1. Introduction
How should we think about the rationality of motivation? In the contemporary literature, Bernard Williams’s approach is helpful: we ask the question, ‘How does one rationally transfer from having one motivational constitution (M1) to another (M2)?’Footnote 1 And the answer, on most anyone’s view, is that one does so by recognizing and appropriately responding to one’s practical and epistemic reasons; a rational transition from M1 to M2 occurs on the condition that the transition is the result of being practically and epistemically rational.Footnote 2 But people disagree on whether we ought to give internalist or externalist analyses of those reasons. If internalists about practical and epistemic reasons are right, Footnote 3 then the reasons to which one responds in rationally transitioning from M1 to M2 are grounded, in some way, in M1. If externalists are right, then those reasons are not grounded in M1.
What it takes to rationally transition from one motivational profile to another, along with the debate between internalists and externalists, is important for at least one very strong reason: we are concerned with whether moral motivation – being motivated, to some extent, to perform moral actions – is rationally required of all rational agents. And this is important because if moral motivation is rationally required of all rational agents then the authority of morality is vindicated. So, the thought goes, we must get clear on practical and epistemic reasons, since those are the two ways we can affect rational transitions from M1 to M2, where M2 includes, one hopes, states of moral motivation.
The above captures, as far as I can see, the current state of play with regard to the debate about rational motivation. And with this debate between internalists and externalists about practical and epistemic reasons in the background, and returning to our question, ‘how should we think about the rationality of motivation?’, we might ask another question: what of the rationality of emotions? For emotions are, on perfectly standard views, motivational states, and so just as having a reason for an action entails a reason to be in a certain motivational state, having a reason for an emotion entails a reason to be in a certain motivational state. But if that is right, then it is quite clear that a rational transition from M1 to M2 could occur by responding rationally to one’s reasons for emotion, viz. by manifesting emotional rationality.
The only ways of denying that this is a way of affecting a rational transition from M1 to M2 is by denying that there are reasons for emotion, but I am not aware of anyone who wants to do this, Footnote 4 and so it seems that we do indeed have three ways of rationally transitioning from M1 to M2.Footnote 5, Footnote 6 The central questions before us, then, are (i) whether we ought to be internalists or externalists about reasons for emotion, and (ii) whether an emotionally rational agent is also a morally motivated agent.Footnote 7
The aim of this paper is to make some headway in answering these questions. I will argue, on the assumption that there are reasons for emotion, that (a) externalism about reasons for emotion entails the view that moral motivation is required of all rational agents, and (b) internalism about reasons for emotion is deeply implausible. I will not, however, give a positive argument for externalism about reasons for emotion. I aim to establish (a) so that we can see what is at stake in a discussion of reasons for emotion, motivation, and the authority of morality.Footnote 8 If the arguments I provide in defense of (a) and (b) are sound, then we should accept that being morally motivated is required of all rational agents. Those who wish to resist that conclusion must either justify internalism about reasons for emotion in a way hitherto unarticulated, or give up on reasons for emotion altogether.
I begin, in Section 2, by articulating two features of emotion that any account of emotions must accommodate, and one feature that any account of reasons for emotion must accommodate. I call this the Basic Account of Reasons for Emotion (BARE) and explain that one can take BARE in externalist or internalist directions. In Section 3 I articulate an externalist development of BARE and argue that an externalist development of BARE entails that moral motivation is required of all rational agents. Lastly, in Section 4, I consider and reject a number of strategies for defending an internalist development of BARE, some of which include attempts to extend strategies employed for defending internalism about practical reason.
2. The basic account of (reasons for) emotion (BARE)
Any account of emotions must include at least two facts about the emotions: (i) they represent their intentional objects in a normative light, and (ii) they are motivational. I begin with a discussion of each of these before turning to a requirement for any account of reasons for emotion.
Emotions are mental states that represent their intentional objects in a normative light. A standard way of capturing this idea is to think that emotions represent their respective intentional objects as having a normative property, e.g. a creature undergoing fear has a representation as of an object’s being dangerous (I use the familiar ‘as of’ to indicate that the representation may not be veridical), anger is a representation as of someone having slighted one, disgust entails a representation as of something contaminated or foul, and so on. But we need not be this strict. We can also allow that the contents of emotions can be non-normative facts presented in a normative light, viz. as reasons for emotion. ‘I can be angry that you didn’t come home when you said you would. I can also be angry with you for not coming home when you said you would [the former indicating anger about a state of affairs, the latter indicating anger with or at a person], ’ Jonathan Dancy says.Footnote 9 The content here does not feature anything explicitly normative; we do not have properties like ‘betrayal’, ‘wrong’, ‘bad’, ‘inconsiderate’, etc. in the content. But that the content is taken by the agent as a reason for emotion shows us that the content, while non-normative, is seen by the agent in a normative light; it is a normative presentation of a non-normative fact. When I say that emotions represent their intentional objects in a normative light, then, I mean that they either have contents with explicit normative content or they have non-normative contents presented as normative reasons. (To anticipate a little, this will turn out to be important when we consider one important strategy for establishing internalism about reasons for emotion).
Emotions are also motivational. I mean this in a very broad sense, such that to go from not having a particular emotion to having a particular emotion entails, inter alia, to undergo a change in motivation.Footnote 10 The particular account of motivation one wants to give, though, is left open. We might give a Humean account of motivation, in which case emotions contain or entail desire, where desires are non-cognitive, essentially motivational mental states that are contrasted with cognitive, essentially motivationally inert mental states (e.g. belief and judgment).Footnote 11 But we might also give an anti-Humean conception of motivation.Footnote 12 In such a case, one might hold that a cognitive state does the motivating. Most obviously, one might think that the representational feature of an emotion is captured by an emotion’s having cognitive content that can motivate, where this is either taken to be something like a belief or judgment that the intentional object of the emotion has a normative property, or that the non-normative content is taken to be a reason and so the agent responds in light of the reason presented.Footnote 13 Whatever one’s theory of motivation, the following is true: if one is not motivated, to some extent, to make amends, one does not feel guilty, if there is no motivation to retaliate or demand apology, one is not angry, and so on. This fact is compatible with all sorts of views on the nature of motivation.
Various accounts of the emotions accommodate these two features (albeit it in different waysFootnote 14) and all such accounts allow that there are reasons for emotion.Footnote 15 Indeed, the recent history of the philosophy of emotion can largely be seen as an attempt to account for just this fact.Footnote 16 And there are at least two reasons for thinking it true. First, it is part of our everyday practices; we routinely think someone is angry when she ought not to be, or is not angry when she ought to be, and that sometimes she is angry but not angry enough or that she is too angry, where the degree of an emotion is at least in part a function of the strength of the motivation that is connected to the emotion. Were there no reasons for emotion, we could never say, ‘You’re being a doormat; you should be indignant, ’ or ‘Don should feel compassion for the suffering man.’ And were there not reasons for having emotions to degrees, we could not say, ‘You should be angry, but not that angry.’
The second reason comes from the shared view across all accounts of emotions: emotions are mental states that represent their intentional objects in a normative light. But if emotions are mental states that represent they can also misrepresent – one’s fear can falsely represent the spider as dangerous, and that you came home later than you said may turn out not to be a reason to be angry (because, say, you stopped on the way home to save someone from their overturned car), and this is, at least in part, what makes it the case that emotions are rationally assessable. This is why Daniel Jacobson and Justin D’Arms characterize reasons for emotions as ‘reasons of fittingness, ’ where an emotion is fitting just in case it accurately represents its intentional object.
Reasons of fit are those reasons that speak directly to what one takes the emotion to be concerned with, as opposed to reasons that speak to the advisability or propriety of having that emotion. So reasons of fit for fear are roughly those that speak to whether or not something is a threat … To think an emotion F (e.g. shame, fear, amusement, etc.) a fitting response to some object X amounts to thinking that X is Φ (shameful, fearsome, funny, etc.).Footnote 17
Reasons for emotion, then, have to do, at least in part, with the veridicality of the representational element of emotion, just as epistemic reasons have to do with the truth of a belief. There are other kinds of reasons to have emotions, as when having them (or not) would be in one’s best interests given one’s circumstances, but moral and prudential reasons are not the sorts of reasons one has for emotions that are intrinsic to the nature of emotion, just as moral and prudential reasons for belief are not intrinsic to the nature of belief.Footnote 18 So, if being angry at one’s friend represents one’s friend as having betrayed one, then one’s reason for being angry is that one’s friend has betrayed one, or that he has, for instance, actively undermined the chances of your success, if being compassionate represents its intentional object as suffering, or as being in a bad way, then a ‘right-kind reason’ for compassion is that someone is suffering, and if being fearful represents its intentional object as being dangerous, a right kind reason to fear something is that it is dangerous.
I am not making the strong claim that reasons for emotion are grounded in the veridicality of emotions and nothing else. I am asserting the weaker claim that any account of reasons for emotion must accommodate the fact that reasons for emotion depend, at least in part, on the veridicality (or truth) of the content of the emotion. One might have a reason to have an emotion with a false content, but that would not be a right-kind reason for emotion; it would be, perhaps, a prudential reason, or a practical reason to get oneself into that state. Call this the veridicality requirement on an account of reasons for emotion.
In sum, there are two requirements on any account of the emotions: that it accommodates the fact that emotions represent their intentional objects in a normative light and that emotions are motivational. And there is one requirement any account of reasons for emotions must meet: the veridicality requirement. Call the combination of these three requirements the Basic Account of (Reasons for) Emotion, or BARE. Anyone seeking to give a plausible account of (reasons for) emotion must meet these requirements – it is the bare minimum one must do.
3. Externalism about reasons for emotion and moral motivation
Let us take a moment to remind ourselves where we are after the detour into the philosophy of emotion. We started with the question, ‘how should we think about the rationality of motivation?’, and we took on board Williams’s suggestion that we think about this in terms of another question, viz. ‘how does one rationally transition from one motivational constitution (M1) to another (M2)?’. There are two standard ways of affecting such a transition – by responding appropriately to one’s practical and epistemic reasons, that is, by exhibiting practical and epistemic rationality – and I have claimed there is a third: by responding appropriately to one’s reasons for emotion, that is, by exhibiting emotional rationality. This led us to ask what sorts of mental states emotions are and what must any account of (reasons for) emotion accommodate. In reply to those questions, I articulated BARE. But all this leaves open whether we should understand reasons for emotion on an internalist or externalist analysis. In this section, I turn to a brief articulation of the distinction between internalism and externalism, and then consider what externalism about reasons for emotion looks like, what its implications are, and why some internalists about practical reasons will want to deny it.
Internalism about a normative reason is the claim that in order for an agent, A, to have a normative reason to x, some motivational fact about A must obtain. Internalism about practical reasons is thus the claim that in order for A to have a normative reason to act, some motivational fact about A must obtain, and similarly, internalism about epistemic reasons/reasons for emotion is the claim that in order for A to have a normative reason to have a belief/emotion, some motivational fact about A must obtain.Footnote 19 Externalism about a normative reason to x is the denial of an internalist claim about a normative reason to x.
Insofar as externalism is a negative thesis, and there are a great variety of internalist theses, it is difficult to establish externalism about a normative reason to x without first articulating the internalist thesis it is denying, or rather, without first establishing all the internalist theses it denies. But this will take us too far afield. So for the sake of brevity, I’ll (provisionally) assume the sort of externalism about practical reasons endorsed by at least some contemporary non-naturalists. On their views (though there are complexities, of course), A’s having a reason to Φ does not depend, in any interesting way, on A’s motivational constitution. In particular, there is nothing about A’s motivational constitution that explains or makes it the case or grounds A’s reason to Φ. Rather, there are certain facts that count as normative reasons, where to be a normative reason is to favor A’s Φ-ing, and this may be, in some cases, a brute fact, or perhaps explained or made the case by certain facts about non-natural value (e.g. that A’s Φ-ing would be, or would bring about something that is, non-instrumentally valuable). On an externalist analysis of reasons for emotion, then, A’s having a normative reason to have an emotion does not depend in any interesting way on M1. Rather, what grounds the reason might be, for instance, the truth of the representational element of the emotion – viz. that the intentional object of the emotion instantiates the normative property it is represented as having (where the normative property is of the non-natural variety) or that the content, though itself lacking any explicit normative content, is genuinely a normative reason to have the emotion – or that having veridical emotions is non-instrumentally valuable (in the way one might suppose having knowledge is non-instrumentally valuable). Again, there are a variety of ways one might articulate an externalist account of reasons for emotion (just as there are a variety of externalist accounts of practical and epistemic reasons), but this broad characterization will suffice for our purposes.
If externalism about reasons for emotion is true then (at least in some cases) moral motivation is required of all rational agents. To demonstrate this I will provisionally assume that if one has a normative reason to x then, ceteris paribus (e.g. there are no countervailing reasons), one is rationally required to x.Footnote 20 With this in mind, suppose A at M1 has no guilt for having Φ-ed, but then sees she has a reason to feel guilty (she sees, for instance, that in Φ-ing she wronged her friend, or broke a promise, or arrived at the surprise party too late). Insofar as she is emotionally rational, and ceteris paribus, she will feel guilt for having Φ-ed; feeling guilty, with the attendant motivational profile of being disposed to apologize, make amends, etc., is present in M2. Notice that, given the truth of externalism, it does not matter what A’s motivational constitution is at M1, and so we can take A to be any rational agent. And since any rational agent can be in this circumstance, all rational agents, insofar as they are emotionally rational and ceteris paribus, will feel guilt for having Φ-ed.Footnote 21
This example happens to be a moral emotion concerning oneself, but we can substitute any moral emotion and get the same result, e.g. a case in which A has a reason to be indignant towards B for having wronged C, or a reason for A to feel compassion for B because B is suffering. Granting the provisional assumption, it is thus a fairly straightforward path from externalism about reasons for emotion to the conclusion that all rational agents have reason to be morally motivated (in those cases in which the emotions for which one has normative reasons are moral emotions).Footnote 22
The provisional assumption, though standard, is controversial. Joshua Gert has argued that there is not one kind of weight normative reasons have, but rather two kinds of weight that correspond to two different kinds of normative reasons.Footnote 23 On the one hand there are justifying normative reasons, which make certain reactions permitted and/or justified but not required. On the other hand, there are requiring reasons, which of course require a certain sort of response. If Gert is right, then we must jettison the provisional assumption and then it is controversial that externalism about reasons for emotion entails that, ceteris paribus, moral motivation is rationally required of all rational agents. That is because the truth of the claim that ‘if one has a normative to x then, ceteris paribus, one is rationally required to x’ depends upon whether the normative reasons are of the permitting or requiring variety. If reasons for emotion are only ever of the former variety and if externalism about reasons for emotion is true, that would only show that all rational agents have permitting, not requiring, reasons for moral emotions and agents would manifest no irrationality in failing to ‘opt in’ to those moral emotions.
There are two things to say here. The first is to highlight that this distinction, if legitimate, highlights that the internalism/externalism debate is only part of the story for vindicating the authority of morality. For, plausibly, in order for morality to be authoritative, it cannot only give reasons to all rational agents, but also must, at least in some cases, give them in a certain way, viz. it must issue requiring reasons. The internalism/externalism debate, however, is simply silent on the issue of the weight of normative reasons; it is about what is required for the existence of a reason, not its weight once it does exist. So even if emotions are always of the permitting variety, that is compatible with thinking that to the extent that externalism as such vindicates the authority of morality, externalism about reasons for emotion vindicates the authority of morality to that extent. It might not go all the way to a complete vindication of the authority of morality, but that is because the internalism/externalism debate is only meant to go part of the way. The second leg of the journey concerns the weight of moral reasons, and that is simply a different issue.
The second thing to say, granting that Gert is right about the metaphysics of reasons, is that, while there are surely cases in which there are permitting reasons for emotion, that does not show that reasons for emotion are always of the permitting kind any more than showing a case of a permitting reason for action shows that reasons for action are always of the permitting kind. One can allow that there are cases of permitting reasons for moral emotions (e.g. compassion for a stranger who lives far away that one has only read about in a newspaper) and cases of requiring reasons for moral emotions (e.g. compassion for the nearby innocent person suffering profound pain), and that would be sufficient to show that externalism about reasons for emotion entails that, at least in some cases, being morally motivated is required of all rational agents. If reasons for emotion were always only of the permitting variety that would be surprising and would need explanation.Footnote 24, Footnote 25
To summarize, externalism about reasons for emotion supports (i) and (ii) or (iii):
(i) Morality is authoritative, at least in part, by virtue of giving all rational agents reasons for moral emotions.
(ii) Given that there is only one kind of normative strength, emotionally rational agents are, ceteris paribus and in morally relevant circumstances, morally motivated.
(iii) Given that there are requiring and permitting normative reasons and there is at least one case in which reasons for moral emotions are of the requiring kind, emotionally rational agents are, ceteris paribus and in morally relevant circumstances, morally motivated.
These are significant claims about the connections between externalism about reasons for emotion and moral motivation that, by themselves, warrant investigation into whether internalism about reasons for emotion is plausible. And at least some internalists about practical reasons will want to avoid externalism about reasons for emotion even though they are not required, on pain of inconsistency, to deny it, and this for three reasons.Footnote 26
First, insofar as a given internalist rejects externalism because it (allegedly) entails unacceptable metaphysical and/or epistemic commitments, Footnote 27 the internalist about practical reason will want to be an internalist about reasons for emotion as well.
Second, being an internalist about practical reasons and an externalist about reasons for emotion introduces a lack of cohesiveness that those philosophers in search of a metanormative theory want to avoid, that is, those who are searching for an account of practical, epistemic, and emotional reasons that gives them all the same treatment, including that they are all internal/external. And indeed, if one allows externalism about reasons for emotion, then one needs an explanation for why externalism about practical reasons is unacceptable.Footnote 28 It may be that such a division is acceptable and there is a good explanation for it, but this is an explanatory burden some internalists will want to avoid.
Lastly, internalists about practical reason who think that one can be fully practically rational without being morally motivatedFootnote 29 and want to continue to maintain that moral motivation is not required of all rational agents will also want to deny externalism about reasons for emotion.
To summarize, in Section 2 I offered an articulation of BARE and in this section I considered what an externalist development of BARE looks like, concluding that it entails i and (ii or iii) above. I also explained that there are at least three reasons why some internalists about practical reasons will want to resist externalism about reasons for emotion. I now want to suggest that, while BARE can be developed in an internalist direction, we already have a prima facie case for an externalist development. That case consists in the rationale behind the veridicality requirement presented in Section 2: we want to ensure that the account of reasons for emotion are of the right kind, that is, that they speak to the fittingness of the emotion and not, as D’Arms and Jacobson highlight, the (prudential) advisability or (moral) propriety of (not) having it. So it is difficult to see why we need to appeal to the motivations of agents to show that they have normative reasons for a given emotion, which is just to say we have a prima facie case for externalism about reasons for emotion. The internalist about reasons for emotion has the burden of demonstrating why the motivations of an agent are relevant to giving the agent reasons of the right kind for emotions.
In the next section I consider, and ultimately find wanting, five strategies for internalist developments of BARE. I should say at the outset that I do not take the arguments in the next section to amount to a refutation of internalism about reasons for emotion. The aim is to show that the burden is on internalists about reasons for emotion to show the position to be plausible, let alone true.
4. Five strategies for internalist developments of BARE
Strategy #1. We came to the conclusion that emotionally rational agents are morally motivated agents in part by thinking of the normative properties emotions represent (or the non-normative properties that are presented as reasons) in the way that externalists think of them, viz. as not depending on the motivational constitution of the agent, so a natural strategy an internalist about practical (and epistemic) reasons might adopt is to give an internalist-friendly account of the properties represented by emotions, that is, an account according to which, roughly, normative properties are instantiated by virtue of standing in some dependence relation to the motivational constitutions of agents. But this strategy for making internalism about reasons for emotion plausible will not prove successful for two reasons.
The first is that emotions do not always have explicit normative content; they do not always represent normative properties. One’s anger that he didn’t come home when he said he would, or one’s anger with him for not coming home when he said he would, don’t have explicit normative contents, and so there are no normative properties to be given an internalist-friendly analysis.Footnote 30
The second reason is more interesting than this. Let us suppose, for the internalist’s sake, that the contents of emotions always have explicit normative content, and that normative properties are to be given an internalist-friendly account. The surprising result is that this still would not get us internalism about reasons for emotion. Take, for instance, the schema for a case of reason for indignation offered above: A has a reason to be indignant towards B because B wronged C. In this case, all the normative properties are grounded in C’s motivational constitution, or perhaps B’s and C’s, but certainly not A’s. So supposing that A is indignant towards B because B and C are spouses and B cheated on C, B’s wronging C consists in, let’s say, harming C by thwarting C’s desires, or perhaps in thwarting the desires of B and C to maintain a good relationship, or something to that effect. Again, it is nothing about A’s desires that ground the normative properties represented in A’s indignation towards B. Or to take the case of compassion, the content of A’s compassion has to do with B suffering, where B’s suffering is a function of B’s desires being frustrated, and not in A’s desires being frustrated.
In both cases it is not A’s desires that are grounding the normative properties that feature in the content of A’s emotion; it is, we are supposing, the desires of B and C. But recall the schema for internalist reasons statements: in order for an agent, A, to have a normative reason to x, some motivational fact about A must obtain. But if the only internalist-friendly feature of our account of reasons for emotion is an internalist-friendly account of normative properties that feature in the contents of emotion then we do not get internalism about reasons for emotion; A has a reason to feel indignant where that reason does not depend on any motivational feature of A. And this point is generalizable to any internalist-friendly account of normative properties. For the objection doesn’t depend on what motivational fact obtains about A, but rather on the fact that A’s motivational constitution cannot ground all the normative properties that feature in the contents of A’s emotions. Indeed, there are no facts about A that can ground all the normative features that figure in the contents of A’s emotions. If we are going to establish internalism about reasons for emotion, then, we will need a strategy that consists in something other (or something more) than supposing that all emotions have explicit normative content and then offering an internalist-friendly account of the normative properties emotions represent.
It might be objected that A’s motivational constitution might come into play in some other way. Perhaps, for instance, A’s finding B’s behavior wrong is required for A to have a normative reason to be indignant, where finding B’s behavior wrong consists in a judgment, the making of which requires a certain motivational constitution. If that is right then we must look to A’s motivational constitution to determine whether A has a normative reason for indignation.Footnote 31
I do not think this is right, though. It cannot be a necessary condition for A’s having a normative reason (be it for action, belief, or emotion) that A judge that A has that normative reason. Prior to A’s judgment that there is such a normative reason, A has that normative reason, which in at least some cases explains why A comes to (truly) judge that A has that reason. Further, if this were a necessary condition, being ignorant of one’s normative reasons would be impossible, since one would not have a reason until one acknowledges it. But all – internalists and externalists alike – allow that one can be ignorant of at least some of one’s normative reasons.
Strategy #2. How else might one argue for an internalist development of BARE? One natural reply to the case of A, B, and C is to claim that if A has, say, no desire for B to be loyal to C, then A will have no reason to be indignant towards B. But that would simply be to assert internalism (and more specifically, Humeanism) about reason for emotion, not to present an argument for it. That reply also fails to discharge the burden articulated at the end of the previous section: given the prima facie case for externalism about reasons for emotion, the internalist needs to explain why we need to appeal to the motivations of an agent in order to ascribe a normative reason for emotion to her. It is, however, worth dwelling on this point in relation to Williams’s well-known (Humean) discussion of an abusive husband who has nothing in his subjective motivational set – no desire, in a broad sense of that term – from which he could deliberate, via a sound deliberative route, to the conclusion that he has a reason to stop abusing his wife.
There are many things we can say to people who lack appropriate items in their S. Suppose, for instance, I think someone … ought to be nicer to his wife. I say, ‘You have a reason to be nicer to her’. He says, ‘What reason?’ I say, ‘Because she is your wife.’ He says – and he is a very hard case – ‘I don’t care. Don’t you understand? I really do not care.’ I try various things on him, and try to involve him in this business; and I find that he really is a hard case: there is nothing in his motivational set that gives him a reason to be nicer to his wife as things are.Footnote 32
Williams concludes that it is false to say that the husband has a reason to be nicer to his wife, though we may say he is ‘ungrateful, inconsiderate, hard, sexist, nasty, selfish, brutal, and many other disadvantageous things.’Footnote 33 But there is no failure of rationality, according to Williams. Putting to the side for now whether the abusive husband has normative reason to act differently, we may ask whether he has normative reason to feel differently, that is, normative reason to be differently motivated in a way that does not depend on his normative reasons for action. The externalist about reasons for emotion will say that indeed he does: he has a reason to feel compassion, viz. that his wife is suffering, or is in a bad way, and this is true even if the normative property represented is given a Humean analysis, since her suffering is grounded in her desires, not her husband’s. Insofar as the husband is emotionally rational and ceteris paribus, he will feel compassion and thus a desire to be nicer to her. Put differently, on the externalist account, there may not now be something in his S (M1) that would motivate him to cease being cruel, but were he emotionally rational, he would have that element (in M2). In resisting the externalist about reasons for emotions, the internalist will have to give us a reason for thinking that, in developing BARE, we should do so in an internalist direction, which would render the externalist’s account of this case misguided.
A natural internalist reply is that the husband ‘really do[es] not care, ’ and if he does not care, then he cannot have a reason to feel compassion. But this is just to assert internalism, not to argue for it. Compassion is a way of caring. It is simply irrelevant to say, in reply to the assertion that he has a reason to care about X, that he does not care about X. It is whether he ought to care, or at least whether he has a normative reason to care, that is at issue. And once again, the internalist must tell us more. Why should we think an appeal to the husband’s motivational constitution is relevant here? And can the internalist appeal to that motivational constitution in a way that does not generate reasons of the wrong kind? Given the prima facie case for externalism about reasons for emotion we need something more than simply an assertion of internalism.
When I put things this way to philosophers in conversation, the standard reply is to alter the case such that the husband not only lacks a desire to be nice to his wife, but also has a desire for her to suffer. ‘Wouldn’t he have a reason to be joyous rather than compassionate?’ it is asked. ‘And if so, doesn’t this speak in favor of internalism about reasons for emotion?’
But even if we present the case as one in which the husband positively desires his wife’s suffering, it will do no good in supporting internalism. We see this by first acknowledging that there is a reason to feel joy when something good is going on (for someone); that is, roughly, what joy represents. And so, if the wife suffering is good in some way for the husband then he has a reason for joy. Now we could deny that this is so – we could deny that her suffering could be good for him – and that would be one way of replying to the rhetorical questions asked in defense of internalism. But this seems implausible, and certainly unlikely to convince an avowed internalist, so we should allow that it is at least possible that someone else’s suffering is good for someone, and we may as well allow that it is at least possible that, in the imagined case, what is bad for the wife is good for the husband. So let us grant to our objector, at least for the sake of argument, that the husband has a normative reason for joy.
But there is a reason to feel compassion when something is going on that is bad (for someone); again, that is, roughly, what compassion represents. The husband case is such a case – his wife is in a state that is bad (for her) – and so he also has a reason to feel compassion. What we have, then, is not a case in which the husband has no reason for compassion, but rather one in which the husband has a reason for joy and a reason for compassion. The case does not show, as our questioner hoped, that there is a reason for joy but not for compassion.
Strategies #3–#5. In this section I consider internalist accounts of reasons for emotion that depend on extending extant internalist arguments for practical reasons. One advantage of this approach is that, insofar as those theories already tell us why appealing to the motivational constitutions of agents is needed to explain their respective practical reasons, they hold out the possibility of discharging the burden of explaining why appealing to motivations of agents is needed in order to explain their reasons for emotion. I do not have the space to discuss each argument and/or variety of internalism about practical reasons and then attempt to see whether it can be extended to an argument for, or turned into an account of, reasons for emotion. What I say here is cursory, and I do not pretend to do full justice to the views to come. Instead, I offer a few remarks on three different accounts in order to provide grounds for skepticism, and to place the ball in the internalist’s court.Footnote 34
Consider, first, Williams’s seminal defense of internalism. On a standard interpretation of that argument, internalism about practical reasons is entailed by two premises: first, reasons must be capable of explaining action, and second, the Humean theory of motivation is true, viz. desires play a special role in the explanation of action. The idea is that for a consideration to be a reason for action it must be possible for that consideration to feature in an explanation of the action it favors, and it can only do so if the action the reason favors bears the right sort of relation to the desires of the agent (e.g. serving or promoting the satisfaction of them).Footnote 35
Even if this is a plausible argument for internalism about practical reasons it is dubious in the case of reasons for emotion.Footnote 36 To see this, let us grant for the sake of argument that for a consideration to be a reason for a token emotion for A it must be potentially explanatory of A’s having that emotion and turn our attention instead to whether it is plausible that desires explain emotions in the way that Humeans think desires explain action.
On the Humean account of explanation of action, desires stand in relation R to action, where R is standardly cashed out as a causal relation. On that view, roughly, Bob’s desire to eat cake causes and (thereby) explains his going to the cake room, and ‘that there is a cake in the cake room’ is capable of being a reason for him to go because it can explain his going there by way of favoring an action that promotes the satisfaction of his desire for cake. R need not be understood as a causal relation, though. On Michael Smith’s view, desires are dispositions (to act) with a mind-to-world direction of fit; a desire to go to the cake room just is a disposition to do that which realizes or makes-true the propositional content of the disposition, e.g. Bob’s desire to eat cake just is the disposition to, for instance, go to the cake room upon learning that there is cake there; doing so will make it the case (or at least make it more probable) that the propositional content of his desire – that he eat cake – will be made true.Footnote 37 Can either of these models be made to work for the explanation of emotion such that the second premise of Williams’s argument can be carried over to an argument for reasons for emotion? I do not think so.
Suppose that Bob desires Alice’s happiness and then tragedy – sudden death, say – befalls Alice. Bob is sad. Does Bob’s desire for Alice’s happiness stand in the same relation to his sadness as his desire for cake stands to his going to the cake room? It would be odd to say that his desire for her happiness causes his sadness in the way that Bob’s desire to eat cake causes him to go to the cake room. Going to the cake room is a way of satisfying the desire to eat cake and so, plausibly, the desire to eat cake causes and (thereby) explains his going. But feeling sadness is not a way of satisfying his desire for her happiness – her death makes promoting that desire impossible – and so cannot explain his sadness in the way his desire to eat cake explains his going to the cake room. Similarly, his desire for her happiness cannot just be a disposition with world-to-mind direction of fit such that he is disposed to feel sad. Sadness is not a way of realizing or making-true the propositional content of his disposition that is his desire for her happiness. So the prospects of extending something like the Humean theory of action to include a Humean theory of emotion are grim.
If we are going to allow that Bob’s desire for Alice’s happiness is explanatorily relevant to his feeling sad then what we should say is not that his desire for her happiness causes his sadness nor that his sadness is a way to make-true the propositional content of his desire, but rather that it is his desire being frustrated that is (part) of the explanation for his being sad. But note, first, that the role Bob’s desire for happiness is playing in this explanation is neither a causal nor a dispositional one. ‘That his desire is frustrated’ explains his sadness in the sense that it rationalizes or makes sense of his sadness and does not stand in the same relation to his sadness as his desire for cake stands in relation to his going to the cake room. And second, not only is the relation different, but so are the kinds of things related. ‘That his desire for her happiness is frustrated’ is a state of affairs but his desire for her happiness is a mental state; these are two sorts of metaphysical entities. Since the explanation for his sadness ‘that his desire is frustrated’ both appeals to a different relation (the rationalizing relation, say) and to different kinds of things being related (a state of affairs in the case of emotion, a mental state in the case of Humean explanations of action), it is quite clear that this sort of explanation of an emotion that involves reference to desire is not at all similar to a Humean explanation of action. And there is nothing strange about this case; it is a perfectly standard case of emotion explanation.
Perhaps this goes too far, though. Surely, it will be thought, there must be some appeal to the mental states of Bob in order to explain his feeling sad where that appeal is to the mental state and not to some fact about Bob (e.g. that his desire is frustrated). More specifically, one might think something like the following must be true: any explanation of an emotion must appeal to the psychology of the agent for whom it is a reason. If that is right then perhaps the internalist can remain neutral on what item of the agent’s mental economy is appealed to, that is, whether it is desire or some other mental state.
But this revised premise, unmoored from a Humean theory of motivation, is too broad to defend an internalist view. We can see this by recognizing that externalists can affirm that an appeal to the agent’s psychology is needed. On Jonathan Dancy’s ‘pure cognitivist’ view, for example, one need only cite cognitive states (e.g. beliefs, particularly those with normative content) in order to explain action, and this is compatible with his externalism about normative reasons.Footnote 38 Williams’s argument was supposed to work because of desire’s special role in action-explanation and not simply because some appeal to an agent’s psychology was required in a full explanation of his action. Jettisoning the Humean theory of motivation and offering a broader account of emotion explanation that appeals to some-mental-state-or-other is thereby disastrous for the argument.
The only way to salvage the Williams-style argument for internalism about reasons for emotion is to specify a mental state, m, where m is to emotion as desire is to action, with the essential caveat that m is not a mental state an externalist could accept as explanatory of emotion. I do not have an argument for the impossibility of such an account, but at this point, it seems implausible. If desires are to be plausibly gotten into the story of a reason for emotion, it will likely have to be in some other way.
This brings us to strategy #4: extending Mark Schroeder’s argument for internalism about practical reasons to an argument for internalism about reasons for emotion. Schroeder rejects Williams’s defense of internalism and rather than appealing to the claim that desires explain action as a premise for an argument in defense of internalism, Schroeder appeals to desire’s ability to explain reasons. That is, in answer to the question, ‘What makes it the case that the fact that there is cake in the cake room constitutes a reason for Bob to go to the cake room?’, Schroeder replies ‘because that fact bears the right sort of relation to Bob’s desire to eat cake.’ More specifically, the answer is that the fact that there is cake in the cake room explains why Bob going to the cake room increases the probability of Bob satisfying his desire to eat cake. In fact, on Schroeder’s view, a normative reason for action just is some fact that explains why performing some act increases the probability of satisfying desire.
Schroeder’s view on practical reasons is prima facie plausible, and that is because it gets at least one plausible intuitive data point exactly right: at least sometimes, what explains why an agent has a reason to do something is that doing it is a means to (or constitutive of) realizing one’s end that is set by desire. That is a fairly commonsense explanation for why someone has a reason to do something, at least in some cases. But this intuitive data point is absent in standard cases of explanations as to why an agent has a reason to token some particular emotion. In explaining Bob’s reason to admire Alice, it is not intuitive to think that what explains Bob’s reason for admiring Alice is that admiring her makes it more probable that he will satisfy at least one of his desires. It is similarly counterintuitive to think that what explains A’s reason for being indignant towards B for how B treated C is that being indignant makes it more probable for A to satisfy at least one of A’s desires. So the intuitive appeal of a view that posits desires as explainers of reasons for actions is simply not present in a view that posits desires as explainers of reasons for emotion.
Why is it not intuitive in the case of reasons for emotions? The obvious problem is that they seem to speak to the advisability or (im)propriety of the emotion as opposed to reasons of fittingness. More specifically, they seem not to speak to whether Alice is admirable or B is a proper object of A’s indignation, respectively. On the other hand, that going to the cake room will promote the satisfaction of Bob’s desire to eat cake does make Bob’s going to the cake room a fitting response.
But Schroeder’s view plausibly has the resources to respond to this worry. Borrowing from moves he makes to defend his account of practical reasons, Schroeder may reply that the sense in which Bob’s desires do not explain his reason for admiring Alice, or A’s reason for being indignant towards B for how B treated C, is that Bob’s and A’s respective desires are not part of their respective normative reasons for admiration and indignation. They are, instead, ‘background conditions’ of the reason. So Bob’s full reason for admiring Alice is ‘that she is an accomplished mathematician’, and A’s full reason for being indignant towards B is ‘that B treated C in such and such a way’, even if some sets of desires of Bob and A, respectively, are part of the full explanations for why those are reasons to admire Alice and be indignant towards B, respectively. If this can be made to work, then a crucial virtue of this account is that it plausibly meets the veridicality requirement and articulates reasons of the right kind: the contents of Bob’s and A’s respective emotions plausibly have veridical content since those emotions are had in light of the reasons there are to have those emotions. This looks promising.
But let us ask, ‘what is the desire that explains why ‘that she is an accomplished mathematician’ is a reason for Bob to admire her? Surely it is not Bob’s desire to admire Alice. It is simply not the case that ‘that she is accomplished’ explains why Bob’s admiring Alice increases the probability of admiring her. The explanation for why Bob’s admiring Alice increases the probability of admiring Alice is simply that admiring her is a satisfaction of that desire; his admiring her just is the truth of the propositional content of the desire; ‘that she is an accomplished mathematician’ is not appealed to in the explanation. We are left wondering what the desire could be that would explain the reason.
Things are no better – indeed, they are worse – if we consider Bob’s desire for Alice’s happiness serving as an explanation of his reason to be sad for Alice when tragedy befalls her.Footnote 39 That desire is by far the most natural candidate for what would explain his reason for being sad, if desires explain reasons for emotion at all. But ‘that Alice has suffered a tragedy’ does not explain how Bob’s feeling sad is a way of increasing the probability of satisfying his desire for her happiness. In fact, if the tragedy that befalls her is sudden death, his feeling sad cannot be a way of increasing the probability of satisfying his desire for her happiness because it is now impossible for that desire to be satisfied. We are left with the shocking conclusion that Bob’s desire for Alice’s happiness cannot explain his reason for being sad when tragedy befalls her.
Schroeder has a move he can make here that he makes when defending his account of practical reasons. He may appeal to some desire or other that Bob has to explain his reason for being sad, even if that desire has nothing to do with Alice or her happiness as such. As Schroeder puts it, an agent’s reasons for action can be ‘massively overdetermined’ in that any one action promotes a great variety of one’s desires.Footnote 40 To take a variant of Schroeder’s surprising example, ‘That Katie is in need of help’ explains why Ryan’s helping Katie promotes his desire for a new pair of shoes, for instance, and so can be his complete reason for helping her. (Perhaps ‘that Katie needs help’ explains why helping her gets Ryan new shoes because Katie gives shoes to people who help her when she is need of help). Similarly, ‘that Alice has suffered a tragedy’ might explain why Bob’s feeling sad is a way of increasing the probability of satisfying his desire for new shoes. (Perhaps Bob’s feeling sad allows him to empathize with Alice’s sister, a woman who expresses her gratitude by purchasing shoes for those who empathize with her).
Let us grant for the sake of argument that something like this is broadly plausible, that is, that Bob’s desire for new shoes can explain his reason to be sad but his sadness would still be had for the right kind of reason ‘that tragedy has befallen Alice’ and his sadness is (thereby) veridical. But despite these assumed virtues, the theory as a whole is still deeply implausible, and certainly not attractive enough to persuade us that internalism about reasons for emotion is more attractive than externalism about reasons for emotion. For the Schroederian approach to internalism about reasons for emotion fails to capture a plausible – I would say indisputable –intuitive data point. If Bob’s desire for Alice’s happiness cannot explain his reason for feeling sad when tragedy befalls her, but his desire for shoes can explain his reason for feeling sad, then something is deeply wrong with this theory. For if we know anything about this case we know that, if desires are relevant at all to an explanation of Bob’s normative reason for sadness, it is Bob’s desire for her happiness that is the relevant one. A theory that denies this is implausible, even if it checks other required boxes.Footnote 41
This brings us to strategy #5, about which I will say the least. The strategy comes to us from David Velleman. On his view, all actions have the (constitutive) aim of understanding oneself and beliefs have the (constitutive) aim of truth or knowledge; those aims ground practical and epistemic reasons, respectively.Footnote 42 If we are to extend this approach to reasons for emotion we will thus need to specify the (constitutive) aim of emotion. But what is the aim of emotion? Do emotions aim at veridicality in the way that beliefs aim at truth?Footnote 43 Do emotions aim at anything at all (other than their intentional objects, of course)? Velleman may very well have much to say in this regard, but it seems to me that skepticism about such a project is warranted (for now, anyway).Footnote 44
5. Conclusion
This paper began with a question about the rationality of motivation and I have argued that an externalist development of BARE entails that all rational agents are, ceteris paribus, required to be morally motivated. If externalism about reasons for emotion is true, then that is one way in which the authority of morality is vindicated. For if the authority of morality is its ability to generate normative reasons for all rational agents, then it does so: it generates normative reasons for moral emotions for all rational agents. I then considered, but found wanting, five strategies for an internalist development of BARE.
Those who wish to resist the conclusion that moral emotions are, at least in some cases, rationally required, have two options. First, they can deny there are reasons for emotion, although as indicated in Section 2, that seems like a significant bullet to bite. Second, they can develop a hitherto unarticulated internalist account of reasons for emotion, perhaps by extending or modifying extant arguments for internalism about practical or epistemic reasons in a way I do not see. A defense of internalism about reasons for emotion along any of these lines would be most welcome.Footnote 45
Disclosure statement
No potential conflict of interest was reported by the author.
