1 Introduction

One fundamental feature of normativity is the action-guiding or practical nature of normative judgment. Normative judgments typically motivate the people who make them. For example, suppose Alex claims that they ought to donate to charity, yet refuses to donate when the opportunity arises just a few minutes later. Many of us would find Alex’s refusal puzzling. We would look for an explanation: perhaps Alex wants to hold off on donating until their next paycheck, when they have a bit more disposable income, or maybe they have a stronger desire to spend their money elsewhere. Without any such explanation, though, we may come to doubt Alex’s sincerity. We may come to think that Alex did not really judge that they ought to donate, but merely claimed to.

The conventional wisdom among metaethicists is that noncognitivism has an especially easy time accounting for and explaining Practicality, and that this is an advantage of the view. Cognitivism and noncognitivism are competing views about what state of mind a normative judgment is. Cognitivists think that normative judgments are beliefs, or belief-like states with cognitive content. On this view, the difference between judging that grass is green and judging that you ought to donate to charity lies in their different contents – the former is about color properties and the latter about normative properties – not in their attitude-types. Noncognitivists deny this. They think that normative judgments are desires, or desire-like states. The difference between judging that grass is green and judging that you should give to charity is that the former is a cognitive, belief-like attitude, and the latter is a noncognitive, desire-like attitude.¹ And this is what is supposed to provide a simple, elegant, and direct explanation of Practicality. Normative judgments motivate because they are desires, and desires motivate because, well, that is just what desires do – they bear some conceptual or necessary connection to motivation.

In this paper, I argue that this conventional wisdom is wrong. The basic problem is that once we specify the phenomenon of Practicality more carefully, we find that every version of Practicality is either too strong or too weak. Stronger versions of Practicality cannot serve as premises in arguments for noncognitivism, since cognitivists can easily reject these specifications of Practicality. Weaker versions are acceptable to cognitivists, but again cannot serve as premises in arguments for noncognitivism, since cognitivists can explain these versions of Practicality just as well as noncognitivists can. Either way, the argument from Practicality to noncognitivism is far less compelling than many have assumed. Although those who already accept noncognitivism can neatly explain Practicality, they cannot argue from Practicality to noncognitivism. The debate between noncognitivism and cognitivism must be settled on other grounds.

2 The Noncognitivist Argument

It is not hard to see why noncognitivists believe they can explain Practicality better than cognitivists can. At first glance, it seems puzzling why normative judgments are practical, when ordinary descriptive judgments are not. But if normative judgments are desires, or desire-like states, then this is not so mysterious. Common sense has it that desire and motivation are inextricably linked – that to desire something is a matter of being motivated to attain it.² In other words, common sense subscribes to the standard theory of desire according to which to desire that P is to be disposed to bring it about that P: providing motivation is precisely what desires are for, or at least one of the things they are for.³ So we can and should explain the Practicality of normative judgment by appealing to this feature of desires. And one way to do that, while holding onto the idea that there is a close connection between normative judgment and motivation, is to hold that normative judgments just are desires, or desire-like states. So there seems to be some sort of argument from Practicality to noncognitivism.

Philosophers often point to David Hume as the first proponent of this argument. Hume’s claim that “morals excite passions, and produce or prevent actions” can be interpreted as a statement of the thesis of motivational internalism (henceforth, “internalism”): the thesis that necessarily, if a person makes a sincere normative judgment of the form ‘I ought to do x’ then they are motivated to do x.⁴ Internalism explains Practicality by tying normative judgment and motivation together with a necessary connection: if someone is not motivated by a judgment they make, then whatever else is true of that judgment, it cannot be a normative judgment, because normative judgments necessarily motivate.

Internalism can be combined with the popular Humean theory of motivation and trotted out as an argument for noncognitivism. The Humean theory of motivation adds to common sense the claim that having a desire is not only sufficient but also necessary for being motivated to perform some action. If the Humean theory of motivation is true, then noncognitivism seems uniquely positioned to explain internalism and so Practicality. Normative judgments are always to some extent motivating because normative judgments just are desires or desire-like attitudes, and such attitudes are necessarily motivating. Cognitivists cannot similarly explain internalism because cognitive attitudes do not themselves motivate.

Thus, we arrive at what Russ Shafer-Landau calls “the noncognitivist argument”:

1. 1.Necessarily, if you judge that you ought to do something, then you are motivated to do it.
2. 2.Taken by themselves, beliefs neither motivate nor generate motivationally efficacious states.
3. 3.Thus, normative judgments are (at least partly⁵) noncognitive states.⁶

Shafer-Landau’s noncognitivist argument is helpful for understanding how different views in contemporary metaethics account for normative motivation. But it is controversial: some deny its validity, and others the Humean theory of motivation. There is, however, a different way of understanding the noncognitivist argument. The thought is that noncognitivism provides the best explanation of a fundamental feature of normativity, namely, its Practicality. The argument between cognitivism and noncognitivism is then to be understood as an argument over which view best explains Practicality. In other words, the noncognitivist argument is an abductive one, which goes more like:

1. 1.Necessarily, if you judge that you ought to do something, then you are motivated to do it.
2. 2.The best explanation of (1) is that normative judgments are noncognitive attitudes.
3. 3.Therefore, normative judgments are (at least partly) noncognitive attitudes.

The first premise of the abductive argument is the same as the first premise of Shafer-Landau’s deductive argument, but the second premise asserts that the truth of this premise is best explained by noncognitivism rather than asserting that it follows from noncognitivism in conjunction with the Humean theory of motivation. The abductive argument is therefore less controversial than the deductive argument and also better captures the sort of reasoning employed by most contemporary defenders of noncognitivism. So it is the argument I will focus on here, though, in the end, my argument against the abductive argument will show what is wrong with the deductive one as well.

3 Strong Internalism

If Strong Internalism is true, then, as a matter of necessity, there is no gap between judging that you ought to do x and doing x. This would be surprising. The kernel of truth in internalism – what seems right about the thesis to many regardless of their metaethical allegiances – is that normative judgments are practical, reliably motivating those who make them. This is a far cry from the claim that it is impossible to fail to act on one’s normative judgments. Nevertheless, Allan Gibbard embraces this version of Strong Internalism for “a crucial sense of ought” while defending quasirealist expressivism, and we find similarly strong theses in R.M. Hare’s defense of prescriptivism and in John Mackie’s famous queerness argument.⁷ Since the abductive argument works best given Strong Internalism, this is a good place to start, even though we will soon find that the thesis is implausibly strong.

To see why noncognitivists have such an easy time accounting for Strong Internalism, take Gibbard’s quasirealist expressivism in Thinking How to Live. Expressivists like Gibbard deny that we can explain the meaning of normative judgments in terms of their truth conditions, content, or character. Instead, the meaning of a normative judgment is explained in terms of the noncognitive, motivational attitude it expresses. For Gibbard, the relevant attitude is a planning state like an intention or conditional preference. A normative judgment like ‘I ought to do x’ expresses a plan to do x, and what it is for me to think that I ought to do x is for me to plan to do x. As Gibbard puts it, “I don’t, then, think that I ought right now to defy the bully unless I do defy him. If I fail to defy him, then as a matter of the very concept of ought, I don’t believe I ought to.” ⁸ Thus, the necessary connection between normative judgment and action falls out of Gibbard’s account of meaning: the attitude that is necessarily connected to motivation just is the one expressed when that normative judgment is sincere.

However, as Gibbard himself recognizes, strong internalism “seems plain wrong.”⁹ It would be nice if sincerely judging that you ought to do something guaranteed that you would do it, but we sometimes fail to do what we sincerely judge that we ought to do. And when we do, it seems wrong to insist that such failure is possible only in cases of insincerity or conceptual confusion. Much more plausible is the claim that motivational failure is possible because we are capable of experiencing conflicting motivations. We are somewhat motivated to do what we judge we ought, but also more motivated to do something else, where that something else is incompatible with acting in accordance with our initial ought-judgment. In other words, sometimes we are strong-willed and do what we judge we ought, but sometimes we are akratic or weak-willed and fail to do so. This is so familiar to our everyday lived experience that it is hard to take seriously a view that rules this sort of thing out as a matter of conceptual necessity. All this spells trouble for the argument from strong internalism to noncognitivism. If it is possible to fail to do what you judge you ought to do, then we should reject any version of internalism on which this is not possible.

Noncognitivism purports to provide the best explanation of a wide range of normative phenomena, including Practicality. But if the explanations on offer commits defenders of the view to implausible positions, then this is hardly an advantage at all. So even if noncognitivism has an explanatory advantage when it comes to accounting for Strong Internalism, this does not count for much! Since accepting Strong Internalism requires denying the possibility of akrasia, it is not the sort of Practicality we want our metaethical theories to account for.

So Strong Internalism is implausible; some weaker version of Practicality is needed. And we shall see that the noncognitivist move is to modify Strong Internalism, qualifying or weakening the thesis into other versions of Practicality that are more plausible and that can avoid various objections and counterexamples. What is important for my argument is the way that Strong Internalism is qualified or weakened, and whether such modifications to the first premise of the abductive argument leave its second premise intact. In the end, we will see that only controversially strong versions of Practicality – which cognitivists can easily reject – can sustain the abductive argument for noncognitivism. Other versions of Practicality do not support noncognitivism, since cognitivism can explain them just as well.

4 Weak Internalism

Whereas Strong Internalism posits a necessary connection between normative judgment and action, Weak Internalism posits one between normative judgment and motivation. According to Weak Internalism, someone who makes a sincere normative judgment is always to some extent motivated to comply with it. Akrasia is possible because being to some extent motivated to do something is different than being all-things-considered motivated to do it. Motivations come in different strengths and are apt to conflict. So, when Bea, who is trying to quit smoking, judges that they ought to refuse the cigarette, according to the weak internalist, they are to some extent motivated to comply; it is just that their weak motivation stands no chance against the strong craving for a cigarette. So, Weak Internalism is supposed to allow us to explain motivational failures by pointing to the possibility of conflicting motivations rather than insisting on insincerity or conceptual incompetence.

Unfortunately, Weak Internalism also faces serious objections, this time concerning the conceptual possibility of amoralism. The amoralist is a person who makes normative judgments while remaining wholly unmotivated to comply with them.¹¹ Many have argued that amoralists are conceivable, and thus, that Weak Internalism must be false: it is not true that necessarily, a person who judges that they ought to do something is to some extent motivated to do it.¹² In fact, many argue that cases of psychopathy fit this pattern, suggesting that amoralism is not merely a conceptual possibility but an actual phenomenon in the world.¹³ So the conceivability of amoralism and the existence of psychopaths put pressure on the claim that sincere normative judgments are always, as a matter of necessity, accompanied by motivation.

Less extreme characters cause trouble for Weak Internalism, too. Many think that certain familiar psychological conditions such as depression, exhaustion, or apathy undermine the supposedly necessary connection between normative judgment and motivation, and that the absence of motivation is plausibly a symptom of such conditions.¹⁴ Take someone suffering from severe depression or extreme exhaustion. Such a person might judge, sincerely, that they ought to get out of bed and make an appearance at their son’s birthday party, but find that they are completely unmotivated to do so. So although Weak Internalism is no doubt much more plausible than Strong Internalism, it is still contentious whether the thesis is weak enough to allow for certain sorts of motivational failures that are important and familiar.

Some Weak Internalists respond by claiming that amoralists, psychopaths, and people suffering from psychological maladies are making normative judgments that they are motivated to comply with, and insist that, appearances notwithstanding, they are still to some extent motivated. In other words, the claim is that such individuals really do make sincere normative judgments that satisfy Weak Internalism, it is just that the motivation they feel to comply with them is very faint and thus easily overridden by competing desires. On this strategy, what ultimately explains the motivational failure is other weighty conflicting desires rather than the absence of a necessity relation between sincere normative judgment and pro tanto motivation.

A second, more popular strategy is to maintain that amoralists, psychopaths, and people suffering from psychological maladies are not really making normative judgments. Someone who appears to make normative judgments while remaining wholly unmotivated to comply with them either lacks competence with normative concepts or else is speaking insincerely. If they are insincere, then they use normative language only in what Hare called the “inverted commas” sense.¹⁵ They are not making a normative judgment themselves, but rather are “alluding to the value-judgments of other people,” claiming that, “so-and-so falls within a class of actions which is generally held (but not by [them]) to be obligatory.”¹⁶ In other words, this strategy is simply to deny that amoralists, psychopaths, and people suffering from psychological maladies make normative judgments. Although such individuals appear to make normative judgments, this response denies that we can take such appearances at face value. The absence of motivation is evidence, on this view, of a disqualifying defect in the judgment.

While it makes sense that someone already convinced of or committed to Weak Internalism would retreat to one of these responses, others are likely to find these moves pretty unsatisfying. The problem posed by amoralists, psychopaths, and people suffering from psychological maladies is supposed to be that these are people who make genuine normative judgments while remaining entirely unmotivated by them. The first response attempts to save the view that normative judgments and motivation are connected as a matter of necessity from this problem by insisting that, appearances notwithstanding, these people are in fact motivated, albeit very, very weakly. The second response instead saves the view by claiming that such people do not really make normative judgments. But if accounting for motivational failure requires defenders of Weak Internalism to posit motivational states almost entirely drained of their characteristic motivational effects, or to claim that apparent normative judgments are not really genuine normative judgments, it seems reasonable to reject such suggestions as ad hoc. Or, in any event, it requires the noncognitivist to give up the abductive argument. Noncognitivists who embrace these suggestions can maintain the internal coherence of their view, arguing from noncognitivism to Weak Internalism. But they cannot argue from this version of Practicality to noncognitivism.

To see why, recall how the abductive argument is supposed to work. The first premise is Practicality, the phenomenon to be explained. The second premise is the claim that noncognitivism best explains the first premise. And the conclusion is that normative judgments are noncognitive states. But once the noncognitivist retreats to a version of the first premise that the cognitivist rejects, the argument rests on shaky ground. We need agreement on an explanandum in order for one side to claim that they explain it better than the other. Otherwise, we have reached a stalemate. Initially, it might have seemed as though the two sides agreed on what needed to be explained, but offered competing explanations. But – and here’s the rub – now it seems like the disagreement has spread from explanans to explanandum. For the argument now seems to be:

1. (1)Weak Internalism, but a version according to which amoralists, psychopaths, and those suffering from psychological maladies that seem unmotivated by their normative judgments are either a tiny bit motivated or do not really make normative judgments
2. (2)Noncognitivism best explains premise (1)
3. (3)Normative judgments are noncognitive attitudes

The problem is that this version of the abductive argument no longer supports noncognitivism since the explanation on offer commits noncognitivists to controversial claims about what the phenomenon is that calls out for explanation. If the first premise is a highly controversial and arguably ad hoc version of Practicality that noncognitivists cannot provide cognitivists with compelling grounds to accept, then defenders of cognitivism can and should deny that this is an apt characterization of the phenomenon that needs explaining. They can therefore reject this explanation, in just the way that they can reject an explanation that commits us to Strong Internalism. This defeats the abductive argument from Weak Internalism, but it also clearly defeats its deductive cousin. Noncognitivism still lacks the explanatory advantage it claims.¹⁷

5 Typicality

Typicality describes the connection that holds between normative judgment and motivation in extremely general terms. It leaves open what sort of connection there is and whether it is necessary and holds as a matter of conceptual truth or contingent and holds as a matter say, of, facts about normative properties or human psychology. Described in this way, the thesis satisfies (a) – there is wide agreement that we at least tend to be motivated by our normative judgments – but (b) is more contentious. The trouble is that cognitivists can explain Typicality just as well as noncognitivists. All they must provide is some explanation of why we tend to be motivated to comply with our normative judgments, and there are lots of ways to explain this tendency that do not involve maintaining that normative judgments are or express noncognitive, desire-like attitudes.

A standard form of cognitivism that is compatible with Typicality holds that normative judgments are beliefs that motivate when they are accompanied by relevant desires.¹⁸ On this sort of view, normative judgments tend to be accompanied by such desires because of the content of normative judgments: they typically concern things we care about. This is obvious enough in the case of instrumental or prudential judgments, but it is even true in the case of moral judgments. Such judgments tend to be about things that promote wellbeing, cause harm to others, enable us to engage in mutually beneficial cooperation, and so on – and most people care about these things. There may be exceptions when it comes to individuals suffering from amoralism, psychopathy, and other psychological maladies who make genuine normative judgments but happen to lack the desires that typically accompany them. But again, these are the exceptions: most people are at least somewhat motivated when they make a sincere normative judgment.¹⁹ On this sort of cognitivism, we explain normative judgments that fail to motivate in terms of absent desires – desires that typically accompany normative judgments, at least in most people, most of the time, but not as a matter of necessity.

Of course, defenders of noncognitivism can accept and explain Typicality, too – in one of two ways. The first is for noncognitivists to insist that Typicality holds because Weak Internalism holds: if desires necessarily motivate, then they at least typically motivate. But as we have seen, this commits them to a stronger version of Practicality than cognitivists are willing to accept, violating (a). The second is for noncognitivists to try to explain why normative judgments only typically motivate the people who make them. Here, they will have to say that normative judgments are desires or desire-like states that are in the business of motivating (and typically do so) but that nevertheless sometimes fail to motivate. However, gone is the distinctive explanatory advantage noncognitivists once had. This is because this explanation eliminates the necessary connection between normative judgment and motivation, and replaces it with a contingent one. And if the connection between normative judgment and motivation is indeed only contingent, then it is far from clear why we should, in light of it, be moved to accept that normative judgments are or express desires.

Behind the intuition motivating the abductive argument was the idea that both normative judgments and noncognitive, desire-like attitudes bear some necessary connection to motivation. This necessary connection is what is supposed to drive the conclusion that normative judgments are or express these noncognitive, desire-like attitudes. But a merely contingent connection is not tight enough to give the noncognitivist the explanatory advantage they need. Cognitivists can plausibly explain this contingent connection just as well as noncognitivists – and perhaps they can explain it even better, since noncognitivists now have the additional burden of explaining why normative judgments should only typically motivate if they are desire-like states. What the noncognitivist needs to make the abductive argument work, then, is a new version of Practicality: one that, again, (a) both cognitivists and noncognitivists can accept, and (b) noncognitivists can explain better than cognitivists. We have now seen that Weak Internalism violates (a), and that versions of Typicality that do not rely on Weak Internalism violate (b). Can the noncognitivist do any better?

6 Conditionality

Conditionality maintains that the necessary connection between normative judgment and motivation holds only under certain conditions.

There are lots of different ways to specify these conditions. Perhaps the most common specification maintains that the connection holds only when the person making the judgment meets some normative standard. For example, Michael Smith has famously defended a version of this view – call it “Normative Conditionality” – according to which, necessarily, if you judge that you ought to do something, then you will be to some extent motivated to do it if you are rational.²¹ On Smith’s view, the connection between normative judgment and motivation is mediated by rationality: it is a criterion of rationality that, if someone believes that they ought to do x, they will have an accompanying desire to do x, which motivates them to do x.²² So, on Smith’s view, normative judgments are beliefs, and they motivate when they are accompanied by relevant desires. In the perfectly rational person, normative beliefs are unfailingly accompanied by relevant intrinsic desires, and thus the perfectly rational person’s normative judgments are unfailingly accompanied by motivation. But, importantly, on Smith’s view, motivational failures are possible. It is just that failing to form the relevant intrinsic desire to do x while nonetheless judging that you ought to do x constitutes an important form of practical irrationality, not unlike failing to form an intention to do something while nonetheless judging that you have sufficient reason to do it.

That Smith provides a cognitivist explanation of Practicality does not, of course, imply that the noncognitivist cannot provide an alternative explanation of the same phenomenon. For one thing, noncognitivists who insist on Weak Internalism can point out that Normative Conditionality (trivially) follows from it. But we have seen that appealing to Weak Internalism cannot give the noncognitivist the explanatory advantage they are after, so let’s consider the noncognitivist who wishes to explain Normative Conditionality without relying on the stronger thesis of Weak Internalism. Here, the noncognitivist will have to locate the irrationality of motivational failures in a slightly different place than Smith. Whereas Smith claims that irrationality can defeat the connection between normative judgments and motivation by resulting in a failure for a normative judgment to be accompanied by a corresponding desire, the noncognitivist will have to claim instead that irrationality results in the noncognitive, desire-like attitudes that constitute our normative judgments failing to play the motivational role they play in rational agents.²³

Thus, both cognitivists and noncognitivists can accept Normative Conditionality and provide an explanation of it.²⁴ The cognitivist can simply claim that, in rational people, normative judgments are accompanied by desires, whereas the noncognitivist can claim that, in rational people, normative judgments motivate. So far, there is no advantage for the noncognitivist nor a disadvantage for the cognitivist. Instead, we find that cognitivists and noncognitivists who accept Normative Conditionality must appeal to different substantive constraints on rationality.

A noncognitivist might counter that their explanation is better, because the cognitivist’s explanation is unsatisfactory or ad hoc. Why is it that, in perfectly rational agents, normative beliefs are necessarily accompanied by relevant desires? Most beliefs do not work that way. So there appears to be a gap in the cognitivist’s story, an explanatory burden unmet. And any story the cognitivist might offer here seems less straightforward and plausible than the noncognitivist’s claim that normative judgments motivate rational agents because they are desires, and desires motivate.²⁵

The problem with this response is twofold. First, the noncognitivist’s explanation is not as simple as it seems. While noncognitivists can very easily explain stronger versions of Practicality by appeal to the idea that normative judgments are desires, we are here considering the view that normative judgments only motivate rational agents. Why is it that, in irrational agents, normative judgments fail to motivate? After all, they are desires, and the noncognitivist now owes some further story about why the motivational force of desires is conditional on rationality in this way. And while different explanations are surely available to the noncognitivist, whatever story they provide will rely on some further bells and whistles, and specifically on controversial views about rationality that the cognitivist may reject.

Second, cognitivists have in fact suggested many ways of explaining the rational connection between normative judgment and motivation that, they claim, are just as straightforward and plausible as the ideas that noncognitivists can appeal to. For example, T. M. Scanlon has argued that there’s nothing especially puzzling about the idea that, in rational agents, normative judgments are accompanied by motivational states. Drawing on his broader picture of normativity, Scanlon argues that normative beliefs motivate because they (unlike descriptive beliefs) are beliefs about reasons. And it is part of our everyday, “commonsense” understanding of rational agency that someone only qualifies as rational if they are motivated by what they take to be reasons.²⁶ For those who do feel some residual puzzle, however, Smith himself has offered a more elaborate explanation, which also draws on the idea that normative judgments are beliefs about reasons.²⁷ On Smith’s view, what you have normative reason to do is what you would desire to do if you were rational, such that to believe that you have normative reason to do something is to believe that you would desire to do it if you were rational. So, if you believe that you have normative reason to do x, you are committed to believing that you would desire to do x if you were rational. And if you fail to desire to do x despite having this belief, you are thus irrational “by [your] own lights”: you fail to desire what you believe you would desire if you were rational.²⁸ This, along with the subsidiary premise that it is irrational not to desire what you believe rationality requires you to desire, implies a rational connection between normative beliefs and accompanying desires.

Of course, the noncognitivist might still reject these explanations, insisting that they identify the correct constraint on rationality, whereas cognitivists identify the wrong one (or that cognitivists’ explanations for their constraints rely on faulty premises, for example, about what further beliefs normative judgments commit one to). But again, there are two problems with this noncognitivist rejoinder. First, noncognitivism is supposed to be a metaethical (or, better, metanormative) view, and one that is compatible with, but neutral between, substantive first-order normative claims about the content of rationality. Thus, while the noncognitivist could perhaps argue that their rational constraint is better than that offered by the defender of cognitivism, most noncognitivists would be loath to rest the defense of their metaethical view on a first-order claim about the substantive content of rationality (and perhaps for this reason, none have tried). Second, while we certainly can engage in a debate about whether noncognitivism or cognitivism offers the best account of the substantive constraints that govern rationality, this is quite a different question than the one we started with: namely, whether Practicality supports noncognitivism. Whether cognitivism or noncognitivism is correct now seems to turn on who can offer better substantive arguments about what rationality requires, rather than on who provides the better explanation of Practicality.

What of other versions of Conditionality? Other versions try to cash out the conditions in non-normative terms, or at least in less explicitly normative terms. One such version takes the relevant conditions to be those in which the person who makes the judgment is psychologically normal.²⁹ On this sort of view – call it “Normal Conditionality” – if someone is psychologically normal, then they will be motivated to comply with their normative judgments. This makes room for the motivational failures that caused trouble for Weak Internalism. Amoralists, psychopaths, and people suffering from psychological maladies might not be motivated to comply with their normative judgments because they have abnormal psychological conditions that defeat the necessary connection between normative judgment and motivation, leaving the judgment without its standard motivational role.

For present purposes, I will set aside doubts one might have about the possibility of cashing out some notion of psychological normality in non-normative terms in a way that does not just amount to a claim of typicality.³⁰ If this cannot be done, then the thesis ultimately collapses into a version of either Normative Conditionality or Typicality, and thus is vulnerable to the objections I have already raised. But even if it can be done, problems remain for Normal Conditionality.

The trouble is that once again both cognitivists and noncognitivists can accept this version of Conditionality and provide a satisfying explanation of it. This is because the claim that in the psychologically normal person, normative judgments are necessarily connected to motivation does not commit us to a particular explanation of why this connection holds. In particular, it is silent on whether it holds because the content of normative judgments is such that they are normally accompanied by desires, or because the states of mind normative judgments express are noncognitive, desire-like states that are assumed to have some sort of conceptual connection to motivation. The cognitivist can simply claim that, in psychologically normal people, normative judgments are accompanied by desires, whereas the noncognitivist can claim that, in psychologically normal people, normative judgments (that is, noncognitive, desire-like attitudes) motivate.³¹ In order for the abductive argument for noncognitivism to go through, though, recall that we need a version of Practicality that (a) both cognitivists and noncognitivists can accept, and (b) noncognitivists can explain better than cognitivists. We abandoned Strong and Weak Internalism because neither could satisfy (a): as far as cognitivists are concerned, both overdraw the connection between normative judgment and motivation by failing to allow for the possibility of certain familiar motivational failures. Conditionality can satisfy (a), but so far, just like Typicality, it looks like no version of it can simultaneously satisfy (b).

There are many varieties of Conditionality, and I cannot go through them all here. What the discussion has revealed, though, is that in explaining any particular version of Conditionality, the cognitivist and the noncognitivist will appeal to different styles of explanation. The cognitivist will claim that normative judgments motivate when they are accompanied by desires. The noncognitivist will claim that normative judgments motivate when the sort of noncognitive attitude that they claim constitutes normative judgment motivates. And this suggests that, ultimately, the cognitivist and the noncognitivist are committed to two different versions of Conditionality:

1. (1)Necessarily, in conditions where you have an accompanying desire to do something, if you judge that you ought to do something, then you are to some extent motivated to do it.
2. (2)Necessarily, in conditions where the sort of noncognitive state that (allegedly) constitutes normative judgment motivates, if you judge that you ought to do something, then you are to some extent motivated to do it.

Here, cognitivists endorse (1) and reject (2), whereas noncognitivists endorse (2).³²

The disagreement between cognitivist and noncognitivists can only be resolved by determining whether endorsing (2) provides one with a better description of the phenomenon of Practicality. The task for noncognitivists hoping to gain support from the abductive argument is to provide an account of the conditions in which desires (or the relevant type of noncognitivist attitude) fail to motivate. For example, if they are willing to budge from the Weak Internalist thesis that desires motivate in all conditions, then they might adopt a theory of desire on which desires essentially involve something other than motivation – for example, a connection to pleasure or to certain forms of learning – and only typically motivate.³³ They would then need to demonstrate that normative judgments motivate in the same conditions that their general theory of desire predicts. Alternatively, they might adopt a more complex account on which a normative judgment that one ought to do something does not strictly speaking express a desire that one do that thing, but some other more complicated noncognitive attitude, such as a desire to accept the rule that one should do that thing, where accepting a rule typically but not always implies motivation to comply with it.³⁴ They would then need to demonstrate that normative judgments motivate in the same conditions as desires that one accept rules do. If some version of this demonstration could work, this would lend support to (2), the noncognitivist-friendly interpretation of Conditionality.

On the other hand, the task for cognitivists hoping to squash the noncognitivists’ appeal to the abductive argument is to show that the noncognitivist cannot make good on this claim. Furthermore, if the cognitivist can identify uncontroversial cases in which normative judgments fail to motivate, but we would expect that the sort of noncognitive attitude that the noncognitivist is appealing to would motivate, this would provide an argument for cognitivism. (But we have seen that such uncontroversial cases are hard to come by.) More generally, we can see that cognitivism and noncognitivism can be distinguished by whether they accept (2). And this, by the way, is one way of making sense of Dreier’s suggestion that, in light of the problem of “creeping minimalism,” we must ultimately distinguish cognitivism from quasirealist versions of noncognitivism by their different explanations of Practicality (see fn. 2).

In any event, whether Practicality supports either noncognitivism or cognitivism therefore turns on whether accepting (2) provides a better description of the phenomenon of Practicality. Are you motivated to do what you judge you ought to do only in conditions where you have accompanying desires to do it, or also in conditions where the sort of noncognitive state that (allegedly) constitutes normative judgment motivates? Until this question is answered, noncognitivists cannot claim that Practicality provides any support for their view; but neither can cognitivists claim that its failure provides the reverse.

7 Conclusion

We look to our metaethical theories to explain the fundamental features of normativity. One such feature metaethicists have latched onto is Practicality: the tight connection between normative judgment and motivation. An alleged advantage of noncognitivism is the ease with which it explains this connection: it claims that normative judgments are (at least partly constituted by) desires, or desire-like attitudes, and that this is all we need to say to explain why such judgments motivate. The problem is that noncognitivists have not provided a suitable characterization of Practicality that allows them to retain the distinctive explanatory advantage they tout. They either posit a strong and controversial connection between judgment and motivation that cognitivists have no trouble rejecting, or they posit a weaker connection that cognitivists can explain just as well.

With their favored versions of Practicality in hand, noncognitivists may tell each other nice, internally coherent stories about how normative judgments motivate. But they should not expect their stories to convince anyone else. Cognitivists have equally good stories to share amongst each other, and the story one prefers will almost certainly depend on what metaethical view one accepts on other grounds.