Knobe's “person as moralist” account provides a novel contribution to the study of human morality. Whereas most research in this domain has examined the features of behavior that guide moral judgment (Cushman Reference Cushman2008; Guglielmo et al. Reference Guglielmo, Monroe and Malle2009; Shaver Reference Shaver1985) or the processes that underlie moral judgment (Greene Reference Greene and Sinnott-Armstrong2008; Haidt Reference Haidt2001), Knobe's target article extends the literature by probing the influence of morality on other psychological judgments.
Despite its promise, however, Knobe's account has several limitations. Knobe neither measures nor defines moral judgment, leaving it unclear precisely what the account posits and how it may be falsified. Nonetheless, any conceptualization of moral judgment consistent with Knobe's account necessarily relies on substantial information about an agent's mental states. Moreover, the results described by Knobe are likewise obtained by instances of entirely non-moral norm violations. Finally, Knobe should clarify why it would be the case that people's moral judgments of badness and blame share no direct relationship.
The crux of Knobe's argument is that moral judgments of badness (hereafter “MJ1”) impact judgments about an agent's mental states and causal role, which thereby impact moral judgments of blame (hereafter “MJ2”). Although MJ2 are often measured, studies of Knobe's account rarely (if ever) measure MJ1. It is therefore critical to know the conditions under which such judgments arise. To this end, Knobe claims to examine “judgment[s] that the agent's action itself is bad.” But this definition does not provide much clarity – if MJ1 are not simply judgments about bad outcomes (sect. 4.1.3, para. 4), are they judgments that an agent caused /knew about/ intended something bad? Absent either a measurement or definition of MJ1, it is unclear precisely what is alleged to influence mental state and causality assessments, and how one could attempt to falsify the account.
In any case, Knobe's account would be most compelling if MJ1 arise in the absence of any considerations of the agent's mental states (which, after all, are proposed to be influenced by MJ1). However, this is clearly not the case. First, the agent's knowledge is relevant to these moral judgments. For example, the harming chairman's action is bad in part because he knew that harm would occur. When agents lack knowledge of the harmful consequences of their action, people no longer view the consequences as intentional (Nadelhoffer Reference Nadelhoffer2006b; Pellizzoni et al. Reference Pellizzoni, Girotto and Surian2010). According to Knobe's account, therefore, such actions must not be bad. But if this is true, then MJ1 require consideration of an agent's knowledge.
An agent's attitude is likewise relevant to moral judgment. The harming chairman's action is bad in part because he displayed absolutely no concern for the environment. When an agent regrets or feels bad about a negative outcome, people are markedly less likely to say the action was intentional (Cushman & Mele Reference Cushman, Mele, Knobe and Nichols2008; Guglielmo & Malle, in press; Phelan & Sarkissian Reference Phelan and Sarkissian2008). On Knobe's account, therefore, such actions also must not be bad, suggesting that MJ1 require consideration of an agent's attitude. Accordingly, MJ1 substantially depend on mental state information, particularly regarding beliefs (that the agent know about a negative outcome) and attitudes (that the agent not care about the outcome). These two elements are widely recognized as essential inputs to moral judgment (Cushman Reference Cushman2008; Darley & Shultz Reference Darley and Shultz1990; Guglielmo et al. Reference Guglielmo, Monroe and Malle2009; Young & Saxe Reference Young and Saxe2009).
Even if one grants that morality impacts mental state judgments, this effect appears to be a special case of norm violation more generally. In fact, the same empirical patterns on which Knobe's account is based are also found for cases of norm violations that have nothing whatsoever to do with morality (Machery Reference Machery2008). For example, people judged an agent's making of black toys to be more intentional when doing so violated, rather than conformed to, the conventional color designation (Uttich & Lombrozo Reference Uttich and Lombrozo2010). People also judged it more intentional to violate a dress code than to conform to one (Guglielmo & Malle, in press). This is because norm violations – whether moral or not – provide diagnostic information about a person's disposition, motives, intentions, and so on (Reeder & Brewer Reference Reeder and Brewer1979; Skowronski & Carlston Reference Skowronski and Carlston1989). Interestingly, Knobe's recent work adopts precisely this explanation, highlighting the impact of non-moral norms on causality judgments (Hitchcock & Knobe Reference Hitchcock and Knobe2009). But this perspective suggests that people are not “moralists” at all; rather, their judgments are sensitive to norms, just as those of a “scientist” would be.
Setting aside the criticisms raised here, Knobe should clarify a puzzling aspect of his proposed account. Knobe distinguishes between early MJ1 (badness judgments) and later MJ2 (e.g., blame judgments). One might expect these judgments to be tightly linked, as they both assess the morality of a given action. However, the connection between them is argued to be fully mediated by non-moral assessments (e.g., regarding mental states and causality, see Figure 5 of the target article). Knobe's account would benefit from a psychological explanation for the existence of such a circuitous path between the conceptually similar MJ1 and MJ2. Why might it be that two moral judgments have no direct relationship to each other?
One possible answer to the puzzle is that MJ1 are not actually moral judgments, but simply judgments about whether an action violated an expectation. Such expectations are sometimes a function of valence – people expect others to bring about positive events and avoid negative ones (Pizarro et al. Reference Pizarro, Uhlmann and Salovey2003). Perceivers may adopt different thresholds for what constitutes a relevant mental state or causal role, depending on the extent to which the action violates expectations. This possibility is largely consistent with Knobe's discussion of default attitude positions (Figs. 8 and 9 of the target article), except that Knobe maintains the threshold is set by moral judgments in particular. Given the discussion here, it is not clear how this can be true. Although expectations (including, but not limited to, those concerning valence) may impact the evidential threshold set by perceivers, moral judgments depend on assessments of an agent's mental states. Accordingly, the claim that such assessments are “suffused with moral considerations” (sect. 5.3, para. 3) is greatly overstated.
Knobe's “person as moralist” account provides a novel contribution to the study of human morality. Whereas most research in this domain has examined the features of behavior that guide moral judgment (Cushman Reference Cushman2008; Guglielmo et al. Reference Guglielmo, Monroe and Malle2009; Shaver Reference Shaver1985) or the processes that underlie moral judgment (Greene Reference Greene and Sinnott-Armstrong2008; Haidt Reference Haidt2001), Knobe's target article extends the literature by probing the influence of morality on other psychological judgments.
Despite its promise, however, Knobe's account has several limitations. Knobe neither measures nor defines moral judgment, leaving it unclear precisely what the account posits and how it may be falsified. Nonetheless, any conceptualization of moral judgment consistent with Knobe's account necessarily relies on substantial information about an agent's mental states. Moreover, the results described by Knobe are likewise obtained by instances of entirely non-moral norm violations. Finally, Knobe should clarify why it would be the case that people's moral judgments of badness and blame share no direct relationship.
The crux of Knobe's argument is that moral judgments of badness (hereafter “MJ1”) impact judgments about an agent's mental states and causal role, which thereby impact moral judgments of blame (hereafter “MJ2”). Although MJ2 are often measured, studies of Knobe's account rarely (if ever) measure MJ1. It is therefore critical to know the conditions under which such judgments arise. To this end, Knobe claims to examine “judgment[s] that the agent's action itself is bad.” But this definition does not provide much clarity – if MJ1 are not simply judgments about bad outcomes (sect. 4.1.3, para. 4), are they judgments that an agent caused /knew about/ intended something bad? Absent either a measurement or definition of MJ1, it is unclear precisely what is alleged to influence mental state and causality assessments, and how one could attempt to falsify the account.
In any case, Knobe's account would be most compelling if MJ1 arise in the absence of any considerations of the agent's mental states (which, after all, are proposed to be influenced by MJ1). However, this is clearly not the case. First, the agent's knowledge is relevant to these moral judgments. For example, the harming chairman's action is bad in part because he knew that harm would occur. When agents lack knowledge of the harmful consequences of their action, people no longer view the consequences as intentional (Nadelhoffer Reference Nadelhoffer2006b; Pellizzoni et al. Reference Pellizzoni, Girotto and Surian2010). According to Knobe's account, therefore, such actions must not be bad. But if this is true, then MJ1 require consideration of an agent's knowledge.
An agent's attitude is likewise relevant to moral judgment. The harming chairman's action is bad in part because he displayed absolutely no concern for the environment. When an agent regrets or feels bad about a negative outcome, people are markedly less likely to say the action was intentional (Cushman & Mele Reference Cushman, Mele, Knobe and Nichols2008; Guglielmo & Malle, in press; Phelan & Sarkissian Reference Phelan and Sarkissian2008). On Knobe's account, therefore, such actions also must not be bad, suggesting that MJ1 require consideration of an agent's attitude. Accordingly, MJ1 substantially depend on mental state information, particularly regarding beliefs (that the agent know about a negative outcome) and attitudes (that the agent not care about the outcome). These two elements are widely recognized as essential inputs to moral judgment (Cushman Reference Cushman2008; Darley & Shultz Reference Darley and Shultz1990; Guglielmo et al. Reference Guglielmo, Monroe and Malle2009; Young & Saxe Reference Young and Saxe2009).
Even if one grants that morality impacts mental state judgments, this effect appears to be a special case of norm violation more generally. In fact, the same empirical patterns on which Knobe's account is based are also found for cases of norm violations that have nothing whatsoever to do with morality (Machery Reference Machery2008). For example, people judged an agent's making of black toys to be more intentional when doing so violated, rather than conformed to, the conventional color designation (Uttich & Lombrozo Reference Uttich and Lombrozo2010). People also judged it more intentional to violate a dress code than to conform to one (Guglielmo & Malle, in press). This is because norm violations – whether moral or not – provide diagnostic information about a person's disposition, motives, intentions, and so on (Reeder & Brewer Reference Reeder and Brewer1979; Skowronski & Carlston Reference Skowronski and Carlston1989). Interestingly, Knobe's recent work adopts precisely this explanation, highlighting the impact of non-moral norms on causality judgments (Hitchcock & Knobe Reference Hitchcock and Knobe2009). But this perspective suggests that people are not “moralists” at all; rather, their judgments are sensitive to norms, just as those of a “scientist” would be.
Setting aside the criticisms raised here, Knobe should clarify a puzzling aspect of his proposed account. Knobe distinguishes between early MJ1 (badness judgments) and later MJ2 (e.g., blame judgments). One might expect these judgments to be tightly linked, as they both assess the morality of a given action. However, the connection between them is argued to be fully mediated by non-moral assessments (e.g., regarding mental states and causality, see Figure 5 of the target article). Knobe's account would benefit from a psychological explanation for the existence of such a circuitous path between the conceptually similar MJ1 and MJ2. Why might it be that two moral judgments have no direct relationship to each other?
One possible answer to the puzzle is that MJ1 are not actually moral judgments, but simply judgments about whether an action violated an expectation. Such expectations are sometimes a function of valence – people expect others to bring about positive events and avoid negative ones (Pizarro et al. Reference Pizarro, Uhlmann and Salovey2003). Perceivers may adopt different thresholds for what constitutes a relevant mental state or causal role, depending on the extent to which the action violates expectations. This possibility is largely consistent with Knobe's discussion of default attitude positions (Figs. 8 and 9 of the target article), except that Knobe maintains the threshold is set by moral judgments in particular. Given the discussion here, it is not clear how this can be true. Although expectations (including, but not limited to, those concerning valence) may impact the evidential threshold set by perceivers, moral judgments depend on assessments of an agent's mental states. Accordingly, the claim that such assessments are “suffused with moral considerations” (sect. 5.3, para. 3) is greatly overstated.