Blame as a sentiment

ABSTRACT

The nature of blame is not to be identified solely with a judgment, or an overt act, or an angry emotion. Instead, blame should be identified with a sentiment: more specifically, a multi-track disposition that manifests itself in various different emotions, thoughts or actions in a range of different circumstances. This paper aims to argue for these two claims. I start by arguing that blame is not solely a judgment, overt act, or an angry emotion. Then I develop the view that blame is a sentiment. In doing so, I also show how viewing blame as a sentiment avoids objections that justifies us in dismissing the previous accounts. In addition, I argue that it significantly affects other inquiries concerning blame. I end by answering a skeptical challenge that there cannot be an illuminating and unifying analysis of blame.

KEYWORDS:

• Blame
• sentiment
• emotion
• disposition
• moral responsibility

1. Introduction

The nature of blame is not to be identified solely with a judgment, or an overt act, or an angry emotion. Instead, blame should be identified as a sentiment: more specifically, a multi-track disposition that manifests itself in various different emotions, thoughts or actions in a range of different circumstances. This paper aims to argue for these two claims. I start by arguing that blame is not solely a judgment, overt act, or an angry emotion (Section 3). Then I develop the view that blame is a sentiment (Section 4). In doing so, I also show how viewing blame as a sentiment avoids the objections that justifies us in dismissing the previous accounts. In addition, argue that viewing blame as a sentiment significantly informs other inquiries concerning blame. I end by answering the skeptical claim that there cannot be an illuminating and unifying analysis of blame (Section 5).

In other words, I believe the problem with previous accounts is that they mistakenly identified the nature of blame with certain manifestations of blame. In that sense, they are over-simplistic, ignore the full complexity of the phenomenon of blaming.

2. Preliminaries

I do not search for a lexical definition of blame. That is, I do not search for how ordinary language or lay people define blame. We can get this by consulting a dictionary or conducting some empirical investigation. Rather, I search for a theory or an analysis of the concept ‘blame’.

A good analysis of blame has certain structural features and does not deviate too much from folk conceptions of blame.¹ To name some structural features, a good analysis is clear, informative, simple, useful in normative investigations and in understanding worldly phenomena. For reasons of transparency and to make it explicit, here are some vital intuitions we have about blame:

Blame is …

1. … about something, i.e. has intentionality,

2. connected to various different emotions, e.g. anger, disappointment, sadness, and guilt,

3. connected to motivations to act or actions, e.g. motivation to demand an apology or explanation, social distancing, and ostracism,

4. connected to thoughts or evaluations, e.g. the evaluation that a particular agent ought to apologize, and the thought that it is appropriate for others to blame a particular agent,

5. relatively long lasting, lasting for several minutes, hours, days or even years, usually until we feel proper amends have been made.

As is evident, it is not entirely unimportant to me how ordinary language and lay people define blame. Instead of letting our folk conceptions determine what blame is, I treat our folk conceptions about blame to count as prima facie evidence for or against certain analyses of blame. If an analysis of blame deviates too much from our folk conceptions about blame, that is a prima facie reason against the analysis (and vice-versa).

At first glance, my endeavor of searching for an analysis of blame may seem uncontroversial. Actually, it is not. There is a growing trend in current blame research which seem to recommend that we should not aim to give an analysis of blame. According to Nussbaum (2016), ‘blame’ is a too hollow concept to analyze. She writes:

In short, while it is very useful to distinguish these different cases, and while we surely learn a lot from the distinctions that these fine philosophers have introduced, human reactions come in many types, and the word “blame” is very imprecise. Maybe it’s not quite as duplicitous as “privacy”, which covers things that have no common thread at all. But it’s pretty empty and uninformative. (Nussbaum 2016, 260)

Fricker (2016) similarly argues that the practice of blame is too heterogenous for there to be an illuminating and unifying analysis of it. She writes:

Take blame. Let us assume that there is an analysis available. The point is we should not expect any such analysis to be very illuminating, owing to the fact that the practice of blame is significantly disunified, and is therefore likely to have distinctive or otherwise central features that may not be present in all instances. (Fricker 2016, 166) [author’s emphasis]

Taken together, these skeptical claims suggest our intuitions about blame are too heterogenous for there to be an illuminating and unifying analysis of it. Now, I believe that this form of skepticism is a last resort. So before accepting it, I want to make sure that no analysis of blame can fit the bill. In the end, I argue that this skeptical conclusion can be resisted.

3. Blame Is Not Solely a Judgment, an Overt Act, or to Be Angry

In the debate about the nature of blame, it is common to analyze blame in terms of one feature. Some identify blame with a judgment, others with an overt act, and some with an emotion, usually an angry emotion.² Below, I critically discuss each of these views in turn. I discuss these and no other views because I believe what I say about these reveal important objections we want a good analysis of blame to overcome. In addition, my aim is not primarily to provide a detailed refutation of all accounts there are on the nature of blame – that would be too lengthy an endeavor. Rather, my aim is to offer an alternative account of blame which I believe avoids important objections and makes better sense of our ordinary intuitions about the nature of blame.

According to what I call the judgment account, to blame is to make a certain judgment. What the crucial judgment consists of, varies from one author to the other. According to the most common suggestions, the particular judgment is, roughly, one of the following:

• the judgment that an agent has ‘stained’ her ‘record’ (about her lifetime moral worth) (Zimmerman 1988, 38), or

• the judgment that an agent has manifested insufficient goodwill (Hieronymi 2004).

Although the suggestions differ in their details, they all analyze blame in terms of a judgment. I will treat all suggestions under the same heading. This is because, as we see shortly, there is an important objection to all accounts that appeal solely to judgments, irrespective of the content of these judgments.

One virtue of the judgment account is that it is simple and neat. To blame is simply to judge that an agent is, more or less, blameworthy. Further, it captures the intuition that blame is about something and connected to thoughts or evaluations. Despite these virtues, I believe the judgment account fails. The main reason for this is that it overlooks the distinction between judging blameworthy and blaming.³ It explains what it is to judge someone blameworthy, not what it is to blame someone.⁴

Below, I provide a case in support of the phenomenon of judging an agent blameworthy without blaming her. If this case is possible, I believe we have a sufficient reason for distinguishing judging blameworthy from blaming.

Wallace (1994) writes one may:

believe that an especially charming colleague who has cheated and lied to you has done something morally wrong, insofar as he has violated a moral obligation not to cheat or lie for personal advantage, and yet you may have trouble working up any resentment or indignation about this case. In a situation of this sort it would perhaps be strange to say that you blame the colleague for what he has done. (Wallace 1994, 76)

According to Wallace, in some circumstances you are not prone to react in any way at all to an agent you judge is blameworthy, by for instance feeling resentment, disappointed or wanting to avoid or confront her. In those circumstances Wallace holds, and I concur, that it would be odd to claim that you blame the agent.

Granting that the above case shows that it is possible to judge someone blameworthy and at the same time deny that one blames her, we should distinguish between blaming and judging someone blameworthy.⁵ In order to arrive at an analysis of blame, we need to understand what it is to blame an agent that is distinct from solely judging her blameworthy.

Moving to what I call the act account. According to the act account, to blame is to perform some overt act, like booing, performing a speech act, or punishing someone. No one, to my knowledge, really defends this analysis of the nature of blame. Some blame scholars refer to it to make helpful comparisons. Scanlon (2008) locate for example his account of the nature of blame between the act account and the judgment account.⁶ In addition, some instrumentalists about blame seem to assume it when discussing the justification of blame, see for example (Arneson 2003; Milam 2021; Svedberg and Tännsjö 2017). Despite all this, I believe it is fruitful to discuss the act account as a potential analysis of blame. Not only because I believe that some version of the act account is what most people ordinarily have in mind, but also because it has a virtue which I believe an adequate analysis of blame should be able to account for. This virtue is that the act account acknowledges the fact that we often do something when we blame. We usually cut relations with those who have wronged us and protest wrongs made by politicians. Despite this virtue, I argue that the act account fails. This is because it overlooks the important phenomenon of private blame and seems ill-equipped to account for the intuition that blame is connected to various different emotions.

We can blame ‘privately,’ that is, without performing any overt act (Wallace 1994, 56; Sher 2005, 74). If that is true, blame cannot be identified solely with an overt act. Consider the following case:

Vaccination: Your friend’s friend refuses to take vaccination against a severe virus. You feel that it is not your place to publicly criticize or boo her for refusing to take the vaccination, so you do not do that. However, you still blame her.

In Vaccination, it is true that the agent blames her friend’s friend but false that she performs any overt act, like booing or punishing her. Rather, she keeps her blame to herself. Regardless of whether or not we believe the details of the case or whether or not we believe her omission is justified, in such a case it is right to say that you blame your friend’s friend. This shows that blame is not just the performance of an overt act.

In reply, one might revise the act account to include ‘private acts.’ Doing so, one might argue, avoids the above counterexample.⁷ Bracketing issues concerning how we should construe private acts and whether it even makes sense to speak of private acts of blaming, even if we avoid the counterexample by modifying the act account to include ‘private’ acts, I still believe it ultimately fails. This is because it seems ill-equipped to account for the intuition that blame is connected to various different emotions. It is not just the case that we often do something when we blame someone; we also often feel something.

The final account I wish to consider says that to blame others is to feel either resentment or indignation against them, and to blame oneself is to feel guilty.⁸ Most defenders of this account identify these affective responses more or less explicitly with Strawson’s reactive attitudes (Menges 2017; Wallace 1994, 2012). Although the accounts differ somewhat in their details, most of them hold that the particular affective response is either resentment, indignation or guilt and that, in the case we blame others, these affective responses motivate hostile behavior, like the motivation to punish or harm the blameworthy. I therefore lump the accounts together and call the account angry blame.

These authors have not said much about the nature of these angry affective phenomena. Menges (2017) claim that an agent may either have an affective episode of anger or guilt, or an affective single-tracked disposition to be angry with someone or feel guilty with oneself.⁹ In addition, he (ibid) also holds that angry emotions need not be identified with a judgment.¹⁰

One virtue of angry blame is that it accounts for the observation that when we blame others, we are usually angry with them, and when we blame ourselves, we usually feel guilty.¹¹ Furthermore, it can account for the observations that blame is about something, as emotions are generally thought to be about something (Deonna and Teroni 2011). In addition, the observation that blame is connected with actions because emotions often motivate us to act (ibid). Despite these virtues, I argue that angry blame is too narrow and that a modified version of it fails to provide a good analysis of blame.

Sher (2005) does not believe that other-blame is always something angry or hostile. He writes:

We may, for example, feel no hostility toward the loved one whom we blame for failing to tell a sensitive acquaintance a hard truth, the criminal whom we blame for a burglary we read about in the newspaper, or the historical figure whom we blame for the misdeeds he performed long ago. (Sher 2005, 88)

I share Sher’s claim that other-blame is not always something hostile. If we reflect on our own lived experiences, we soon find that it supports Sher’s claim. I believe our own lived experiences support another claim as well: namely, that blame – both other-blame and self-blame – is connected to various different emotions, not just for example hostile emotions in the case of other-blame. A view of the nature of blame strictly in terms of angry emotions or guilt therefore seems too narrow.

Instead of rejecting the fact that various different emotions can be instances of blaming, defenders of angry blame can modify their view to include all emotions we believe can be instances of blaming. Such a view could look like follows: to blame is to feel guilt, or shame, or resentment, or indignation, or disappointment, or hurt feelings, or disgust, or contempt, or sadness, or any other emotion in the vicinity.

Even though the modified angry blame account seems better in terms of co-extensionality than the non-modified angry blame account, it fails on other grounds. The modified account provides us with a disjunction consisting of different emotions. These emotions differ in, among other things, their phenomenology and which kind of behavior they tend to prompt: resentment generally attack tendencies, guilt usually reparation tendencies, and disappointment generally withdrawal tendencies (Prinz 2009). Not only are they different, we can experience the emotions included in the disjunction without blaming someone. For example, I can feel disgusted with my food. Consequently, we want an explanation of when an instance of these emotions counts as an instance of blaming and, more generally, what ties instances of all these different emotions together to count as instances of blaming. We can call this worry the unity worry.¹²

In the literature on the nature of blame, there is no clear or generally accepted answer to this worry. Below, I elaborate on two possible answers and argue that none of them is successful (at least not yet).

One way of unifying the emotions is by arguing that they have the same behavioral upshot. Drawing inspiration from Rosen (2015), we can argue that instances of affective episodes or dispositions that prompt sanctioning behaviors are instances of blaming.¹³ Drawing inspiration from Skorupski (2010), on the other hand, we can argue that instances of affective episodes or dispositions that prompt withdrawal of recognition, social distancing, ostracism, are instances of blaming.

Generally, angry emotions tend to prompt sanctioning behavior, not withdrawal tendencies. Further, disappointment, guilt, shame and hurt feelings are generally thought to not prompt sanctioning behavior, but rather withdrawal or social distancing. Accepting the Rosen or Skorupski inspired way of unifying the blame emotions would therefore probably exclude some of the emotions our ordinary thoughts suggest can be instances of blaming. For that reason, these ways of uniting the emotions we believe can be instances of blaming seem to fail.

Another way to unite the emotions is by claiming they have the same content. There are no clear suggestions about what that content should be. To elaborate on one suggestion. If the content is ‘x is blameworthy’, that poses theoretical problems for some who want to analyze blameworthiness in terms of fitting blame.¹⁴ More precisely, the analysis would turn out circular if we analyzed blameworthiness by reference to an emotion with the content ‘x is blameworthy’.¹⁵ In addition, accepting this suggestion risks over-intellectualizing blame, excluding people whom we think can blame from being able to blame. For example, young children, as it is questionable whether they have the concept of blameworthiness.

Thus, it seems that we have ended up in a dead end. If we accept the Rosen or Skorupski-inspired way of unifying the blame emotions, we fail to unite all the emotions we believe can be instances of blaming. The other way is not promising either, at least not yet. If we accept the modified version of angry blame without answering the unity worry in a plausible way, we end up lacking a neat and informative analysis of blame, i.e. a good analysis of blame.

Moving to a final worry with both angry blame and the modified angry blame account, it is questionable whether it even makes sense to view blame as an affective episode. Intuitively, blame is relatively long-lasting. Blame typically lasts for several minutes, hours, days, or even years. And we do not normally say that we experience a ‘pang’ or ‘episode’ of blame, while we commonly say that we experience a ‘pang’ or ‘episode’ of anger or fear. Finally, we do not stop to blame someone after having experienced a bout of anger or disappointment in, for instance, the presence of the blameworthy agent – we continue to blame her after that experience, usually until we feel proper amends have been made. This is not a knock-down objection to angry blame. Some defenders of this view claim, remember, that we may have a disposition to be angry, and dispositions are usually taken to be relatively long-lasting and exist in the agent after its manifestation has disappeared.

In sum: to blame is not solely to make a certain judgment. To blame is not solely to perform some overt act either. And, to blame is not solely to just feel angry or guilt.¹⁶ Next, I argue that blame is a sentiment. More precisely, blame is a multi-track disposition that manifests itself in various different emotions, thoughts or actions in a range of different circumstances. Put in other words, I believe the problem with previous accounts is that they mistakenly identified the nature of blame with certain manifestations of blame. This failure is similar to identifying love with certain manifestations of love. To love is not solely to give roses or to only feel happy in the presence of the beloved: love is a sentiment, that is, it is a multi-track disposition that in a range of different circumstances manifests itself in various different emotions, thoughts or actions.

4. Blame as a Sentiment

In this section, I (i) argue that blame is a type of sentiment.¹⁷ More precisely, blame is a multi-track disposition that manifests itself in various different emotions, thoughts or actions in a range of different circumstances. In this sense, it can be viewed as a hybrid account: an account of blame that combines the judgment account, act account and modified angry blame account. In addition, I (ii) argue that by viewing blame as a sentiment, we avoid the problems earlier raised to the other accounts of blame. Finally, I (iii) show that viewing blame as sentiment significantly affects other inquiries concerning blame.

I take a sentiment to be a multi-track disposition (Ben-Ze’ev 2000; Deonna and Teroni 2009; Vendrell Ferran forthcoming). Below, I say more about the nature of dispositions. It is a disposition that in a range of different circumstances manifest itself in various different emotions, thoughts or actions. Depending on the sentiment we are concerned with, the manifestations and triggering conditions will differ. Paradigmatic examples of sentiments include ‘love’, ‘hate’, ‘like’, ‘dislike’, and ‘care’. To illustrate how sentiments manifest themselves in various different ways, consider what Prinz (2004) writes about the sentiments ‘like’ and ‘dislike’:

If you like someone, then you experience joy in her presence. But you may also experience amusement when she makes a joke, excitement when anticipating your next encounter, sadness when you are apart, distress when she is harmed, and so forth. If you dislike someone, you may experience anger, disgust, or contempt in her presence. You may even experience Schadenfreude when she falls victim to misfortune. (Prinz 2004, 189)

It makes perfect sense to say that when you blame someone, you are in a range of different circumstances disposed to experience various different emotions, thoughts or actions. Characteristic emotions, actions and thoughts include, but are not limited to, the following:

Characteristic emotions: resentment, indignation, guilt, shame, disappointment, contempt, sadness, hurt feelings, disgust.

Characteristic actions: demand an excuse or explanation, withdraw, ostracism.

Characteristic thoughts: thinking that it is fitting for others to blame her, that she ought to take responsibility for her action.

The blame sentiment is usually triggered when the agent is confronted with the particular blameworthy agent and her action in some way, such as when the agent is contemplating, hearing about, being reminded about, being confronted by, seeing, or perceiving the particular blameworthy agent and her action. To illustrate all this, you might experience resentment when thinking about the fact that your partner has cheated, disgust when seeing that she is unaffected by the fact she cheated, sadness when thinking about the fact that your relationship with her has changed due to the fact that she cheated, think that she owes you an excuse when contemplating the fact that she cheated, or be disposed to demand an excuse from her when discussing the fact that she cheated with her.

Importantly, sentiments are taken to remain in the agent after its manifestation have disappeared. This is a difference between sentiments and affective episodes. For example, I still love my partner after having experienced an affective episode of joy in, say, her presence. As hinted at in the previous section, I still blame someone after having experienced an episode of resentment when being reminded about the blameworthy agent’s action.

Further, sentiments are commonly taken to have the property of intentionality: they are about something, usually agents or specific objects (Naar 2018). I love my cat, care for my child, and dislike a particular politician, for example. Like other sentiments, blame is also about something.

Finally, sentiments, in contrast to affective episodes, are relatively long-lasting (Ben-Ze’ev 2000; Deonna and Teroni 2011; Naar 2018; Tappolet 2016). By ‘relatively long-lasting’ I mean lasting for several minutes, hours, days, weeks, or even years. Blame is commonly taken to last for a relatively long period of time: several minutes, hours, days, weeks and even years. Usually until we feel proper amends have been made.

Now we can see how viewing blame as a species of sentiment avoids the objections discussed in the previous section.

The problem for the judgment account is that it fails to capture that we react in certain ways when we blame someone – we do not just judge someone blameworthy. Viewing blame as a sentiment avoids this problem, as blame on this account is manifested in various different reactions. The problem for the act account is that blame is connected to overt acts, but cannot be identified with only an overt act. We can blame ‘privately.’ On this account, to blame is to be disposed to perform, among other things, overt acts. When the blame disposition is not manifested or manifested in something other than overt acts, e.g. an emotion, we can say that you blame privately. It also avoids the too narrow worry presented to angry blame in virtue of identifying blame as a disposition that is manifested in a plurality of emotions. How about the unity worry? How do we explain whether an instance of disappointment is an instance of blaming? And how do we explain what ties instances of all the characteristic emotions, thoughts and actions together to constitute instances of blaming? I believe defenders of the sentiment account can answer this worry.¹⁸ More precisely, I believe defenders of the sentiment account can answer this worry differently depending on, among other things, which view of the nature of dispositions they subscribe to. Below, I illustrate how a realist about emotional dispositions might go about answering this worry. In order to understand that reply better, it is helpful to first say what I take realism about emotional dispositions to be.¹⁹

According to Naar (2013, 2018), emotional dispositions like sentiments are actual properties of persons. These properties are connected to or directed towards their manifestations but are not to be identified with them. They are distinct. Consequently, a number of disassociations are possible between them. An agent might in suitable circumstances behave in a way that is characteristic of a person who blames someone without having the disposition which we call blame. Reversely, an agent might have the disposition which we call blame without in suitable circumstances behaving in a way that is characteristic of a person who blames someone. Further, these properties are taken to play a causal role in producing the manifestations they are connected.²⁰

Granting such a rough realist conception of dispositions, we can answer the unity worry. We can say that an instance of disappointment is an instance of blaming when it is a manifestation of the blame sentiment, that is, produced by the disposition which we call blame. Likewise, what ties instances of all the characteristic emotions, thoughts and actions together to constitute instances of blaming is that they all are manifestations of the blame sentiment, that is, produced by the disposition which we call blame.²¹ Related to this, it seems that some of blame’s characteristic manifestations are also characteristic manifestations of other sentiments. For example, disgust might manifest hate, not just blame. How do we decide whether an instance of disgust is an instance of blaming or hating? The answer to this worry is similar to the answer I just gave above: disgust is an instance of blaming when it is a manifestation of the blame sentiment, that is, produced by the disposition which we call blame, and an instance of hating when it is a manifestation of the hate sentiment, that is, produced by the disposition which we call hate.²²

Over and above the virtues mentioned above, the way I view blame as a sentiment is not committed to the claim that blame always manifest itself in aggressive or hostile behavior. This is an important upshot. If blame is not always something hostile, it is not equally evident, as some philosophers argue, that it is morally bad or that we ought to discourage it (Nussbaum 2016; Pereboom 2009). Thus, accepting my view of blame is likely to significantly affect other inquiries concerning blame.

5. Conclusion

Recall the skeptical challenge presented at the beginning of this paper: our intuitions about blame are too heterogenous for there to be an illuminating and unifying analysis of it. Is analyzing blame as a sentiment an illuminating and unifying analysis? I believe that it is. As showed in the previous section, it unites all our intuitions about blame to a single analysis. On the account I develop, blame has intentionality, is connected to various thoughts, actions and emotions because blame manifests itself in various different thoughts, actions and emotions, and relatively long-lasting.

Regarding whether the account is informative or illuminating, we do not need to know all the details in order to have an illuminating or informative analysis of blame. We only need to know enough. I believe the account of blame I have developed in the previous section is sufficiently illuminating and informative. It tells us what blame is and explains how we, by viewing blame as a sentiment, avoid problems associated with other popular accounts of blame. What more information do we as philosophers need in order to have an illuminating analysis of blame? To summarize and conclude, the nature of blame is not to be identified solely with a certain judgment, some overt act, or an angry emotion. Instead, blame is to be identified with a sentiment.

Acknowledgments

For helpful feedback on earlier drafts, I would like to thank in particular David Alm, Julien Deonna, Björn Petersson, András Szigeti, Matthew Talbert, Fabrice Teroni, Jakob Werkmäster, the anonymous reviewer, participants at the Higher Seminars at Lund University, members of the LGRP and members of Thumos, the Genevan research group on emotions, values and norms.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

¹ I take this to be a fairly uncontroversial claim and therefore assume without argument.

² It should be noted that it has not always been entirely clear to me whether the authors I discuss below intend to provide an analysis of blame or something else. With the risk of being uncharitable to some, I assume they intend to provide an analysis of the nature of blame.

³ For simplicity, I write ‘judging blameworthy’ instead of for example ‘judging the agent has manifested insufficient good will’ or “judging the agent has stained her moral record’. I believe we can change ‘judging blameworthy’ to any of the particular accounts presented and have the same problem.

⁴ This distinction echoes the general distinction pointed out by Kubala (2017) and Scheffler (2010) between judging valuable and valuing. Kubala (2017) argues for the psychological possibility of valuing without believing valuable. He states ‘[a]s a matter of psychological fact, it is possible to value something […] without believing it valuable.’ (Kubala 2017, 60) Scheffler (2010) has similarly defended the distinction between valuing and believing valuable. He states ‘[b]ut the proposal that to value X is simply to believe that X is valuable is unsatisfactory in any case, for it is not only possible but commonplace to believe that something is valuable without valuing it oneself. There are, for example, many activities that I regard as valuable but which I myself do not value, including, say, folk dancing, bird-watching, and studying Bulgarian history. Indeed, I value only a tiny fraction of the activities that I take to be valuable.’ (Scheffler 2010, 21).

⁵ Some argue that it is even possible to blame an agent without judging her blameworthy (Menges 2017; Pickard 2013; Portmore 2022). It is common to refer to this phenomenon as irrational blame or recalcitrant blame (Menges 2017; Pickard 2013; Portmore 2022). The discussion of recalcitrant blame draws much inspiration from D’Arms and Jacobson’s work on recalcitrant emotions (D’arms 2003; D’Arms and Jacobson 2000). The idea here is that judgements and emotions can come apart in the sense that we may experience an emotion despite having made a judgment that conflicts with it. In some circumstances we may for example judge that a monster in a movie does not pose a danger to us and still fear it or judge that an agent is not loveable, but rather cruel, and still love her. Similarly, these blame scholars argue we may sometimes judge an agent not blameworthy and still blame her. Here, I do not take a stance regarding the feasibility of this phenomenon. For my purposes, it is sufficient that we do not believe that simply judging an agent blameworthy is to blame her.

⁶ Scanlon summarizes his account nicely as follows: ‘Briefly put, my proposal is this: to claim that a person is blameworthy for an action is to claim that the action shows something about the agent’s attitudes toward others that impair the relations that others can have with him or her. To blame is to judge him or her to be blameworthy and to take your relationship with him or her to be modified in a way that this judgment of impaired relations holds to be appropriate.’ (Scanlon 2008, 128–129).

⁷ Thanks are owed to the anonymous referee and András Szigeti for introducing this reply to me.

⁸ Wolf writes: “Let me return to the first and perhaps the main point that I wish to defend – namely, that the range of attitudes and related activities that I am used to referring to when I use the word ‘blame’ is distinct from the range to which Scanlon refers, and has a potentially valuable role in our lives. The range of attitudes I have in mind, as I have said, is a range that includes resentment, indignation, guilt, and righteous anger – they are emotional attitudes that involve negative feelings toward a person, arising from the belief or impression that the person has behaved badly toward oneself or to a member (or members) of a community about which one cares and which tend to give rise to or perhaps even include a desire to scold or punish the person for his bad behavior. I shall refer to the range of attitudes I have in mind as ‘the angry attitudes’ and the kind of blame that is characterized by these attitudes as ‘angry blame.’ (Wolf 2011, 335–6).

⁹ There are two crucial distinguishing properties between affective episodes and affective dispositions. First, while affective episodes are usually thought to be short lived, lasting for minutes, or in rare cases even hours, affective dispositions are usually thought to be long-lived, lasting for several minutes, hours, days, weeks, months, even years. Second, while affective episodes are commonly thought to have a phenomenology, a ‘what-it-is-likeness’, affective dispositions are not commonly thought to have a phenomenology (at least not by themselves). To have an episode of fear involves experiencing various bodily changes, e.g. increased heart rate and feeling warm. By contrast, having a disposition to be afraid of certain objects does not involve bodily feelings. Rather, affective dispositions can be viewed as tendencies to experience certain affective episodes with bodily feelings. For more about this distinction, see (Deonna and Teroni 2011; Tappolet 2016).

¹⁰ For defenders of the view that emotions are identified with judgements, see (Nussbaum 2001; Solomon 1976). For critique of such views, see (Deonna and Teroni 2011). If we accept that the angry blame emotions need not be identified with a judgment, this can be interpreted as a virtue because they have the tools to make sense of irrational or recalcitrant blame, see fn 5.

¹¹ A further criticism with the judgment account is that it seems ill equipped to account for this intuition and the intuition that we often do something when we blame someone.

¹² Pickard (2013) writes similarly: ‘If blame is like an emotion, then which emotion is it like? For, there is no “basic” emotion of blame. Indeed, it seems that blame can be connected to a range of different emotions. Most obviously, these include anger, hate, and resentment. But the range can plausibly be extended to conclude certain other states that have an affective dimension without being uncontroversially identifiable as emotions, such as, for instance, disappointment, indignation and contempt. Moreover, as expected given this range, blame’s manifestations can be equally various. Alongside punishing, blame can also be manifest in berating, attacking, humiliating, writing off, rejecting, shunning, abandoning, and criticizing, to name but a few behaviours. There is thus a challenge facing the suggestion that blame is like an emotion. The challenge is to unite these various emotions and manifestations thereof into a single account of blame. For, given that they can occur without counting as instances of blame, we must explain what makes them count, when they do, as instances of blame.’ (Pickard 2013, 622–3).

¹³ It should be mentioned that Rosen (2015) does not aim to provide an analysis of blame. Despite this, I believe what he writes about blame and how he unifies the blame emotions in his 2015 paper is helpful in this section.

¹⁴ Which is common to do, see for instance (Ewing 1948; Skorupski 2010).

¹⁵ For defenders of circular fitting attitude analyses of value, see (Garcia 2018; Tappolet 2016).

¹⁶ Again, this should not be read as saying that the accounts discussed above are adequately refuted and cannot be made to work. Rather, it should be read as saying that the worries they are subject to are severe enough to motivate the search for an alternative account of the nature of blame that can account for the presented worries.

¹⁷ In the literature on emotions, the term ‘sentiment’ has been used in many different ways, compare for instance how the following philosophers use the term (D’Arms and Jacobson 2000; Deonna and Teroni 2009; Helm 2009; Hume 2000; Naar 2018; Vendrell Ferran forthcoming). Some use it to mark an affective episode, others a disposition. Among those marking it as a disposition, there are disagreements regarding how to spell out the details. Feel free to call my view the disposition view if you believe it does not coincide perfectly with your preferred understanding of sentiments.

¹⁸ Thanks are owed to the anonymous referee for inviting me to clarify the unity worry and how defenders of the sentiment account can answer it.

¹⁹ An antirealist – antirealism about dispositions is the thesis, roughly put, that emotional dispositions are nothing over and above their manifestations – might unite instances of the characteristic emotions, actions and thoughts by claiming that they are part of the same rationally connected sequence of emotions, thoughts and actions characteristic of blaming. An instance of disappointment is an instance of blaming when it is part of such a sequence. I do not have the space to compare these kinds of explanations, and more generally, argue for whether or not we should be realists or antirealists about sentiments.

²⁰ For other realist theories about dispositions, see (Molnar 2003; Mumford 1998; Heil 2003).

²¹ In the context of love, Naar (2013) writes something similar: ‘What ties all the events of a given sequence together as constituting an expression of love? The answer, on the dispositional account, is simply that a disposition, love, is their common origin. And we now have a way to tell whether or not a given event is part of the relevant sequence: we need to consider the disposition that produced it; if it is the disposition with which we identify love, then the event is an expression of love, and if it is not, then the event is not an expression of love. What distinguishes an episode of joy towards another person’s embarrassment as an expression of love from a similar episode that is an expression of cruelty is thus that the former originates in love while the latter originates in cruelty. (Perhaps certain cases can be interpreted as mimicking cases.) This point is trivial only superficially, however: whatever sort of thing love-the-disposition is – so far, the account is silent on this more specific question – an event that counts as an ’expression of love’ must come from that thing.’ (Naar 2013, 353).

²² I do not think it is a unique flaw with my account that it includes manifestations that are manifestations of other sentiments as well. It is common that sentiments share some characteristic manifestations with each other. Consider, for instance, the sentiments ‘love’ and ‘care’, and ‘hate’ and ‘dislike’. The emotion ‘joy’ might be a characteristic manifestation of both love and care. Likewise, ‘disgust’ might be a characteristic manifestation of dislike as well as hate. When we individuate sentiments, or multi-track dispositions more generally, I think it makes sense to look at all characteristic manifestations (or, rather, all characteristic manifestations we know of) instead of just some of them. If we consider all manifestations we know about, we see for instance that even though hate and blame share some characteristic manifestations, they do not share all. For example, to demand an excuse, feel guilt or thinking it is fitting for others to blame the blameworthy agent are not characteristic manifestations of hate.