1 Introduction

In this essay, I raise a puzzle concerning rational emotions: emotions that a rational agent tends to have or lack in response to her taking there to be normative reasons for or against them.Footnote 1 I’ll use a case of pity throughout. But we can, I believe, generate similar results for other emotions. Here is the case:

Horrible Birthday: Today is your birthday. Everyone forgets. Your boss makes you work overtime. When you finally arrive home, you realize you’ve left your house key in your office and the building is now locked. It’s now near midnight. As you wait for the locksmith in the rain, you reflect on your day. It is obvious that your birthday makes you pitiable. But you wonder just how much pity you should feel, because you know that your feeling pity with a certain intensity will also be part of your already miserable birthday.

Two features of this case are worth clarifying.Footnote 2 First, what you are wondering about is how much pity you should feel in response to your birthday as a whole.Footnote 3 This includes not only what you know has happened but also what you know will happen on that day.Footnote 4 Second, the subject of your deliberations—what specifically you are wondering about—is the precise intensity of the pity you should feel.Footnote 5 In other words, the question you are trying to settle does not concern whether you ought to feel pity at all, rather the focus concerns what intensity you ought to feel it with.Footnote 6

With these features of the case in mind, consider again the question: In Horrible Birthday, how intensely ought you to feel pity? The trouble in answering this question, which we are assuming you fully appreciate, is that feeling pity with a certain intensity is a response to how things go for you on your birthday, but your birthday is not yet over. How things go for you on your birthday will include the pity you feel. As a result, the pity you will feel with a certain intensity forms part of what the pity is directed toward.Footnote 7 And all of the following claims seem plausible:

Ought Depends on Fitting-Feelings: Whether you ought to feel pity about your birthday with intensity X near midnight, in part, depends on whether feeling pity with intensity X near midnight fits the extent to which you are pitiable on account of your birthday.

Fitting-Feelings Depend on Value: Whether feeling pity with intensity X near midnight fits the extent to which you are pitiable on account of your birthday, in part, depends on the badness of your birthday.

Value Depends on Feelings: The badness of your birthday, in part, depends on whether you feel pity with intensity X near midnight.

Ought Invariance: That you ought to feel pity with intensity X near midnight does not depend on whether or not you in fact feel pity with intensity X near midnight.

These claims form an inconsistent set. To see why, consider two ways your birthday might pan out assuming the truth of Ought Depends on Fitting-Feelings, Fitting-Feelings Depend on Value, and Value Depends on Feelings. Here is the first version:

More Intense: Things pick up where they left off in Horrible Birthday. You end up feeling a violent wave of pity.

Since this pity brings with it intense negative feelings, this makes your birthday worse than had you felt nothing instead. But let us assume that this increased badness makes your feeling pity perfectly fit the value of your birthday. And so it turns out that, in virtue of the fact that you felt pity with great intensity, you felt as you ought. Your violent wave of pity in More Intense vindicates itself. Now the second version:

Less Intense: Things pick up where they left off in Horrible Birthday. You end up feeling intense pity but not nearly as intense as in More Intense.

As before, this pity brings with it negative feelings that make your birthday worse than had you felt nothing instead. But, since these negative feelings are not as intense, they make your birthday worse to a lesser extent than in More Intense. And so, let us assume that, in virtue of the fact that you felt less intense pity which turned out to fit the lesser badness of your birthday, you felt as you ought. Your not-so-violent wave of pity in Less Intense vindicates itself.

Ought Depends on Fitting-Feelings, Fitting-Feelings Depend on Value, and Value Depends on Feelings thus tell us that in Horrible Birthday whether you ought to feel pity with a certain intensity depends on your feeling pity with just that intensity. In More Intense, it was only because you felt pity with great intensity that feeling pity with great intensity is how you ought to have felt. And, in Less Intense, it was only because you felt pity less intensely that feeling pity less intensely is how you ought to have felt. But these results conflict with Ought Invariance. For this principle tells us that the intensity of the pity you ought to feel does not depend on your feeling pity with just that intensity. It cannot be that part of what makes your emotional response the one you ought to have is the very fact that you end up having it.

Which of the four claims concerning Horrible Birthday made above is mistaken? I am unsure. I thus can be brief. What I hope to do in the remainder is make vivid that this is a genuine puzzle. There is much to be said in favor of each of these claims.

2 Ought Depends on Fitting-Feelings

Ought Depends on Fitting-Feelings is a specific application of the general claim that, for certain rational emotions, whether a subject ought to have this emotion depends on whether the constituent feelings are fitting. That is, the fittingness of feelings is one source of reasons for rational emotions. This claim is attractive. To see why, we can turn to Broad, whose account of the emotions makes the idea that we have reasons to feel with certain intensities particularly compelling.Footnote 8

Broad takes rational emotions to have a two-part structure. He writes:

Every emotion is an epistemologically objective or intentional experience, i.e., it is always a cognition, either veridical or wholly or partly delusive. But every emotion is something more than a mere cognition. An emotion is a cognition which has one or more of the specific forms of a certain generic kind of psychical quality which we will call emotional tone. (1954: 205)

Every rational emotion has a cognitive aspect which represents its object more or less accurately. Due to this aspect, we have reasons to think: reasons for our rational emotions that are related to the fact that the thoughts they implicate can be more or less accurate.Footnote 9 A fact is a normative reason for having a certain emotion if it serves as evidence for the truth of the proposition implicit in this emotion.Footnote 10 Pity, for example, tries to accurately represent its object as pitiable, and hence facts that make probable the accuracy of this representation are normative reasons to pity.

But, for many emotions, there is also a non-cognitive, affective aspect that is felt with a certain intensity. Presumably, fearing, desiring, respecting, admiring, hoping, pitying, being guilty, being angry, and the like involve having certain positive or negative feelings more or less intensely. And the intensity of these feelings can fit their objects more or less well. Here again is Broad:

As we have seen, there are two aspects to every emotion. In its cognitive aspect, it is directed towards a certain object, real or imaginary, which is cognised, correctly or incorrectly, as having certain qualities and standing in certain relationships. In its affective aspect, it has an emotional quality of a certain kind and of a certain degree of intensity. Now some kinds of emotional quality are fitting and others are unfitting to a given kind of epistemological object. […] A degree of fear which would be appropriate to what one took to be a mad bull would be inappropriate to what one took to be an angry cow. (1954: 209)

For an emotion to be rationally held, it is not enough for it to accurately represent its object. The constituent feelings must also fit this object. Imagine, for example, that you accurately represent your birthday as pitiable. Yet you feel nothing; your cognition lacks any accompanying emotional tone. Broad’s suggestion is that your response is unfitting. The intensity of your negative feelings ought to fit the extent to which your birthday is pitiable. But, if you feel nothing, it doesn’t. You have reason to feel more. As Svavarsdóttir argues, if we limit ourselves only to reasons to think we’ve missed something important. As she writes, “The crucial mistake of those who value an object of no value is hardly that of representing the object falsely. Rather, in some sense, they misplace or waste their emotional […] energies on the object in question” (2014: 101). So it seems we also can have reasons to feel: Reasons for our rational emotions that are related to the fact that the intensity of the feelings they involve can be more or less fitting.Footnote 11

If we accept that we have reasons to feel, then it is hard to resist the thought that such reasons are present in Horrible Birthday. Thus it seems that, if the foregoing is on the right track, whether you ought to feel pity toward your birthday with a certain intensity near midnight, in part, depends on whether feeling pity with that intensity near midnight fits the extent to which your birthday is pitiable.

3 Fitting-Feelings Depend on Value

Fitting-Feelings Depend on Value is a specific application of the general claim that, for certain rational emotions, whether a feeling fits its object in part depends on the goodness or badness of the emotion’s object. In line with the account suggested above, most recent theories of the rational emotions hold that such emotions represent their objects as having certain evaluative properties.Footnote 12 For example, pity takes its object to be pitiable, desire takes its object to be desirable, regret takes its object to be regrettable, envy takes its object to be enviable, fear takes its object to be fearsome, and so on.Footnote 13 And the evaluative properties the object possesses are salient for the fittingness of the feelings had toward it. As D’Arms and Jacobson write, “[O]ne can nevertheless urge that an emotional response is unfitting because it is an overreaction. Thus your envy might be too large for the circumstances, if what you have is almost as good as your rival’s. Then you would not be warranted in being much pained over such a trifling difference” (2000: 74).Footnote 14 It thus seems that what determines whether a feeling of a certain intensity is fitting or not is dictated by the evaluative properties of the object.

We can flesh out this idea in a relatively uncontroversial way. One common idea concerning feelings is that some are positive and others negative.Footnote 15 So one natural proposal is to link our positive feelings to goodness and our negative feelings to badness. We then claim that the intensity of these positive (or negative) feelings is modulated by the amount of goodness (or badness) of the object. As a rough approximation, we might hold that if the goodness of an object, O1, is greater than another object, O2, then, all else equal, the intensity of the positive feelings fitting in an emotion (e.g., awe) whose object is O1 is greater than the intensity of the positive feelings in same the emotion type (e.g., awe) whose object is O2. And similarly, if the badness of an object, O3, is greater than another object, O4, then, all else equal, the intensity of the negative feelings fitting in an emotion (e.g., pity) whose object is O3 is greater than the intensity of the negative feelings in the same emotion type (e.g., pity) whose object is O4.

This is very rough. It is silent on what triggers the all else equal clause. And, more importantly, it treats the fittingness of the feelings as mapping on to the goodness (or badness) of the object in a straightforward way. This appears to yield the correct result in cases like Horrible Birthday. The intensity of the negative feelings which are fitting for pity seem to track the badness of your birthday. Fear, regret, desire, respect, admiration and many other emotions seem to work like this. However, other emotions involve more complicated evaluations. For example, envy represents your rival as having some good that you lack. Thus, all else equal, the greater the evaluative gap the more intense the negative feelings that are fitting.

Still, glossing these complications, the basic contours of this proposal seem plausible. And so long as something in the vicinity is correct, whether feeling pity with a certain intensity near midnight fits the extent to which your birthday is pitiable, in part, depends on the badness of your birthday.

4 Value Depends on Feelings

Value Depends on Feelings is a specific application of the general claim that the goodness or badness of an emotion’s object sometimes in part depends on the subject’s feelings. Whether we feel good or bad matters. When you feel positively, this makes an evaluative difference. When you feel negatively, this makes an evaluative difference.Footnote 16

These claims are extremely weak. First, accepting such claims does not commit us to holding that only feelings make an evaluative difference. There can be other goods and bads. Second, accepting such claims does not commit us to holding that positive feelings always make a positive evaluative contribution and negative feelings always make a negative evaluative contribution. For example, we can, consistent with retributivism, still hold that when the vicious experience an episode of negative feelings this is good. Third, accepting such claims does not commit us to any particular view concerning evaluative magnitudes. For example, we are not wedded to the view that feeling positively with a certain intensity is strictly proportional to this episode’s evaluative contribution.

Given the claim’s modesty, it is very hard to deny that the badness of your birthday, in part, depends on whether and to what extent you feel pity on that day. Waiting for the locksmith in the clutches of the negative feelings involved in an intense bout of self-pity makes some difference to your birthday’s evaluative status.

5 Ought Invariance

Ought Invariance is a specific application of the general claim that deontic property possession does not depend on whether or not the possessor obtains. This general claim is known as the

Principle of Deontic Invariance: For any subject, S, and possessor of deontic properties, φ, if S’s φ-ing has the property of being what ought to obtain, then (i) if S were to φ, then S’s φ-ing would have the property of being what ought to obtain, and (ii) if S were not to φ, then S’s φ-ing would have the property of being what ought to obtain.Footnote 17

This principle is intuitive.Footnote 18 The deontic properties of an action, for example, are not counterfactually dependent on its performance. As Prichard writes, “[T]he existence of an obligation to do some action cannot possibly depend on actual performance of the action […] the obligation cannot itself be a property which the action would have, if it were done” (2002: 99; cf., Broome, 2004: 74). Similarly, that you ought to have an emotion with a certain intensity does not depend on whether you end up having that emotion with that intensity.

Beyond its intuitiveness, a powerful argument can be marshaled in favor of the Principle of Deontic Invariance.Footnote 19 First notice that, a theory that violates this principle is committed to the claim that

What’s Done is a Reason to Do: It is possible for the fact that S will φ to contribute to making it the case that S ought to φ.Footnote 20

But this claim is incompatible with a widely accepted idea concerning the intimate connection between reasons and deliberation, namely, the

Deliberative Constraint: For any agent, S, fact, F, and possessor of deontic properties, φ, if F is a reason for S to φ, then it would be intelligible for S to use F as a premise in deliberating about whether to φ, if S knew F.Footnote 21

Why accept this constraint? Here’s Shah:

[C]laiming that a consideration is a reason for an agent to φ implies that it is capable of being a reason for which the agent φs. Claiming that a consideration is a reason for which an agent φs in turn implies that the consideration guided the agent in its capacity as a reason. A consideration could not guide an agent to φ in its capacity as a reason unless the agent were capable of φing on the basis of his recognition of the consideration as a reason to φ. Deliberation, or reasoning, is the process in which agents recognize reasons, and then φ on the basis of this recognition. So something could not be a reason for an agent to φ unless it was capable of swaying him towards φing in his deliberation about whether to φ. (2006: 485-486)

This is a compelling line of thought.Footnote 22 It connects reasons with reasoning.Footnote 23 For a fact to be a reason for you to have an emotion with a certain intensity it needs to be the sort of thing that you, if you knew it, could use in your reasoning your way to that emotion with that intensity. If some fact, were it known by you, is barred from intelligible use in deliberation, then it isn’t a reason for you.

The Deliberative Constraint spells trouble for any theory that violates the Principle of Deontic Invariance. For such theories countenance the possibility that What’s Done is a Reason to Do. But the Deliberative Constraint rules out this possibility. If S knew she will φ, then the question as to whether or not to φ is closed for S. That is, once it’s known by S that she will φ deliberation concerning φ-ing is over. To put it metaphorically, you cannot stand at the crossroads and reason about what to do if you already know what you are going to do. As Carlson notes, “It is conceptually impossible to deliberate about what to do in a certain situation, if you already know what you will do in this situation” (1995: 100). The fact that S will φ, if known, cannot intelligibly function as a premise in S's deliberations concerning whether or not to φ. And hence the Deliberative Constraint tells us that this fact cannot be a reason.

We can now apply this idea to Horrible Birthday. Recall the case concludes with your wondering about the following question: How intensely ought I to feel pity in the moments before midnight? If the Deliberative Constraint is correct, then one fact that cannot determine the intensity of the pity you ought to feel is the fact that you will feel pity with a certain intensity in the moments before midnight. Why? Because for this fact to be part of what determines the intensity of the pity you ought to feel—i.e., for this fact to be a reason—it needs to be the sort of thing that, if you knew it, you could intelligibly use in settling your question. And the fact that you will feel pity with a certain intensity on your birthday cannot intelligibly be one of the facts that you use in settling your question. For if you knew the intensity with which you will feel pity, the issue is already settled. You don’t have the opportunity to use the fact about what you will feel as a premise in arriving at what you should feel, because once this fact is known by you there’s nothing to deliberate about.Footnote 24

Since you cannot reason with facts about the intensity with which you will feel pity, such facts cannot be reasons for you to feel pity with a certain intensity. The Deliberative Constraint delivers Ought Invariance.

6 Conclusion

We cannot jointly hold Ought Depends on Fitting-Feelings, Fitting-Feelings Depend on Value, Value Depends on Feelings, and Ought Invariance. Yet what I hope to have shown is that each of these claims is independently plausible. One must go. I am genuinely perplexed as to which it should be.