Some Of Our Books


« A worry regarding the necessity of defensive force | Main | Reminder: Ethics discussions at PEA soup - Jaworska & Tannenbaum on moral status »

February 14, 2014


Feed You can follow this conversation by subscribing to the comment feed for this post.

Thanks Justin and Daniel. Great post. I just had one question about the framework. Here seems to be the two theoretical alternatives:

1. Shadow-sceptics who think that the relevant properties are response-independent and that the relevant emotions at best track the response-independent properties.

2. Sentimentalists like you who think that the relevant properties are response-dependent properties and that the relevant emotions and especially their fittingness play a much more constitutive role in what the properties are.

Now, I've always loved the contrast between these positions. It also seems implausible or at least controversial that either the shadow sceptic or the sentimentalist would be right about all property-reaction pairs. At least we need some general standards for evaluating which one of the views is right in any given case.

You give two interesting standards: the sensitivity of the emotion and whether the property can be picked out without talking about the relevant property.

This second crux seems mistaken to me. The orthodox view (following Wright - page 120 onwards in Truth and Objectivity), Miller ch. 7 in Intro to Metaethics and others) that seems right to me is that the sentimentalist as a defender of response-dependence has to be able to describe the relevant emotion without mentioning the property the emotion is a response to. So, the relevant independence is that of the emotion and not that of the property, and task of the sentimentalist to argue for the independence.

In contrast, if we need to use the property to explain what the relevant reaction is, then we violate the independence constraint and the property in question is not response-dependent one. I actually think that this is a real challenge for the sentimentalist in the case of fear.

This way of looking at things seems to carve the joints at the right places at least in metaethics. McDowell's view where you can't describe the property or the emotion without the reference to the other more easily fits non-naturalist realist views than response-dependence accounts such as Lewis's (as Miller nicely explains). In contrast, your independence condition of the property would put McDowell in metaethics amongst the sentimentalists and defenders of response-dependence which seems wrong to me.

So, the question is, why understand the role of the independence constraint in the new way? And, do you think of your view just as McDowell's and then think that people have been wrong in thinking that his view isn't a sentimentalist response-dependence view?

Thanks, Jussi. We agree with you (and Wright) that "the sentimentalist as a defender of response-dependence has to be able to describe the relevant emotion without mentioning the property the emotion is a response to." So we think we have to shoulder the burden of giving an account of fear that does not appeal to danger, and similarly for the rest of the natural emotions.

Which is to say that we, unlike McDowell and Wiggins, do not hold a no-priority view of the relation between response and property -- at least, not when it comes to the natural emotions! (We do think this is right about some properties other than the sentimental values, but that's another story.) In our taxonomy these no-priority sensibility theorists still come out as sentimentalists, but that's a matter of semantics and perhaps not so interesting. We aspire to meet Wright's challenge.

What we haven't done here is tell you anything about how that story about the natural emotions will go. Maybe Justin will have something more to say about it! Or I'll try this afternoon when I have more time.

Yes, thanks Jussi. Just one more thing about independence. As Dan noted, we accept the (Wright/Miller/Suikkanen) independence burden on us, and that's why we're limiting these claims about sentimental values to the natural emotions, where we think we can meet it. But we also think there is a parallel burden on the shadow skeptical opponent, that we are trying to argue that he can't meet. That is, he has to be able to say what dangerousness is in a way that avoids appeal to fear. So the independence constraint we were adverting to for the shadow skeptic is the flip side of the one we accept as a constraint on us. The no-priority view would be the view that neither side can meet the relevant independence constraint, I guess.

Dan & Justin,

I'm just here to make my usual trouble, and if you'd prefer to focus on this newer stuff, please don't feel obligated to get into this. But I thought it might be worth rehashing some things I suspect you, Dan, have heard (probably from Christian) in case there are new things to say, or others want to weigh in.

I take it you think the fitting and meriting relations are normative, in that they provide normative reasons: We have reason to fear the dangerous, be amused by the humerous, etc. I'm skeptical of this. I like the general picture of sentiments as representative, and I think they play an important role for us, I just don't think their fittingess provides us with reasons to feel them. Here are three quick points in favor of my view:

Potential Redundancy
Suppose I'm in the woods and come upon a poisonous snake. I am capable of fully representing the snake's danger to me, and being suitably motivated, without feeling fear. I can't imagine why, in that case, I would still have reason to feel fear. This is because if I did have a reason to fear, it would stem from my needing to fear the snake in order to respond appropriately to it. But ex hypothesi, I don't.

(This might seem similar to Shadow Skepticism, but I take it that it's different given that we might well be able to represent even attitude-dependent values without deploying our sentiments.)

Lack of Motivation from the Analogy with Language
The above problem isn't surprising when you think of the sentiments as representative in something like the way words are. The word "dog" is fitting for my dog Silke. This does not provide me with any reason to think or say "dog" when I see her, unless that would help me achieve some goal. More generally, there is simply no reason to go around representing things.

It's More Plausible that Unfittingness Is Normative
While I don't have any reason to think or say "dog" when I see Silke, you might argue that I do have reason not to think or say "cat"—i.e., not to misrepresent. I'm not sure that's true, but it certainly sounds more plausible than the idea that I have reason to go around representing things.

Of course, you could claim that both fittingness and unfittingness are normative. But assuming each sentiment is either fitting or unfitting for each object—either the sentiment accurately represents the object or it doesn't—this amounts to ruling out error theory by fiat. For instance, every time I see an animal (say) I either have reason to fear it or reason to not fear it. The possibility that I have no reasons at all has vanished. I don't think we should be comfortable with that result.

Hi Guys, I am very sympathetic. Two thoughts meant in a constructive spirit.

(1) It seems to me that there are other cases that cast doubt on "the liable to affect self-interest" model. One that springs to mind: someone foresees that they will be harmed but thinks the harm will be wrongly inflicted on them and they think they have strong moral reasons to pursue the course that will bring on the harm. Martin Luther King speaks out, knowing he will be harmed by racists. Part of his courage was his not fearing their reprisals. But the ensuing harms might have been certain and quite seriously affected his self-interest. But maybe you guys think this is to moralistic and that it would have been fit for King to feel fear (although less morally admirable)? I suppose you then need to hold that it is not unfit for him to lack the fit response (e.g. that fitness is permission like).

(2) Maybe the threshold idea is too simple and we need a contextual one? If Jim is living high on the hog but finds out he is likely to suffer a non-trivial, but not super serious harm it is fit for him to fear it, but if Stan is in a bad environment and regularly suffering serious harms, it is not fit for him to fear a non-trivial but not serious one. Given his situation, the harm is less worth taking seriously, so it is at least less fit for fear. For example, on this view, people who are in war, for example, should not fear the kinds of small harms that they rightly fear in better times. Not sure this is right, but if it is, it might help you make your case?

To press the contextual line, I should have said that trivial harms might well be less fearsome in war (fit for *less* fear rather than always unfit for fear at all). It seems to me that such thinking about degrees of fearsomeness fits your model and that only an ad hoc move would allow the skeptic to mirror it. But that is just a hunch!

Ok one more thought....not sure about this one either, but here goes. It is common to fear a harm the first time but to have one's fear lessen when the same harm comes again and again. Sometimes this looks unfit/pathological (being beaten down by mistreatment) but perhaps in some cases we think it is fit/healthy. And it seems hard to see how the shadow skeptic will account for any fit cases let alone the difference between the pathological and healthy ones. It seems you would need a good case to drive this home (and that it related to the stuff about actual and perceived risk).

Great post. I am worried about the fitting attitude analysis of the dangerous as that which merits fear, because some things seem to merit fear which are not at all dangerous. For example, viewing the shower scene in 'Psycho', a magician doing a sawing the person in half type of trick, or going on a roller coaster ride. Whether things merit fear, to put it bluntly, seems to have more to do with how dangerous they seem to a fully informed and otherwise generally normal observer than with how dangerous they actually are. It would be no surprise, of course, to learn that 'seemingly dangerous' is a response dependent property. So you need to prove that your response-dependent analysis is of the underlying property of actual dangerousness, if you want to show there's a surprising and significant kind of response dependence here.

Hi David. Our inability to satisfy you previously does not imply that we don’t take the worry seriously. We’ve started thinking further about the issue recently, though, so we’ve got some new things to say. But I’m going to let Justin take the first swing at it, as it’s stuff that he is more on top of than I am. He should be dropping by here shortly.

Brad: Thanks, this will help us explicate our view, although you've anticipated some of our answers. Maybe I can say a bit more to make them plausible.

We do want to say what you expect about your MLK example. Surely fear is fitting in his circumstances (there were death threats, etc.) but, arguably, he was more virtuous for not being afraid — if indeed he wasn’t afraid. As a historical matter, I have no idea whether he was or not! He might have been afraid often, just not cowed into silence. One can act despite fear. Do you know more about the facts of the case, or might you be seduced into thinking that the very fact that MLK spoke shows that he did not fear reprisal? I would rather say that it shows only that he did not fear them so much as to render him incapable of acting virtuously.

It seems like what you’re worried about here is that you anticipate, correctly, that we want to say that his not being afraid — if indeed he wasn’t — is unfitting. That result doesn’t bother us, because we are prepared to say that in some cases the fitting feeling is contrary to virtue, and the virtuous feeling (or lack of feeling) might be unfitting.

We agree with your contextualist maneuver, though we left that out of our (already long) post as an added complication. If you’re living in suburbia, then a swerve that puts your car out of control for a moment merits (some) fear, whereas if you’re in a war zone, maneuvers of your vehicle that are (objectively speaking) more risky do not merit fear. We haven’t thought about this recently but that is what we’ve been inclined to say. That doesn’t seem problematic for us, though it’s an additional complication, but there might be a problem I haven’t thought of in the neighborhood.

This was a response to your first two posts, Brad. The third one brings up another topic that we've thought about -- and which we've discussed a bit with Antti Kauppinen, who has some very interesting sentimentalist thoughts of his own about it. We refer to this phenomenon as the instability of affect. More soon on that and further comments. Thinking at the speed of Soup -- Justin gets credit for that one -- proves difficult. But these are all great comments.

@Simon: There is a lot to respond to in that short post. (Thanks!) Let me start slowly. Those are interesting and problematic cases, which are well worth thinking about, but we might have different intuitions about them. You wrote,

...some things seem to merit fear which are not at all dangerous. For example, viewing the shower scene in 'Psycho', a magician doing a sawing the person in half type of trick, or going on a roller coaster ride. Whether things merit fear, to put it bluntly, seems to have more to do with how dangerous they seem to a fully informed and otherwise generally normal observer than with how dangerous they actually are.

As a general methodological point, I don't think we should construct our theory of fitting emotions around what (I think) are atypical cases. But that doesn't mean we shouldn't consider them. I wonder how much of your worry about our claim, and your positive proposal, hangs on those sort of cases. Here are some first thoughts about them, and then a request for clarification.

1) Psycho. Emotional response toward fictions are notoriously tricky; there’s a cottage industry devoted to them. Ken Walton was my dissertation advisor and is now my colleague, and he has convinced me that there is a great deal of pretense — at least — involved in these aesthetic response to (what are known to be) fictions. That said, there are sub-fear responses, such as anxiety and the startle reflex, which are certainly at play in the experience of watching that scene. (This is consistent with Ken's view, btw.) But I wouldn’t want to build my theory of when emotions are fitting around fictional cases, for several reasons.

2) The magician. Are we really afraid as audience members, do you think? That’s not the phenomenology I would expect. (Don’t we all know that the magician’s assistant isn’t really going to sawed in half? This seems the common thread here to these cases, but it would help me for you to draw out why you think they challenge our theory.)

3) Roller coaster rides. I love them! I go on them intentionally. I don’t think they merit fear, and I suspect that I would act differently if I did. I think it’s more dangerous driving to the amusement park. Now: do I feel fear when slowly riding up that first hill? I’m not sure about that (see the sub-fear response point), but suppose I grant that I, or most people, do. What is the problem for us in allowing this to be a case of predictable unfitting fear?

I’m not sure I understand your positive proposal. Is the suggestion that a fully informed, normal observer would hold that they seem dangerous, even though they really aren’t dangerous? I need a bit more help to understand this. What does “seem dangerous” mean other than: they are the sort of things that people are disposed to fear? (As noted, I think this is more plausible about the last case than the first two, but maybe getting away from the most controversial cases will help me think about your proposal.)

Hi David, Thanks for pressing this again, it is helpful. And we will be curious to see how you and others react on this point. Here are a couple of things we have wanted to say to support the claim that considerations of fit are reasons for emotions.

First, distinguish reasons to feel from practical reasons to do things intentionally so as to affect one’s feelings (for instance, by bringing feelings about, suppressing them, preserving them, etc.). We see this as analogous to a distinction between reasons to believe and practical reasons to affect one's beliefs intentionally. The claim that there are reasons to feel afraid of a poisonous snake in your path is not the claim that there are reasons to do anything to see to it that you fear the snake. Much less is it the claim that the balance of reasons favors seeing to it that you fear the snake if you can. It is rather the claim that there are reasons supporting fear, and indeed they constitute an adequate justification of a certain sort, even if you would be better off without fear. One of the ways in which human beings can be sensitive to reasons is by feeling in ways that are supported by these sorts of reasons, and this is a form of reason sensitivity we don’t have to do anything intentional to exercise--it's like sensitivity to reasons to believe in that respect. If you fear a mouse, in contrast, you do not have available a certain kind of defense of yourself that would have been available were it a snake you were afraid of.

Actually we think there are also some reasons to do things to bring about or preserve fitting feelings, but I’m saying that considerations of fit would still be reasons to feel even if they were not also practical reasons. Considered as practical reasons, we readily grant that reasons of fit are often outweighed. But we are inclined to embrace an ideal of how to be according to which there is always something to be said for registering affectively the things that matter. Consider the case of sadness over a loss. It’s painful, and that is a reason to take steps to avoid or suppress it if one can. But there is something to be said against that too. There is something good about experiencing the loss in that human way we have, by being sad about it. And that is a practical reason against suppressing sadness. So it seems like a practical reason in favor of sadness. But we don’t have much of an argument for this further claim about practical reasons supplied by considerations of fit, beyond the appeal to an ideal you may not find attractive.

I’ll try to speak more directly to your other points tomorrow.

Hi Dan, thanks. It's great to be able to meet virtually and have these conversations. Yes, I think 'seemingly dangerous' does mean roughly 'the sort of thing people are disposed to fear', which is why it will be no surprise that it's response dependent. Now should we understand 'merits fear' as (1) coextensive with 'the sort of thing that psychologically healthy, fully-informed people are disposed to fear', or as (2) coextensive with 'dangerous', or as (3) both of these? You are committed to either (2) or (3). The kind of examples I raised suggest that (1) and (2) come apart, so (3) can't work. And if my (I think) non-theory-driven intuitions about 'merits fear' are right, the correct answer must be (1). Other examples of the same kind: watching a trapeze artist perform over a safety net, the haunted house at the fairground. I could have used examples of things that are dangerous but seem not to merit fear, such as smoking cigarettes, or persistently failing to eat your 5 a day of fruit and veg. I suspect you'll find these less convincing. If you say all such cases where 2 and 3 seem to come apart are of predictable unfitting attitudes, it looks like you are in danger of begging the question about what 'merits fear' is coextensive with. I note that in your quick response about the roller coaster, you incautiously wrote 'I don’t think they merit fear, and I suspect that I would act differently if I did. I think it’s more dangerous driving to the amusement park.'. (Rather than the theory-neutral, and at least to me, much less plausible: 'I think driving to the amusement park merits more fear'). This suggests to me that your judgement about whether roller coasters merit fear might be a theory-driven one.

I mistakenly said near the end of my last comment 'If you say all such cases where 2 and 3 seem to come apart are of predictable unfitting attitudes, it looks like you are in danger of begging the question about what "merits fear" is coextensive with'. I should have said 'cases where (1) and (2) seem to come apart'. I neglected to address your previous response as well. You suggested a different way of dealing with at least some of these cases: they are not really cases of fear at all, but of something else, maybe the occurrence of 'sub-fear responses, such as anxiety and the startle reflex', or some kind of pretence (I'm not sure exactly what you have in mind here). A worry about your response: to capture the phenomenology of the cases accurately, we will have to characterize the relevant 'sub-fear responses' in at least partially sentimental terms. What we experience in the cases I have in mind often feels at least a lot like fear. But doesn't the proposed reply then lead to an unnecessary multiplication of entities? You would have to describe a 'sub-fear' sentiment, whose being merited would not seem to correspond to a property that we have any any particular reason to care about, and then 'fear' proper, whose being merited would correspond to the dangerous. Why not just take the simpler route and say that fear itself is felt in these cases, and that is is a sentiment whose being merited does not always correspond to the dangerous?
I'd say what's distinctive about the cases I have in mind is that either (i) they arouse fear, even when we know they are not really dangerous and our fear-system is functioning well, or in the opposite case (ii) they fail to arouse fear, even when we know they are really dangerous and our fear-system is functioning well. Why does this make them distinctive? Because what it is (or at least part of what it is) for us to fear something is for it to seem dangerous to us. What we believe to be the case and what seems to be the case usually lines up. But, as in the Muller-Lyer illusion (where the lines seem to be different lengths even when our visual system is functioning well and we know they are the same length), what seems to be the case can sometimes come apart from what we believe to be the case.

Hi Simon. Thanks for this, and nice to hear from you.

A central feature of your worry, if I am tracking, is that you think we are stretching the natural notion of merit to fit our theory when we say roller coasters don’t merit fear. You think there is a more intuitive notion of ‘merit’ on which fear is merited at things that people are naturally disposed to fear when their fear system is not malfunctioning, even when they know those things are not really dangerous. Now ‘merit’ is not an expression with a lot of ordinary language currency, at least as applied to emotions, so we thought it a term of art that we could appropriate. But we do want to be using it as our term for one natural and familiar way that people really do assess their emotions, and arguing that this assessment can explain certain kinds of evaluative thinking. So while we would not be bothered if there is another issue that has as good claim to being expressed with that term in English, we would be bothered if our claim that fear of rollercoasters is not merited (insofar as they are not dangerous) did not look like a familiar and natural thing one could say as a criticism of fear in such cases. But we think it is. We think it is a familiar question, not one we are inventing, that people ask when they wonder whether some normal emotional tendency is merited in a particular case or even in general.

Another normal emotional tendency is the tendency to be ashamed in the face of the contempt of others. And perhaps the shame system is functioning as natural selection built it to when one feels such shame. But there are cases where people can be ashamed in such circumstances despite knowing that the thing they are being contemned for is not shameful. (Say it’s a teenager being abused for being gay.) If so, there is a stance that they have toward their shame, according to which it lacks support by certain sorts of reasons. And that is the stance we have from outside as well (though we can realize sympathetically that it is quite understandable to feel that way). That is the kind of stance we call thinking the emotion unmerited. Surely we are not making up that phenomenon. If someone said that it seemed intuitive to them to say that the person’s shame was merited in such a case, I would wonder what they meant. If they grant that it makes sense to take the sort of stance toward that shame that I am talking about, but they want to say it is merited anyway because they think ‘merit’ means something like normal (it’s to be expected that people would feel it and without any malfunction in their shame system), this is a merely verbal issue over which we don’t want to fight. If they disagree that there are strong reasons against being ashamed of being gay, we have a normative dispute. If they claim not to understand the issue we are talking about, then I guess we have to keep trying to point to the phenomenon in other cases and persuade them this is a form of assessment they go in for as well.


Justin has spoken to what we take to be the leading thought in your comments. Let me try to clear up some things around the margins. (In a separate post, I’ll make a suggestion about another of your central thoughts, the cases where people commonly do not fear what they judge to be dangerous, such as smoking.)

First, about the cases. Since we grant that there are cases where people often do feel fear despite not believing themselves in danger, my worries about some of your examples are not so pertinent. That said, on the case of emotions with fictional objects and the point about pretense, I recommend Walton’s “Fearing Fictions” paper.

Second, also briefly, about what I called “sub-fear” responses. Anxiety is generally considered a mood rather than an emotion, insofar as it is object-less. And startle is a response, with a cause but not an object. We want to accept the common distinction between being anxious or startled and being afraid, but again this is more important for some of your specific examples than for your central point. We agree that there can be genuine fear without judgment of danger, and that this can be a common and even a normal response.

Finally, about words like danger, contamination, and incongruity. Those words offer the best ordinary-language “gloss” of what the emotions of fear, disgust, and amusement can be said to be about. They are rough-and-ready, and in some cases there isn’t a single word — e.g., we gloss shame and pride as concerning what reflects badly / well on you — but we think it inevitable that, in trying to makes sense of themselves and their emotions, people will reach for such descriptions. And that some are better than others.

The general point is that ordinary language is supple, and there are perfectly good uses of “dangerous” and “contaminated” (etc.) that are not response-dependent. We do not mean to be arguing against the semantic propriety of these uses, and we use the terms this way sometimes ourselves. What we want to insist upon is something like this (now moving to the case of disgust and contamination):

1. It is fine to say that the best way to gloss what disgust concerns, in English, is with “contaminated.”

2. There is (also) a sense of “contamination” on which something one might call the Germ Theory of Contamination is true. In the context of pathology, for instance, contamination means by germs (or perhaps germs or toxins or parasites).

Our commitment is not to denying either (1) or (2), exactly, but to denying their conjunction. That is, we deny that the Germ Theory (or even a germ, toxin, or parasite theory) of the disgusting is true. We hold that something can be disgusting without being germy (etc.) — for instance, because of how it looks. And we hold that something can fail to be disgusting (and hence not merit disgust) even though it is so contaminated. For instance, a glass of water into which a sterilized cockroach has been dipped is, arguably, disgusting. There are cases think far more decisive, but this is one that borrows from the literature.

While this puts us at odds with the most renowned scientists of disgust, such as Paul Rozin, we assert that there is no science of the disgusting. This is one of the central points of our paper, “Sentimentalism and Scientism,” which is linked in our original post.

This may display my unfittingness to do philosophy. But I am having trouble seeing the views as rivals, at least at the level of what they say about properties. Why can't both of the following be the case? (1)Some property (say danger) is what merits a certain response and this is why that property is the one that counts as danger. And (2)there be some other way to pick this property out that does not trade on its fear-response meriting role.

I'll apologize ahead of time if I don't come back to reply to any reply right away as I will only be on the computer a little bit this weekend.


Justin (and Dan),

It is rather the claim that there are reasons supporting fear, and indeed they constitute an adequate justification of a certain sort, even if you would be better off without fear.

Do you find the same story plausible about language? Suppose that, for some reason, I'm set up such that when I think the word "snake," I freeze up. I come upon a snake and, recognizing it, could just turn and run. But then I think "SNAKE!!!" Do you want to say that the combination of psychology (we just tend to think words of things we see, without intending to) and representational accuracy (if it was a mouse, I'd have made a mistake) similarly provides me with a kind of justification?

If yes (the cases seem the same to me, certainly), that just drives home for me the absence of normativity. I don't think I had any reason to think "snake" in that case. I just think there's a sense in which the naturalness and accuracy of my doing so inclines us away from characterizing me as making a mistake, and there's a temptation to express that in terms of a kind of justification. Nevertheless, where genuine normative reasons are concerned, I had none: Nothing that actually matters spoke in favor of my feeling fear or thinking "snake."

Hi Mark. As we understand the dialectic, the shadow skeptic wants to insist that it doesn’t much matter whether emotions are fitting; rather, people are interested in these other properties that have nothing to do with the emotions but are about risk and harm (or germs, etc.) So even if there is some other way of picking out those properties that does not trade on their emotion-meriting role, if what is of interest about such properties lies in their emotional-regulating role, then the shadow skeptic loses the argument.


We think that you have reason to believe that there is a snake in the path, even if it would be disastrous to have that belief as it happens.

There are vexed questions about the role of language in belief, but we certainly don’t think you have reason to use language in various other ways to represent the snake, simply in virtue of your evidence that it is there. No reason to write ‘snake’ or say it, for instance, even though those would be accurate representations. We are not sure what to say about whether you have reason to “think ‘snake’” because we are not sure what sort of activity that is--whether that is just a way of believing there’s a snake, or some sort of action, or what.

That's really helpful. So, is the story something like this? Sentiments aim at accurate representation of their objects just as beliefs aim at truth. There can be epistemic reasons to believe regardless of what practical reasons there are to believe or even to not believe—(roughly) reasons having to do with the truth of the thing believed. So, too, can there be fittingness reasons to feel regardless of what practical reasons there are to feel or even to not feel—(roughly) reasons having to do with the accuracy of representing the object as the sentiment does.

If that's right, I think I see and understand the view. I still want to resist it, but that's because I don't think epistemic reasons are normative reasons, either, except in the case where we have reasons to think about what's true. Similarly, I wouldn't think fittingness reasons would be normative except in cases where we have reasons to represent via the sentiments. In the snake case I describe, I don't think this condition is met.

This is the promised followup on Simon’s cases where people are typically (and normally) not afraid of what they judge dangerous — his example, which is also the one we’ve used in thinking about this issue, is smoking. Here are two possibilities:

1. Our initial gloss of fear was inadequate; it’s really about imminent dangers. Since smoking will harm you, if it does, only years down the road, it is not imminently dangerous and, hence, does not merit fear.

That accords with what some philosophers have wanted to say about fear, but right now Justin and I are inclined to think it’s a mistake. Here’s why:

Suppose that exposure to some poison will certainly kill you — there is no antidote — but not for x years. Suppose too that this poison has been put into your coffee mug by Dr. Moriarty. Thinking about your situation, it seems like I can be literally afraid on your behalf. (This is to suppose that I care so deeply about your fate that I am prone to fear for your life when it is in jeopardy.) Maybe you will have another sip and seal your fate. Or maybe you will leave the rest of the mug unfinished, and be safe. It seems like I can be afraid regardless of whether you will die instantaneously or in x years, and we see no reason to think it fitting only in the first case (as the imminence requirement implies).

2. We think there’s something else going on in the smoking case. In the first place, no single cigarette will kill you. One might imagine a case where either you will smoke that first cig, and then become addicted and later die; or else you won’t smoke it and will not die (as soon, of smoking related causes). Maybe this is more plausible if we change the case to smoking crack.

But even if it is true that, having taken that first hit of crack, you are doomed — still, your fate still goes through your own agency. People don’t always go down the slippery slope, and you could have stopped yourself later. So it still seems odd to think that the first hit killed you, non-imminently, on analogy with the dose of poison.

Hence we are inclined to think that the plausibility of this case (and most realistic ones) trades on a subtle slide surrounding “Cigarettes are dangerous to your health.” Yes, smoking (or cigs) are, as an activity or type of object, dangerous. But no particular cig, or act of smoking one, really is. Which is why we didn’t include imminence in our gloss of fear, originally.

What do you think?

Thanks Dan!

So the crucial bit is that the shadow skeptic thinks thatit doesn’t much matter whether emotions are fitting. OK, I see that one would not want to say that if one thought that the property was danger because it is fit to respond to it in the specified way. So is the issue of being able to specify that property in non-fitting attitude terms or not just evidence for or against this claim? If you can't specify in in non-fitting attitude terms the FA theorist wins, but they don't lose if you can specify it that way; it is a standoff.

Hi Justin and Dan,

I think I've heard this argument from Justin in seminar back in the day (and maybe had it from Dan in conversation). I'm in agreement with what you say. Here a quick comment and question:

(1) I wonder if cross-cultural work on what people take to be fearsome would buttress the point. Top reported fears in various areas will differ significantly, I suspect, making it even less likely that there's a common empirical property shared by all things thought to be fittingly feared.

(2) You're just giving examples of natural emotions, above, when you mention shame, amusement, pride, disgust, and fear. But you don't mention anger/resentment, even in passing in the post. That makes me wonder if the omission is intentional and for some reason you reject a sentimentalist account of anger/resentment and what I would say is it's corresponding value: wrongness/injustice.

Hi, Mark. I’m guessing that was intended for me, not Dan. (It would not be the first time someone swapped our names.) The important point for us is not about in principle specifiability in response-independent terms. We suspect that any accurate response independent specification of dangerousness (should one be forthcoming) will lack an interesting response-independent rationale for grouping together the things it does. And if that is right, it seems to us a good reason to suppose that what people care about in this vicinity is what it’s fitting to fear.

You seem to be suggesting that we are setting things up in such a way that FA can only win and cannot lose, but I don’t think that’s true. The shadow skeptic tries to characterize danger (contamination, incongruity, etc.) in response-independent terms, and suggests that he has captured what people care about in the vicinity of fear (disgust, amusement, and so forth). If so then the shadow skeptic wins and the sentimentalist loses. So I don’t think the fight is fixed. We need to make a case that people are really interested in what feelings are fitting, and that it is what fixes what things count as dangerous, insofar as danger is the important property it often seems to be. And we try to do that about danger here, and to make a more abstract case about the importance of questions of fit elsewhere in the book we’re working on.

Hi Zac. Thanks for the helpful comment.

Interesting point about disparity in elicitors. Our view about that is that although there is wide cross-cultural difference at the superficial level, there is often broader underlying agreement about the kinds of thing that merit emotions, and that these similarities can be captured by using the admittedly vague terms with which we describe the emotions’ concern. So although cultures (as well as individuals) differ about what counts as an insult, they seem to agree that insults merit anger. We find supporting evidence in the general accounts of the emotions that are given by Aristotle, e.g., which are quite familiar despite vast specific differences between antiquity and modernity.

We do think the diversity is more evidence for sentimentalism over shadow skepticism, for something like the reason you suggest.

As for anger and resentment, we didn’t bring them up in part because we’re unsure of exactly what we want to say about them. In particular, we are agnostic about the prospects for a sentimentalist theory of morality (along the lines suggested by Mill and Gibbard). We do think anger is a natural emotion, but we think it too broad to do the trick, and we suspect that resentment is not a natural emotion. This is more or less the view we put forward in “Expressivism, Morality, and the Emotions.”

Hi Dan,

First, I meant the contextual stuff I mentioned (and what you call the instability of affect) as points in favor of your views (over the shadow skeptics)

Second, I am not sure about the psychological claim about King, although I seem to remember it being brought up in James Cone's "Malcom, Martin, and America." However, I think the interview linked below is interesting in this context, because King explicitly links increased courage to stand up for one's dignity with a reduction of fear, and that is probably explained by the moralizing story towards which I gestured. You can skip to 4:22..King mentions the reduction of fear right off the bat - i.e. when he begins to discuss the nature of the "new negro".

Hi Justin,
Thanks for both responses and sorry about the name change. I actually wasn't meaning to suggest that you were biasing the setup. It could well be that some arguments are must wins for one side and not for the other. I'm really just trying to get a feel for the positions by asking questions.

Thanks, Dan and Justin, for your very thoughtful replies to my comments, most of which I find very persuasive (I apologize for my delay in replying as I'm rather busy at the moment). I think your reply to the cigarettes example is very compelling. I'm not entirely sure about about your claim that there's a natural way of evaluating emotions according to which, for example, roller coasters don't merit fear. To try to get a getter handle on this, I'd like to paraphrase a part of your post and ask what you'd like to say about the following claim:

To think something red is to think it fitting to see it redly. Equivalently, this is to think the object merits being seen redly.

Thanks for the really interesting post, Justin and Dan. I was away and just discovered a terrific and dauntingly long thread. Just a couple quick reactions.

I think you are right about the need to invoke the emotion of fear to determine precise thresholds for an evaluative property like danger. So I’d grant it is not possible to give a fully response-independent account of that property. However, this doesn’t mean a fitting-response account of danger is the correct one. It is one thing to say that a response plays a role in determining a threshold, it is quite another to say that the nature of the property can be simply cashed out in terms of the fittingness of that response. The core nature of tallness is a matter of height, but the threshold is determined by contextually salient comparison classes. Similarly, one could grant Gibbard that human emotional propensities like guilt and resentment play a role in determining thresholds for moral wrongness. But this seems consistent with cashing out the core nature of moral wrongness in terms of equal respect for persons, or ideal agreement, or whatever. I think this is a plausible position in the case of moral wrongness and something similar may hold for danger. Laura and I wrote a paper that makes this point a while back for the AAP – in case you are interested here is a link:

A small clarification about my position. I am not a teleofunctionalist like Jesse Prinz. I don’t think emotions are sensitive to evaluative properties – although they may track them in the weak co-variation sense, which is very cheap: emotions track a lot of things in this sense. You suggest that without supposing emotions are sensitive to value properties I would be committed to claiming that one should ignore emotional concerns, because these are matters of indifference. I don’t see why this follows. Even if they are not sensitive to values, emotions are reliably correlated with values and can play very important roles in our psychology. But this is a minor point.

Justin, I'm picking up on the exchange with Mark van Roojen. It sounds like you are staking out this position: any response-independent characterization of being fearsome (which I think we should distinguish from *the* fearsome - the set of things that are fearsome) fails to rationalize that way of grouping things. The grouping only makes sense with a response-dependent characterization of being fearsome (or the fearsome). This seems to put all the action on the epistemic side of things rather than the metaphysical side of things, as it were. I just want to see if that is right, and if that is where you want to draw the response-dependent, response-independent distinction.

I say this because one could have the view that being fearsome just is being some-particular-response-independent-property, but given the ways we have of thinking about that property, we cannot really understand the grouping save in response-dependent terms. (Perhaps some other sorts of cognizers could but we cannot.) Such a view is aptly characterized as response-independent so long as we are there labelling the metaphysical aspect of the view. But it is not the kind of view you are taking issue with. For you, this would count as a response-dependent view because we need to appeal to responses to make sense of the grouping. Is that right? (I have similar sorts of questions for sensibility theorists like McDowell and Wiggins, if that helps see what I'm getting at.)

Thanks very much for all these comments, everyone. I'm afraid I need to turn back into a pumpkin at least for the next few days, but I have really enjoyed the dance. Matt, I think what you say is right, and very helpful.
Francois, you may well be right that there is more to the concept of danger than just what merits fear, even if we are right that danger can't be adequately explained without appeal to what merits fear. I think that would be ok with us because compatible with the importance of the question what merits fear (i.e. the importance of fearsomeness), but we will have to think more about where that leaves things.
Simon, I think Dan may get back to you about your last post.
This has been extremely helpful, I really appreciate all the comments. Pea Soup and the Soupers are great!

Here are a few thoughts, sorry that they've been delayed.

@Brad: Thanks for the video, it's very interesting. And I did take your suggestion (about context) as helpful, though I'm not sure that I can defend the claim that the shadow skeptic cannot accommodate it.

@Matt and Mark: I think we (Justin & I) need to be more careful about what our claim is, with respect to properties and the property of being fearsome, as Matt helpfully puts it. It is not that we really care if the property can be specified, characterized, etc in some non-RD way, post hoc. It is rather that what is of interest about the property, or what makes it that property rather than another, has to do with something about human (emotional) responses. In this respect, the situation seems analogous (at least to me) to the situation with being red. (There is another respect, mentioned below in my response to Simon, which is disanalogous.) Even if a purely physical characterization of being red can be given, as I gather most philosophers think, that property only matters because of human color vision.

This sounds like what you say, Matt, about failing to rationalize that way of grouping things. Does that count as all on the epistemic side, as you see it? (I confess that my grasp of the distinction between M&E breaks down here.) I do think the view you describe counts as response-dependent, in the sense we mean to be getting at. Perhaps then we should be talking about what rationalizes the property of being fearsome, rather than about how to characterize that property? It would be very helpful to us to know how best to get at what we care about and not raise the issues we don't care about.

@Simon: You ask what we would say about this:

(*) To think something red is to think it fitting to see it redly. Equivalently, this is to think the object merits being seen redly.

I guess my answer is that I'm not sure why a simpler and more straightforward sort of dispositional account isn't adequate for redness, in terms of normal conditions and standard observers, rather than in terms of merit/fittingness.

If you have idiosyncratic color responses (that cannot be vindicated by something like Humean delicacy, as we can vindicate normal responses over those of the colorblind), there is no point insisting that the things normal people see as red aren't red. This is partly because there is no prospect of convincing them with reasons, and partly because the point of the concept RED is to allow (most) people to describe the world in convenient ways. Not to evaluate it.

But "The normal isn't normative" when it comes to value, I'm inclined to say. It makes good sense, and might be true, to say: Although most people respond differently, and I have no account of normality or the non-standardness of their conditions, I dispute their value judgments. Although most people are amused (or what have you), it isn't funny. Of course one then needs a theory of error to explain their responses. And we have the beginnings of such a theory, which we see as our contribution to sensibility theory. But that's another and longer story!

However, unlike some other sentimentalists, we agree with you (if I've got you right) that pervasive patterns of response offer strong evidence for the fittingness of those responses. We just think this evidence is defeasible, and that there are circumstances in which rational justification can unseat even very common tendencies by giving an unjustifying explanation of them. (If one looks back to common but now repudiated emotional dispositions, one can find examples of this phenomenon.)

But I fear that here I am drifting into obscurity. Or already out to sea in it.

@Francois: Thanks for your comments. Let's talk more about this! I just might be able to find time to get a response together here, but it'll probably have to be by email.

Hi Dan, thanks. What worries me about the answer you just gave (I thought you might give it) is that it seems we can get the same kind of criticism of 'seeing redly' going now as you were pressing in the case of fear. So, if a dispositional analysis of redness seems right, a dispositional analysis of fear ought to be equally plausible. More specifically, like dangerousness and what merits fear, redness and what merits being seen redly seem capable of coming apart. I'll avoid finkishness, though I'm afraid I'll have to resort to a science fictional example. Imagine a new type of computer display screen has been invented. Although it is the size and shape of an ordinary screen, it doesn't reflect or emit light (well, it does reflect light, but only in uniform dull metallic way). To display a coloured image to a user, it sends out some very precise magnetic pulses that directly produce electrical impulses in the optic nerve of the viewer. When it outputs a plain red computer image, then, it has the disposition to be appear redly (to normal observers in standard conditions). Intuitively, it merits being seen redly in one clear sense (the sense which picks out how an observer with a well-functioning visual system would see it). But intuitively, we don't believe it to be red. So in a different sense, because it isn't red, it doesn't merit being seen redly. If such an invention (or similar features of the natural world) existed, perhaps this would be a natural and familiar way to criticize our visual impressions. But none of this would show, as far as I can see, that the point of the concept RED is not to describe the world but to evaluate it.

Hi Daniel and Justin

sorry about not having had time to get back to you (it's been hectic) and thanks for the response. It seems like we agree. There's three views on the table:

1. sentimentalism. They satisfy the independence condition for the emotion but not for the property.

2. No priority view. They fail the independence condition for the property and for the emotion. This is a view different from sentimentalism - here there is no one-way explanation of the property in terms of the emotion.

3. Shadow scepticism (which I would call realism proper). They accept independence on both sides.

What I was objecting to in the post was that if the independence condition fails on the property side the consequence is that sentimentalism is true: "Without [property description] independence, the theory is sentimentalist". But now we seem to agree that this isn't quite right because the this particular independence condition is not satisfied by the no priority view and this is definitely not a sentimentalist view. I agree though that you can still make the case against the shadow sceptic in the same way.

The comments to this entry are closed.

Ethics at PEA Soup

PPE at PEA Soup

Like PEA Soup

Search PEA Soup


  • Unless otherwise indicated, the views expressed in any given post reflect the opinion of only that individual who posted the particular entry or comment.