Printer Friendly

Blaming as a social process: the influence of character and moral emotion on blame.



When we assign blame for something bad that happened, we are doing something social--we are identifying another human being who caused harm, without justification or excuse. A window broken by a hurricane elicits a story about cause, but not a story about blame; a window broken by a person elicits blame attribution. Once a human agent is identified, we naturally turn our attention to blame severity, a complex judgment shaped by several different concerns. A window broken by a child's stray baseball is assessed differently from a window broken by a vandal, or by a burglar, or by a white supremacist. Assessing blame involves not only determining the badness of the harm (for example, property damage versus injured person), but also the badness of the actor's mental state (for example, accident versus intentional), and perhaps even the badness of the actor's motive (for example, general destructiveness versus racial hatred). When viewed this way, we see that blame--as a psychological matter--involves attributions about other people and the extent to which they intend to harm us or otherwise pose a threat to the social order. In this sense, blaming is social because it is about attributions of other people and their intentions.

The law takes account of each of the blame dimensions just mentioned. The criminal law reserves more severe offense categories for more severe harm. It imposes more serious liability when mens rea is more culpable, all else being equal. And sometimes, as is the case with hate crimes, it explicitly takes into account the actor's motive for causing harm. But there is another possible influence on blame not yet mentioned--an influence which the law has sought to minimize. The moral character of the actor, apart from that actor's motive or reasons for acting, might play an important role--as a descriptive matter--in the psychological process of blame. Yet, for the most part, the law eschews the role of moral character in legal blame. (1)

In a previous article, my colleague and I explored empirically the question of motive for acting, and how it produces inferences about moral character that influence blame judgments. (2) By contrast, the first overarching goal of the current article is to provide experimental evidence supporting the hypothesis that psychological blame is influenced by perceptions of the actor's overall virtue or lack thereof, even apart from the actor's reason for acting in the specific instance. This article marshals experimental evidence to support the idea that a person with a flawed moral character is blamed more for causing harm than a person who is otherwise virtuous. Thus, we are likely to blame more severely a drug-addicted high school dropout who knocks down ten rural mailboxes with a baseball bat than an A-student who is on the chess team who engages in the same action. Similarly, we are likely to blame more severely an abusive parent who drives recklessly and unintentionally hits a pedestrian than a model parent who performs the same act with the same mens rea. The experiment reported here implements a rigorous test of this hypothesis and shows that even mildly unpleasant character traits, such as unreliability, can lead observers to blame more harshly, and to bolster these harsh blame judgments with increased perceptions of the actor's causal role and his intent to cause harm.

The findings reported here echo research examining criminal cases suggesting that the defendant's prior criminal record can influence jury verdicts. For example, in cases where evidence is weak, there is a positive correlation between the jury learning that the defendant had committed prior crimes and the likelihood of conviction. (3) This suggests that in the absence of compelling evidence tending to prove guilt, juries sometimes use the fact that defendants had committed prior crimes as a reason to think they might be guilty of the crimes in question. Existing experimental research also suggests that information about prior crimes can increase the likelihood of criminal conviction. (4) These findings are important, but they focus on the narrow question of the influence of prior crimes on verdicts in criminal cases. By contrast, the focus of this article is more expansive. First, the focus here is on the influence of moral character broadly, not just on inferences drawn from prior crimes. (5) Instead of examining the effect of an emotionally provocative criminal violation of the social contract, such as the influence of the defendant's prior armed robbery, this article focuses on subtle cues about the actor's character traits, such as generosity and reliability, and shows that even mild virtue deficits lead to more severe blame judgments. Second, the focus of this article is on blame generally, not just on verdicts; it is on the basic social-psychological question of how blaming processes work, both in court and out of court, and both within the law and in everyday social life. To that end, rather than examining the influence of a person's prior criminal record, as previous scholars have done, I instead examine the influence of subtle cues about character, such as whether someone works hard or goofs off on the job.

In addition to testing a more expansive notion of moral character, a second goal of this article is to explore the extent to which the influence of moral character varies with the actor's mental state. For example, is a negligent actor with a bad character blamed just as much as a reckless actor with a good character? Mental states, after all, are often difficult to know with precision. A driver who kills a small child after he runs a stop sign while he is dialing his cell phone is reckless (6) if he is conscious of and disregards the risk that someone might die; he is negligent (7) if he was merely careless without consciously disregarding this risk. (8) But it is difficult to glean another person's mental state with such precision. If we think the driver might have been aware of the risk of death, knowing that the driver is a person of poor moral character might be enough to push us toward a more severe blame judgment. Conversely, knowing the driver is an otherwise virtuous person might pull us in the other direction, toward less severe blame. In this way, moral character might serve as a kind of proxy for mental state, so that a person with a bad character is blamed as if he were reckless, whereas a person with a good character is blamed as if he were not reckless. The experiment reported here tests this idea.

A third focus of this article is moral emotion as the mechanism through which moral character influences blame. (9) When we observe a moral violation, we often react emotionally. When we hear about a thug who snatches a purse from a vulnerable elderly victim, we feel angry. When we hear about the trauma experienced by a sexual assault survivor, we feel sympathy. When we learn about the sadistic murder of a child, we feel disgust. Experiencing moral emotions can cause us to make harsher moral judgments. (10) The experiment reported in this article seeks first to establish the link between perceiving someone's moral character and experiencing emotion. Specifically, the experiment tests the idea that, compared to an otherwise virtuous person who causes harm, an otherwise bad person who causes harm makes us feel angrier and more disgusted, which in turn leads to more punitive attributions of blame. These emotion-charged blame judgments have potentially important implications for basic social-psychological mechanisms involving blame, and can inform our understanding of blame both within the legal system and outside of it in everyday social life. This process of inculpation, infused with emotion and driven by character inferences, can help to explain the charging decisions of prosecutors, the claiming decisions of injured people, the settlement behavior of litigants, and the punitive damages decisions of juries, among many other examples.

Prior empirical research on the influence of prior conviction on legal judgments of guilt has focused on the "whodunit" question: Given the defendant's prior record and other evidence presented, what is the probability that he committed the offense in the instant case? By contrast, the focus of the study presented in this article is quite different: Given that the actor undisputedly performed a specific act causing a particular result, with a specifically described mental state, what is the marginal influence, if any, of the actor's moral character on observers' judgments of blame? Whereas earlier empirical work on prior conviction focused on the probability that the defendant committed the offense, the current study focuses on character-driven, emotionally-infused blame judgments, given that the actor did cause the posited harm. The results of the current study suggest that perceptions of moral character and resulting moral emotions not only drive blame judgments but also the putatively fixed judgments of causation and intent as well.



Consider the following scenario (11): John was speeding to get home, driving 40 miles per hour in a 30-miles-per-hour zone. He came to an intersection and applied the brakes but was unable to stop in time because of an oil spill on the road. (12) John hit another car in the intersection, injuring the other driver. John was speeding home in order to:

(a) Hide from his parents an anniversary present for them that he had left out in the open, OR

(b) Hide from his parents a vial of cocaine that he had left out in the open.

When people think about this scenario, those who learn that John was hiding cocaine think that he was more responsible for the accident, on average, than those who learn that John was hiding a present. (13) From the perspective of legal responsibility, this is an odd result, given that John's behavior, the other driver's behavior, and the accident scene are the same in each scenario. John's state of mind, of course, was different--in the sense that the object that he sped home to hide was different. But comparing the two scenarios, John's state of mind was not different in any legally relevant way. If John is liable for the accident, it is because he was negligent--he should have foreseen that his speeding caused a risk of harm. The foreseeability of the harm is the same in each scenario, as is John's failure to exercise reasonable care.

Besides differences in responsibility judgments, there is another oddity that emerges in reactions to this story. Not only do people think John is more responsible, but they also think he is more of the cause of the accident when the object he was thinking about hiding was cocaine, rather than a present. (14) From the perspective of legal causation, this also presents something of a puzzle. The chain of events between the accident and the injury is exactly the same in each scenario. The risk of colliding and injuring another person does not differ from one scenario to the next, and neither does John's breach of the standard of due care. What does differ, however, is John's motivation for engaging in the act that led to the collision. In the scenario involving the cocaine, John's motivation for speeding was nefarious; in the scenario involving the present, John's motivation for speeding was laudable (despite the fact that it does not outweigh the foreseeable risk of injury).

In light of the drug-related nature of John's nefarious goal in the cocaine scenario, one might object that, in fact, people might have imagined his conduct to be different from the anniversary present scenario. Perhaps people thought that John was under the influence of cocaine at the very moment of the collision, thus making him more responsible. Or, perhaps John's goal of hiding cocaine was more distracting than the goal of hiding a present, so that in the cocaine scenario people interpret John as having paid less attention. Or, perhaps in the cocaine scenario people imagined John driving faster than in the anniversary present scenario--even though the speed was specified to be the same in both versions--so that John seemed more responsible and his actions more causal for those reasons.

There is further evidence, however, supporting the hypothesis that an actor's bad motive can result in harsher judgments of blame and causation for harm. In previous work, (15) we presented people with a scenario involving a man--Sam--who stored oxygen tanks in his backyard shed. When a cigarette butt landed near the shed, the tanks accidentally exploded and a neighborhood youth was killed. Just as blame and causation judgments were harsher for John the driver when he had a bad motive for speeding, judgments of blame and causation in this study were harsher for Sam when he had a bad motive for storing oxygen tanks in his shed. Specifically, if Sam stored oxygen because he was a high school football coach who cheated by giving oxygen to his players, he was judged to be more responsible, more blameworthy, and his role more causal in the accidental death of the youth compared to if he stored oxygen to care for his sick daughter. (16)

In the oxygen study, it is implausible that participants perceived that there was any greater danger or risk of oxygen explosion in the football scenario than in the sick daughter scenario. Rather, it must have been Sam's bad motive that prompted people to perceive him as more blameworthy, responsible, and causal in the football scenario. Perhaps, however, these differences in blame and perception of causation are not so anomalous from a legal perspective. After all, the social value of the conduct of storing oxygen tanks to care for a sick child is greater than that of cheating at football. And the social value of speeding home to surprise loved ones with a gift is greater than that of hiding illegal drugs. When the social value of conduct is sufficiently great, this can justify the risk that is assumed by the actor. In tort law, a reasonable person considers the risk of harm in light of the utility of the conduct. (17) Thus, a vaccine that carries a risk of harming 1 in 1000 children might be considered reasonable given the social value of the vaccine; a piece of candy that carries the same risk of harm would not.

Like tort law, criminal law takes into account the extent to which the actor's risky behavior was justified when assigning liability for crimes involving recklessness. The Model Penal Code test focuses on "the nature and purpose" of the actor's conduct and its surrounding circumstances. (18) Depending on the nature and purpose of the conduct, the actor's disregard of the associated risk might be considered consistent with the standards of a law-abiding person, on the one hand, or a gross deviation from those standards, on the other. In the latter case, the actor would be deemed to have acted with a reckless state of mind. For example, a surgeon who disregards a substantial risk of death to the patient in performing a dangerous operation might be justified in doing so if the purpose of the conduct is to save the patient's life, and there is no safer way to do so. On the other hand, if the purpose of the conduct is something other than saving the patient's life (for example, testing a new surgical technique), it is unlikely that the risk of harm would be justified.

In light of the legal standards for risky conduct in tort law and criminal law, it is perhaps not so anomalous that in the two studies described, people would consider John the cocaine hider more blameworthy than John the present hider, and Sam the football cheater to be more blameworthy than Sam the father of the sick child. Putting to the side the issue of blameworthiness, it is perhaps still puzzling why people would see an actor with a bad motive more causal in the chain of events leading to harm. As a general matter, there is evidence that causal proximity influences blame judgments. (19) Thus, harm that occurs by touching or other personal contact with the victim is judged to be less morally permissible than the same harm that occurs without personal contact. (20) But in the cocaine--present study, and in the oxygen tank study, causal proximity does not seem to vary between the good motive and bad motive scenarios. To the contrary, in both versions of the scenario in both studies, there is really no question about the extent to which the actor caused the harm. In the cocaine-present study, John failed to stop his car at the stop sign and collided with another car. To be sure, there were other causal factors contributing to the accident. (21) But under any plausible account of causal explanation for the harm, John's conduct was a primary cause of the harm, and the proximity of his actions to the harm was the same in each version of the scenario. Similarly, in the oxygen tank study, Sam's conduct was a primary cause of the harm, and the proximity of his actions to the harm was the same in each version of the scenario. In both studies, therefore, there seems to be no obvious, straightforward explanation as to why having a bad motive would transform conduct into a stronger cause of harm.

Mark Alicke's theory of culpable control represents one attempt to explain this puzzle. The theory posits that when people assess blame, they try to gauge the extent to which the actor exercised control over the outcome. (22) If an actor acts intentionally and that action directly causes the outcome, then the actor is perceived to have high control; conversely, if there is an absence of intent and there was a long chain of events that caused the outcome, then the actor is perceived to have low control. How much intent did the actor have? How strongly causal was the actor's role in the harm? It is well understood that these judgments often are made under uncertainty. (23) Alicke argues that perceptions of intent and harm are directly influenced by our initial affective reaction to the harm situation. (24) Once strong negative reactions are evoked by harmful conduct, then people's desire to blame kicks into gear, and their assessment of factors like foreseeability, intent, and causation is colored by their motivation to understand the conduct as highly blameworthy. (25) On this account, people blame early, then justify the blame assessment by pointing to corresponding levels of foreseeability, intent, and causation. Thus, John the cocaine hider evoked a strong negative reaction, leading to perceptions of high blameworthiness for the accident and injury, and correspondingly high perceptions of causation to justify the blame attribution after the fact. John the present hider evoked a reaction that was less strongly negative, so less extreme judgments of blame and causation followed. Similarly, on this account, Sam the football cheater evoked a strong negative reaction, leading to perceptions of high blameworthiness for the youth's death, and correspondingly high perceptions of causation to justify the blame attribution; the reactions to Sam the father of the sick daughter were less negative. On this account we engage in "blame validation": We make blame attributions spontaneously according to how strongly negative our gut reaction is then we validate our blame assessment by tuning evaluations of causation and intention accordingly.

Alicke's blame-validation model posits that we have an immediate negative reaction to an actor with a bad motive who caused harm, which leads to a fast, automatic initial blame judgment. (26) This initial blame judgment then guides subsequent perceptions about the actor's causal role in producing the harm. The extent to which this type of blame-early account accurately describes moral decisionmaking is a matter of dispute among social psychologists and experimental philosophers. (27) Alicke's model is a version of a blame-early model. By contrast, the blame-late models posit that once a perceiver detects a harmful event, she assesses causation and mental state before coming to a conclusion about blame. (28) More research will be required in the future to sort out the causal order of blame and related attributions. Regardless of whether blame in fact comes early or late, there remains strong evidence that an actor's bad motive for acting (for example, hiding drugs, cheating at football) can influence perceptions of causation.

We have seen that bad motives can increase attributions of blame and causation. We have also seen that law takes into account motive as sometimes relevant to whether an actor is justified in disregarding a risk of harm. Therefore, at least with regard to blame, the law takes account of people's natural inclination to blame according to the actor's motive, even if the motive-causation connection is not explicitly accounted for in law. But reason for acting is only one example of how people's initial evaluation of an actor might push them toward harsh judgments of blame or pull them toward lenient judgments. There are many other variables that enter into the blame equation. Some variables are more familiar, such as the severity of the harm, or the actor's intent. But other potential antecedents of blame are considered more problematic, at least by some theorists. On the one hand, it is well settled that an actor's race, for example, or nationality, or religion should not influence our perceptions of blame for harm caused by that actor. (29) On the other hand, law sometimes permits consideration of an actor's motive. (30) In this vein, consider another characteristic: the actor's moral character. The fact that John the cocaine hider was perceived as more blameworthy and causal than John the present hider might have been as much a function of John being derogated as a drug user as John's motives being derogated as illegitimate. Thus, an actor's bad motive might simply establish that the subsequent harm was not justified by the nature and purpose of the actor's conduct, as the law explicitly recognizes. Alternatively, an actor's bad motive might serve as a proxy for that actor's bad moral character, which is derogated and leads perceivers to blame more harshly.

There is evidence for the proposition that out perception of a person's motive is important for our understanding of that person's moral character. For example, a soccer player who intentionally spikes a player on the opposing team is seen as being less moral as a person when his motive was for personal gain (trying to win the game) than when his motive was reactionary (retaliation for being recently bumped and taunted by the other player). (31) Similarly, a person who administers an electric shock to another person for monetary gain is rated less moral than a person who shocks another person to avoid being shocked himself. (32) The influence of motive on perceptions of moral character can sometimes be dramatic, transforming an outlaw into a hero, or at least an anti-villain. Consider, for example, the television character named Omar Little on the HBO series The Wire. Omar robs and kills drug dealers for a living, but at the same time upholds his own moral code, which entails not harming law-abiding citizens, and no thuggery on Sundays. Because there is a kind of purity in his motives, Omar garners more moral admiration than condemnation. Motive, therefore, can make or break perceptions of moral character.

In everyday social life, sorting the "bad guys" from the "good guys" plays an important social function. When we encounter a new person, we immediately size up her "perceived intent"--traits like friendliness, trustworthiness, and helpfulness signal whether the other is a friend or foe. (33) We make these judgments remarkably quickly. In one study, for example, when a photo of a face was flashed for a fraction of a second, people were most reliable at judging trustworthiness, compared to other traits like competence and likeability. (34) We are especially keen to size other people up first on moral traits, and only later on non-moral traits (such as competence and intelligence), perhaps because the friend-or-foe question is an especially important one when encountering a stranger. (35)

The primacy of moral character judgments suggests that these concerns might infuse the process of blaming. That is, when we are deciding the extent to which to blame an actor who has caused harm, we might be implicitly asking ourselves, "To what extent is this actor a bad person?" rather than (or perhaps in addition to), "To what extent is this particular action wrong?" (36) Because blaming might be focused to a large extent on perceptions of the actor's moral character, then anything that influences out perceptions of the actor's moral character is likely to influence our judgments of blame for harm. Character perceptions might be gleaned from the act itself. For example, consider a large man walking down a narrow hallway who bumps a passerby with his shoulder, and then calls the passerby a derogatory name. Those actions, in themselves, are likely to trigger negative perceptions of the actor's moral character. If blame judgments derive primarily from character judgments, then the blame process is relatively simple: only a bad person would do such a thing, so this actor is to blame. If blame judgments are constructed by assessing various factors surrounding the act, then the process is quite different: we examine the extent to which the harm was intentional, the extent to which the actor caused it, and the severity of the harm, and then decide whether the act was wrong, and finally whether the actor had any excuse or justification. The latter process can be thought of as act-based blaming, as opposed to character-based blaming. The result may be the same (at least in some cases), but the focus is quite different. The main point to note here is that act-based blaming does not consider the actor's moral character.

The study of John the driver suggests that we do, in fact, engage in character-based blaming, at least insofar as information about John's motive for speeding (hiding cocaine or hiding a present) informs our perception of John's moral character. The case of Sam--the man who stored the oxygen tanks in his shed--illustrates even more clearly that we engage in character-based blaming. In our experiment, not only did we find that people blame Sam more harshly when his motive was cheating rather than caretaking, but we also round that Sam's motive influenced perceptions of Sam's moral character: Sam who stored oxygen to cheat was judged to have less of a moral character, be less trustworthy, and be more of a bad person than Sam who stored oxygen to take care of his sick daughter. (37) Thus, to judge moral character people sometimes use information derived from the act itself (as in the shoulder bump example), and they sometimes use information derived from the motive for the act (as in John the driver and Sam the oxygen storer). In either case, to the extent that the blaming process focuses on the actor rather than the act, inferences about character will inform perceptions of blame, no matter how they were derived.

In this article, I expand the inquiry begun earlier regarding the relationship between moral character and blame. Specifically, I investigate the extent to which information about moral character influences blame even when that moral character information is independent from the harmful act. For example, if Joe kills a man in a bar fight, do we blame him less if we also learn that Joe spends his free time as a volunteer tutor for underprivileged youth? Conversely, do we blame him more if we learn that Joe is a greedy miser who refuses to help support his elderly mother? There is already some evidence that suggests that independent information about moral character can influence blame judgments. (38) In one study, participants read a story about Jack, who argued with and then pushed another man. The man slipped and suffered serious injuries. Jack was presented as someone who helped a stranded motorist, gave extra time off to a worker, and volunteered at a homeless shelter, or alternatively, as someone who drove by the motorist, denied the time off to the worker, and made an excuse to not volunteer at the homeless shelter. People rated "bad Jack" as more to blame for the victim's injuries than "good Jack."

I sought to replicate and extend this finding. First, in the experiment reported in this article, I focused on the influence of independent evidence of character on judgments of blame when the severity of the harm is extreme--that is, death. It is possible that when the resulting harm is grave, the impulse to blame becomes so strong that it overwhelms any potential influence of independent information about moral character. The plausibility of this hypothesis is supported by numerous experimental studies that consistently find that actors are blamed more harshly for severe harm than for mild harm, even when other circumstances are held constant. (39)

An alternative, competing hypothesis is that independent information about moral character influences blame judgments even for severe outcomes such as death. This hypothesis is supported by the results of the study involving Sam, who was round more blameworthy for an accidental death when his motive for storing the oxygen was bad than when it was good. In that study, bad motive led to perceptions of bad moral character, which led to increased blaming. By contrast, in the experiment reported here, moral character is derived not from motive, but from information largely independent of reasons for acting. According to this competing hypothesis, information about moral character will influence judgments of blame for a severe outcome regardless of whether the moral character inference derives from the actor's motive or from an independent source.

The second aim of the experiment reported in this article was to examine variations in the actor's state of mind, to ascertain whether the influence of moral character information varies according to the actor's level of awareness of the risk of harm (for example, reckless versus negligent). It is well established empirically that blame judgments generally increase with increasingly culpable states of mind, (40) so that an actor who causes harm recklessly, for example, is perceived as more blameworthy than one who causes harm negligently. (41) The joint influence of bad moral character and culpable mental state might be additive. It might be the case, for example, that a negligent actor with a bad moral character is blamed just as harshly as a reckless actor--in this sense, moral character becomes a kind of proxy for a more culpable mens rea. In addition, a reckless actor with good moral character might receive more lenient treatment in terms of perceptions of blameworthiness, and be blamed only as much as a negligent actor. The experiment reported here tests these possibilities.

A related hypothesis regarding moral character and mental state is that people's perceptions of mental state themselves are influenced by the actor's moral character. For example, people might perceive an act as intentional precisely because it was performed by an actor with a bad moral character, whereas that same act might be perceived as merely reckless if the actor's character is good (or if character information is absent). After all, mental state is rarely if ever completely knowable. In making the uncertain determination about mental state, people might use information about character to fill in the blanks. (42) As Pizarro and Tannenbaum put it, "Bad people, by definition, are likely to desire and intend bad things." (43) The epistemic elusiveness of the mental states of others suggests the possibility that our judgments about those mental states are vulnerable to a variety of influences. Experimental philosopher Joshua Knobe argues that intentionality judgments are more than just factual descriptions of others' mental states--they are also normative judgments about praise and blame. (44) Experimental evidence supports this view, and suggests that moral judgments about an act and its outcome can affect perceptions about the extent to which that act was performed intentionally. For example, people are more likely to say that a company executive intentionally caused an outcome if the outcome was negative (harming the environment) than if the outcome was positive (helping the environment). (45) This is true despite the fact that in both versions of the vignette, the executive said that he did not care about harming or helping the environment, he only cared about making a profit. This suggests that people perceive the executive who caused the negative outcome as a bad person who must be punished, and in an effort to justify that punishment, they interpret his act as intentional. (46) The possibility that intentionality judgments are infused with moral judgments suggests that the moral character of the actor influences judgments of intentionality. Unlike Knobe's intentionality studies, we vary the moral character of the actor, rather than the valence of the outcome.

Finally, in this study we investigate the influence of emotion on blame. Moral violations provoke emotional reactions in observers. (47) For example, when considering the moral quandary regarding whether one ought to push a large stranger off a footbridge to his death in order to save five people from a runaway trolley, people become very emotionally engaged at the thought of pushing someone to their death. (48) By comparison, people are less emotionally engaged at the thought of flipping a switch to redirect a runaway trolley to save rive people, even though, as before, one person will be killed as a result. (49) Depending on the features of the situation, thinking about harm can engage a variety of emotionally laden processes.

Witnessing behavior that poses a threat to the integrity of the social order typically triggers contempt, anger, disgust, or some combination of these. (50) These moral emotions are associated with disapproval of others' behavior, and they arise spontaneously upon observing an actor causing harm. The experience of contempt, anger, or disgust triggers the response tendency to blame and punish. (51) In fact, when blame and punishment goals are frustrated, "moral spillovers" can occur in which observers who learn about an unpunished transgressor become more punitive toward unrelated acts of harm, (52) or even engage in deviant behavior themselves. (53) Thus, another aim of this study is to demonstrate that a harmful outcome caused by an actor with a bad moral character provokes in observers negative emotional reactions in the form of anger, contempt, and disgust, which then lead to an increased tendency to blame and punish.

In the experiment that follows, I investigated the extent to which moral character, apart from motive or reasons for acting, influences judgments of blame, intent, causation, emotion, and punishment. To accomplish this, I varied two factors: the moral character of the actor (good or bad) and the mental state of the actor (aware or not aware of the risk of harm).



The experiment was designed to test the hypothesis that bad moral character influences perceptions of blame, and that this influence increases as mens rea weakens. A group of adults volunteered to participate in an online questionnaire that varied both the moral character of the wrongdoer and the wrongdoer's mental state. The case is loosely based upon People v. Hall, (54) a Colorado Supreme Court case that held that a skier who causes death can be tried for reckless manslaughter. Some of the facts of the actual Hall case were retained, but fictional details were added to permit variation of moral character and mental state.

A. Participants

A total of 201 participants were recruited through Amazon's Mechanical Turk web service, which allows assignment of simple tasks to a large population of online users. Participants were paid $1.50 for completing the survey, which took about five minutes. Participants were assured that their responses would remain anonymous and that identifying information would not be collected. Seven participants were excluded because they failed to correctly respond to an instructional manipulation check, (55) leaving a sample size of 194 participants. Of these, 83% identified as White, 4% as Black, 2% as Hispanic, 6% as Asian Pacific, 4% as South Asian, 1% as Middle Eastern, and 1% as other. Sixty-seven percent were college educated. The mean age was about 41 years. Women comprised 61% of the sample. (56)

B. Design and Materials

Participants were randomly divided and placed into one of four groups based on the two independent variables. The first variable, moral character, had two levels: good or bad. The second variable, recklessness, also had two levels: low or high. The variables were incorporated into the vignette that follows. The experimental design was two by two between subjects, so that each participant saw only one of the four versions of the vignette, reproduced below.
   Nathan is a 24 year old man who works as a ski lift operator on
   Vail Mountain in Colorado. He is an experienced skier and was a
   member of his high school skiing team.

   [Good Character]: Nathan is considered a model employee. He always
   arrives on time for his shift, and works hard. His supervisor
   considers him responsible and reliable. From May through September,
   when the ski slopes are closed, Nathan spends his time volunteering
   at a local animal shelter, as well as helping his family run their
   small business.

   [Bad Character]: Nathan is not considered a good employee. He often
   arrives late for his shift, and is sometimes caught absent from his
   post. His supervisor considers him irresponsible and unreliable.
   From May through September, when the ski slopes are closed, Nathan
   spends his time loafing around town and watching TV. His family bas
   a small business but he rarely helps them out with their work.

   One day in April, after finishing his shift, Nathan left his post
   at the top of the mountain and headed down the hill. He was skiing
   very fast, ski tips in the air, his weight back on his skis and his
   arms out to his sides. It was late in the day, and there were only
   a few skiers on the slope.

   [Low Recklessness]: Nathan felt confident that he could avoid
   anyone in his path, but then he lost control.

   [High Recklessness]: Nathan was feeling reckless that day, and he
   knew there was a risk he might hit someone but he didn't care, and
   then he lost control.

   He was skiing out of control for some time when he took flight off
   of a knoll, and was unable to stop when he saw people on the slope
   below him. Nathan's ski collided with the head of one of those
   people, Alan.

   Unfortunately, Alan suffered a severe brain injury and died a few
   days later. A test of Nathan's blood determined that there were no
   drugs or alcohol in Nathan's system at the time of the accident.

Each participant read only one version of the vignette. (57) After reading the vignette, participants were asked to provide their own personal opinion about Nathan and his role in the death of Alan: to what extent he is responsible; how negatively he should be judged; how much he is to blame; to what extent he caused the death; how intentional were his actions; and how foreseeable was death from Nathan's perspective. All questions were measured on a scale ranging from 1 (not at all) to 7 (very much). The order of presentation of these questions was random.

As a check on the character manipulation, participants rated three items that were presented in random order: Nathan's good moral character (1=not at all; 7=very much), Nathan's trustworthiness (1=not at all; 7=very much), and the extent to which Nathan is a good person (1=bad person; 7=good person). As a check on the recklessness manipulation, participants rated two items presented in random order: the extent to which Nathan believed he could kill someone because of the way he was skiing (1=not at all; 7=very much), and the extent to which Nathan believed he could injure someone because of the way he was skiing (1=not at all; 7=very much).

To gauge participants' emotions experienced as a result of reading the story, we first asked them how angry they were feeling right now (1=not at all angry; 7=very angry) and how upsetting the story was (1=not at all upsetting; 7=very upsetting). Participants were then asked "When you think about Nathan, to what extent do you feel ...?" Following this question stem, participants rated separately the extent to which they felt disgust, contempt, and compassion (1=not at all; 7=very much).

Finally, participants assessed deserved punishment by indicating how severely Nathan should be punished for the death of Alan (1=not at all; 7=very severely), and what kind of prison term should Nathan receive for killing Alan (1=no prison at all; 4=average imprisonment; 7=life imprisonment).

C. Results (58)

1. Manipulation Checks

Judgments of Nathan's moral character, trustworthiness, and goodness were highly correlated, (59) and were combined into a single measure of perceived overall moral character consisting of the mean of these ratings. As predicted, perceptions of Nathan's overall moral character depended on whether he was described as a responsible worker, volunteer, and son (good Nathan) or an irresponsible worker and son (bad Nathan). (60) This provides evidence that perceived moral character was successfully manipulated.

The vignette also varied according to Nathan's awareness of the risk of harm. Judgments of the extent to which Nathan believed he might kill and believed he might injure someone because of the way he was skiing were highly correlated, (61) and were combined into a single measure of awareness of risk consisting of the mean of these judgments. As predicted, perceptions of Nathan's awareness of risk depended on whether he was described as being aware or not aware of the risk of injury or death. (62) This provides evidence that the extent to which Nathan was judged to be aware of the risk of causing injury or death was successfully manipulated.

2. Responsibility Judgments

Did moral character and recklessness influence the extent to which Nathan was perceived as responsible, blameworthy, and judged negatively for the death of Alan? The three dependent measures were highly correlated, so they were combined into a single measure of Overall Responsibility. (63) The means are illustrated in Figure 1. Nathan was perceived as having more overall responsibility for Alan's death if his character was bad rather than good; (64) Nathan was also perceived as more responsible if he was aware of the possibility of death or injury, compared to if he was not aware. (65) These two main effects were qualified by a significant interaction, (66) depicted in Figure 1. Nathan's overall responsibility depended on his moral character, but this difference was more prominent when mens rea was weak--that is, when Nathan was not aware of the risk of injury or death. Planned contrasts nevertheless confirmed that "bad Nathan" was perceived to be more overall responsible than "good Nathan," both when Nathan's mens rea was reckless (67) and when it was not reckless. (68)


3. Judgments of Causation

Participants rated the extent to which Nathan caused the death of Alan. Nathan was judged to be a stronger cause of Alan's death when he was aware of the risk before he acted, compared to when he was not aware. (69) As depicted in Figure 2, there was no character-driven difference in perceived causation when Nathan's mens rea was reckless. (70) The pattern of means depicted in Figure 2 suggests that when Nathan was not aware of the risk of death or injury "bad Nathan" was perceived to be a stronger cause of Alan's death than "good Nathan." However, a planned contrast testing this difference did not reach conventional levels of statistical significance. (71) Therefore, it is unclear whether moral character played any role in perceptions of causation in this experiment, although the pattern of means suggests that the actions of "bad Nathan" might have been perceived as more causal than "good Nathan" when mens rea was not reckless. Further research will be required to investigate this relationship.


4. Judgments of Intent

Participants rated the extent to which Nathan acted intentionally in the death of Alan. When Nathan's moral character was bad, he was judged to have acted more intentionally than when his moral character was good. (72) In addition, and perhaps not surprisingly, when Nathan was aware of the risk of death or injury, he was judged to have acted more intentionally than when he acted without awareness. (73) Planned contrasts indicate that when Nathan acted recklessly, "bad Nathan" was perceived to have acted more intentionally than "good Nathan." (74) Similarly, when Nathan acted without recklessness, "bad Nathan" was perceived as acting more intentionally than "good Nathan." (75) The intention judgment means are depicted in Figure 3.


5. Judgments of Foreseeability

Finally, participants rated the extent to which Alan's death was foreseeable from Nathan's perspective. Not surprisingly, the death was judged to be more foreseeable when Nathan was described as being aware of the risk, compared to when he was not. (76) More importantly, participants judged Alan's death to be more foreseeable from "bad Nathan's" perspective than from "good Nathan's" perspective. (77) Planned contrasts indicate that the character-driven difference just described emerges only when Nathan acted recklessly; (78) when Nathan acted without recklessness, there was no reliable difference in foreseeability between "bad Nathan" and "good Nathan." (79) The foreseeability means are depicted in Figure 4.


6. Punishment Judgments

Judgments of how severely Nathan should be punished and of what length prison term he should receive were highly correlated, (80) so they were combined into a single measure, in which a rating of "1" indicates no punishment and "7" indicates severe punishment. Participants judged Nathan as deserving more severe punishment when he acted with a reckless mental state compared to when he was not reckless. (81) In addition, there was a main effect for character, such that "bad Nathan" was perceived to deserve more severe punishment than "good Nathan." (82) Planned contrasts confirmed that the increased punishment severity for "bad Nathan" (compared to "good Nathan") proved statistically significant in both the reckless (83) and not reckless conditions. (84)


7. Moral Emotion

Participants indicated how angry they were feeling, how upsetting the story was, and when they thought about Nathan to what extent they felt disgust, contempt, and compassion. The compassion item was reverse coded. All five items were correlated, (85) and were combined into an index of negative moral emotion.

The two independent variables--mens rea and character--influenced participants' negative moral emotion. Specifically, when participants learned that Nathan acted with a reckless state of mind, participants' moral emotions were more negative than if Nathan's state of mind was described as not reckless. (86) Similarly, reading about "bad Nathan" led participants to experience more negative emotion than reading about "good Nathan." (87) There was no statistically significant interaction between state of mind and character. (88)

The fact that mens rea and moral character each influenced both moral emotion and overall responsibility raises the question of whether negative moral emotion mediates the relationship between character and overall responsibility, as well as the relationship between mens rea and overall responsibility. Assuming that negative emotion could mediate the relationship between moral character and perceptions of overall responsibility, (89) we investigated its statistical significance by first testing whether there is a relationship between character and participants' experiences of negative emotion. We found that, indeed, these two factors were related. (90) A second regression showed that character influenced perceptions of overall responsibility. (91) Then we regressed perceptions of overall responsibility on character and moral emotion simultaneously, and found that moral emotion predicts perceptions of overall responsibility after controlling for character. (92) Character no longer predicted overall responsibility after controlling for moral emotion. (93) The indirect nature of the effect of character on perceptions of responsibility was further confirmed using a bootstrapping procedure. (94) It is therefore possible that learning about harm caused by an actor with a bad moral character leads to negative moral emotion, which in turn leads to strong attributions of blame and responsibility.

At the same time it is also possible that stronger mens rea (that is, recklessness) leads perceivers to experience negative moral emotion, which in turn leads to greater attributions of blame and responsibility. We tested the statistical significance of this hypothesized relationship by first testing whether there is a relationship between Nathan's mens rea and participants' experience of negative emotion, and found that indeed, these two factors are related. (95) A second regression showed that Nathan's mens rea influenced perceptions of overall responsibility. (96) Then we regressed perceptions of overall responsibility on mens rea and moral emotion simultaneously, and round that moral emotion predicts perceptions of overall responsibility after controlling for mens rea. (97) However, mens rea still predicted overall responsibility after controlling for moral emotion. (98) Thus, it seems implausible that moral emotion mediates the relationship between mens rea and attributions of blame and responsibility.

One question hot yet addressed is the extent to which people consciously endorse the idea of using the moral character of the actor to gauge judgments of blame, causation, and intent. It is possible that when we judge blame and responsibility for harm, we deliberately take into account the actor's moral character. It might be, for example, that we explicitly infer that a person with a bad character has caused undetected harms in the past, so that he deserves to be blamed and punished more now that he is caught. On the other hand, it is possible that when we take account of moral character in blame judgments, we do so without even realizing it. It might be, for example, when confronted with two identical harms committed by actors with identical mental states, we would disavow blaming and punishing differently based upon differences in the actors' moral character.

8. Illuminating the Influence of Moral Character

To investigate the extent to which people explicitly endorse using moral character information to assess an actor's responsibility, intentionality, and deserved punishment, a follow up experiment was conducted. The experiment was identical to the first experiment presented in this article, with one key exception: this time, each participant made judgments about both "good Nathan" and "bad Nathan." The recklessness manipulation remained, as in the first experiment between subjects. Thus, each participant received two versions of the story. In the reckless version, participants assessed the blameworthiness of reckless "good Nathan" and reckless "bad Nathan." In the not-reckless version, participants assessed the blameworthiness of non-reckless "good Nathan" and non-reckless "bad Nathan." The order of presentation of good and bad moral character was counter balanced, so that hall the participants rated "good Nathan" first and then "bad Nathan" and half did the reverse. Using this method, we assessed the extent to which participants would try to keep their assessments of Nathan's blameworthiness consistent across the two versions of the story that differed only according to Nathan's moral character.

The participants in this follow-up study--Experiment 2--were recruited in the same method as the first study, using Amazon's Mechanical Turk web service. A total of forty-eight participants completed the questionnaire, and were paid $1.50 each. (99) The results revealed that the differences in perceptions of responsibility, cause, intent, foreseeability, and punishment demonstrated in the first experiment disappeared when each participant judged both "good Nathan" and "bad Nathan." (100) This is illustrated in Figure 6, which depicts overall responsibility judgments in Experiment 1 and Experiment 2. When participants judged the overall responsibility of just one version of Nathan in Experiment 1, their judgment was influenced by Nathan's moral character. This was true regardless of whether Nathan's state of mind was described as reckless (left panel of Figure 6) or not reckless (right panel of Figure 6). Note that these statistically significant differences in overall responsibility judgments disappeared in Experiment 2, when each participant read two stories (about "good Nathan" and "bad Nathan") and made two overall responsibility judgments. In this context, participants now perceived Nathan's overall responsibility roughly the same, regardless of whether his moral character was good or bad. The statistically significant differences in perceived responsibility based on the actor's moral character disappeared when differences in moral character were made explicit. (101)




To a large degree, liability in the criminal law is based on the notion of blameworthiness. For most crimes, liability requires a morally blameworthy act. For example, a person who uses force against another person in self defense is not criminally liable because under the circumstance, the act is considered not worthy of blame. Legal blame in criminal law is predicated on the idea that the actor performed a prohibited act, accompanied by a specific mental state, without justification or excuse. Yet, the results reported in this article provide further support for the idea that the blaming process is infused with motivation and emotion, and not dictated solely by individual acts and their consequences. Humans are social beings, and blame is a social process. When we observe a harmful outcome, our first reactions are emotional, and those emotions are informed by our immediate assessment of what kind of person could have caused this harm. On this account, a person with a bad moral character who causes a harmful outcome is a person who disrespects the community's way of life. (102) As observers and community members, we react to such disrespect with moral outrage, and we experience the urge to blame and punish. (103) Conversely, we are more willing to exculpate, at least partially, an otherwise virtuous person who causes harm, because his prior good deeds have in some sense licensed the transgression. (104)

In this respect, there is some tension between the process of legal blame and the psychology of moral blame. The experimental results indicate that, as a psychological process, moral blame is sometimes informed by emotion. Our emotional reactions are not only a product of the act and the outcome, but also a product of inferences about the general virtuousness of the person who performed the act that caused the harm. Reading about "bad Nathan" made participants feel angry, disgusted, and full of contempt; these emotions then led to blaming and punishing "bad Nathan" more severely than "good Nathan." The legal blame process is limited to considerations of mental state, conduct, and result; the psychological blame process includes those considerations but seems to involve much more, including the emotions of the perceiver and the moral character of the actor. Other variables, not examined here, can doubtless come into play in the psychological blame process, including perceptions of victim characteristics. (105)

Legal and psychological blame processes also are in tension regarding assessment of the actor's mental state. Procedures for legal blame assume an assessment of mental state that is independent of the moral character of the actor. But the results reported here show that perceptions of intent, foreseeability, and possibly causation can be colored by independent reasons for thinking the actor is a bad person (Figures 3 and 4). Remarkably, this result holds true even when, as here, the mental state of the actor was clearly specified--Nathan was described as being either aware or unaware of the possibility of harming someone. Therefore, even when mental state is "knowable," we construct mental state from more than just inferences surrounding the act itself, but also inferences about the goodness or badness of the person who performed it. We give the benefit of the doubt to a person with a virtuous character who causes harm: we perceive his actions as less intentional and perhaps even less causal, and the harm less foreseeable than if his character is flawed. Thus, compared to the unreliable loafer, the reliable employee who helps the local animal shelter and pitches in when his family needs help is blamed and punished less harshly, and we align our perceptions of his mental state and the harm's foreseeability accordingly. Remarkably, it seems that we do not deliberately use character information to inform responsibility judgments, for when differences in character are made explicit, as in Experiment 2, we moderate our responsibility judgments so that we hold the virtuous harmdoer equally responsible as the ignoble harmdoer.

It is not clear why we would want to disavow the influence of the actor's character in our judgments of blame and responsibility. After all, a person who is an unreliable worker and an irresponsible son has demonstrated that he holds "an attitude of indifference to the welfare of others." (106) Our impulse to more harshly blame "bad Nathan" for causing the death of the skier perhaps reflects the inference that just as he was indifferent to his employer, to the patrons he served in his employment, and to his parents, "bad Nathan" was also indifferent to the welfare of the person with whom he collided when he decided to ski in a dangerous fashion. This inference may or may not accurately reflect "bad Nathan's" mental state, but it is an inference that is arguably reasonable to draw.

Recall, as well, that this inference appears to be driven by our emotional reaction to Nathan. Compared to "good Nathan," when Nathan was portrayed as having an ignoble character, his actions triggered more negative emotions involving anger, disgust, and contempt. That Nathan seemed to be indifferent to the welfare of the people around him made us emotionally upset, and it is these moral emotions that appeared to drive our impulse not only to blame and punish more harshly, but also to perceive Nathan as more aware of and intentional toward the harm, and possibly to have played a more causal role in it. Thus, our own emotional reactions to harm might inform out perceptions of the harmdoer's mental state. Our perception of the harmdoer's mental state can in turn inform our perception of the severity of the harm. In one study, when people received shocks they thought were intentionally given, they reported them to be more painful than shocks they thought were unintentional, even though the magnitude of the shocks was actually the same. (107) Further research is necessary to determine whether emotional reaction or perceived intention is primary, or alternatively, whether each informs the other in a feedback loop.

The results of the experiment reported here are consonant with findings regarding the effect of prior criminal record on actual verdicts. One study examined over 300 criminal trials and found that in cases where the evidence was weak, there was a positive correlation between the jury learning that the defendant had committed prior crimes and the likelihood of conviction. (108) A juror who hears that the defendant had committed prior crimes might infer that the defendant is a bad person, which might lead to inferences about mental state, causation, and blame that we observed in the experiment reported here. Earlier experimental research also suggests that jurors who hear that the defendant committed prior crimes are more prone to convict, especially when the defendant's prior crimes were serious in nature or similar to the current accusation. (109) These results are consistent with the notion that when we size someone up as a bad person, we perceive his acts as more causal, his mental states more intentional, his creation of risk more foreseeable, and his blameworthiness greater, than a similarly situated good person.

It is important to note, however, that the focus of this article is not on the narrow question of the influence of prior crimes--it is on the broader question of the influence of moral character writ large, however those inferences about moral character are made, whether through prior crimes or otherwise. Note that the basis for inferring bad moral character need not be extreme. In contrast to the prior work examining the role of serious transgressions like having committed a prior robbery or burglary, the experiment reported here manipulated relatively subtle character traits, such as being a responsible worker and son. The results demonstrate that we perceive badness not only in people who have seriously harmed others in the past, but also in those who tend to free ride within their own social groups; this perceived badness triggers evaluative emotions that push us toward more punitive attributions of mental state and blame.

Moreover, unlike prior research on the influence of information about prior crimes on verdicts, this article focuses not specifically on verdicts, but rather on the more basic social-psychological question of how blaming processes work. The knowledge gleaned from investigating basic psychological processes of blame can indeed inform our understanding of how juries and judges decide on guilt in criminal cases and liability in tort cases. But it can do much more than that. When a prosecutor decides which criminal offenses to charge in the indictment, and later, how lenient or harsh the plea agreement will be, these decisions may be influenced, without her even realizing it, by her perceptions of the moral character of the defendant. When an injured person decides whether or not to make a claim in tort for compensation, that potential plaintiff's decision will likely be influenced by the extent to which he blames the other person, (110) and these attributions of blame can be intensified to the extent that the potential plaintiff perceives the other person as a "bad guy." When a litigant in a business contract dispute refuses to settle despite the fact that the offer on the table exceeds his own lawyer's estimate of the expected value of going to trial, that stubborn refusal may, in some cases, relate to the litigant's perception that the litigant on other side is a bad actor or is acting in bad faith. And, when a jury awards punitive damages in an amount that reflects its indignation at the defendant's conduct, that indignation may derive from the perception that the individuals involved in the harm were putting profits ahead of human welfare, (111) leading to perceptions that these were immoral actors. (112) The psychological process of moral blame is a fundamentally social process, and our judgments of blame are often attempts to address threats to the social order.

(1.) See, e.g., FED. R. EVID. 404(a) (permitting admission of evidence of the defendant's bad character only when it is used in specific rebuttal to evidence of good character submitted by the defense). By contrast, the criminal law embraces the consideration of the defendant's character in sentencing judgments. For example, the U.S. Code provides that "[n]o limitation shall be placed on the information concerning the background, character, and conduct of a person convicted of an offense which a court of the United States may receive and consider for the purpose of imposing an appropriate sentence." 18 U.S.C. [section] 3661 (2006). In practice, the Presentence Investigation Report "may contain a wide range of information about the defendant, all of which may be considered by the court in determining the sentence." Sentencing Guidelines, 38 GEO. L.J. ANN. REV. CRIM. PROC. 681, 725 n.2161 (2009). For a dissenting view, see Ekow N. Yankah, Good Guys and Bad Guys: Punishing Character, Equality and the Irrelevance of Moral Character to Criminal Punishment, 25 CARDOZO L. REV. 1019 (2003). The focus in this article is primarily on blame, rather than punishment judgments.

(2.) Janice Nadler & Mary-Hunter McDonnell, Moral Character, Motive, and the Psychology of Blame, 97 CORNELL L. REV. 255 (2012).

(3.) See Theodore Eisenberg & Valerie P. Hans, Taking a Stand on Taking the Stand: The Effect of a Prior Criminal Record on the Decision to Testify and on Trial Outcomes, 94 CORNELL L. REV. 1353, 1388 (2009) (noting that data suggest that disclosure of prior crimes increases the chances of conviction in close cases).

(4.) This relationship appears to be strongest when the defendant's prior crimes were serious in nature or similar to the current accusation. See Edith Greene & Mary Dodge, The Influence of PriorRecord Evidence on Juror Decision Making, 19 LAW & HUM. BEHAV. 67, 69 (1995) (summarizing prior research indicating that evidence of prior convictions affects chances of conviction, especially if "the prior conviction was for an offense similar to the one jurors were deciding"); Valerie P. Hans & Anthony N. Doob, Section 12 of the Canada Evidence A ct and the Deliberations of Simulated Juries, 18 CRIM. L.Q. 235, 237-38 (1976) (discussing potential ways that revealing a defendant's criminal record might alter a verdict); Sally Lloyd-Bostock, The Effects on Juries of Hearing About the Defendant's Previous Criminal Record: A Simulation Study, 2000 CRIM. L. REV. 734, 753-55 (2000) ("The results clearly confirm that evidence of previous convictions can have a prejudicial effect, especially when there is a recent previous conviction for a similar offence."); Roselle L. Wissler & Michael J. Saks, On the Inefficacy of Limiting Instructions: When Jurors Use Prior Conviction Evidence to Decide on Guilt. 9 LAW &. HUM. BEHAV. 37, 42 (1985) (noting that disclosure of similar prior crime increases the likelihood of conviction more than disclosure of dissimilar prior crime).

(5.) The mechanism by which information that the defendant committed prior crimes increases the likelihood of conviction is still not entirely clear. It is likely that perceivers treat prior crimes as evidence of propensity to commit the current offense that operates within a specific category--the idea being, for example, "once a burglar, always a burglar." Support for this idea has emerged from several experiments, which demonstrate that mock jurors are more likely to convict when the defendant's prior crimes are similar to the current offense, compared to when the prior crimes are dissimilar to the current offense. See, e.g., Wissler & Saks, supra note 4, at 43 ("[T]he highest conviction rate occurred when the prior conviction was the same as the present charge."). It is also possible that the positive correlation between prior crimes and likelihood of conviction is partially accounted for by credibility concerns--a testifying defendant who is impeached with evidence that he committed prior crimes involving deceit may be more likely to be convicted. Additionally, part of the variance might be accounted for by jurors' inferences about the defendant's dangerousness, which is arguably another version of the propensity thesis, in which the propensity inference is made across the broad category of violent acts. The important point here, however, is that none of these explanations for the association between prior crime and likelihood of conviction is focused on the broad idea of moral character, which is the focus of this article.

(6.) I am assuming, for the sake of this example, recklessness as defined by the Model Penal Code (MPC). See MODEL PENAL CODE [section] 2.02(2)(c) (1962).

(7.) See id.

(8.) This example appears in Kimberly Kessler Ferzan, Opaque Recklessness, 91 J. CRIM. L. & CRIMINOLOGY 597, 600 (2001). Ferzan explores the problematic space in between recklessness and negligence. Using the current example, suppose we can infer only that the driver recognized, on a very general level, that driving while dialing a cell phone is a dangerous thing to do, but nothing more. This state of mind--dubbed "opaque recklessness" by Ferzan--does not, strictly speaking, meet the MPC definition of recklessness because the driver did not consider and disregard the risk of death. According to the hypothesis discussed in the text herein, the opaquely reckless driver would be blamed, in part, according to available information about his moral character. In fact, moral character information might have an especially strong inference when mental state is ambiguous or is located in the liminal space between two culpability categories, as arguably is the case with opaque recklessness.

(9.) Moral emotions can be thought of as those emotions that are evoked by a threat to or violation of a personally valued moral principle. See C. Daniel Batson, What's Wrong with Morality?, 3 EMOTION REV. 230, 233 (2011).

(10.) See, e.g., Jennifer S. Lerner et al., Sober Second Thought: The Effects of Accountability, Anger, and Authoritarianism on Attributions of Responsibility, 24 PERSONALITY & SOC. PSYCHOL. BULL. 563, 570 (1998) (noting that participants that were anger-primed made more punitive attributions than other participants); Thalia Wheatley & Jonathan Haidt, Hypnotic Disgust Makes Moral Judgments More Severe, 16 PSYCHOL. SCI. 780, 780 (2005) ("[M]oral judgments can be made more severe by the presence of a flash of disgust.").

(11.) This scenario and results described herein are reported in Mark D. Alicke, Culpable Causation, J. PERSONALITY & SOC. PSYCHOL. 368, 369 (1992).

(12.) Alternatively, the stop sign was obscured by a large tree branch. Id. The pattern of results described below is the same in the oil and tree branch variations. Id. at 370.

(13.) Id. at 370.

(14.) Id.

(15.) Nadler & McDonnell, supra note 2: see also Keith J. Holyoak & Dan Simon, Bidirectional Reasoning in Decision Making by Constraint Satisfaction 128 J. EXPERIMENTAL PSYCHOL.: GENERAL 3, 11-13 (1999) (demonstrating that when an investor was motivated by malicious greed, perceivers were more likely to positively assess arguments supporting liability for libel, compared to when he was motivated by honest concern).

(16.) In a third neutral condition, the man stored oxygen for delivery as part of his home health care business. In this scenario, judgments of blame and causation were roughly midway between those in the football cheating scenario on the one hand, and the sick daughter scenario on the other. Nadler & McDonnell, supra note 2.

(17.) RESTATEMENT (SECOND) OF TORTS [section] 291 (1965); see also United States v. Carroll Towing Co., 159 F.2d 169, 173 (2d Cir. 1947).

(18.) MODEL PENAL CODE [section] 2.02(2)(c) (1962).

(19.) See Fiery Cushman et al., The Role of Conscious Reasoning and Intuition in Moral Judgment: Testing Three Principles of Harm, 17 PSYCHOL. SCI. 1082, 1086 (2006) ("Subjects rated ... harm involving physical contact as morally worse than harm without contact.").

(20.) See id.; Joshua D. Greene et al., An fMRI Investigation of Emotional Engagement in Moral Judgment, 293 SCIENCE 2105, 2105-06 (2001) (noting that most people find pushing a stranger off a bridge to be more troubling than flipping a switch causing a runaway trolley to hit a stranger, even though both actions resulted in five lives being saved).

(21.) In one version of the scenario the road was slippery because of a prior oil spill; in another version, the stop sign was partially obscured by a tree branch. There were no notable differences in the results between these two versions of the scenario.

(22.) Mark D. Alicke, Culpable Control and the Psychology of Blame, 126 PSYCHOL. BULL 556, 557 (2000).

(23.) See Alicke, supra note 11, at 368.

(24.) Alicke, supra note 22, at 558.

(25.) The study of motivated cognition in law has received increased attention in recent years. See, e.g., Dan M. Kahan, Culture, Cognition, and Consent: Who Perceives What, and Why, in Acquaintance Rape Cases, 158 U. PA. L. REV. 729 (2010); Dan M. Kahan et al., "They Saw a Protest": Cognitive Illiberalism and the Speech-Conduct Distinction, 64 STAN. L. REV. (forthcoming 2012), available at id=1755706; Dan M. Kahan et al., Whose Eyes are You Going to Believe? Scott v. Harris and the Perils of Cognitive Illiberalism, 122 HARV. L. REV. 837 (2009); Avani Mehta Sood & John M. Darley, The Plasticity of Harm in the Service of Punishment Goals: Legal Implications of Outcome-Driven Reasoning, 100 CA. L. REV. (forthcoming 2012), available at

(26.) Alicke, supra note 22, at 558.

(27.) Compare Alicke, supra note 22, and Jonathan Haidt, The Emotional Dog and Its Rational Tail: A Social Intuitionist Approach to Moral Judgment, 108 PSYCHOL. REV. 814 (2001), with Fiery Cushman, Crime and Punishment: Distinguishing the Roles of Causal and Intentional Analyses in Moral Judgment, 108 COGNITION 353 (2008), and Bertram F. Malle et al., Moral, Cognitive, and Social: The Nature of Blame, in SOCIAL THINKING AND INTERPERSONAL BEHAVIOR (Joseph P. Forgas et al. eds., forthcoming 2012).

(28.) Malle et al., supra note 27.

(29.) See, e.g., 18 U.S.C. [section] 3593(f) (2006).

(30.) See supra note 18 and accompanying text.

(31.) Glenn D. Reeder et al., Inferences About the Morality of an Aggressor: The Role of Perceived Motive, 83 J. PERSONALITY & SOC. PSYCHOL. 789, 792-94 (2002).

(32.) Id. at 794-95.

(33.) See Susan T. Fiske et al., Universal Dimensions of Social Cognition: Warmth and Competence, 11 TRENDS COGNITIVE SCI. 77, 77 (2007) (describing the fundamental nature of the two dimensions of warmth and competence in social cognition).

(34.) Janine Willis & Alexander Todorov, First Impressions: Making Up Your Mind After a 100-Ms Exposure to a Face, 17 PSYCHOL. SCI. 592,594 (2006).

(35.) See id.: Bogdan Wojciszke et al., On the Dominance of Moral Categories in Impression Formation, 24 PERSONALITY & SOC. PSYCHOL. BULL. 1251, 1254-55 (1998) (finding that subjects round moral traits to be more important than competence traits when forming impressions of strangers).

(36.) See David A. Pizarro & David Tannenbaum, Bringing Character Back: How the Motivation to Evaluate Character Influences Judgments of Moral Blame, in The SOCIAL PSYCHOLOGY OF MORALITY: EXPLORING THE CAUSES OF GOOD AND EVIL 97 (Mario Mikulincer & Phillip R. Shaver eds., 2011) ("[P]roposing a person-based character approach as an alternative to the act-based theories.").

(37.) Nadler & McDonnell, supra note 2.

(38.) Mark D. Alicke, Evidential and Extra-Evidential Evaluations of Social Conduct, 9 J. SOC. BEHAV. & PERSONALITY 591, 595-601 (1994).

(39.) For a review, see Jennifer K. Robbennolt, Outcome Severity and Judgments of "Responsibility": A Meta-Analytic Review, 30 J. APPLIED SOC. PSYCHOL. 2575 (2000) (demonstrating that people attribute greater responsibility for a harm when the harm is severe than when the harm is minor).

(40.) That intentionally doing something bad is worse than doing it unintentionally is understood by children, see John M. Darley et al., Intentions and Their Contexts in the Moral Judgments of Children and Adults, 49 CHILD DEV. 66, 67-68 (1978), as well as by adults, see John M. Darley et al., Doing Wrong Without Creating Harm, 7 J. EMPIRICAL LEGAL STUD. 30 (2010); Lawrence M. Solan & John M. Darley, Causation, Contribution, and Legal Liability: An Empirical Study, 64 LAW & CONTEMP. PROBS. 265 (Autumn 2001).

(41.) See Francis X. Shen et al., Sorting Guilty Minds, 86 N.Y.U.L. REV. 1306 (2011).

(42.) See Pizarro & Tannenbaum, supra note 36, at 11-12 ("For instance, if there is evidence that an individual is a bad person, the inference that he or she intended a negative outcome seems reasonable.").

(43.) Id. at 12.

(44.) See Joshua Knobe, Person as Scientist, Person as Moralist, 33 BEHAV. & BRAIN SCI. 315. 317 19 (2010).

(45.) Id.

(46.) See id.

(47.) See Greene et al., supra note 20.

(48.) Id. at 2106.

(49.) Id.

(50.) See Paul Rozin et al.. The CAD Triad Hypothesis: A Mapping Between Three Moral Emotions (Contempt, Anger, Disgust) and Three Moral Codes (Community, Autonomy, Divinity), 76 J. PERSONALITY & SOC. PSYCHOL. 574, 575 (1999) (arguing that contempt, anger, and disgust are related "emotional reactions to the moral violations of others").

(51.) See Daniel Kahneman & Cass R. Sunstein, Indignation." Psychology, Politics, Law, (Univ. of Chi. Law & Econ., Olin Working Paper No. 346, 2007) (examining the relationship between indignation and the response tendency to punish the guilty actor), available at; Lerner et al., supra note 10, at 570 (noting the role that anger plays in justice judgments).

(52.) See Julie H. Goldberg et al., Rage and Reason: The Psychology of the Intuitive Prosecutor, 29 EUR. J. SOC. PSYCHOL. 781, 787-89 (1999) (examining the relationship between anger over injustice and determinations of who deserves punishment).

(53.) See Elizabeth Mullen & Janice Nadler, Moral Spillovers: The Effect of Moral Violations on Deviant Behavior, 44 J. EXPERIMENTAL SOC. PSYCHOL. 1239, 1243 (2008) (finding that observing outcomes that violate strongly held moral beliefs can lead to engaging in deviant behavior); Janice Nadler, Flouting the Law, 83 TEX. L. REV. 1399, 1440 (2005) (arguing that perceptions of injustice in the law can sometimes lead people to display a greater willingness to disregard the law).

(54.) 999 P.2d 207 (Colo. 2000).

(55.) See Daniel M. Oppenheimer et al., Instructional Manipulation Checks: Detecting Satisficing to Increase Statistical Power, 45 J. EXPER1MENTAL SOC. PSYCHOL. 867 (2009) (describing instructional manipulation checks as a question embedded within the experiment requiring the participant to confirm reading the instruction). In out experiment, we asked, "According to the story, Nathan spent most of his free time doing what?" and asked respondents to choose one of four options: Watching television, Volunteering, Reading, or None of the above.

(56.) Although we did not collect additional demographic information in this particular experiment, similar recent experiments conducted by the author drawing from the same Mechanical Turk population yielded samples that were: 95% U.S. residents for at least the past 12 years, 56% with a household income of less than $50,000 per year, 48% liberal or very liberal, 28% moderate, and 25% conservative or very conservative.

(57.) The four versions of the vignette were good moral character / low recklessness; good moral character / high recklessness, bad moral character / low recklessness; bad moral character / high recklessness. Each participant read only one of these four versions.

(58.) All analyses were conducted using Analysis of Variance (ANOVA) unless otherwise noted. An analysis of variance measures for statistical differences between the means of groups whose data are categorical (as opposed to continuous). WILLIAM L. HAYS, STATISTICS 376-81 (5th ed. 1994). Throughout the results section of this article, "significantly" refers to statistical significance, which denotes the rejection of the null hypothesis--the possibility of no differences between the various groups--at a probability level indicated by the p-value reported. Thus, "p" is defined as the probability of finding a difference or relationship between two groups as large as that observed if there were, in fact, no difference or relationship between them. Id. at 267-82.

(59.) Cronbach's alpha = .95. Cronbach's alpha measures the internal consistency of a set of items, and ranges between 0 and 1, with higher numbers indicating higher consistency. RICK H. HOYLE ET AL., RESEARCH METHODS IN SOCIAL RELATIONS 83-84 (7th ed. 2002).

(60.) F(1,193) = 331.92, p < .001, [[eta].sub.p.sup.2] = .634; Mean (Good Character) = 5.41; Mean (Bad Character) = 2.69.

(61.) Cronbach's alpha = .87.

(62.) F(1,191) = 112.60, p < .001, [[eta].sub.p.sup.2] = .372; Mean (Aware) = 4.73; Mean (Not Aware) = 2.54.

(63.) Cronbach's alpha = .91.

(64.) F(1,193) = 33.61, p < .001, [[eta].sub.p.sup.2] = .123.

(65.) F(1,193) = 45.19, p < .001, [[eta].sub.p.sup.2] = .166.

(66.) F(1,193) = 7.58, p < .05, [[eta].sub.p.sup.2] = .017.

(67.) z = 2.57, p = .01.

(68.) z = 5.63, p < .001.

(69.) F(1,193) = 13.68, p < .001, [[eta].sub.p.sup.2] = .066; Mean (Reckless) = 6.43: Mean (Not Reckless) = 5.76. We did not detect a statistically significant main effect of moral character on causation. F(I. 193) = 3.74, p = .127, [[eta].sub.p.sup.2] = .011. Mean (Good Character) = 5.97, Mean (Bad Character) = 6.23. Similarly, there was no significant interaction between character and recklessness on judgments of causation. F(I, 193) = 1.78, p <.29, [[eta].sub.p.sup.2] = .006. Planned contrasts reveal effects of character within each mens rea category, as detailed in the text.

(70.) A planned contrast shows no difference between good and bad character within the reckless mens rea group, z = 0.34, p = .74.

(71.) z = 1.83. p = .067.

(72.) F(1,193) = 17.70, p < .001, [[eta].sub.p.sup.2] = .077; Mean (Good Character) - 1.51: Mean (Bad Character) = 2.17.

(73.) F(1,193) = 331.92, p < .001, [[eta].sub.p.sup.2] = .634: Mean (Good Character) = 1.46: Mean (Bad Character) = 2.20. The interaction between character and recklessness was not statistically significant. F(1, 193) = 1.43; p = .23.

(74.) z = 3.82, p < .001.

(75.) z = 2.13, p = .03.

(76.) F(1,193) = 32.79, p < .001, [[eta].sub.p.sup.2] = .139; Mean (Reckless) = 3.62; Mean (Not Reckless) = 2.39.

(77.) F(1,193) = 12.03, p < .001, [[eta].sub.p.sup.2] = .051; Mean (Good Character) = 2.65; Mean (Bad Character) = 3.38. The character-reckless interaction did not reach statistical significance. F(1,193) = 2.39, p = .12.

(78.) z = 3.54, p < .001.

(79.) z = 1.36, p < .174.

(80.) Cronbach's alpha = .88.

(81.) F(1,193) = 54.81, p < .001, [[eta].sub.p.sup.2] = .198; Mean (reckless) = 4.46: Mean (not reckless) = 3.15.

(82.) F(1,193) = 32.87, p < .001, [[eta].sub.p.sup.2] = .119; Mean (bad) = 4.32; Mean (good) = 3.33.

(83.) z = 3.45, p < .001.

(84.) z = 4.66, p < .001.

(85.) Cronbach's alpha = .84.

(86.) F(1,190) = 27.11, p < .001, [[eta].sub.p.sup.2] = .092; Mean (reckless) = 4.21; Mean (not reckless) = 3.34.

(87.) F(1,190) = 79.24, p < .001, [[eta].sub.p.sup.2] = .269; Mean (bad) = 4.55; Mean (good) = 3.04.

(88.) F(1,190) = 1.89, p = .17.

(89.) See Klaus Fiedler et al., What Mediation Analysis Can (Not) Do, J. 47 EXPERIMENTAL SOC. PSYCHOL. 1231 (2011) (describing how statistical mediation analysis can "examine the degree to which a third variable (Z) accounts for the influence of an independent (X) on a dependent variable (Y) conditional on the assumption that Z actually is a mediator").

(90.) [beta] = 1.51, t(190) = 8.24, p < .001.

(91.) [beta] = 1.04, t(190) = 5.06, p < .001.

(92.) [beta] = .62, t(190) = 9.04, p < .001.

(93.) [beta] = .09, t(190) = .46, p = .64.

(94.) See Kristopher J. Preacher & Andrew F. Hayes, Asymptotic and Resampling Strategies for Assessing and Comparing Indirect Effects in Multiple Mediator Mode&, 40 BEHAV. RES. METHODS 879 (2008). Using 5000 bootstrapped resamples of the data, the analysis revealed a 95% confidence interval of .64 to 1.23. Bootstrapping is a nonparametric resampling procedure which does not impose the assumption of normality of the sampling distribution. Id. at 880.

(95.) [beta] = -.88, t(190) = -4.29, p < .001.

(96.) [beta] = -1.22, t(190) = -6.12, p < .001.

(97.) [beta] = .56, t(190) = 9.53, p < .001.1

(98.) [beta] = -.76, t(190) = -4.37, p < .001.

(99.) Because the design of this experiment was within-subjects, we were able to achieve sufficient statistical power for hypothesis testing using a smaller number of participants.

(100.) Causation(reckless): Mean(good) = 6.50; Mean(bad) = 6.62. Causation(not reckless): Mean(good) = 5.95; Mean(bad) = 6.09. Intent(reckless): Mean(good) = 3.16; Mean(bad) = 3.31. Intent(not reckless): Mean(good) = 1.64; Mean(bad) = 1.82. Foreseeable(reckless): Mean(good) = 3.85; Mean(bad) = 4.04. Foreseeable(not reckless): Mean(good) = 2.73; Mean(bad) = 2.86. Punish(reckless): Mean(good) = 4.73; Mean(bad) = 4.98. Punish(not reckless): Mean(good) = 3.33; Mean(bad) = 3.75. All ts < 1.0; all ps> .60.

(101.) Reckless condition: t(49) = -0.64, p = .74. Not Reckless condition: t(42) = -0.52, p = .70.

(102.) See Philip E. Tetlock et al., People as Intuitive Prosecutors: The Impact of Social-Control Goals on Attributions of Responsibility, 43 J. EXPERIMENTAL SOC. PSYCHOL, 195, 196 (2007) (describing disrespect for a way of tire as a source for outrage).

(103.) See Kahneman & Sunstein, supra note 51; SAMUEL H. PILLSBURY, JUDGING EVIL: RETHINKING THE LAW OF MURDER AND MANSLAUGHTER 33 (1998) ("[P]unishment is deserved according to the wrongdoer's choice to disregard another's value.").

(104.) See Daniel A. Effron & Benoit Monin, Letting People Off the Hook: When Do Good Deeds Excuse Transgressions?, 36 PERSONALITY & SOC. PSYCHOL. BULL. 1618, 1621-23 (2010) (showing that good deeds reduce condemnation when they are in a different domain (for example, crusading against drugs) than the subsequent transgression (for example, committing sexual harassment)).

(105.) See Mitchell J. Callan et al., The Consequences of Victim Physical Attractiveness on Reactions to Injustice: The Role of Observers' Belief in a Just World, 20 SOC. JUST. RES. 433, 441 (2007) ("[W]e found clear evidence that the death of a physically attractive versus less attractive victim increased participants; punishment reactions."); Mary R. Rose et al., Appropriately Upset? Emotion Norms and Perceptions of Crime Victims, 30 LAW & HUM. BEHAV. 203 (2006) (examining the possibility that victims expressing less emotion may have less punishment assigned to the offender); Janice Nadler & Mary R. Rose, Victim Impact Testimony and the Psychology of Punishment, 88 CORNELL L. REV. 419, 434-37 (2003) (noting that victim characteristics tan influence punishment judgments).

(106.) See PILLSBURY, supra note 103, at 161.

(107.) Kurt Gray & Daniel M. Wegner, The Sting of Intentional Pain, 19 PSYCHOL. SCI. 1260, 1261 (2008).

(108.) Eisenberg & Hans, supra note 3, at 1388.

(109.) See Greene & Dodge, supra note 4, at 69: Hans & Doob, supra note 4, at 237-38: Lloyd-Bostock, supra note 4, at 753-55; Wissler & Saks, supra note 4, at 47.

(110.) See William L.F. Felstiner et al., The Emergence and Transformation of Disputes: Naming, Blaming, Claiming ..., 15 LAW & SOC'Y REV. 631, 641 (1980) ("[People] are more likely to [see an

experience as injurious or voice a grievance about it] if blame can be placed upon another.")

(111.) See Kahneman & Sunstein, supra note 51.

(112.) See Reeder et al., supra note 31.

Copyright [c] 2012 by Janice Nadler.

This article is also available at

JANICE NADLER, American Bar Foundation and Northwestern University School of Law: Thanks to the American Bar Foundation for financial support. The author is grateful to Omri Ben-Shahar, Lisa Griffin, and Richard McAdams for useful suggestions, and to participants in the Duke Law and Contemporary Problems symposium conference, Adjudicating the Guilty Mind (May 10, 2011) and workshop participants at the University of Chicago Law School for helpful feedback.
COPYRIGHT 2012 Duke University, School of Law
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Adjudicating the Guilty Mind
Author:Nadler, Janice
Publication:Law and Contemporary Problems
Date:Mar 22, 2012
Previous Article:Adjudicating the guilty mind.
Next Article:Public perceptions of white collar crime culpability: bribery, perjury, and fraud.

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |