’re people who are set up in such a way that they don’t have the right kinds of moral desires, and therefore cannot do what’s good for the right kinds of reasons. My account here says that these people are not subject to moral oughts or moral reasons, and I’m more than happy to agree with that. (They might become subject to them at a later date. Perhaps with a good education or a good friendship, they might change.)

Before moving on I want to emphasise that this account is not trying to be particularly revisionary. In fact, I’m arguing for the opposite: that this account of morality, including of who is subject to moral oughts and who isn’t, is in keeping with what we should expect. In the same way that it wouldn’t be fair to view (most) very young children as being subject to normative moral requirements, the same is true of the rare cases of agents with no moral desires. The desire-based account is stronger for its way of explaining these marginal cases.

Anyone who claims that morality applies categorically will have some caveats, but there’s just some disagreement about what those caveats are. Presumably for now it only applies to things that are alive. Perhaps, for some, it applies to creatures with a certain ability to reason or a rational capacity. For the desire-based account, it’s creatures with certain desires. I’ve shown here that these accounts aren’t so different.

3.2. Moral Condemnation

Another important feature of the moral system is moral condemnation. I’ll begin by explaining why it’s taken to be an important part of moral discourse, and why my opponents might worry that such criticism isn’t possible on my account. This time I’ll respond to the worry in three parts.

Usually when we see an immoral action we condemn it, we think it worthy of our criticism. There’s a moral standard (that can be universal, as just discussed) and if an agent in question fails to meet that standard then we can criticise them on that basis. It seems like something more than just personal disapproval of that agent’s actions, but a more substantive criticism that’s supported by a real morality. 20 This kind of criticism might seem like an important part of moral discourse.

Moral realists might worry about the status of moral condemnation if we accept a moral system that’s only made up of weakly categorical imperatives, instead of strongly categorical imperatives. The latter, after all, are those that apply to agents regardless of what they desire, and that come with some kind of normative force. If moral ‘oughts’ only apply in a strongly categorical way to people who have moral desires, then my opponent might worry that we cannot morally criticise exactly the people who are the most in need of our criticism.

Here I’ll begin by conceding a point. For the agent who has no desires at all to be a good person, it might just not be the case that they ought to do what’s right. There are a range of criticisms unavailable to us about such an agent: she’s not being irrational, for example, and she’s not failing to do what she ought to do. We cannot criticise them in the kind of way that Kiesewetter seems to want to, when he talks about moral criticism and says,

Criticising someone involves more than the judgment that the criticised person has violated some standard; it also involves the judgment that the standard is authoritative for her. […] this means that the person has decisive reason to conform to this standard. (Kiesewetter 2017, 25)

He goes on,

It seems blatantly incoherent to maintain a criticism while accepting that the person criticized had sufficient reasons for what she is criticized for. (Kiesewetter 2017, 29)

This objection also comes up against desire-based theories of what we have reason to do. Williams, arguing in favour of such a theory (‘reasons internalism’) discusses such a case, of a man who has no desire to do the right thing. Williams says,

There are many things I can say about or to this man: that he is ungrateful, inconsiderate, hard, sexist, nasty, selfish, brutal and many other disadvantageous things. (…) There is one specific thing that the external reasons theorist wants me to say, that the man has a reason to be nicer. (Williams 1995, 39)

But I still think that accounts of desire-based normativity have enough tools to deal with this problem, and I’ll show how in three ways.

Response One: Most Agents Will Still have Moral Oughts

My first response is this: agents have a variety of desires, and what they desire most overall will often be different to what they feel most strongly at any given time. This has become a familiar theme in this paper: I don’t stop wanting to be a productive philosopher when I’m in bed in the mornings, and I don’t stop wanting to do what’s right when I’m faced with an incredibly difficult moral decision. Because of this, we can often criticise people for failing to do what’s right, and we can do so on the grounds that they’re not correctly adhering to their desires or to what they ought to do. When an agent fails to do something good we can criticise them because they’re too busy paying attention to their shorter term desires over their longer term ones, choosing the easy options over those that will help them fulfil what they want the most overall.

Brink makes an important point here, following Hume. He says,

If, for example, sympathy is, as Hume held, a deeply seated and widely shared psychological trait, then, as a matter of contingent (but “deep”) psychological fact, the vast majority of people will have at least some desire to comply with what they perceive to be their moral obligations, even with those other-regarding moral obligations. Moral motivation, on such a view, can be widespread and predictable, even if it is neither necessary, nor universal, nor overriding. (Brink 1989, 49)

It seems, fortunately, like most people do have moral desires, even if we’re often bad at acting on them, bad at prioritising them, or bad at seeing how to act on them. It seems like it just happens to be a fact that most people, therefore, ought to act morally, even if what they ought to do is contingent on what they desire.

At other times, the reason why people fail to do what’s good will be down to an epistemic mistake of some kind. I might have moral desires but be really bad at recognising good actions when I see them, for example. Here it would still be the case that I ought to do what’s actually right, because, presumably, I’d still want to do what was right and I’d be failing to do it. If someone mistakenly thinks that the right thing to do is to catcall the woman walking past (“I’m such a thoughtful person,” they think, “I’ll just give her a quick boost to her self-esteem”) then it’s still open to us to criticise them for not getting the facts right.

Response Two: Criticism on Other Grounds

Perhaps that was a bit too optimistic an answer, to declare that we do have strong moral desires most of the time. But luckily I have two more responses in stock. Secondly, then, even when an agent’s desires, correctly weighed, don’t give them the most reason to do what’s morally right, we can still criticise them on other grounds. Indeed, I’ll argue that on these occasions it would be far more appropriate to criticise them on these other grounds, rather than on the grounds of irrationality.

As a reminder, we can’t criticise these agents for not following their reasons correctly, for being irrational or for failing to do what they (normatively) ought to do. (We can, of course, still say that they ‘failed to do what they ought to do’, but in a way that acts more like a description of moral rules or of our own preferences than a normative prescription—the rules that apply to them in a ‘weakly categorical’ sense.) But there are other significant criticisms that we can make. Perhaps most notably, we can criticise them for not having the right (moral) desires in the first place. 21 We can call them cruel, thoughtless, callous, selfish, mean-spirited. The most relevant criticisms that we tend to use in these times seem like they’re descriptions of the agent’s psychological state: criticisms of their priorities, their desires.

Suppose we want Marlene to join us at the protest today, but she doesn’t care about the political issue that’s riled us up. Morally speaking, it seems to be the case that she has an obligation to get involved—she’s a member of a privileged group and there is a clear injustice that she could help to prevent by attending. But, on balance, she just doesn’t want to help—and certainly not more than she wants to stay at home and re-organise her bookshelf. She has no good excuses or mitigating factors: she’s not socially anxious, she’s not tired, and she would be in a good position to make a difference. Just because we can’t criticise her for being irrational, we can still criticise her in a variety of more important ways. Marlene is being selfish, she’s failing to adequately care about the plight of those less fortunate than her, and she doesn’t have a strong enough desire to do what’s right. It doesn’t matter that she’s not failing, by her own lights and according to her own desires, but it matters that she’s failing to meet our standards, and failing to meet the objective moral standards that are important to us.

This doesn’t seem to be an uncommon view; it’s what we do when we call people callous, rude, or selfish. Brink makes this point,

(…) Moral requirements would still apply to agents independently of their contingent and variable desires, even if they would not provide agents with reasons for action independently of their desires. Thus, we could still charge people who violate their moral obligations with immorality, even if we could not always charge them with irrationality. (Brink 1989, 75)

Foot, too, makes a similar point, when she says “The fact is that the man who rejects morality because he sees no reason to obey its rules can be convicted of villainy but not inconsistency” (Foot 1972, 310). Being convicted of villainy seems like plenty, for the limited number of situations in which that’s all we can do.

This response is also relevant to other forms of criticism that might seem at first to be criticisms of irrationality. Take ‘thoughtlessness’ for an example. My opponent might worry that on my view it’s difficult to criticise moral agents as being thoughtless, because it’s a criticism that seems to aim at agents who have failed to notice certain things, failed to think or deliberate properly. The one kind of criticism that my system can’t account for is to criticise people on the basis of failing to follow their own reasons when they have none of the relevant desires. But to fail to think and attend properly to certain things can be a result of not having moral desires. Arpaly and Schroeder (2014, 227), for example, list four ways that desire can affect cognition other than through directly affecting action.

  1. Through involuntary shifts in attention

  2. Through changing dispositions to learn and recall

  3. Through changes in subjective confidence

  4. Through distortion by emotions and wishes

Because our desires affect the way we learn and recall things, our confidence, our attention, etc. they affect what we do beyond just affecting what choices we consciously make. Criticising someone on the basis of having the wrong desires encompasses a lot.

Perhaps this response just puts the problem one step back. How do we criticise an agent for their desires, when it’s not the case that they ought to have different desires? But here we can just put our feet down. We can criticise others without needing to criticise them by their own lights, in a way that is authoritative for them (as Kiesewetter thought was important). We can just criticise them, full stop. And there are still standards by which these agents can be wrong, other than the standards of rationality. The immoral agent with no moral desires is doing something wrong by moral standards, even though she’s not doing something irrational. We can still use weakly categorical oughts when we talk about them, it’s just that this is describing a set of rules (such as the moral law) that they’re failing to meet, rather than saying what they normatively ought to do.

It might be the case that some people think of this as something more akin to a descriptive claim, rather than criticism as such. But this seems to have too narrow a view of what criticism is. Think of other things I can criticise: I can criticise a bird for being too loud, an artwork for being boring, or a tree for being ugly. I can criticise a person for being bad at sports, or a storm for causing damage to the trees in my local park. It seems unnatural to say that I am merely describing these things, when what I am really trying to do is to be critical, to say something negative about them. Even when none of the targets of my criticism ever had a reason to be otherwise, and don’t care about what I have to say—the words can still serve the purpose of expressing my own disapproval. I think that our moral criticisms are often closer to this kind of criticism, and that’s ok.

Furthermore, our criticisms can often have an important effect even when we’re not using them to persuade people to follow their moral obligations. For example, suppose I see someone throwing litter on the street and I call them out for it. This person might not care at all about clean streets or nature or keeping the local area beautiful for people and animals to enjoy, etc. But my criticism might serve to motivate them in other ways. Perhaps by criticising them, I might appeal to reasons they do have — such as reasons to avoid looking bad in public. It’s unlikely that this would be an appeal to a moral desire, and that they would therefore be acting moral as a result, but my moral criticism here can still serve a purpose of getting them to pick up their litter. Perhaps it might even lead to a change in attitude down the line that eventually leads them to have the right moral desires after all. If enough people show you what they value, perhaps that can give you reason to consider taking those values on yourself.

Response Three: When the Criticism is Inappropriate

I have one final defence of the ability to morally criticise people under a system of hypothetical imperatives. I’ve already argued that we can criticise agents with no (or not enough) moral desires in a number of ways, and that we can still criticise people in a lot of cases for failing to do what they ought to do. Finally, I’ll say that the people for whom the last category doesn’t apply—the people without the correct moral desires—are simply not the people who normative moral imperatives should apply to. This is something that I argued for more in the previous section, but is also relevant here.

When we criticise a very young child for causing someone pain, we’d be wrong to criticise them as being irrational or for failing to follow their own reasons. It’s not the case that the moral law already applied to them and they just failed to act in accordance with it. After all, they don't have the moral maturity yet to have moral reasons. And it’s plausible to say, I think, that this is because they don’t have the right kinds of moral desires yet. In criticising people without adequate moral desires for being irrational we’d be making the same kind of mistake. Better to criticise them in a different way, and/or do our best to instill and encourage the right kinds of desires in them in the future.


The aim of this paper is to convince my opponents that we can have a plausible, familiar, and real moral system while still understanding moral imperatives as being hypothetical, rather than categorical. I did this by addressing two important features of morality—the two that might most have been in danger if we understood morality to be a system of hypothetical imperatives—and I showed how they could both be accounted for.

Although this paper should be of interest more generally to people who care about ethics, meta-ethics, and moral psychology, I also had two main opponents: firstly, some of my opponents are moral non-realists. That is, people who think that moral realism is implausible because moral imperatives cannot be strongly categorical. Secondly, my paper also targeted moral realists, who would rather give up the idea of morality as a system of hypothetical imperatives than morality itself. Instead, I hope to have shown, we can have our morality and eat it, too.


I am incredibly grateful to a number of people who have helped me with this paper over the years. As well as the paper’s referees, in particular I'm grateful to Aloysius Ventham, Alex Gregory, Fiona Woollard, Neil Sinclair, Brian McElwee, James McGuiggan, Ben White, Sophie Keeling, Olof Leffler, Eline Gerritsen, and Joe Slater. I've also had some wonderful feedback presenting these ideas at conferences, and so would also like to thank audiences at the UK Kant Society Conference 2019, The Braga Meetings on Ethics and Politics 2019, The 2019 International Conference in Ethics, and the 2018 Irish Philosophical Society Conference. Thanks also to two anonymous reviewers for the European Journal of Analytic Philosophy for their comments on an earlier version of the paper.


[1] Kant (2012) is specifically interested in what we will, rather than what we desire. Since I’m interested

in the wider concept (desires construed broadly) I’ll stick with that. This is also a move that Wedgwood

(2011) and Smith (2004) make, according to Kolodny and Brunero: “Some suggest that this focus, on intentions and beliefs about necessary means, inspired by Kant’s initial discussion of hypothetical imperatives (…) is overly narrow (…). Not simply intentions, but also desires, should be considered (…)”. (Kolodny and Brunero 2020)

[2] In normal language it’s often the case that the former, conditional, part of the imperative is left implicit. Finlay explains this in detail in (2014, 146-175).

[3] I say ‘might’ to remain neutral on the relevance of the agent’s epistemic perspective. It might be the

case that hypothetical imperatives apply to an agent if the action will bring about something that the agent desires, or it might be the case that they apply to an agent if the agent believes that the action would bring about something they desire.

[4] Or that they have reason to do, depending on the specific theory.

[5] See Mackie (1977) and Joyce (2001). Bedke (2010) is an example of an opposing view here, he disagrees that desire-based theories of normativity are any less weird than objective theories. He argues that the desire-based theorist and the objective theorist have a “sincere disagreement” (see Bedke 2010, 50) when they discuss what counts as a normative reason, and therefore they can’t mean such completely different things by the term that, for the desire-based theorist, it’s conceptually necessary that all normative reasons must have a relation to the ends of the agent. I disagree here, as other desire-based theorists might – I think this part of the definition of a normative reason is, after all, the crux of much of this debate.

[6] Another example of the kinds of broad range of desires I’m after is when Williams (1981, 105) refers to a subject’s “motivational set”. He includes in this set “dispositions of evaluation, patterns of emotional reaction, personal loyalties and various projects (…) embodying commitments of the agent”. I take it that many of these will tend to coincide with what we desire.

[7] The former two theories here are discussed, along with a number of others, in Schroeder (2017). The feeling-attraction theory is argued in Smithies and Weiss (2019).

[8] It might be the case that some accounts of desire are better placed to explain why desires provide the normative force that they do. This might be the case with accounts that explain desire in terms of feeling attraction, for example, such as the accounts you might find in Chang (2004) or Smithies and Weiss (2019). I’m grateful to an anonymous referee for their insights here.

[9] Some good discussion on this topic – the ‘moral worth’ of an action, and its relation to desire, can be found in e.g. Arpaly (2002) and Sliwa (2016).

[10] This compatibility with moral realism, and with certain important features of morality, is what makes my project distinct from others who want to defend desire-based views of morality but who don’t see this as being compatible with such features, such as Harman’s (1975) moral relativism or Street’s (2008) constructivism. Instead of arguing for a moral relativism, I’m arguing for a relativism about normativity, and objectivism about morality.

[11] I should note that not everyone who finds this part of morality puzzling is most commonly known as an error theorist. Williams, for example, argues for a desire based theory of normativity in Williams (1981; 1995) and then against the strongly categorical nature of morality in Williams (2011).

[12] See Shafer-Landau (2003, 165-167) in particular. Another moral realist, (Brink 1989), is at least sympathetic to the possibility of questioning the strongly categorical nature of moral imperatives, even though he doesn’t go down that route himself.

[13] Williams (2011), for example, discusses morality’s inescapability in Williams. Joyce also (2001, 51) talks about hypothetical imperatives not being inescapable. Sinclair in his (2016) talks of the inescapability of moral reasons.

[14] It might be an additional worry to some that our moral desires, although they might be continually present, might often not be our strongest desires, and so the moral course of action might rarely be what our desires would point us towards overall. I don’t see this as a particularly big problem for the desire-based normative theorist. They could respond in three ways: firstly, by denying that desire strength is proportional to normative strength (instead it might just be the case that some desire is needed to get normativity off the ground at all). Secondly, they might gesture to the fact that subjective strength of desire is going to be complicated – and not necessarily correlated to how strong a desire feels to an agent. The kinds of desires that moral desires are (persistent, and the kind that we’re likely to endorse) might mean that they tend to be quite strong after all, and have a lot of normative weight. Thirdly, they can bite the bullet, and say that it just is the case that we won’t always/often have the most reason to do the morally best action.

[15] “Anyone who offers an account of the morality of right and wrong is bound to be asked whether he

is claiming that there are ‘universal’ moral principles” (Scanlon 2000, 328).

[16] Given the broad understanding of desire here, I take it that an agent is (in at least some sense) not able to act without having a desire to do so.

[17] This is a different question to whether babies have moral status. Presumably they’ll be worthy of our moral attention long before they start having moral obligations of their own.

[18] A better laid out explanation of this can be found in Arpaly and Schroeder (2013). Aristotle (2004), too, spoke about how taking pleasure and pain in the right things is something that can be brought about through moral education. This is mentioned in Homiak (2016). Mark Schroeder agrees, saying that “Virtue (…) involves desiring the right things, and to the right degree” (2004, 177). For more discussion of when an agent might qualify for moral obligations see, for example, Alvarez and Littlejohn (2017) who talk about a distorted capacity for moral thinking. This is compatible with my view that it’s dependent on an agent having moral desires, since a lack of moral desires might do just that. This is similarly the case with Rosen’s (2004) suggestion of brain anomalies, or of being badly taught.

[19] It’s worth noting that even if agents do reliably qualify for moral obligations at the same time that they get moral desires, that doesn't necessarily mean on its own that the moral desires caused them to become moral agents. It could be, for example, that both arrive at the same time because of a third factor that causes them—such as the agent being able to recognise moral reasons. I don't mean for this correlation to be definitive proof of desire-based theories of normativity (and I want to thank an anonymous referee for this point, which I had missed on my own). I do, though, take it to be a point in the theory's favour over accounts that don't have a way to explain moral agency as easily. I also take it to be a point of defence for the theory, against those who think that desire-based theories of normativity don't do a good job of tracking who it is that moral obligations apply to.

[20] Smith, Lewis and Johnston, for example, describe a “panic” at the idea that there isn’t an objective rationale for morality because, for one thing, our disapproval at each other’s moral views wouldn’t be the same kind of serious thing we took it to be (Smith, Lewis and Johnston 1989, 103-104).

[21] Arpaly and Schroeder (2013) argue that to be good (and virtuous) is to have the right kinds of desires.



Alvarez, Maria, and Clayton Littlejohn. 2017 “When Ignorance is No Excuse.”In Responsibility—The Epistemic Condition. edited by Philip Robichaud and Jan Willem, , editor. Oxford: Oxford University Press.;


Alvarez, Maria. 2009 “How Many Kinds of Reasons?”. Philosophical Explorations. 12(2):181–193.


Aristotle. 2004 The Metaphysics. London: Penguin Books.;


Arpaly, Nomy. 2002 “Moral Worth.”. The Journal of Philosophy. 99(5):223–245.


Arpaly, Nomy, and Timothy Schroeder. 2013 In Praise of Desire. Oxford: Oxford University Press.;


Bedke, Matthew. 2010 “Might All Normativity Be Queer?”. Australasian Journal of Philosophy. 88(1):41–58.


Brink, David. 1989 Moral Realism and the Foundations of Ethics. New York: Cambridge University Press.;


Chang, Ruth. 2004 “Can Desires Provide Reasons For Action.”In Reason and Value: Themes from the Moral Philosophy of Joseph Raz. edited by R. Jay Wallace, Philip Pettit, Samuel Scheffler, and Michael Smith, , editor. p. 56–90. Oxford: Oxford University Press.;


Finlay, Stephen. 2014 A Confusion of Tongues. Oxford: Oxford University Press.;


Foot, Philippa. 1972 “Morality as a System of Hypothetical Imperatives.”. The Philosophical Review. 81:305–316.


Goldman, Alan. 2009 Reasons from Within. Oxford: Oxford University Press.;


Harman, Gilbert. 1975 “Moral Relativism Defended.”. Philosophical Review. 84(1):3–22.


Homiak, Marcia. 2016 “Moral Character.”The Stanford Encyclopedia of Philosophy. 2016 Edition),. edited by Edward N. Zalta. Available online at:.


Joyce, Richard. 2001 The Myth of Morality. Cambridge: Cambridge University Press.;


Joyce, Richard. 2006 The Evolution of Morality. Cambridge, MA.: MIT Press.;


Kant, Immanuel. 2012 Groundwork for the Metaphysics of Morals. Cambridge: Cambridge University Press.;


Kiesewetter, Benjamin. 2017 The Normativity of Rationality. Oxford: Oxford University Press.;


Kolodny, Niko, and John Brunero. 2020 “Instrumental Rationality.”In the Stanford Encyclopedia of Philosophy. 2020 Edition),. edited by Edward N. Zalta. Available online at:.


Lofitis, Kipros. 2020 “Standard and Alternative Error Theories about Moral Reasons.”. Ratio. 33:37–45.


Mackie, John. 1977 Ethics: Inventing Right and Wrong. London: Penguin.;


Manne, Kate. 2014 “Internalism about Reasons: Sad But True?”. Philosophical Studies. 167(1):89–117.


Markovits, Julia. 2014 Moral Reason. Oxford: Oxford University Press.;


Olson, Jonas. 2014 Moral Error Theory: History, Critique, Defence. Oxford: Oxford University Press.;


Pettit, Philip, and Michael Smith. 1990 “Backgrounding Desire.”. Philosophical Review. 99(4):565–592.


Rosen, Gideon. 2004 “Skepticism about Moral Responsibility.”. Philosophical Perspectives. 18(1):295–313.


Scanlon, Thomas. 2000 What We Owe to Each Other. London: Harvard University Press.;


Schroeder, Mark. 2004 Slaves of the Passions. Oxford: Oxford University Press.;


Schroeder, Timothy. 2017 “Desire.”In Stanford Encyclopedia of Philosophy. 2017 Edition),. edited by Edward N. Zalta. Available online at:.


Shafer-Landau, Russ. 2003 Moral Realism: A Defence. Oxford: Oxford University Press.;


Sinclair, Neil. 2016 “Reasons, Inescapability and Persuasion.”. Philosophical Studies. 173(10):2823–2844.


Sliwa, Paulina. 2016 “Moral Worth and Moral Knowledge.”. Philosophy and Phenomenological Research. 93(2):393–418.


Smith, Michael. 2004 “Instrumental Desires, Instrumental Rationality.”Proceedings of the Aristotelian Society (Supplementary. 78(1):p. 93–109


Smith, Michael, David Lewis, and Mark Johnston. 1989 “Dispositional Theories of Value.”. Proceedings of the Aristotelian Society. 63:89–174.


Smithies, Declan, and Jeremy Weiss. 2019 “Affective Experience, Desire, and Reason for Action.”. Analytic Philosophy. 60(1):27–54.


Street, Sharon. 2008 “Constructivism about Reasons.”In Oxford Studies in Metaethics. 3: edited by Russ Shafer-Landau, , editor. p. 207–247


Wedgwood, Ralph. 2011 “Instrumental Rationality.”In Oxford Studies in Metaethics volume 6. edited by Russ Shafer-Landau, , editor. p. 280–309


Williams, Bernard. 1981 “Internal and External Reasons.”In Moral Luck. Cambridge: Cambridge University Press.;


Williams, Bernard. 1995 “Internal Reasons and the Obscurity of Blame.”In Making Sense of Humanity. Cambridge: Cambridge University Press.;

This display is generated from NISO JATS XML with jats-html.xsl. The XSLT engine is libxslt.