The Virtues of Reactive Attitudes

For the purposes of this paper I will assume rather than argue for this position, namely a form of metaethical constructivism (in the Humean rather than Kantian tradition). I argue for the position in Tessman (2015; 2017). See also: Lenman (2010); Lenman and Shemmer (2012); Street (2008; 2012). Both humans and other animals engage in the activities of valuing, but I focus only on humans.

The common phenomenon of “blaming the victim” is based on unfair normative expectations; this is evident, for example, when a woman is held responsible for being raped because she was walking alone at night or dressed in a way that the rapist found to be provocative, or a Black man is held responsible for being subject to excessive force during an arrest because he acted in a manner that caused a police officer to feel threatened.

One can argue (on the basis of a claim about the function of morality, given certain assumptions about human nature and about what human cooperation requires) that any moral system that does not have some norm of fairness or reciprocity, and some recognition of the need for justification of shared normative expectations, is inadequate. David Wong, in developing a constructivist position that he calls pluralistic relativism, argues that there are some universal constraints on any adequate morality, including that any adequate morality must have some norm of reciprocity and some requirement that norms be justifiable to those who are to be governed by them. He claims that “some form of reciprocity is a norm for all cultures we know, where reciprocity is conceived as a fitting and proportional return of good for good” (Wong 2006, 47), noting that, “while reciprocity in a very general sense is the norm for human cultures, the specific form it takes varies a great deal across cultures” (Wong 2006, 50). Regarding the constraint of justifiability, he points out that “as a system for promoting cooperation, morality works through a large degree of voluntary acceptance of its norms… If conformance to its norms…depended solely on the threat of force or coercion, the costs would detract greatly from the benefits of social cooperation itself… voluntary acceptance of moral norms came to be seen as based on their justifiability to those governed by them. Hence another constraint on moralities is that justification for following the norms and reasons of an adequate morality cannot crucially depend on falsehoods” (Wong 2006, 59).

See Slote (1984; 1992). I am denying that the self-other asymmetry is due to some shortcoming or vice related to an improper balance between self-regarding and other-regarding traits. I don’t think that the asymmetry is necessarily due to any shortcoming. Rather, I will be arguing that we may rightly hold ourselves responsible for things that no one else can hold us responsible for.

Strawson develops the concept of reactive attitudes in the context of a discussion of determinism—a context that I will leave to the side. Strawson argues that accepting “the truth of a general thesis of determinism” (1962, 195) would have no effect on our practices of responsibility—it would not stop us from having reactive attitudes that express how and in what way other people’s regard for us (and for others) matters to us. It would not lead us to excuse people from responsibility for their actions (where we tend to excuse when a bad action turns out not to have been based on any underlying ill will), and it would not lead us to exempt all people from counting as responsible agents (as, for instance, we would exempt certain categories of people, such as young children) and thus to see people as “an object of social policy…as a subject for what…might be called treatment…as something…to be managed or handled or cured or trained” (1962, 194). Instead, regardless of the truth of determinism, the fact that we are committed to having interpersonal relationships in which other people’s quality of will matters to us means that while it would be logically possible to adopt this “objective attitude” towards others rather than responding to them as a participant in an interpersonal relationship (i.e. responding with the reactive attitudes), it is “practically inconceivable” (1962, 197) that we would do so; our valuing of interpersonal relationships precludes it.

Theorists in the Strawsonian tradition have refined what Strawson calls “quality of will” into several different aspects of our wills. For instance, Shoemaker (2015) argues that we sometimes have a reactive attitude in response to someone’s “quality of character,” sometimes in response to their “quality of judgment” and sometimes in response to their “quality of regard.” Given this division, my focus is on quality of regard.

It is important to note that while Strawson claims that it is natural to react emotionally to other people’s regard for us (or for others) in interpersonal relationships, he is not claiming that certain specific emotions are the natural ones to have. Which particular emotions reflect how someone’s regard matters is determined by many complex factors and, of course, is culturally variable. See Flanagan (2017).

See, for instance, D’Arms and Jacobson (2000a; 2000b); Robinowicz and Ronnow-Rasmussen (2004); Hieronymi (2005; 2019).

What is important about this for them is that it poses a problem for theories such as (neo)sentimentalism that rely on what they call the “response dependency thesis,” namely the thesis that “to think that X has some evaluative property Φ is to think it appropriate to feel F in response to X” (D’Arms and Jacobson 2000a, 729). The problem is that it may be inappropriate to feel F in response to X for reasons that have nothing to do with whether or not X has the evaluative property Φ, so it is impossible to know whether or not X has the evaluative property Φ simply on the basis of whether or not it is “appropriate” to feel F.

For the purposes of this example, I am accepting D’Arms’ and Jacobson’s stipulation about this case: that the grieving spouse would be unable to experience deep grief without conveying this grief to the children in a way that would traumatize them. This means accepting that the grieving spouse cannot insulate the children from the grief by compartmentalizing--setting some of the grief to the side when interacting with the children--and that if the children were to witness the deep grief they would indeed be traumatized by it. I grant that many real cases would differ from this case because (some, perhaps even most) people can fittingly grieve the death of a beloved spouse without being incapacitated from caring for their children.

Thanks to John Hacker-Wright for pointing out that the person grieving their spouse would think one thought too many by asking whether it was right to grieve.

In her insightful discussion of what is wrong with thinking one thought too many, Susan Wolf (2015) is critical of the way that impartial morality is assumed to override other evaluative considerations. (Interestingly, the concept of the moralistic fallacy comes from a similar concern, namely, a concern about morality overstepping its boundaries so that morally evaluative responses engulf and thus determine the rightness of all other evaluative responses.) Love, for instance, is undermined if it must always pass the test of impartial morality. There is an important difference, however, between Williams’, and Wolf’s, points and mine. Williams and Wolf have a narrower understanding than I do of morality; I count both impartial values and many partial values as moral. As a moral value pluralist, I believe that not only are there a “variety of values” (to use Wolf’s [2015] book title), but that there are a variety of irreducibly different kinds of moral values—for instance, those related to justice and those related to care. Some moral values that are not impartial are “sacralized,” namely experienced as values that one is non-negotiably required to protect, whose value cannot be traded off for the sake of other values, and that one degrades by subjecting to justification, since the implication of seeking justification is that one’s commitment to the value is contingent upon finding adequate justification. I think that Williams and Wolf would agree that the kinds of values that tend to be sacralized are those about which one mustn’t have one thought too many--but they contrast these values with moral values, rather than, as I do, recognize that the sacred values may also be moral values, though of a different kind than impartial values. When sacred values conflict with each other, for instance when grieving the death of one’s beloved spouse conflicts with caring for one’s children, one experiences what is called a “tragic trade-off” (as contrasted with a taboo trade-off, namely a conflict between a sacred and a non-sacred value) [see Tetlock et al 2000; Tetlock 2003]. The person who is grieving the death of their beloved spouse thinks one thought too many if they question whether or not their grief is appropriate—if they make their grief pass any kind of evaluative test by checking for extrinsic reasons not to grieve; but they also think one thought too many if, when responding to their children, they rationally reflect on the permissibility of compromising their ability to care for their children (by, for instance, allowing themselves to be incapacitated by grief).

Though what I am presenting here are claims made by metaethical constructivists, I am also critical of some aspects of how the process of construction is conceived by other (Humean) metaethical constructivists. For my critique of Sharon Street’s position, see Tessman (2015).

In this way, a reactive attitude such as resentment is not about the one who is being resented, in the sense that it is not about whether or not they deserve the resentment; rather, it is about the one who resents, in the sense that it is about what matters to them (Hieronymi 2019).

Valerie Tiberius gives a detailed account of how we might critically examine our own (or help a friend examine their own) “ultimate values” within the context of what she calls a “value-fulfillment theory of well-being.” However, the critical questions that she proposes we ask mix constitutive and extrinsic reasons for valuing:

Intersubjectively constructed values are the values that are assumed to be relevant to responsibility in, for instance, most contractualist moral theory; here, some kind of (hypothetical) agreement is what justifies a set of values or normative expectations (see Scanlon 1998; Darwall 2006). In order to draw the contrast between subjective and shared values, I am skipping over all of the details about the kind of agreement (and the justification for it) that contractualists think are necessary for shared values; these details create important disagreements amongst different versions of contractualist theory. Descriptive work on social norms similarly assumes that norms, and the values that they embody, emerge through some kind of social process of acceptance, though the description of that process varies too (see Ullman-Margalit 1977; Bicchieri 2006; 2017; Brennan, Eriksson, Goodin and Southwood 2013). Philosophical work on responsibility tends in a contractualist direction, so there is very little thought in this literature given to what one might hold oneself responsible for (through self-reactive attitudes) on the basis of subjective values. Theorists have been prone to dismiss self-reactive attitudes as misguided whenever they diverge from interpersonal reactive attitudes. An exception is Kathryn Norlock, who argues (in an unpublished manuscript entitled “If It’s Excusable, Then Why Do I Feel So Bad?: Accounting for Rational Self-Forgiveness When No One Blames Us”) that there are cases in which taking responsibility for one’s own actions (and even blaming oneself) is appropriate even though it is also appropriate for other people to excuse one’s actions.