Crime and punishment: Distinguishing the roles of causal and intentional analyses in moral judgmentIntroductionGeneral methodsData collectionStatistical methodsExperiment 1MethodsResultsDiscussionExperiment 2MethodsResultsDiscussionExperiment 3MethodsResultsDiscussionExperiment 4MethodsResultsDiscussionGeneral discussionConclusionsAcknowledgementsReferencesCrime and punishment: Distinguishingthe roles of causal and intentionalanalyses in moral judgmentFiery CushmanDepartment of Psychology, Harvard University, 1120 William James Hall,33 Kirkland Street, Cambridge, MA 02138, United StatesReceived 9 March 2007; revised 29 February 2008; accepted 4 March 2008AbstractRecent research in moral psychology has attempted to characterize patterns of moral judg-ments of actions in terms of the causal and intentional properties of those actions. The presentstudy directly compares the roles of consequence, causation, belief and desire in determiningmoral judgments. Judgments of the wrongness or permissibility of action were found to relyprincipally on the mental states of an agent, while judgments of blame and punishment arefound to rely jointly on mental states and the causal connection of an agent to a harmful con-sequence. Also, selectively for judgments of punishment and blame, people who attempt butfail to cause harm more are judged more leniently if the harm occurs by independent meansthan if the harm does not occur at all. An account of these phenomena is proposed that dis-tinguishes two processes of moral judgment: one which begins with harmful consequences andseeks a causally responsible agent, and the other which begins with an action and analyzes themental states responsible for that action.Ó 2008 Elsevier B.V. All rights reserved.Keywords: Morality; Moral psychology; Punishment; Attribution theory; Intentional action; Theory ofmind; Causation0010-0277/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.cognition.2008.03.006E-mail address: [email protected]/locate/COGNITAvailable online at www.sciencedirect.comCognition 108 (2008) 353–3801. IntroductionOn a snowy January Sunday, Hal and Peter watch football and share beers at alocal bar. Both drive away intoxicated, and both lose control of their cars on theslick roads. Hal collides with a neighbor’s tree, but Peter collides with a young girlplaying in the snow. In the state of Massachusetts, Hal can expect a $250 fine fordriving under the influence of alcoh ol. Peter faces a minimum of 2.5 years in prison– and up to 15 years – for vehicular manslaughter.Cases like this have long puzzled philosophers and legal scholars (Hall, 1947; Hart& Honore, 1959; McLaughlin, 1925; Nagel, 1979; Williams, 1981), and it is easy tosee why. Hal and Peter seem to have engaged in equally wrongful behavior, but poorluck in the case of Peter leads to a punishment that is radically more severe . Yetwhile we might be tempted to propose equal punishments for both drivers, it doesn’tseem right to let Peter off with a $250 ticket for killing a girl, or to send Hal to prisonfor 2.5–15 years for hitting a tree. What cases like this reveal is a complex interactionof two different factors in our moral judgments: the assessment of causal responsibil-ity for harm versus the assessment of intent to harm (along with related mental statessuch as beliefs, desires, and negligence). In particular, our judgments of the moralwrongness of a behavior seem to rely principally on an agent’s mental state, whileour judgments of deserved punishment show greater sensitivity to the harms actuallycaused by the agent.This conceptual distinction between causal and intentional factors is recapitu-lated in the psychological literature on moral judgment. Researchers in the cogni-tive development tradition of moral psychology pioneered by Piaget haveemphasized the dominance of intentional factors over causal factors in adult moraljudgment (e.g. Hebble, 1971; Baron & Ritov, 2004; Shultz, Wright, & Schleifer,1986; Yuill & Perner, 1988; Zelazo, Helwig, & Lau, 1996). The adult pattern ofjudgments is the outcome of a developmental shift: the moral judgments of youngchildren are dominated by a causal analysis of harmful actions, while the moraljudgments of older children and adults focus instead on the intention to produceharm. Critically, harmful intentions alone are found to be sufficient to warrantmoral reprobation in mature children and adults, even in the absence of any harm-ful consequence.This characterization of mature moral judgment stands in contrast to the basicmodel proposed in attribution theory, which owes greatly to a foundational workof Heider (1958). Attribution theorists have typically suggested that adult moraljudgment begins by analyzi ng causal responsibility, only subsequently proceedingto an analysis of intention (Darley & Shultz, 1990; Fincham & Roberts, 1985; Heid-er, 1958; Shaver, 1985; Shultz, Schleifer, & Altman, 1981; Weiner, 1995). A centralclaim of this tradition is that, absent harmful consequence, malicious intentions areinsufficient to trigger moral judgments of moral responsibility, blame and punish-ment in adults (reviewed in Weiner, 1995). As Darley and Shultz (1990) write, ‘‘judg-ments of moral responsibility presuppose those of causation. If the protagonist isjudged not to have cau sed the harm, then there is no need to consider whether heis morally responsible for it.”354 F. Cushman / Cognition 108 (2008) 353–380Notably, the cognitive development and attribution literatures have adopted dif-ferent dependent measures as their primary focus. Research in the attribution tradi-tion usually asks subjects to evaluate the level of blame, moral responsibility andpunishment deserved by an agent (e.g. Darley, Klosson, & Zanna, 1978; Fincham& Jaspers, 1979; Fincham & Shultz, 1981; Shultz et al., 1981). By co ntrast, researchin the Piagetian tradition is more likely to ask subjects to judge whether an agent hasbehaved badly, wrongly, or naughtily (e.g. Hebble, 1971; Imamoglu, 1975; NelsonLe Gall, 1985; Piaget, 1965/1932; Wellman, Cross, & Bartsch, 1986). Of course, thisparallels our intuitions in the cases of Hal and Peter: each seems to have actedequally wrongly, but on the basis of consequences their behaviors merit differentpunishments.Experiments 1 and 2 of the present study build on this background, directly test-ing whether judgments of blame and punishment show enhanced sensitivity