1. Introduction
There is clearly something appealing about the neo-sentimentalist idea that some evaluative concepts are affective in some way. It is very intuitive to think that some of our evaluative concepts are affective at base.Footnote 1 For instance, it seems that our concept of the disgusting is linked to the emotion of disgust and understanding that concept involves disgust in some way. This paper offers a way of thinking about evaluative concepts that vindicates the intuitive pull of the view.Footnote 2
This paper argues that if we read neo-sentimentalists as claiming that some evaluative concepts are identical to some affective concepts, then their view is unstable. I begin with two well-known problems that put neo-sentimentalism under tension. First, one natural formulation of neo-sentimentalism yields the wrong account of reasons for evaluative judgments, and secondly, the most popular reformulation that sidesteps this problem makes the view problematically circular. I then raise a new problem: if we accept that concepts are individuated at the level of Fregean sense, then it cannot be true that some evaluative concepts are identical to some affective concepts.
The paper shows how we can avoid this new problem by giving an account of evaluative concepts which is neo-sentimentalist in spirit, but which doesn’t make the identity claim. A theory of concepts like Christopher Peacocke’s, according to which they have multiple structural components, allows neo-sentimentalists to claim that the evaluative concepts they are targeting have affective components without embracing the identity claim. If we accept Peacocke's view of concepts, there are two ways of thinking of the evaluative concepts. I argue that one of these views is preferable, then show how it avoids the tension created by the two prominent objections to neo-sentimentalism. I end by considering the worry that avoiding these objections makes my view incompatible with anti-realism about value.
2. Neo-sentimentalism
Some thick evaluative concepts like disgustingFootnote 3 seem to have affective dimensions.Footnote 4 If we are interested in their natures, then a natural starting-point is a neo-sentimentalist view of evaluative concepts which is effectively the traditional sentimentalist view of properties, but applied to concepts. Traditional sentimentalists claim that certain evaluative properties, like disgustingness, are identical to certain affective properties. To be disgusting just is to elicit (enough) disgust or to merit disgust. Since neo-sentimentalists are explicitly making their claim at the level of concepts, not properties, we can read them as claiming that some evaluative concepts are the same as some affective concepts. A concept like disgusting just is the concept of meriting disgust in the same way that the concept bachelor just is the concept unmarried man of marriageable age. The two concepts are identical in being the same way of thinking about their targets – any thoughts involving one can be specified using the other without a change in meaning.Footnote 5
Since neo-sentimentalism is explicitly a view about our evaluative concepts, not evaluative properties (Tappolet Reference Tappolet and Bagnoli2011, 121), it characterizes our evaluative outlook rather than evaluative features of the world. It is often formulated in terms that suggest the identity view. Christine Tappolet (Reference Tappolet and Bagnoli2011, 117) says that:
Neo-sentimentalism is the view, roughly, that to judge that something has an evaluative property is to judge that some emotional response is fitting or appropriate with respect to it.
Here she is equating two judgments. Since judgments are constructed out of concepts (and not properties) it would be natural to think she is identifying the two concepts that seem to differ between the two judgments – the one about the evaluative property and the one about the appropriateness of the emotional response.
Similarly, D’Arms (Reference D’Arms2005, 3) says that:
to apply a response-dependent concept ϕ to an object X … is to think it appropriate … to feel an associated sentiment F towards X.
One might naturally think that applying the first concept is just the same thing as applying the second concept because they are the same concepts.
Let's call the view that the two concepts are identical identity neo-sentimentalism. Identity neo-sentimentalism is appealing, in part, because it captures the neo-sentimentalist intuition that our evaluative notions are dependent on features of our affective lives and that our evaluative outlook on the world would be different if we had radically different emotions. Some of our evaluative concepts only matter to us, and can only be grasped, because of our affective constitution.Footnote 6 For instance, if nobody was amused by anything, we would not know when amusement was merited and we would not need the concept funny. Our grasp of the evaluative features is fundamentally affective – we cannot understand the evaluative property in the way that others do without grasping the affective dimension. As Justin D’Arms and Jacobson (Reference D’Arms and Jacobson2000, 722) put it, ‘evaluation, and in particular moral evaluation, is somehow grounded in human sentiment.’ This would all be secured if the two concepts were identical.
Many variants of neo-sentimentalism also aim to reduce axiological normativity to deontic normativity. It can seem mysterious or spooky that objects can have properties that call for certain responses (Mackie Reference Mackie1977, 38). If we have a better handle on responses being called for, merited or fitting, then we can use this to explain the normative side of evaluative properties. If axiological properties just are deontic properties that govern the responses we ought to have, then there is nothing to explain in addition to the deontic notion governing the response we ought to have. Likewise, evaluative concepts seem precariously poised between the realm of the is and the realm of the ought. It seems weird that their primary use is attributive, yet their attribution entails ways we ought to respond. The identity view explains away the is-ness and shows that axiological concepts are deontic concepts in disguise, alleviating this weirdness. If the concept disgusting is nothing more than a concept concerning the appropriateness of disgust, then we can believe that Dave is disgusting without committing ourselves to the existence of any weird properties (we are only committed to there being appropriate and inappropriate emotions). This is akin to the way that believing that Dave is a bachelor commits us to nothing more than Dave being an unmarried man of marriageable age.
2.1. Internal tension
Neo-sentimentalism faces two well-discussed problems that, as Michael Brady (Reference Brady2008) points out, pull it in opposite directions. Dealing with one makes it harder for the view to deal with the other. This internal tension makes it difficult to formulate an interesting and stable view that vindicates our intuitions about the affective nature of some evaluative concepts by explaining them wholly in terms of affective-deontic concepts.
The first problem, called the wrong kinds of reason problem, immediately arises if someone claims, for example, that disgustingness is identical to the concept meriting disgust. The problem is that there is slippage between an object-directed emotion being merited and the object satisfying the corresponding concept. Something might merit an emotional response for all sorts of reasons, and some will be irrelevant to any axiological claims about it. If a demon will punish us for not admiring him, this is a reason to admire him, but not a feature that makes him admirable (Rabinowicz and Rønnow-Rasmussen Reference Rabinowicz and Rønnow-Rasmussen2004). It might be morally best, and best all things considered, not to be amused by an offensive joke, but this doesn’t make the joke any less funny (D’Arms and Jacobson Reference D’Arms and Jacobson2000).
If we take axiological normativity to reduce to deontic normativity, then we predict that some things are axiological reasons when they are not. If our concept of the disgusting just is the concept of meriting disgust, anything that is a reason to think that something merits disgust would be a reason to think that it is disgusting. But having a merely prudential reason to be disgusted by something is not a reason to think it is disgusting.
Advocates of neo-sentimentalism can avoid the wrong kinds of reason problem by reformulating their claims using a more specific deontic relation than meriting. The slippage between an emotion's normative status and the applicability of the concept is alleviated if the emotion's normative status depends on the property being instantiated. Casting the neo-sentimentalist claim in terms of fittingness closes the gap. Where something can be merited or not for extrinsic reasons, fittingness is matter of a match between the response and the object. Whether a response fits an object only depends upon the way that the response lines up with the object. As D’Arms and Jacobson put it, ‘To judge F fitting is to endorse the response in the relevant way, which constitutes taking the circumstances to be [as the response presents them]’ (Reference D’Arms and Jacobson2000, 746). Fittingness is something like correctness – emotions that fit their objects are correct responses to those objects and their correctness cannot be outweighed by extrinsic reasons not to feel them. Disgust might be a fitting response to bodily waste, even in cases where it is overwhelmingly counterproductive.
Formulating neo-sentimentalism in terms of fitting responses is somewhat circular though. The move from ‘meriting’ to ‘fitting’ avoids the wrong kind of reason problem precisely because something is a fitting target of an evaluative response only if it has the right evaluative property. The proposal dodges the irrelevant reasons by making the deontic dimension of the concept something like accurate emotional representation. As a result, understanding the emotional concept requires understanding the evaluative property. Understanding the fit of a representational state requires understanding what it represents, so we cannot have a handle on the fittingness of the emotion without a handle on the evaluative property itself.Footnote 7
Rather than recasting their view in terms of a different deontic relation, neo-sentimentalists can deny that the ‘wrong kinds of reasons’ for the relevant emotions contribute to an emotion being merited. However, combining Pamela Hieronymi's influential version of this proposal with identity neo-sentimentalism will result in circularity. If we understand ‘merit’ using her account of reasons, we get the view that an attitude is only merited if there are compelling reasons to be committed to the truth of that attitude's content (Hieronymi Reference Hieronymi2005, 444). This means that understanding the concept merited disgust requires understanding disgust's content, namely disgustingness. So grasping merited disgust requires some grasp of the property disgustingness, since the only thing that makes disgust merited are reasons to take its object to be disgusting – the circularity is back and it isn’t clear that concept is affective at base.Footnote 8
There are philosophically interesting views which accept this circularity (Wiggins Reference Wiggins and Wiggins1987), but their view of evaluative concepts is not affective at base. If the fittingness of an emotion amounts to the emotion correctly representing its object as having an evaluative property, then we cannot grasp the fittingness without grasping the property. The emotion doesn’t play a grounding role in our grasp of the evaluative concept and it is just one way of representing the evaluative property, among others.
2.2. Concept identity problem
In addition to the tension caused by the two problems discussed above, identity neo-sentimentalism face a further problem: concepts are too finely individuated for it to be literally true that disgusting is identical to fitting target of disgust.
Identity neo-sentimentalism depends for its plausibility on being a claim about concepts rather than properties. Traditional sentimentalists make a corresponding claim about the properties themselves and risk making the properties too subjective by making their existence depend on human reactions. By contrast, the way we think about evaluative properties is dependent on our reactions. So our evaluative concepts can depend on our affective orientations, even if evaluative properties are wholly objective.
Neo-sentimentalism also depends on concepts being more finely individuated than their referents. It is not plausible to think that we couldn’t have any concept of disgustingness without an affective life featuring disgust. Creatures who are affectively different to us might have different concepts of the same evaluative properties. Neo-sentimentalists who think that concepts are more finely individuated than properties can accommodate this because those creatures can have concepts of the same properties as us, without having the same concepts as us. the morning star and the evening star refer to Venus, yet someone who has one concept needn’t have the other. In the same way, our concept disgusting might refer to the same property as another creature's concept disgusting*, yet the two can be distinct concepts.
The problem I want to raise for identity neo-sentimentalism follows from the fact that it is a claim about concepts and the fact that concepts are more finely individuated than properties. Most views of concepts that individuate them more finely than their referentsFootnote 9 take them to be individuated at the level of Fregean sense.Footnote 10 They are modes of presentation of the objects or properties that they are about. If two modes of presentation are identical, then one cannot understand them and still be informed by statements that tell us that one is identical to the other. (We cannot find ‘Venus is Venus’ informative, though we can find ‘The Morning Star is Venus’ informative.) If someone can possess both concepts and yet have different attitudes towards them, then they are distinct concepts. Accepting this is accepting that concepts are individuated at the level of cognitive significance.
Concepts, in the sense I am using, can be different even when it would be irrational to deny that they corefer (Peacocke Reference Peacocke2008, 147). The concepts two and square root of four are distinct, even though reflection reveals that they must corefer. On the Fregean view, all concepts that aren’t lexically identical, obvious synonyms, or definitions that express what someone must grasp to employ the concept, are in fact distinct.
Identity neo-sentimentalists about disgusting claim that our concept of the disgusting is the same as our concept fitting target of disgust. To think something is disgusting is to think that disgust is a fitting response to it. However, even though both concepts apply to all and only the same objects, and refer to the same property, this doesn’t establish that they are identical.Footnote 11
If concepts are individuated at the level of Fregean sense, disgusting and fitting target of disgust differ, even though it is rationally required to think that they corefer. This is because it is possible to be competent with both concepts, yet doubt the truth of the identity claim. Someone might accept the biconditional x is disgusting iff disgust is a fitting response to x but then balk at accepting the identity claim, just as one could accept that a shape is triangular iff it is trilateral but not accept that trilaterality is triangularity. Identity neo-sentimentalism is not trivially true in a way that makes it obvious to anyone who possesses the concepts, so if concepts are individuated at the level of cognitive significance, it is false.
The concept bachelor is identical to the concept unmarried man of marriageable age because nobody grasping both could doubt that they are identical; nor could they have different attitudes towards two propositions differing only in that one employs bachelor and the other employs unmarried man of marriageable age. By contrast, it is possible to have different attitudes towards disgusting and fitting target of disgust and still grasp both concepts. Suppose Jane rejects sentimentalism because she doubts that things are disgusting because disgust is a fitting response to them. This does not make it true that Jane doubts that disgust is a fitting response to something because disgust is a fitting response to it. She has different attitudes towards two propositions, which differ only in that one employs disgusting and the other employs fitting target of disgust.Footnote 12 This shows that neo-sentimentalists should not hold the identity view.Footnote 13
So maybe neo-sentimentalists should only claim that concepts like disgusting are necessarily coreferential with concepts like fitting target of disgust? While this claim is much safer, it fails to capture the spirit of neo-sentimentalism, since the two concepts could necessarily corefer simply in virtue of facts about disgust rather than facts about our way of thinking about disgustingness. If disgust has evaluative content that attributes disgustingness to its objects, then its fittingness will depend on its objects being disgusting.Footnote 14 This guarantees that disgusting and fitting target of disgust necessarily corefer, but doesn’t entail neo-sentimentalism.
The concepts square and fitting target of square beliefs are necessarily coreferential, but this does not reveal anything about the concept square, certainly not that the concept is doxastic in the way that neo-sentimentalists think that some evaluative concepts are affective. Any concept is such that representations of things satisfying it are fitting iff the things satisfy the concept. The neo-sentimentalist claim is that there is something distinctive about the relevant evaluative concepts which makes their connection to emotions reveal an important part of the way that we think about things when we use them. Since representational theories of emotions suffice for the necessary coreference claim, without claiming anything about the natures of the evaluative concepts, neo-sentimentalism must claim more than necessary coreference.
Neo-sentimentalists could claim that the necessity of coreference reveals something about disgusting, not disgust, if they claim that the two concepts necessarily corefer because the property of disgustingness is response-dependent. On such a view disgustingness needs to be understood in terms of the emotion disgust. This would be reflected in the concept itself, so that there would be an affective dimension to understanding it. But this is a claim at the level of properties, not concepts, so it surely isn’t entailed by neo-sentimentalism. Instead, the neo-sentimentalist should claim that features of the concept disgustingness makes it the case that the two necessarily corefer. We will want to know which features. The next two sections try to answer such a question – they explore the options for giving emotions a role in evaluative concept reference determination without making the identity claim.
2.3. Peacocke
If we accept that concepts are individuated at the level of Fregean sense and still want a recognizably neo-sentimentalist view of evaluative concepts, we must show how emotions contribute to the natures of some evaluative concepts without committing ourselves to conceptual identity claims. Arming ourselves with a story about concepts that has some moving parts and finding a place for emotions within their structure offers a good way of doing this.
To that end, I will adopt Christopher Peacocke's Reference Peacocke1992 view of concepts. He advocates a conceptual role semantics that effectively gives a use-theory of concepts. According to Peacocke, concepts are individuated by the conditions a thinker must meet to count as possessing them.Footnote 15 These possession conditions are patterns in belief transitions that the thinker must find primitively compelling – that is, transitions in thought that seem appropriate to the thinker in a way that doesn’t rely on other commitments that the thinker has (Peacocke Reference Peacocke1992). This has a functionalist feel – concepts are constituted by their place in a set of input and output relations in rational thought.Footnote 16
Peacocke uses the concept conjunction to illustrate his claim. One must satisfy two requirements to possess our concept of conjunction. Firstly, when one believes p as well as believing q, one must find the transition to believing p and q primitively compelling. Secondly, when one believes p and q one must find the transition to believing that p (as well as to believing that q) primitively compelling.Footnote 17 Exhibiting this pattern is necessary and sufficient for possessing the concept conjunction.
Some concept possession conditions require transitions to or from non-doxastic states. According to Peacocke, possessing observational concepts requires a thinker to find it primitively compelling to transition from perceptual acquaintance with a specific observable feature to a belief that the observable feature is instantiated in the perceived environment. To possess the concept red someone must find it primitively compelling to go from reddish patches in their visual field to the belief that something they can see is red (and, plausibly, if they believe that something is red, they must find it primitively compelling to think that it is coloured, among other things).Footnote 18 As with all transitions mentioned in possession conditions, the compulsion is defeasible. Thinking lighting conditions are abnormal defeats the inclination to take reddish experiences to indicate the presence of red things. Even undefeated compulsions needn’t always compel us all the way to belief. If my dog suddenly looks brown, then my antecedent belief that he is grey can outweigh my inclination to believe otherwise, even if my visual experience goes undefeated.
Other concepts have non-doxastic output conditions. Possessing delicious might require being moved to eat anything we believe to be delicious (defeasibly moved, and only to some degree). Someone who is disinclined to eat things that they believe are delicious has a different concept to ours since it lacks the appetitive dimension.
It is worth emphasising that this is a use-theory of concepts. It explains concept-possession wholly in terms of patterns of concept-use and claims that a concept's possession conditions determine its nature (Peacocke Reference Peacocke1992, 5). Being reductive in this way makes it a suitable framework for a reformulation of neo-sentimentalism because it provides a way of giving explanatory priority to the emotions in an account of evaluative concepts. If we show that the natures of some evaluative concepts are determined by patterns in the way that emotions contribute to our thinking, we vindicate a neo-sentimentalist explanatory ambition. This allows a picture on which affective patterns together with the patterns specified in the rest of the possession conditions jointly suffice for the possession of some evaluative concepts. We do not, in addition, need thoughts about the reference of the concept themselves – we do not need a theoretical grasp of the evaluative properties. This captures something of the spirit of neo-sentimentalism because it shows that these concepts are affective at base – in the final analysis we understand disgustingness in terms of disgust.Footnote 19
Although thinkers meeting these possession conditions needn’t explicitly grasp them, this account of concept possession is quite demanding. Someone who has eccentric views, is fundamentally confused, or is ignorant about a topic probably won’t meet the possession conditions for the relevant concepts (since they will fail to find some necessary transitions compelling). As such, they will not fully grasp the concept. Yet we often still want to say that their thoughts employ those concepts, because it seems they are disagreeing with us about those very concepts. Peacocke (Reference Peacocke1992, 29) handles such incomplete grasp of concepts by saying that a thinker needn’t have full possession of a concept in order for their thoughts to employ it. I might think that I have arthritis in my thigh, using the concept arthritis, even though this thought displays a lack of full possession of arthritis (the fact that it refers to an ailment of the joints must be reflected in its possession conditions). According to Peacocke, my thought uses the concept arthritis because I am inclined to defer to experts about arthritis, so the patterns in the thinking of those I defer to reflect the fact that arthritis is necessarily a conditions of the joints.
Peacocke gives a theory of conceptual reference determination that relies on his theory of possession conditions, claiming that concepts have whatever referents vindicate the patterns of transition that feature in their possession conditions. Conjunction refers to the conjunction function because that is the function that makes true all the beliefs formed by undergoing these transitions. Both input and output conditions thus help to determine conceptual reference because both restrict which items could vindicate the pattern of transitions.Footnote 20
If there is a pattern of thinking associated with a purported concept which fails to single out a property or object, then it is a spurious concept (i.e. not a concept, just like faux-leather is not leather). A pattern can fail to single out an item by being either too restrictive (if the input and output conditions preclude anything vindicating the pattern), or too liberal (if the input and output conditions are satisfied by too many things). Any concept-individuating pattern is constituted by truth-preserving transitions. These transitions will have a normative force. When the input conditions obtain, this will be a reason to deploy the concept in the relevant way.
To see how the reference determination story works, consider François Schroeter's worry that traditional versions of neo-sentimentalism cannot achieve reference determination. If all we know about X is that it merits fear, this does not guarantee that X is dangerous rather than fearsome. When we want to work out if something is dangerous, we ask whether it causes harm, not whether it merits fear. Meriting fear isn’t criterial for danger (Schroeter Reference Schroeter2006, 355).Footnote 21 If Schroeter is right, and if there is nothing more to the concept danger than the idea of something meriting fear (i.e. if the evaluative concept just is the affective-deontic concept), then it cannot have a determinate reference. In Peacocke's terms that means it is not a concept at all, but a spurious concept. Schroeter doesn’t think that this rules out the neo-sentimentalist story about possession conditions for the concept – possessing danger may require thinking that fear is a fitting response to anything that is dangerous. Instead he thinks that we should accept that the possession conditions for a concept needn’t determine its reference. He claims that emotions play a role in the first but not the second.Footnote 22
However, we can agree that no single link to fear is sufficient to guarantee that the concept danger is about the corresponding property, and yet insist that emotions can contribute to determining the reference of the concept. Meriting an emotion doesn’t singlehandedly determine the concept's reference, but it might still be part of what determines reference. Suppose that the concept danger has an affective output condition, and we must believe that something merits fear whenever we believe it to be dangerous. This doesn’t determine whether the concept refers to danger or fearsomeness (since something being fearsome makes it merit fear). But this is not a problem. On the Peacockean picture, both directions of transition are always required to determine the reference of a concept.
The concept red does not have determinate reference thanks to either its input or output condition alone. All colour concepts require that someone possessing them are inclined to transition from believing that something has that colour to believing that it is coloured. Nevertheless, this helps determine the reference of the concept red because the concept refers to whatever vindicates the entire pattern of thinking given in the possession conditions.
If we move from identity neo-sentimentalism to a Peacockean version, we acquire more resources to achieve reference determination, namely the other possession conditions for the concept in question. It is a condition for the adequacy of a proposed input condition that it, together with the output condition, does determine the right reference.
In the next two sections I discuss two ways of exploiting this machinery to create the links between emotions and evaluative concepts necessary to capture the neo-sentimentalist intuitions. If we want to show that emotions play a reference-determining role in our evaluative concepts, and that the evaluative concepts are therefore fundamentally affective, we can do that either via the input or the output conditions for those concepts. In Section 3 I formulate and discuss a view that locates the emotions on the output end of evaluative concepts. This reformulation preserves much of the spirit of identity-based normative neo-sentimentalism. In Section 4 I formulate and discuss a view which locates the emotions on the input end of the concepts. This, it will emerge, captures some of the spirit of perceptual theories of the emotions, which stress analogies between emotions and perceptual experiences of value. I will argue that the latter does a better job of giving plausible possession conditions for evaluative concepts.
3. Output neo-sentimentalism
Given Peacocke's picture of concepts, we can secure a necessary connection between emotions and evaluative concepts if possessing those concepts requires output transitions involving emotions. Perhaps possessing the concept disgusting involves being compelled to feel disgusted by anything we encounter that we believe to be disgusting. (Suppose that the input conditions require finding it primitively compelling to move from thinking that something could contaminate one to thinking it is disgusting.Footnote 23) This is squarely in the spirit of traditional neo-sentimentalism since it uses the element of compulsion built into the possession conditions to express the neo-sentimentalist doctrine that evaluative properties are captured by rational pressure to undergo emotions. However, it risks the possession conditions being too hard to meet.
This is easiest to see if we look at a proposal that fails for another reason. Suppose that an output condition for the concept disgusting requires that any time we believe that something we encounter is disgusting we feel disgust, and the disgust seems right.Footnote 24 The problem with this proposal is that even those who clearly possess the concept aren’t always inclined to be disgusted when they encounter things they believe to be disgusting. They may occasionally be too happy to be at all moved by the thing, or too depressed to have any affective response to it.
Concept-possession should be fairly stable though, not coming into and out of existence with minor changes in the agent. Moods can totally remove the inclination to have emotional responses, not merely overpower or defeat them. If concept possession depended upon actual affective inclinations, our moods would change which concepts we possess.
Rational compulsions are more stable than affective inclinations. Many things that halt the psychological inclination to form a belief do not eradicate the underlying rational compulsion. A rational compulsion doesn’t go away when it is neglected, overpowered or defeated. Consider the concept disjunction. To possess it I must be rationally compelled to believe p v q any time I believe p. This task is endless, so I do not actually form many disjunctive beliefs. But even when I feel no actual disposition to form the belief, I’m still compelled to assent to p v q if asked, so I still have the relevant compulsion. When a rational compulsion is outweighed, the inclination to believe remains, it is merely opposed by another inclination. When a rational compulsion is defeated, the inclination is held in abeyance by the defeater.
A normative-affective output condition allows that the compulsion survives even when we are not inclined to feel an emotion. Suppose the possession conditions require that on encountering something we believe to be disgusting we must be compelled to find disgust merited (or fitting), whether or not we feel it.
However, it is hard to make sense of thinkers without reasonably sophisticated metareflective capacities meeting this condition in cases where they don’t feel the emotion. That makes it too demanding, since thinkers without the capacity to think of potential mental states as merited seem to still be able to have first-order evaluative beliefs. It seems like children can believe that food is disgusting without being able to think of disgust (or any other mental state) as being merited. The two thoughts seem to employ different cognitive abilities – one object-directed and the other metareflective. The metareflective thought seems to require a theory of mind, where the evaluative thought plausibly doesn’t. This view is supported by Shaun Nichols, who cites empirical evidence that children can typically make evaluative judgements and infer relevant things from those judgments years before they can make judgements about emotions being warranted, or even correctly anticipate the emotions others are likely to feel (Nichols Reference Nichols and Sinnott-Armstrong2008).Footnote 25
It isn’t clear that there is a weaker alternative for output neo-sentimentalists. If the requirement is that disgust seems Footnote 26 merited, even when we have no affective compulsion towards disgust, the output condition still demands that we sometimes find ourselves rationally compelled by a mental state that we aren’t experiencing. It is hard to see how someone could meet this condition without metareflective capacities. If we do not feel the emotion, how is it supposed to figure in our thinking clearly enough for it to seem merited, other than via metareflection?
Perhaps output neo-sentimentalists can insist that possession of disgusting does require certain metareflective thoughts. I have argued that someone can grasp the concept without having those metareflective thoughts, but this claim about concept possession might just borrow its appearance of plausibility from the claim that someone can have thoughts about disgustingness without any metareflective thoughts. Maybe a thinker without the right metareflective inclinations can have thoughts using disgusting because they meet the lower standards that Peacocke requires for having thoughts that employ concepts (i.e. the right deference to experts), not because they possess the concept.
I don’t think this diagnosis is correct. It needs to show that the children can have thoughts that employ the relevant concepts, but there are cases that show that putting metareflective demands in the possession conditions can rule out even this. This is easier to see in cases of eccentric understanding, but the point carries over to our case. Suppose Emily is a stoic of a particular sort. She doesn’t believe that emotions can be merited, and is never inclined to take an emotion to be merited.Footnote 27 But Emily is also a realist about evaluative properties, and thinks that disgustingness really exists. Her beliefs about what is disgusting are influenced by her feelings of disgust (even though she denies that disgust is ever merited), she believes that paradigmatically disgusting things are disgusting, she makes standard inferences from things being disgusting, and she avoids contact with them. I think we should grant that Emily is in full possession of the concept, Footnote 28 but the output neo-sentimentalist must deny that Emily even has thoughts involving the concept. Since Emily does not defer to others about the merit of disgust, she does not meet the lower standards required to have thoughts involving the concept. Denying that she has such thoughts seems untenable.Footnote 29
Since neither the occurrence of the emotion nor beliefs about the emotion's merit fit neatly in the output role, we should consider a way of tying emotions to evaluative concepts via the other half of a concept's possession conditions.
4. Input neo-sentimentalism
Input neo-sentimentalism is the view that evaluative concepts are individuated by input conditions that mention emotions. Emotional experiences are canonical inputs for certain evaluative concepts. Just as our concept red is partially individuated by reddish visual experiences inclining us to form red beliefs, some of our evaluative concepts are partially individuated by emotions inclining us to form evaluative beliefs. This proposal has something of the character of a perceptual theory of the emotions, but it is a theory of the concepts, not the emotions.
I will illustrate input neo-sentimentalism using the example we have been working with (disgusting), then briefly consider how far this might generalize to other evaluative concepts Section 4.1. I claim that one necessary condition for a thinker to count as possessing the concept disgusting is that when they experience disgust at x, they find it primitively compelling to believe that x is disgusting. There may be other input conditions: to possess the concept one might need to move from the belief that x is bodily waste to the belief that x is disgusting.Footnote 30 The output conditions for the concept will involve avoidance behaviour, and perhaps other beliefs (e.g. that the disgusting item might contaminate one). If any concept is dependent on our emotional dispositions for its nature, then disgusting is. Without some acquaintance with disgust it is hard to see why anyone would have the concept in questionFootnote 31 and disgust clearly plays a formative role in our normal way of grasping the concept disgusting. So it's at least intuitive that the concept is partially individuated in this way.
This proposal avoids the problems about metareflective capacities from Section 3. Satisfying the possession conditions for this concept does not require metareflection. I needn’t be able to have beliefs about disgust to experience it, nor for these experiences to make me believe that things are disgusting. Plausibly, one can only feel rational pressure to do something that one can conceptualize. By contrast, a nonconceptual experiential state can cause one to feel rational pressure to have some sort of response. Peacocke's own account of the concept red relies upon this sort of transition – the reddish experiences that form the inputs to our beliefs about redness are nonconceptual episodes. Their concept-individuating role only requires the capacity to respond differentially to them (by forming a belief), not conceptualize them. In this regard, the transition definitive of input neo-sentimentalism is unlike the affective transition suggested by output neo-sentimentalism.
The proposed possession conditions for disgusting constitute a recognizably neo-sentimentalist view. If what I have said is right, possessing the concept requires treating an emotion as a canonical input for evaluative beliefs; it is built into the concept that the emotion is an experiential mode of access to the property. Observational concepts (like red) are observational because possessing them requires sensory input transitions. The experiential pathway to red beliefs is a core transition required for possession of the concept, not merely one among indefinitely many ways of forming red beliefs, so having reddish experiences is part of our particular perspective on the colour red – the nature of the concept is partially constituted by such a transition. In the same way, if input neo-sentimentalism is right, evaluative concepts like disgusting are affective by virtue of having affective input transitions.Footnote 32 Moving from disgust to disgustingness beliefs is part of our particular perspective on the evaluative property so disgust necessarily plays an input role for our concept. Thinking in accordance with this input role, as well as with the other patterns required by the possession conditions, suffices for possession of our concept disgusting. This is part of what it is to have such a concept and so the concept is affective at base – constituted in part by affective transitions. This captures the spirit of neo-sentimentalism.
4.1. Scope
The sort of story I have told about disgusting is meant to apply to other evaluative concepts. The most obvious candidates are the sorts of thick evaluative concepts which Tappolet (Reference Tappolet and Bagnoli2011) thinks neo-sentimentalism characterizes: concepts like funny, boring, sexy, obnoxious and so on that are picked out by obviously affective terms. In many cases the story presented here will apply mutatis mutandis (amusement playing an input role for funny, boredom playing an input role for boring, and so on).
The proposal looks less compelling when it comes to thinner evaluative concepts. No particular emotion inclines us to believe that things are good, for instance. The neatest approach to accommodating concepts like good is to see approval as playing an input role for the concept, where approval is understood as a determinable affective reaction for which the familiar positive emotions are more determinate cases.
Even if this story cannot accommodate thin evaluative concepts, I do not think this damages the interest of the overall picture. I am attracted to Elizabeth Anderson's (Reference Anderson1993, Chapter 1) view that our evaluative experience of the world typically employs thick evaluative concepts. It rarely strikes us that someone is just bad, we normally find them boorish, or sleazy, or some such. Evaluations as good or bad are abstractions from these more specific evaluations. If this is right, the sorts of evaluative concepts that are subject to input neo-sentimentalist treatment have far-reaching effects on our evaluative stance.
Relatively objective evaluative concepts, like danger, might seem less plausibly affective since something can be dangerous without causing fear, and fear only contingently detects danger. We might think that given this one can have the concept danger without any particular emotional dispositions. But that is a non-sequitur. Concepts can have modal features that their referents lack. It is not in the nature of Venus that it is the first star to appear in the evening. Yet to possess the concept evening star we must find it compelling to think it is the first star to appear in the evening. In the same way, our concept danger could require that finding something frightening compels us to think it is dangerous, even though this is not a necessary part of the property. (Other creatures might have concepts of danger that are not affective at all.)
Finally, the plausible scope of the input neo-sentimentalist claim depends on the rational roles that emotions play. If Sabine Döring (Reference Döring and Goldie2010) is right that we (rationally) endorse the representational content of our emotions by default, then thinking in accordance with the input conditions like those proposed for disgusting is ubiquitous and many evaluative concepts are plausibly input neo-sentimentalist in shape. By contrast, if Brady (Reference Brady2013, Chapter 3) is right that we do not endorse our emotions’ content until we have found independent reasons to do so, then we rarely meet the input conditions I have proposed, so they are implausible.
I think Döring's view is plausible, and though I do not have space to convincingly argue for it, I take it to be the majority view.Footnote 33 At the very least, it should be uncontroversial that we are more inclined to believe that something is dangerous when we are afraid of it than when we aren’t, and not just because fear helps us find reasons to be afraid. Suppose I wake in the night, afraid of whatever caused a noise, and I can find no reasons for or against taking the noise to indicate any danger. I am at least more likely to believe I am in danger than I would be if I was woken by a noise and felt no fear.Footnote 34 Even if this is all that can be said for Döring's default mode, it might still suffice for input neo-sentimentalist accounts of danger.
5. The wrong kind of reason problem
We saw above (Section 2.1) that the wrong kind of reasons problem puts identity neo-sentimentalism under tension. Emotions can be merited in a variety of ways, so if we claim that evaluative concepts are identical to the concepts of merited emotions, then those reasons to feel the emotions which don’t bear on the evaluative features still end up being counted as reasons to believe evaluative claims. Identity neo-sentimentalists can claim that the affective concepts they target are concepts of fitting emotions, but this makes their account circular.
On my view, anyone with the concept disgusting who is disgusted by x will feel some rational pressure to believe that x is disgusting. My claim puts any old disgust on the input end (rather than fitting disgust) because adding a fittingness requirement would rob me of my advantage with respect to the metareflective worries (Section 3). We can be disgusted by something for the wrong reasons and if we possess the concept disgusting this will put us under some rational pressure to believe that the object is disgusting. So it looks like my view is subject to the problem. However, this is not the end of the story about the way that the concept applies pressure. If the concept has output conditions mentioning contamination, then believing that the object cannot contaminate us puts us under pressure not to believe that the object is disgusting. The wrong kind of reasons objection is particularly damaging to identity neo-sentimentalism because there is no such countervailing force. On their view, if something merits disgust that is the end of the story about its disgustingness. My view builds in potentially countervailing forces via the other possession conditions.
We might worry that although I have more resources to respond, they are just more resources of the same kind. If we can cook up cases where disgust is merited by something that isn’t disgusting, surely we can cook up cases where I’m disposed to make all the transitions that individuate the concept disgusting, but where we don’t think I have reason to take anything to be disgusting. If the only norms governing evaluative beliefs are compliance with the concept possession conditions, then a sufficiently baroque example should generate the wrong kinds of reason verdict.
However, the difference between my view and more traditional neo-sentimentalisms cuts deeper. Most neo-sentimentalists aim to get all of the axiological normativity out of the deontic norms for emotions and this makes it uncomfortable to embrace the circularity that comes from focusing on fittingness. That circularity becomes vicious when combined with the reductive agenda. By contrast, my account doesn’t have the same ambition. I don’t claim that the normativity of axiological beliefs involving concepts like disgust reduces to the rational pressure applied by the possession conditions for disgust. I rely on conceptual normativity to pick out the evaluative property that vindicates the pattern of thought.
I do not claim that an evaluative concept applies to an object because it is thought about in a way that satisfies the concept-individuating pattern of transitions. If we have misleading experiences, we can form nonveridical beliefs by deploying our concepts in their canonical way. I can believe that something is red, in response to a reddish experience, and still be wrong if the lighting is subtly distorting. First-order thought involving a concept isn’t made true by our practices. It is made true by relations between the referents of the thoughts. As such, the property itself does normative work. As well as vindicating the pattern of core transitions that individuates the concept, the property also stands in all sorts of evidential and normative relations not captured by the pattern. A door handle can be disgusting because it is slimy. Because someone is disgusting, I shouldn’t want to share a taxi with them. None of this follows from norms governing concepts, but from the relations that the property itself stands in with other features of the world.
6. Anti-realism
Many versions of sentimentalism have an anti-realist flavour, but we might worry that input neo-sentimentalism is incompatible with anti-realism about values. Conceptual reference is determined by whatever satisfies the pattern of transitions. What if nothing satisfies the pattern of transitions I have offered for disgustingness? If there is no property to refer to, the concept has no normative force. So far so good according to me. If our concept is spurious, it shouldn’t have normative force. If we cannot find a property that vindicates the transitions given for disgust's possession conditions then I may have given the wrong possession conditions, or perhaps the pattern of transitions is indefensible because nothing is disgusting. Neither would be fatal to the overall project – giving the shape of thick evaluative concepts.
This proposal is incompatible with neo-sentimentalist projectivism. It cannot be that our conceptual treatment of objects makes evaluative claims about them true. A more traditional neo-sentimentalism allows us to make sense of axiological normativity given deontic normativity, even if there are no real evaluative properties, as traditionally conceived. If an axiological claim is true just because a deontic claim is true, then we don’t need evaluative properties to make our evaluative claims true. By contrast, input neo-sentimentalism cannot avoid error theory in the face of value scepticism.
On this view, if one accepts scepticism about traditionally conceived evaluative properties, one can embrace error theory about evaluative claims or give an alternate account of the properties themselves. The latter option amounts to showing that some properties satisfy the patterns of transition that individuate our evaluative concepts. Perhaps the properties that vindicate these patterns are sufficiently unlike our traditional conception of evaluative properties to escape the scepticism.Footnote 35 When it comes to giving such an account of the properties, input neo-sentimentalism does not rule out many ambitious claims. For instance, it might turn out that input neo-sentimentalism is true because evaluative properties are response-dependent, like secondary qualities. This would help explain why a pattern of thought that involves moving from disgust to beliefs about disgustingness picks out the property disgustingness. My claims about evaluative concepts are compatible with any of these options.
7. Conclusion
I have proposed that emotions help to individuate some of our evaluative concepts. The emotions are canonical inputs to beliefs that attribute the corresponding evaluative properties. Possessing our very concepts requires finding these transitions compelling. My view is not quite neo-sentimentalism, but it explains why emotions are central to our ways of thinking evaluatively. It captures the intuitive appeal of neo-sentimentalism without running into some of the theoretical problems facing more ambitious views of evaluative concepts.
Funding
This work was supported by University of Geneva [grant number FNS 100012_150265/1].
Acknowledgement
I am immensely grateful for my time at Thumos and I received extremely valuable feedback from Julien Deonna and Fabrice Teroni on early drafts as well as from audiences at the Swiss Centre for the Affective Sciences, especially from Elena Cagnoli, Mary Carmen and Richard Dub, who also commented on drafts of the paper. Special thanks to Kim Frost who gave extremely helpful comments on another draft and to Laura Schroeter who gave very helpful advice.
