This rich book should be read by psychologists and philosophers alike, because it introduces a wealth of relevant research, ideas, and references. Doris's (Reference Doris2015b) thesis is that judgments of moral responsibility are relatively independent of notions of freedom and determinism (he's a compatibilist), do not depend on accurate self-knowledge, but do depend on social negotiations and social context. This means that often there is no objectively “correct” or single answer to questions of who is responsible; it depends on context. Some effects of context on moral judgments (e.g., nonconscious priming) cannot be explained to others, while other effects can or might be. But this depends on finding common ground with others. We comment on four core ideas – collaboration, dual process theories, self, and identity – and suggest that social coherence in reflections about moral agency may depend on culture. Finally, we ask whether social coherence confers logical coherence.
“Collaboration” (Ch. 5) refers to the thesis that human reasoning, including reasoning about morally responsible agency, is social and negotiated rather than principled and based on mental states. Thus, accurately reading mental states is less relevant than reflectivists contend. Haan made a similar point and contrasted it with Kohlberg's rationalist view of moral development. “[M]oral truth is based on agreements moral agents achieve about their common interest and is not predetermined by rules or principles, that is, truth is to be achieved, not revealed” (Haan Reference Haan1978, p. 289). This idea is supported by work on priming culture (context writ large) among bicultural participants. Incidental cultural icons can switch people from making causal attributions as members of a culture with “independent” self-concepts, to members with “interdependent” self-concepts (Markus & Kitayama Reference Markus and Kitayama1991) and vice versa, quickly and unconsciously (Hong et al. Reference Hong, Morris, Chiu and Benet-Martinez2000). Presumably cultural differences in moral judgment and responsibility attributions follow (e.g., Miller Reference Miller1984 on differences between attributions by Indian and American children).
Further, Shimizu et al. (Reference Shimizu, Lee and Uleman2017) showed that cultural differences in memory effects of “spontaneous trait inferences” reside entirely in automatic (unconscious, implicit) rather than controlled (conscious, explicit) processes. Spontaneous personality trait inferences are those made without intentions or awareness (Uleman et al. Reference Uleman, Rim, Saribay and Kressel2012). Because traits are commonly understood as causes (cf. Uleman Reference Uleman, Stroessner and Sherman2015), this shows a process by which cultural (i.e., collaborative) differences in causal attributions may emerge. Using Jacoby's (Reference Jacoby1991) process dissociation procedure (PDP; see next paragraph), Shimizu et al. (Reference Shimizu, Lee and Uleman2017) argue that most important cultural differences occur through automatic processes.
Dual process theories reflect many different dualities. In chapter 3 Doris highlights the one between automatic and controlled processing from semantic priming studies, in which these two processes yield conflicting results in cognition or behavior. Note that there are many kinds of priming (e.g., repetition, goal, procedural, and perceptual), each with its own properties (e.g., Förster et al. Reference Förster, Liberman, Friedman, Morsella, Bargh and Gollwitzer2009), although this does not alter Doris's main point. There are also many definitions of “automatic” and “controlled.” Processes may be called automatic if people are unaware of them and/or unaware of their effects; they are not intended; they have short reaction times (<500 ms); they require little or no cognitive capacity; concurrent cognitive tasks do not interfere with them; or they are uncontrollable (see Bargh Reference Bargh, Wyer and Srull1994). Jacoby's (Reference Jacoby1991) PDP differs by estimating how much control exists when participants actually attempt to control cognition or behavior. We prefer Jacoby's definition because it does not label processes automatic if they have any (but rarely all) of the above characteristics, which do not always co-occur. The errors that occur in the false recognition paradigm (that has become standard for detecting spontaneous inferences) occur in spite of participants' efforts to control the unconscious intrusion of these inferences into task performance (Todorov & Uleman Reference Todorov and Uleman2004). And they provide the bases for estimating the simultaneous operation of automatic and controlled processes in spontaneous inferences.
Doris also recognizes the distinction between holding people accountable for behavior prompted by events of which they are not aware (e.g., subliminal priming) and events that operate uncontrollably in spite of awareness and goals (e.g., addictions). These are both taken as challenges to a reflective position. Several varieties of self-control and its absence are examined in a fine collection of 27 chapters edited by Hassin et al. (Reference Hassin, Ochsner and Trope2010), from the neural to the social level.
“Selves” recognizes that there are many (Ch. 8) and that their effects differ. Selves define the threats and values that determine one's morally responsible agency. Remarkably, even the implicit accessibility of various selves affects thought and behavior. Blaming the victim (whether of rape, non-sexual assault, or natural disaster) is commonly attributed to the “belief in a just world” (BJW): believing that people deserve what they get and get what they deserve, and that the world is morally predictable. Rather than relinquish this belief when they are faced with apparently unfair victimization, people may deal with their distress by blaming the victim, thereby restoring a predictable and fair world in which only the deserving encounter harm. However, research has produced inconsistent results: sometimes victims elicit compassion and support. In past research on victim blame, BJW and self-relevance were often confounded. So we manipulated them independently in a series of six studies (Granot et al., Reference Granot, Uleman and Balcetisunder review). We found that the classic phenomenon of blaming the victim only occurs when both justice concerns and relevance to the self are activated.
In all of these studies participants read vignettes describing a victimization: a newspaper account of a hurricane victim, a fictional account of a young adult assaulted after a party, and genuine accounts of a sexual assault and an armed robbery on campus. Relevance to the self was manipulated by asking participants to assume a first-person or a third-person observer perspective; by displaying a photograph of the participant or a confederate on the participant's computer screen throughout the study; or by presenting “personal safety tips” to half of the participants in the case of assaults. While these manipulations of self-relevance did not ask participants to adopt “selves” that were unfamiliar to them, they activated the self in both explicit and subtle ways.
In each study, and across studies in a meta-analysis, blaming the victim only occurred when both self and justice concerns were high. Morally responsible agency was differently attributed in the same vignettes on the basis of relatively incidental changes in self concerns. These findings not only clarify the basis for inconsistencies in prior BJW findings, but they also illustrate the contextual malleability of moral judgments.
Should anyone doubt that personal relevance affects morality judgments relative to a more evenhanded god's-eye-view, Ham and Van den Bos (Reference Ham and Van den Bos2008) showed how the two perspectives can differ from each other even within one participant. In two studies, participants read brief vignettes describing unjust events. Some vignettes were more relevant to the self than others, for example, “You and your colleague do the same work. You make(s) 1400 Euros a month and your colleague makes 4100 Euros a month” (Ham & Van den Bos Reference Ham and Van den Bos2008, p. 700). Spontaneous justice inferences were measured through response times (RTs) to justice-related words in a probe-recognition paradigm, i.e., were measured implicitly. Justice concerns arose most strongly (i.e., RTs were slowed most) to unjust vignettes involving the self (or to a relevant friend in study 2), compared to vignettes involving strangers. That is, injustice activated justice concerns only under high self-relevance. Explicit judgments were also obtained from these same participants, of how just events were in similar vignettes. Here self- (or friend-) relevance made no difference. Thus, participants were simultaneously of two minds. Implicitly (spontaneously), self-relevance was taken into account in activating justice concerns, but explicitly it was not. Not only are there multiple selves, but also the same self can have different effects for implicit and explicit judgments.
Identity (Ch. 8) affects how visual information is encoded and processed, in ways directly relevant to morally responsible agency. Doris posits the possibility that cultural identities might lead people to attend to different aspects of agency, but even attending to the same things does not ensure such information is similarly processed. Granot et al. (Reference Granot, Balcetis, Schneider and Tyler2014) asked participants to judge responsibility and blame from videotapes of altercations between two parties. In some cases, these were from dashboard cameras of police officers stopping motorists for traffic violations; in other cases, these were staged fights purportedly from security cameras. Participants' identification with one of the parties was either measured through self-reports (identification with police) or manipulated through assigning the parties to otherwise neutral in-groups and out-groups. These identifications, interacting with attention, affected judgments of blame and responsibility in a counterintuitive way, one that Doris would recognize as difficult to justify. Those who viewed the judgment target briefly were unaffected by identification – all blamed the target similarly. But those who studied the target more thoroughly were polarized in their judgments of blame and responsibility. If the target was a police officer with whom one identified, or a member of one's in-group, blame and responsibility were less; if the target was the out-group party, blame and responsibility were greater.
“Collaboration” again. Is the expectation of logical consistency in judging morality culturally relative? Perhaps social collaboration produces logically consistent systems of moral thought, at least in local linguistic communities. Many prominent Western philosophers (e.g., Kant; see Uleman Reference Uleman2010) have sought to develop logically consistent systems of moral thought. But Nisbett et al. (Reference Nisbett, Peng, Choi and Norenzayan2001) suggest this may be a particularly Western concern. They note broad cultural variations in systems of thought, and contrast holistic (traditional Chinese) with analytic (traditional Greek) thought:
We define holistic thought as involving an orientation to the context or field as a whole . . . an emphasis on change, a recognition of contradiction and of the need for multiple perspectives, and a search for the “Middle Way” between opposing propositions. We define analytic thought as involving detachment of the object from its context. . . . Inferences rest in part on the practice of decontextualizing structure from content, the use of formal logic, and avoidance of contradiction. (Nisbett et al. Reference Nisbett, Peng, Choi and Norenzayan2001, p. 293, italics added)
Historically, Chinese culture valued holistic thought whereas Greek culture valued analytic thought and its requirement of logical consistency. Nisbett et al. (Reference Nisbett, Peng, Choi and Norenzayan2001) cite extensive evidence for the persistence of this cultural difference between Eastern and Western systems of thought, with Westerners more concerned with resolving than transcending contradictions.
This suggests that “collaboration” may resolve logical contradictions only when the cultural “system of thought” requires a resolution. Otherwise, multiple perspectives are embraced and not found wanting. It can also be argued that moral reasoning (and reasoning in general) is in the service of self-justification (Mercier & Sperber Reference Mercier and Sperber2011). While this is collaborative in that it necessarily involves others, it is unlikely to produce a god's-eye-view of moral responsibility.
You have
Access
This rich book should be read by psychologists and philosophers alike, because it introduces a wealth of relevant research, ideas, and references. Doris's (Reference Doris2015b) thesis is that judgments of moral responsibility are relatively independent of notions of freedom and determinism (he's a compatibilist), do not depend on accurate self-knowledge, but do depend on social negotiations and social context. This means that often there is no objectively “correct” or single answer to questions of who is responsible; it depends on context. Some effects of context on moral judgments (e.g., nonconscious priming) cannot be explained to others, while other effects can or might be. But this depends on finding common ground with others. We comment on four core ideas – collaboration, dual process theories, self, and identity – and suggest that social coherence in reflections about moral agency may depend on culture. Finally, we ask whether social coherence confers logical coherence.
“Collaboration” (Ch. 5) refers to the thesis that human reasoning, including reasoning about morally responsible agency, is social and negotiated rather than principled and based on mental states. Thus, accurately reading mental states is less relevant than reflectivists contend. Haan made a similar point and contrasted it with Kohlberg's rationalist view of moral development. “[M]oral truth is based on agreements moral agents achieve about their common interest and is not predetermined by rules or principles, that is, truth is to be achieved, not revealed” (Haan Reference Haan1978, p. 289). This idea is supported by work on priming culture (context writ large) among bicultural participants. Incidental cultural icons can switch people from making causal attributions as members of a culture with “independent” self-concepts, to members with “interdependent” self-concepts (Markus & Kitayama Reference Markus and Kitayama1991) and vice versa, quickly and unconsciously (Hong et al. Reference Hong, Morris, Chiu and Benet-Martinez2000). Presumably cultural differences in moral judgment and responsibility attributions follow (e.g., Miller Reference Miller1984 on differences between attributions by Indian and American children).
Further, Shimizu et al. (Reference Shimizu, Lee and Uleman2017) showed that cultural differences in memory effects of “spontaneous trait inferences” reside entirely in automatic (unconscious, implicit) rather than controlled (conscious, explicit) processes. Spontaneous personality trait inferences are those made without intentions or awareness (Uleman et al. Reference Uleman, Rim, Saribay and Kressel2012). Because traits are commonly understood as causes (cf. Uleman Reference Uleman, Stroessner and Sherman2015), this shows a process by which cultural (i.e., collaborative) differences in causal attributions may emerge. Using Jacoby's (Reference Jacoby1991) process dissociation procedure (PDP; see next paragraph), Shimizu et al. (Reference Shimizu, Lee and Uleman2017) argue that most important cultural differences occur through automatic processes.
Dual process theories reflect many different dualities. In chapter 3 Doris highlights the one between automatic and controlled processing from semantic priming studies, in which these two processes yield conflicting results in cognition or behavior. Note that there are many kinds of priming (e.g., repetition, goal, procedural, and perceptual), each with its own properties (e.g., Förster et al. Reference Förster, Liberman, Friedman, Morsella, Bargh and Gollwitzer2009), although this does not alter Doris's main point. There are also many definitions of “automatic” and “controlled.” Processes may be called automatic if people are unaware of them and/or unaware of their effects; they are not intended; they have short reaction times (<500 ms); they require little or no cognitive capacity; concurrent cognitive tasks do not interfere with them; or they are uncontrollable (see Bargh Reference Bargh, Wyer and Srull1994). Jacoby's (Reference Jacoby1991) PDP differs by estimating how much control exists when participants actually attempt to control cognition or behavior. We prefer Jacoby's definition because it does not label processes automatic if they have any (but rarely all) of the above characteristics, which do not always co-occur. The errors that occur in the false recognition paradigm (that has become standard for detecting spontaneous inferences) occur in spite of participants' efforts to control the unconscious intrusion of these inferences into task performance (Todorov & Uleman Reference Todorov and Uleman2004). And they provide the bases for estimating the simultaneous operation of automatic and controlled processes in spontaneous inferences.
Doris also recognizes the distinction between holding people accountable for behavior prompted by events of which they are not aware (e.g., subliminal priming) and events that operate uncontrollably in spite of awareness and goals (e.g., addictions). These are both taken as challenges to a reflective position. Several varieties of self-control and its absence are examined in a fine collection of 27 chapters edited by Hassin et al. (Reference Hassin, Ochsner and Trope2010), from the neural to the social level.
“Selves” recognizes that there are many (Ch. 8) and that their effects differ. Selves define the threats and values that determine one's morally responsible agency. Remarkably, even the implicit accessibility of various selves affects thought and behavior. Blaming the victim (whether of rape, non-sexual assault, or natural disaster) is commonly attributed to the “belief in a just world” (BJW): believing that people deserve what they get and get what they deserve, and that the world is morally predictable. Rather than relinquish this belief when they are faced with apparently unfair victimization, people may deal with their distress by blaming the victim, thereby restoring a predictable and fair world in which only the deserving encounter harm. However, research has produced inconsistent results: sometimes victims elicit compassion and support. In past research on victim blame, BJW and self-relevance were often confounded. So we manipulated them independently in a series of six studies (Granot et al., Reference Granot, Uleman and Balcetisunder review). We found that the classic phenomenon of blaming the victim only occurs when both justice concerns and relevance to the self are activated.
In all of these studies participants read vignettes describing a victimization: a newspaper account of a hurricane victim, a fictional account of a young adult assaulted after a party, and genuine accounts of a sexual assault and an armed robbery on campus. Relevance to the self was manipulated by asking participants to assume a first-person or a third-person observer perspective; by displaying a photograph of the participant or a confederate on the participant's computer screen throughout the study; or by presenting “personal safety tips” to half of the participants in the case of assaults. While these manipulations of self-relevance did not ask participants to adopt “selves” that were unfamiliar to them, they activated the self in both explicit and subtle ways.
In each study, and across studies in a meta-analysis, blaming the victim only occurred when both self and justice concerns were high. Morally responsible agency was differently attributed in the same vignettes on the basis of relatively incidental changes in self concerns. These findings not only clarify the basis for inconsistencies in prior BJW findings, but they also illustrate the contextual malleability of moral judgments.
Should anyone doubt that personal relevance affects morality judgments relative to a more evenhanded god's-eye-view, Ham and Van den Bos (Reference Ham and Van den Bos2008) showed how the two perspectives can differ from each other even within one participant. In two studies, participants read brief vignettes describing unjust events. Some vignettes were more relevant to the self than others, for example, “You and your colleague do the same work. You make(s) 1400 Euros a month and your colleague makes 4100 Euros a month” (Ham & Van den Bos Reference Ham and Van den Bos2008, p. 700). Spontaneous justice inferences were measured through response times (RTs) to justice-related words in a probe-recognition paradigm, i.e., were measured implicitly. Justice concerns arose most strongly (i.e., RTs were slowed most) to unjust vignettes involving the self (or to a relevant friend in study 2), compared to vignettes involving strangers. That is, injustice activated justice concerns only under high self-relevance. Explicit judgments were also obtained from these same participants, of how just events were in similar vignettes. Here self- (or friend-) relevance made no difference. Thus, participants were simultaneously of two minds. Implicitly (spontaneously), self-relevance was taken into account in activating justice concerns, but explicitly it was not. Not only are there multiple selves, but also the same self can have different effects for implicit and explicit judgments.
Identity (Ch. 8) affects how visual information is encoded and processed, in ways directly relevant to morally responsible agency. Doris posits the possibility that cultural identities might lead people to attend to different aspects of agency, but even attending to the same things does not ensure such information is similarly processed. Granot et al. (Reference Granot, Balcetis, Schneider and Tyler2014) asked participants to judge responsibility and blame from videotapes of altercations between two parties. In some cases, these were from dashboard cameras of police officers stopping motorists for traffic violations; in other cases, these were staged fights purportedly from security cameras. Participants' identification with one of the parties was either measured through self-reports (identification with police) or manipulated through assigning the parties to otherwise neutral in-groups and out-groups. These identifications, interacting with attention, affected judgments of blame and responsibility in a counterintuitive way, one that Doris would recognize as difficult to justify. Those who viewed the judgment target briefly were unaffected by identification – all blamed the target similarly. But those who studied the target more thoroughly were polarized in their judgments of blame and responsibility. If the target was a police officer with whom one identified, or a member of one's in-group, blame and responsibility were less; if the target was the out-group party, blame and responsibility were greater.
“Collaboration” again. Is the expectation of logical consistency in judging morality culturally relative? Perhaps social collaboration produces logically consistent systems of moral thought, at least in local linguistic communities. Many prominent Western philosophers (e.g., Kant; see Uleman Reference Uleman2010) have sought to develop logically consistent systems of moral thought. But Nisbett et al. (Reference Nisbett, Peng, Choi and Norenzayan2001) suggest this may be a particularly Western concern. They note broad cultural variations in systems of thought, and contrast holistic (traditional Chinese) with analytic (traditional Greek) thought:
We define holistic thought as involving an orientation to the context or field as a whole . . . an emphasis on change, a recognition of contradiction and of the need for multiple perspectives, and a search for the “Middle Way” between opposing propositions. We define analytic thought as involving detachment of the object from its context. . . . Inferences rest in part on the practice of decontextualizing structure from content, the use of formal logic, and avoidance of contradiction. (Nisbett et al. Reference Nisbett, Peng, Choi and Norenzayan2001, p. 293, italics added)
Historically, Chinese culture valued holistic thought whereas Greek culture valued analytic thought and its requirement of logical consistency. Nisbett et al. (Reference Nisbett, Peng, Choi and Norenzayan2001) cite extensive evidence for the persistence of this cultural difference between Eastern and Western systems of thought, with Westerners more concerned with resolving than transcending contradictions.
This suggests that “collaboration” may resolve logical contradictions only when the cultural “system of thought” requires a resolution. Otherwise, multiple perspectives are embraced and not found wanting. It can also be argued that moral reasoning (and reasoning in general) is in the service of self-justification (Mercier & Sperber Reference Mercier and Sperber2011). While this is collaborative in that it necessarily involves others, it is unlikely to produce a god's-eye-view of moral responsibility.