Following Peter Strawson, John Doris claims that the ‘right way’ of thinking about agency should attend to those practices where we tend to ascribe moral responsibility. These practices are usually signaled by the presence of reactive attitudes. Reactive attitudes (e.g., gratitude, resentment, indignation, anger, guilt) are peculiar kinds of emotions whose expression we recognize as proper for some typical and paramount interpersonal relationships (Scanlon Reference Scanlon, Coates and Tognazzini2013). This line of thinking highlights two points, which are half descriptive and half normative. Reactive attitudes are so deeply embedded in our psychological evolved nature and social interactions that attempts to revise these attitudes must be seen as carrying the burden of the proof in a cost-benefit analysis. Furthermore, insofar as reactive attitudes are affects and emotions aimed to regulate our behavior within a set of interpersonal relationships, they offer a natural path to ground normativity on a factual basis, including a genuine feeling of ‘to-be-doneness’ attached to them (Mackie Reference Mackie1977).
We agree on the fecundity of thinking about moral responsibility through this lens, and we find Talking to Our Selves to be one of the most refreshing books on these issues that we have read in recent years. However, we want to comment on the way Doris models agency through reactive attitudes. For Doris, reactive attitudes are symptoms of morally responsible agency. In particular, the aptness of a reactive attitude toward X is a symptom of X's agency. Thus, if you can justifiably express a reactive attitude toward X, then X will surely be a morally responsible agent.
We think that some aspects of this modeling, although rich and suggestive, are importantly undertheorized. In what follows, we will briefly develop two points on the relationship between reactive attitudes and agency. First, reactive attitudes can sometimes be apt even if they are not tracking morally responsible agency. Aptness of reactive attitudes and hence characterizations of moral responsibility seem to fluctuate between different levels of what “apt” is. Second, in his characterization of agency, Doris sidetracks this slippery slope of the aptness condition with a reference to the values toward which our emotional attitudes react. As we will see below, this is equally problematic.
As we have just mentioned, sometimes reactive attitudes are apt even when we are not tracking agency by exemplifying them. In fact, it has been recently pointed out that some societies tend to discount intentionality for purposes of assigning moral responsibility (Barrett et al. Reference Barrett, Bolyanatz, Crittenden, Fessler, Fitzpatrick, Gurven, Henrich, Kanovsky, Kushnick, Pisorf, Scelzaa, Stichl, von Ruedenn, Zhaoh and Laurence2016). Doris himself recognizes this possibility by mentioning cases of ‘strict liability’ (Doris Reference Doris2015b, pp. 24, 154–55). In cases of strict liability (e.g., warfare atrocities, catastrophic slips), we get moral responsibility – and associated reactive attitudes – directed toward targets that lack core features of agency (e.g., intentionality, knowledge of the outcomes associated to the relevant behavior). Additionally, many times, the object of a reactive attitude is not straightforwardly related to issues of merit and lack thereof (Levy Reference Levy2011). We can love someone even if we recognize that he or she does not deserve our affection. The same general point might apply for admiration, respect, shame, pride, and so on.
To further examine the possibility of having reactive attitudes directed toward clear instances of non-agency, let us be a bit of a gadfly and argue against Doris's proposal with his favorite example of the bee. Our anger toward the bee could be, in fact, an apt reaction in terms of the goal we want to facilitate in our relation with the bee; that is, not being harmed by the bee or being undisturbed by insects while on a picnic. Even if the bee is clearly something of an agential blob in terms of moral agency, the anger we direct against such an organism (and not against an action which has not happened yet and we try to avoid) can be seen as an efficient way of facilitating the above goal.
Our point is simply that in judging the aptness of the reactive attitudes there is easily confusion of a sort that has already been noticed (d'Arms & Jacobson Reference D'Arms and Jacobson2000). Although we assume Doris is not unaware of that danger, we consider the following under a different angle. When judging the appropriateness of reactive attitudes it is not obvious how to adopt a neutral or impartial perspective. Elements of reactive attitudes (e.g., guilt, resentment, admiration, contempt) that are part of the examined system of responsibility practices might supply functions not only of a different kind but also at different levels. This implies that we should be extremely clear about the level from which we are appraising the aptness of a reactive attitude in a given case.
In its most basic form, there are at least two different levels of appraisal for evaluating the aptness of reactive attitudes. From the individual level, reactive attitudes are a key element for regulating our social interactions and for signaling our status in partner-choice selection in both biological (Debove et al. Reference Debove, Baumard and André2015) and cultural evolution (Fessler & Holbrook Reference Fessler and Holbrook2013). As Robert Frank (Reference Frank1988) claimed 30 years ago, moral emotions supply to the individual a selective tool for maximizing her chances of being included in mutually cooperative enterprises. Thus, moralistic tendencies are plausibly psychologically evolved tools apt for uses in intra-group competition (Gavrilets Reference Gavrilets2012).
Normative discourse in philosophy on the appropriateness of reactive attitudes usually takes place at a different level, which is also a higher level than the individual or even subpersonal level just described. It is usually at a group level of analysis in which the function of the responsibility system is the proper target of discussion. From this perspective, reactive attitudes would be of paramount importance in normative and evaluative processes of negotiation aimed at a group's stability and cohesion (how inclusive that group is, is another question). It is easy to recognize this default view on scholarly work on reactive attitudes, for instance, in Strawson's well-known appeal to the value of reactive attitudes for “human life.” Returning to the case of the bee, feeling anger could be “fitting” for the person who, by following such a pedestrian strategy, promotes their goal of being undisturbed while having lunch in the park. Less clear is whether feeling anger toward a bee will be adaptive from a certain group standpoint (e.g., right now we are rather short of bees for pollination . . . ). Shall this make us doubt Doris's premise that the aptness of reactive attitudes is a symptom of morally responsible agency? The general point, again, is that both levels of analysis can deliver opposite verdicts about the aptness of a peculiar reaction or trait. When Doris (Reference Doris2015b) writes, for instance, “there's pretty good reason for you to be angry with me for what I did, if what I did is a function of my mean-spirited matterings” (p. 169), he conflates how mean-spiritness is an attribute at the individual level of analysis in a partner-choice framework (e.g., “you don't wanna be friends with me pal?”), whereas the justification of the practices of the responsibility system should take place at a higher, group level of analysis (e.g., “should we accept and encourage anger among conflictual relationships in our group?”). We would not like to be uncharitable to Doris's account. Our point is rather that it is not easy to not succumb to this slippery slope of aptness. “Aptness for whom?” should be a slogan here as well.
Let's now attend to our second concern with Doris's view of the relationship between reactive attitudes and agency. In Doris's theoretical framework it is the “reference to each man's values [that] explain why they deserve the attitudes I subject them to” (Doris Reference Doris2015b, p. 37). We believe, however, that such an emphasis on the primacy of values of the morally responsible agent is infelicitous. That is, the case not only because explaining human behavior values can sometimes be epiphenomenal to other characteristics that are not really under the control or guidance of the agent, but also because, more generally, reactive attitudes might be inappropriate responses to those very same values that excite them in an agonistic fashion. In other words, reactive attitudes can be a functional response to competing values (it might be their individual level function to respond when confronted with specific values) without these reactions being adequate to limiting or canalizing those competing values in socially desirable ways.
Consider the case of reactive attitudes inside the family. For millennia, it was a deep-seated belief that it was justifiable to corporally punish children, even using extreme violence. Cultural prevalence and persistence of proverbs in very different societies are a sign of the enduring attractiveness of this component of the responsibility system. “He that spareth the rod hateth his son. But he that loveth him chasteneth him betimes.” “Better to beat your child when small than to see him hanged when grown” (see Pinker Reference Pinker2011). The moralistic rationale of these kinds of practices was to protect a hierarchical relationship between children and other adults. Such a relational model was presumed to be essential in forming the underage being (Fiske & Rai Reference Fiske and Rai2014). To state it delicately, no meta-analysis of the available scientific evidence has found any positive effect of this kind of moralistic violence either on children's development or on the quality of the relationship between children and their elders.
Distinguishing among non-converging levels is a crucial step in understanding that social and cultural evolution might cause certain reactive attitudes to appear as justifiably more apt without that being a real proof of their efficacy against the values that trigger them. For instance, Jennifer Jacquet (Reference Jacquet2015) underlined how recent social trends have promoted the reactive attitude of guilt when facing environmentally pernicious consumerism. Jacquet argued that such emotional sign of the times is not very efficient in solving large-scale cooperation dilemmas. Instead, coordinated practices of shaming the most environmentally pernicious agents would prove much more effective. Laboratory experiments in cooperative dilemmas suggest that contribution to public goods can be heightened up to 50% when the practice of shaming uncooperative agents remains a possibility (Jacquet et al. Reference Jacquet, Hauert, Traulsen and Milinski2011). Private and public initiatives conducive to shaming the most disruptive agents in real-world settings have actually led to impressive results in saving water or decreasing tax fraud.
The “apt for whom?” question can also help in visualizing what could be termed the Machiavellian challenge in the appropriateness of moral attitudes when reacting against other agents' values. Indeed, there is a natural tendency toward the formation of coalitions (Weeden & Kurzban Reference Weeden and Kurzban2014) that claim that certain reactive attitudes against other groups are justified while defending their inclusive interests. Even though those coalitions can be broad and end up promoting something akin to the “public interest,” a theory of morally responsible agency must be reflective on these issues. The impact of the widespread justification of responsibility practices on public policies is well-documented. Reactive attitudes toward merit, effort, and luck are strong predictors of different levels of redistributive social spending across a wide range of societies (Alesina & Angeletos Reference Alesina and Angeletos2005; Benabou & Tirole Reference Benabou and Tirole2006). To sum up, reactive attitudes, because they are biologically and evolutionarily anchored on an individual (and even subpersonal) rationality, do not necessarily react efficiently against other values that excite them.
When drawing conclusions from these considerations one could remind oneself not only that values evolve through economic and technological pressures (Morris Reference Morris2015) – and here we applaud Doris when he points out that “[morally responsible] agents are negotiations” (2015b, p. 148) – but also it is essential to keep in mind that some coalitions and groups can impose a disproportionate burden of the responsibility system on other groups. A good characterization of moral agency should make room for this fact. Think simply of how motherhood, as compared to fatherhood, plausibly has been and continues to be accompanied by very demanding responsibilities and expectations. Or think simply about how legal and judicial systems have been periodically criticized since the Enlightenment for imposing stronger responsibility requirements on some groups than others. As we can grasp how these demands and requirements have evolved, shall we conceive of morally responsible agency as evolving as well?
You have
Access
Following Peter Strawson, John Doris claims that the ‘right way’ of thinking about agency should attend to those practices where we tend to ascribe moral responsibility. These practices are usually signaled by the presence of reactive attitudes. Reactive attitudes (e.g., gratitude, resentment, indignation, anger, guilt) are peculiar kinds of emotions whose expression we recognize as proper for some typical and paramount interpersonal relationships (Scanlon Reference Scanlon, Coates and Tognazzini2013). This line of thinking highlights two points, which are half descriptive and half normative. Reactive attitudes are so deeply embedded in our psychological evolved nature and social interactions that attempts to revise these attitudes must be seen as carrying the burden of the proof in a cost-benefit analysis. Furthermore, insofar as reactive attitudes are affects and emotions aimed to regulate our behavior within a set of interpersonal relationships, they offer a natural path to ground normativity on a factual basis, including a genuine feeling of ‘to-be-doneness’ attached to them (Mackie Reference Mackie1977).
We agree on the fecundity of thinking about moral responsibility through this lens, and we find Talking to Our Selves to be one of the most refreshing books on these issues that we have read in recent years. However, we want to comment on the way Doris models agency through reactive attitudes. For Doris, reactive attitudes are symptoms of morally responsible agency. In particular, the aptness of a reactive attitude toward X is a symptom of X's agency. Thus, if you can justifiably express a reactive attitude toward X, then X will surely be a morally responsible agent.
We think that some aspects of this modeling, although rich and suggestive, are importantly undertheorized. In what follows, we will briefly develop two points on the relationship between reactive attitudes and agency. First, reactive attitudes can sometimes be apt even if they are not tracking morally responsible agency. Aptness of reactive attitudes and hence characterizations of moral responsibility seem to fluctuate between different levels of what “apt” is. Second, in his characterization of agency, Doris sidetracks this slippery slope of the aptness condition with a reference to the values toward which our emotional attitudes react. As we will see below, this is equally problematic.
As we have just mentioned, sometimes reactive attitudes are apt even when we are not tracking agency by exemplifying them. In fact, it has been recently pointed out that some societies tend to discount intentionality for purposes of assigning moral responsibility (Barrett et al. Reference Barrett, Bolyanatz, Crittenden, Fessler, Fitzpatrick, Gurven, Henrich, Kanovsky, Kushnick, Pisorf, Scelzaa, Stichl, von Ruedenn, Zhaoh and Laurence2016). Doris himself recognizes this possibility by mentioning cases of ‘strict liability’ (Doris Reference Doris2015b, pp. 24, 154–55). In cases of strict liability (e.g., warfare atrocities, catastrophic slips), we get moral responsibility – and associated reactive attitudes – directed toward targets that lack core features of agency (e.g., intentionality, knowledge of the outcomes associated to the relevant behavior). Additionally, many times, the object of a reactive attitude is not straightforwardly related to issues of merit and lack thereof (Levy Reference Levy2011). We can love someone even if we recognize that he or she does not deserve our affection. The same general point might apply for admiration, respect, shame, pride, and so on.
To further examine the possibility of having reactive attitudes directed toward clear instances of non-agency, let us be a bit of a gadfly and argue against Doris's proposal with his favorite example of the bee. Our anger toward the bee could be, in fact, an apt reaction in terms of the goal we want to facilitate in our relation with the bee; that is, not being harmed by the bee or being undisturbed by insects while on a picnic. Even if the bee is clearly something of an agential blob in terms of moral agency, the anger we direct against such an organism (and not against an action which has not happened yet and we try to avoid) can be seen as an efficient way of facilitating the above goal.
Our point is simply that in judging the aptness of the reactive attitudes there is easily confusion of a sort that has already been noticed (d'Arms & Jacobson Reference D'Arms and Jacobson2000). Although we assume Doris is not unaware of that danger, we consider the following under a different angle. When judging the appropriateness of reactive attitudes it is not obvious how to adopt a neutral or impartial perspective. Elements of reactive attitudes (e.g., guilt, resentment, admiration, contempt) that are part of the examined system of responsibility practices might supply functions not only of a different kind but also at different levels. This implies that we should be extremely clear about the level from which we are appraising the aptness of a reactive attitude in a given case.
In its most basic form, there are at least two different levels of appraisal for evaluating the aptness of reactive attitudes. From the individual level, reactive attitudes are a key element for regulating our social interactions and for signaling our status in partner-choice selection in both biological (Debove et al. Reference Debove, Baumard and André2015) and cultural evolution (Fessler & Holbrook Reference Fessler and Holbrook2013). As Robert Frank (Reference Frank1988) claimed 30 years ago, moral emotions supply to the individual a selective tool for maximizing her chances of being included in mutually cooperative enterprises. Thus, moralistic tendencies are plausibly psychologically evolved tools apt for uses in intra-group competition (Gavrilets Reference Gavrilets2012).
Normative discourse in philosophy on the appropriateness of reactive attitudes usually takes place at a different level, which is also a higher level than the individual or even subpersonal level just described. It is usually at a group level of analysis in which the function of the responsibility system is the proper target of discussion. From this perspective, reactive attitudes would be of paramount importance in normative and evaluative processes of negotiation aimed at a group's stability and cohesion (how inclusive that group is, is another question). It is easy to recognize this default view on scholarly work on reactive attitudes, for instance, in Strawson's well-known appeal to the value of reactive attitudes for “human life.” Returning to the case of the bee, feeling anger could be “fitting” for the person who, by following such a pedestrian strategy, promotes their goal of being undisturbed while having lunch in the park. Less clear is whether feeling anger toward a bee will be adaptive from a certain group standpoint (e.g., right now we are rather short of bees for pollination . . . ). Shall this make us doubt Doris's premise that the aptness of reactive attitudes is a symptom of morally responsible agency? The general point, again, is that both levels of analysis can deliver opposite verdicts about the aptness of a peculiar reaction or trait. When Doris (Reference Doris2015b) writes, for instance, “there's pretty good reason for you to be angry with me for what I did, if what I did is a function of my mean-spirited matterings” (p. 169), he conflates how mean-spiritness is an attribute at the individual level of analysis in a partner-choice framework (e.g., “you don't wanna be friends with me pal?”), whereas the justification of the practices of the responsibility system should take place at a higher, group level of analysis (e.g., “should we accept and encourage anger among conflictual relationships in our group?”). We would not like to be uncharitable to Doris's account. Our point is rather that it is not easy to not succumb to this slippery slope of aptness. “Aptness for whom?” should be a slogan here as well.
Let's now attend to our second concern with Doris's view of the relationship between reactive attitudes and agency. In Doris's theoretical framework it is the “reference to each man's values [that] explain why they deserve the attitudes I subject them to” (Doris Reference Doris2015b, p. 37). We believe, however, that such an emphasis on the primacy of values of the morally responsible agent is infelicitous. That is, the case not only because explaining human behavior values can sometimes be epiphenomenal to other characteristics that are not really under the control or guidance of the agent, but also because, more generally, reactive attitudes might be inappropriate responses to those very same values that excite them in an agonistic fashion. In other words, reactive attitudes can be a functional response to competing values (it might be their individual level function to respond when confronted with specific values) without these reactions being adequate to limiting or canalizing those competing values in socially desirable ways.
Consider the case of reactive attitudes inside the family. For millennia, it was a deep-seated belief that it was justifiable to corporally punish children, even using extreme violence. Cultural prevalence and persistence of proverbs in very different societies are a sign of the enduring attractiveness of this component of the responsibility system. “He that spareth the rod hateth his son. But he that loveth him chasteneth him betimes.” “Better to beat your child when small than to see him hanged when grown” (see Pinker Reference Pinker2011). The moralistic rationale of these kinds of practices was to protect a hierarchical relationship between children and other adults. Such a relational model was presumed to be essential in forming the underage being (Fiske & Rai Reference Fiske and Rai2014). To state it delicately, no meta-analysis of the available scientific evidence has found any positive effect of this kind of moralistic violence either on children's development or on the quality of the relationship between children and their elders.
Distinguishing among non-converging levels is a crucial step in understanding that social and cultural evolution might cause certain reactive attitudes to appear as justifiably more apt without that being a real proof of their efficacy against the values that trigger them. For instance, Jennifer Jacquet (Reference Jacquet2015) underlined how recent social trends have promoted the reactive attitude of guilt when facing environmentally pernicious consumerism. Jacquet argued that such emotional sign of the times is not very efficient in solving large-scale cooperation dilemmas. Instead, coordinated practices of shaming the most environmentally pernicious agents would prove much more effective. Laboratory experiments in cooperative dilemmas suggest that contribution to public goods can be heightened up to 50% when the practice of shaming uncooperative agents remains a possibility (Jacquet et al. Reference Jacquet, Hauert, Traulsen and Milinski2011). Private and public initiatives conducive to shaming the most disruptive agents in real-world settings have actually led to impressive results in saving water or decreasing tax fraud.
The “apt for whom?” question can also help in visualizing what could be termed the Machiavellian challenge in the appropriateness of moral attitudes when reacting against other agents' values. Indeed, there is a natural tendency toward the formation of coalitions (Weeden & Kurzban Reference Weeden and Kurzban2014) that claim that certain reactive attitudes against other groups are justified while defending their inclusive interests. Even though those coalitions can be broad and end up promoting something akin to the “public interest,” a theory of morally responsible agency must be reflective on these issues. The impact of the widespread justification of responsibility practices on public policies is well-documented. Reactive attitudes toward merit, effort, and luck are strong predictors of different levels of redistributive social spending across a wide range of societies (Alesina & Angeletos Reference Alesina and Angeletos2005; Benabou & Tirole Reference Benabou and Tirole2006). To sum up, reactive attitudes, because they are biologically and evolutionarily anchored on an individual (and even subpersonal) rationality, do not necessarily react efficiently against other values that excite them.
When drawing conclusions from these considerations one could remind oneself not only that values evolve through economic and technological pressures (Morris Reference Morris2015) – and here we applaud Doris when he points out that “[morally responsible] agents are negotiations” (2015b, p. 148) – but also it is essential to keep in mind that some coalitions and groups can impose a disproportionate burden of the responsibility system on other groups. A good characterization of moral agency should make room for this fact. Think simply of how motherhood, as compared to fatherhood, plausibly has been and continues to be accompanied by very demanding responsibilities and expectations. Or think simply about how legal and judicial systems have been periodically criticized since the Enlightenment for imposing stronger responsibility requirements on some groups than others. As we can grasp how these demands and requirements have evolved, shall we conceive of morally responsible agency as evolving as well?
ACKNOWLEDGMENT
Funding for this work comes from the project FFI2015-67569-C2-2-P, MINECO/FEDER: “The constitution of the subject in social interaction. Identity, norms and the sense of action from an empirical perspective.”