Research on affective influences on social judgment has developed over the last 20 years to yield not only general models of how those influences occur [34, 36, 37] but also considerable sub-literatures on affect and particular kinds of social judgment. One such topic of research is the role of emotions on legal judgments. The growing recognition of the importance of emotions in the making and practice of law [3] indicates the need for a better understanding of existing empirical research, so as to inform both descriptive and prescriptive issues. Moreover, an important recent model of moral judgment emphasizes the role of affect-driven intuitions over moral reasoning [45]; as a species of moral judgment, legal judgment might also be predicted to be strongly influenced by such intuitions, which makes it worthwhile to explore in more detail how emotions may affect legal judgments. As yet, however, this research has not been systematically reviewed, nor has any model been proposed to account for the various empirical findings. In this paper, we critically review the available research on moods, emotions, and attributions of responsibility and blame. In particular, we identify and explain several types of affective influence on such attributions and propose a model that accounts for all of them. We also discuss the extent to which legal decision makers can correct for any affective influences they perceive to be undesirable. Finally, we briefly examine how this research can inform the normative question of whether emotions ought to affect legal judgments, if at all.

We start with a few ideas about affect that seem well established in the literature. Emotion includes feelings, cognitions, and actions (or inclinations to act, whether or not realized; e.g. [70]). The primary function of emotions is to signal changes in the environment that are important to the person experiencing the emotion, and to help that person choose among and coordinate competing goals and values [17, 73]. Although some of this processing occurs pre- and subconsciously, the signaling function is inescapably cognitive. The cognitive theory of the emotions (e.g. [62]), which most research on emotions and social judgment takes to be valid, explains that each emotion depends on an (ordinarily) implicit appraisal of the significance for the person of changes in that person's environment (see [72]). Appraisals that differ along various dimensions—for example, whether the person feels relatively certain or uncertain about the circumstances or whether the person feels that the circumstances are subject to his or her own control, another's control, or no one’s—are reliably associated with different emotions [80]. Accordingly, researchers have posited that emotions, or groups of them, can be differentiated from one another on the basis of the general cognitive structures of these appraisals, also known as their “appraisal structures” or “core relational themes.” For instance, the cognitive structure of anger is “disapproving of someone else's blameworthy action and being displeased about the related event” ([62], p. 148); its core relational theme is to have perceived “a demeaning offense against me and mine” ([54], p. 164).

Footnote 1

Moods may generally be distinguished from emotions as being less intense, more diffuse, relatively enduring, and tending to lack a readily identifiable source (e.g. [18]). The distinction, however, is sometimes difficult to draw. Moods also fulfill a signaling function, but are concerned primarily with the valence of the affective experience (i.e., whether it is positive or negative). This makes it more difficult to conduct detailed analyses of how the specific contents of moods influence subsequent attributions of responsibility and blame.

Emotions and moods can influence legal judgments in at least three kinds of ways. They can affect people's strategies for processing information—the extent to which people's processing of information tends to be “top-down” or schema-driven versus “bottom-up” or data-driven—that can in turn affect legal judgments. For instance, anger may incline jurors to use stereotypes (an instance of top-down processing), and thus to attribute responsibility in conformance with the stereotype, for example, by being more likely to hold a criminal defendant liable for a stereotype-consistent than a stereotype-inconsistent offense (see [9]). Moods and emotions can also bias judgment in a direction consistent with the valence of the mood—a mood-congruency effect. Jurors in a negative mood, for instance, may perceive more negative information about a party, recall more negative information about that party, and thus be influenced by that biased data set when judging that party's liability. Finally, emotions, and to a lesser extent moods, can directly or indirectly affect how people make particular kinds of judgments by offering informational cues to the proper attribution of responsibility or blame (as we discuss in detail later in this paper).

These types of affective influence are not wholly distinct. For instance, the informational theories of emotion effects, discussed in more detail later, could be considered as species of mood-congruency effects, in that the decision maker's emotion (or mood) inclines the decision maker toward judgments that are consistent with the emotion (or the valence of the mood). In addition, mood effects on information processing style may be mediated by the informational cues the mood provides (i.e., negative mood signals to the person that the environment is problematic, motivating the person to engage in more careful, “bottom-up” processing) [7].

Fig. 1
figure 1

Paths of mood and emotion effects on attributions of legal responsibility and blame. (a) Appraisal tendency—incidental (Path 3(a)): The experience of incidental emotion makes that emotion's cognitive structure more accessible and thus more likely to be used in construals of case features and judgments of responsibility. (b) Appraisal tendency—integral (Path 3(b)): The experience of emotion aroused by case features makes that emotion's cognitive structure more accessible and thus more likely to be used in further construals of case features and judgments of responsibility. (c) Affect-as-information—incidental (Path 4(a)): Incidental emotion is taken as directly informative of judgments of responsibility. (d) Affect-as-information—integral (Path 4(b)): Emotion aroused by case features is taken as directly informative of judgments of responsibility

Full size image

Figure 1 depicts the various paths of affective influence on these sorts of legal attributions. We will discuss each of these types of influence in turn. First, however, we set the stage with two recently proposed models purporting to describe the role of emotions in people's moral and legal judgments.

Emotions and moral/legal judgments: affective valence models

Two recently proposed models stress the fundamental role of intuitive, affective responses in the formation of moral judgments. Haidt [45, 46] argues that moral judgments arise from quick and automatic moral intuitions that include an affective valence (good or bad, like or dislike); moral reasoning is then enlisted to generate post hoc justifications for one's judgments that can be used to persuade oneself and others that the judgments are correct. Reasoning only rarely overrides the initial intuition. Similarly, Alicke [1] offers a culpable control model in which “relatively unconscious, spontaneous evaluations … [which] are affective reactions to the harmful event and the people involved” drive judgments of blame ([1], p. 558).

The notion that intuitive judgment is the default process is consistent with several dual-process theories of cognition and social judgment ([45], p. 820; see [15]). Cognitive-Experiential Self-Theory (CEST) [25, 26, 63], for instance, posits that people process information through an affect-driven experiential system in which responses are automatic and more rapid than those produced through people's more effortful and logical rational system. Haidt's ([45], p. 822) argument that reasoned explanation serves primarily to construct rationalizations for intuitively reached judgments is consistent with Nisbett and Wilson's [61] and other research on the post hoc nature of causal reasoning ([45], p. 822). The centrality of moral emotions rather than moral reasoning to moral action is supported by research by Damasio [17] and others on the inability of persons with prefrontal cortex damage to form appropriate moral judgments, and by Batson [5] and others on the role of empathy in motivating altruistic behavior ([45], p. 824).

The analysis of affective influences on legal judgments of responsibility and blame that we propose below is to some extent consistent with Haidt [45] and Alicke [1]. Like both the moral intuitionism and culpable control accounts, we find that observers’ intuitive affective responses may influence their attributions of legal responsibility and/or blame, although judgments of legal responsibility differ from those of moral or ordinary blaming in being circumscribed by rules of substantive and procedural law, including formal burdens of proof and rules of evidence that exclude from consideration matters that judges of responsibility outside the law may consider. Thus, purely rationalist explanations of judgments of legal responsibility and blame are probably inadequate.

More importantly, the research we review also indicates divergences from these affective valence models. For instance, although the research has shown that decision makers’ emotions can exert significant effects on their judgments of legal responsibility and blame, those effects are limited in size and do not always extend to all legally relevant dependent measures (e.g., damage awards), and so they do not necessarily support Haidt's [45] broad claim that it is usually affective intuition alone, as opposed to non-affective intuitions and/or conscious reasoning, that drives consequent judgments. Moreover, specific factors such as accountability ([55]; for a review, see [58]) and perceived injustice [44] have been shown to attenuate the effects of emotion on legal judgments.

Pizarro and Bloom [69] have argued that Haidt's [45] moral intuitionist model understates the role of reasoning in moral judgments, for instance, by ignoring how cognitive appraisals give meaning to intuitive emotional responses and how people can educate their intuitions by selective exposure to relevant stimuli. We do not directly address whether the moral intuitionist or culpable control models understate the role of reasoning in attributions of legal responsibility, although this may fairly be inferred from our sense that those theories overstate the causal force of spontaneous affective reactions to attributional tasks. (The model depicted in Fig. 1, for instance, allows for the infusion of affective influence at any one or more of several steps in the decision making process; by implication, the decision maker could [as indicated by Path 7] proceed from construal of target features to attribution of responsibility to appropriate punishment or damage award without any significant affective influences at all.) We now proceed to review some of the complex roles that affect plays in judgments of legal responsibility.

Emotion effects on information processing

The first way in which affect can influence judgments of legal responsibility and blame is through its effect on decision makers’ information processing strategies. Research has uncovered robust effects of moods and emotions on information processing, including receptivity to persuasive messages. Many studies, for instance, have shown that people in moderately positive moods tend to think more creatively and to be better at drawing associations and at inductive reasoning than people in a neutral mood, whereas people in moderately negative moods tend to be better at analytic and deductive reasoning [35, 38, 50]. Happy moods tend to increase reliance on heuristics, an instance of “top-down” information processing [7, 64]; negative moods tend to lead to more deliberate, bottom-up information processing [39, 64, 66], and hence to less reliance on peripheral and heuristic cues, according to the prominent “Elaboration Likelihood Model” (ELM) of attitude change [65]. These effects of mood on information processing can be moderated by, among other things, people's need for cognition and the way the information is framed [86].

An increasing amount of research traces the influence of affect on information processing strategies not (only) to the valence of the affective state but to particular qualities of the specific emotion experienced. For instance, some studies have found that although anger and sadness are both negatively valenced emotions, only anger leads to less systematic information processing (i.e., greater reliance on heuristics). This effect is because of what has been labeled the appraisal tendencies of the respective emotions (e.g. [83]). Specifically, some emotions (e.g., anger, disgust, and happiness) are typically associated with a greater sense of certainty; others (e.g., hope, anxiety, and some forms of sadness) are typically associated with uncertainty (cf. [62, 80]). The more certain people feel, the less inclined they are to process information systematically, because they are more confident that they already know what they need to know to address the task at hand. Accordingly, Tiedens and Linton [83] found that the higher degree of certainty associated with anger, as opposed to sadness (or fear), leads to greater susceptibility to heuristic cues. In one experiment, for instance, participants were asked to indicate how much they agreed with an essay on grading policies. In one condition the essay was presented as a published work and attributed to an education professor (expert); in the other, it was presented in typewritten format, authored by a student (nonexpert). Participants who had been induced to feel angry (or another emotion associated with certainty) were influenced by the expertise cues (they registered greater agreement with the “expert” essay); those induced to feel worry were unaffected by these cues. Tiedens and Linton [83] also found that angry participants were less able to distinguish substantively stronger from weaker arguments. Other researchers have similarly found that anger leads people to consider fewer factors when making judgments [55] and makes them more likely to be influenced by stereotypes in making related social judgments [8, 9]. (In addition to accounting for emotion effects on information processing style, appraisal tendency has also been offered to explain how emotion may provide an informational cue to decision making. We discuss this and other informational effects of emotion on judgment later.)

Information processing style would be predicted to mediate the effect of emotions on attributions of legal responsibility and blame differently in different situations. Anger, for instance, may enhance or mitigate blaming of a target person, depending on whether the peripheral processing that anger increases favors or discourages attributing blame to that person. Thus, undergraduate participants in whom anger had been induced were likelier to find a peer guilty of a stereotype-consistent than a stereotype-inconsistent offense [9].

All of the effects discussed so far are produced by moods or emotions that are incidental to the judgment task, as opposed to moods or emotions provoked by responses to the target of judgment. They are indicated by Path 1 in Fig. 1. It is also possible that integral moods or emotions provoked by features of the judgment target may yield information processing effects with implications for attributions of responsibility or blame, although this has not yet been clearly shown (see [77]).

Footnote 2

Directional processing

People's moods can not only affect the type of processing in which they engage when pondering information leading to social judgments; moods can also incline people to construe that information in a direction consistent with the valence of the mood, a mood-congruency effect. People in positive moods tend to make more positive evaluations of ambiguous information; people in negative moods tend to interpret the same information more negatively [12, 40, 41, 66]. Directional processing of this sort would be expected to affect legal decision making. Jurors in a negative mood, for instance, would be predicted to perceive more negative information about the judgment target, to recall more negative information about the target, and thus to be influenced by that biased data set when forming ultimate judgments of responsibility. And, as is the case with regard to affective influences on processing strategies, recent research has identified these kinds of directional effects for specific emotions as well as more general moods. For instance, DeSteno et al. [19] found that inducing anger in participants led them to judge angering events to be more likely to occur than sad events, whereas inducing sadness led them to estimate sad events to be more likely to occur. All of these directional processing effects are produced by incidental affect, that is, the person's affective state before encountering target-relevant information. They are indicated by Path 2 in Fig. 1.

Both mood effects on information processing strategy and mood-congruency effects can sometimes be reduced if the person making the judgment is sufficiently motivated and able to recognize the possible bias and to correct for it, as explained by the “Flexible Correction Model” [67, 85]. Correction processes can, however, result in overcorrection (e.g., a mood-incongruence effect) [6] or in an unwitting exacerbation of the initial congruency bias [66]. We discuss these and other effects of attempts to correct for perceived emotional biases later in this paper.

Informational effects of emotions on attributions of legal responsibility and blame

Researchers using various experimental designs have identified several ways in which emotion affects and/or is affected by attributions of legal responsibility and blame, as well as the actions or inclinations to act (such as punishing or awarding compensation) associated with those attributions. These are indicated as follows in Fig. 1:

  • Path 3: From emotional state indirectly to attributions of responsibility (and associated actions or inclinations), via construal of features of the judgment target (appraisal tendency). These effects could be provoked by incidental (3(a)) or integral (3(b)) affect.

  • Path 4: From emotional state directly to attributions of responsibility (and associated actions or inclinations) (affect-as-information). Theses effects could be provoked by incidental (4(a)) or integral (4(b)) affect.

  • Path 5: From construal of target features to attributions of responsibility to emotional response (and associated actions or inclinations to act).

  • Path 6: Feedback loops from emotions provoked by construals of target features (4(b)) or attributions of responsibility (5) back to further construals of the target (3(b)) or to the attribution of responsibility (4(a) or 4(b)).

Footnote 3

Footnote 4

In an aggregate, these paths show various relationships between decision makers’ emotions and their attributions of legal responsibility and blame. First, although all these emotion influences fall under the general rubric of informational effects-–in legal judgment tasks as in other situations, emotions can be functional to the extent they inform the person experiencing the emotions about relevant features of the environment and motivate the person to respond appropriately [17]—the research suggests that there may be meaningful differences between the appraisal tendency account of indirect effects [56, 57] and the affect-as-information account of direct effects (e.g. [16]). Second, Fig. 1 indicates that one such difference is that the effects of emotion on attributions may be either indirect (Paths 3(a) and 3(b)), as in appraisal tendency, or direct (Paths 4(a) and 4(b)), as in affect-as-information. Emotion influences attributions indirectly when the effects are mediated by some other process or judgment, such as the person's construal of features of the judgment target. Emotion influences attributions directly when nothing mediates that influence. Third, attributions of legal responsibility and blame may be affected by moods and emotions that are incidental or extrinsic to the judgment target or task(Paths 3(a) and 4(a)) and/or by moods and emotions that are elicited by the target or task, and should thus be considered integral (Paths 3(b) and 4(b)). Fourth, the model accommodates complex influences via feedback loops, in which emotions generated in response to the target in turn affect further construals of the target (thus indirectly affecting attributions of responsibility) and emotions generated by attributions in turn shape those attributions either indirectly (through further target construals) or directly. Finally, we note that nothing in the research reviewed excludes the role of nonemotional cognitions, including cognitive biases resulting from the use of heuristics such as availability and representativeness, in judgments of responsibility and blame; these are represented by Path 7 in Fig. 1.

In Path 3(a), incidental emotion affects the way people construe the relevant features of the target, which in turn affects their subsequent judgments. For instance, a consistent finding in the research is that people who are angry tend to blame more. Lerner et al. [55] found that participants who viewed an anger-provoking video clip and then read several vignettes of accident cases blamed the defendants who caused the injuries more than did participants who had watched an emotion-neutral video. Similarly, Keltner et al. [52] found that angry participants tended to attribute more responsibility to the person than the situation regarding ambiguous social mishaps; sad participants did the opposite. Keltner and his colleagues found that anger and sadness affected participants’ attributions of causal responsibility, which in turn affected blaming.

Footnote 5

The process or mechanism posited to explain this influence for incidental emotion on subsequent judgments of responsibility is appraisal tendency, the same process offered to explain the effects of specific emotions on information processing mentioned previously. Experiencing an emotion makes features of that emotion's cognitive structure more accessible [12, 13] and thus more likely to be utilized (consciously or not) in subsequent perceptions and judgments. (This is true whether the emotion induction itself involves cognitive appraisals of causal responsibility [e.g., stories meant to provoke anger or sadness] or is purely physical [e.g., adoption of facial expressions and postures indicated by anger and sadness] and thus largely noncognitive, proving that it is the experience of the emotion rather than the cognitive appraisals alone that influence subsequent attributions; [52], pp. 748–749.) Thus, angry people blame more because the cognitive structure of the anger they experience (“disapproving of someone else's blameworthy action and being displeased about the related event”) makes salient the role of dispositional factors of other people (as opposed to situational factors) as causes of harm, which in turn engenders blame. Thus, emotion influences blaming judgments indirectly: Path 3(a) proceeds from incidental emotion to construal of case features and thence to attribution of causal or legal responsibility or blame.

Footnote 6

One especially interesting feature of the appraisal tendency process is that even where people are aware that the source of their emotional state has nothing to do with the judgment target, the emotion continues to affect their judgments [60]. Anger, for instance, has been shown to persist past the emotion-provoking episode in the form of a residual arousal or excitation, which may then influence subsequent, unrelated decisions [92]. Apparently, people remain at least partly unaware of the ways in which that emotion has primed them to construe the target ([55]; cf. [91]).

As yet there has been no research showing appraisal tendency effects from integral, as opposed to incidental, emotion. Neveretheless, we hypothesize the existence of Path 3(b) because there seems to be no reason in principle why emotion elicited in response to features of the judgment target or the initial attribution of responsibility (as opposed to emotion induced by manipulations extrinsic to the judgment task) might not influence further evaluations of target features, and thence, attributions. (We discuss this possibility below with respect to “feedback loops.”)

In Path 4, people take their experience of an emotion as directly informative about the target of their judgment. Indeed, Path 4 is described in the literature by the affect-as-information model [16, 73–76). Emotions can have direct effects on ultimate judgments when the emotions are incidental, i.e., substantively irrelevant, to the judgment target or task (Path 4(a)). This happens when people misattribute their emotional response to the target instead of its true source. For instance, in a classic study [75], people asked on rainy days to gauge their life satisfaction gave more negative responses than did people asked on sunny days. When the attention of the former group of respondents had been called to the weather, however, the difference disappeared. That is, people took their current mood (negative or positive) as informative about the judgment target (“How satisfied am I with my life?”). In effect, they misattributed the experienced emotion (provoked by something not relevant to the target of judgment) to the target. When the misattribution was corrected by identifying the true source of their emotion, the effect disappeared. Recently, Dunn and Schweitzer [21] found that direct incidental emotion effects on judgments of trust were moderated not only by the valence of the emotion but also by the secondary appraisal dimension [80] of control. Incidental emotion affected how much participants trusted another person (a co-worker) in a direction consistent with the valence of the emotion (e.g., happy participants expressed greater trust than sad ones), but in addition, emotions with a control dimension consistent with the judgment task affected trust more than did judgment task-inconsistent emotions: Anger, which is associated with other-person control, affected trust of another person more than did a similarly valenced emotion (e.g., sadness) not associated with other-control.

Incidental emotion effects on many sorts of decisions, from judgments of life satisfaction [75] to risk perceptions [19] to levels of trust of another person [21], have been explained in terms of the affect-as-information mechanism. Incidental affect has recently been shown to affect participants’ judgments about the justice of distributions of rewards to themselves and others who engaged in a task, where participants were uncertain about information important to their justice judgments [84]. Affect-as-information has also been invoked to explain incidental emotion effects on judgments of blame [42]. To the best of our knowledge, however, there are no studies that directly test the process using such judgments as the dependent variable.

Emotion that is integral to the judgment task, that is, provoked by a consideration of features of the judgment target, can also directly affect attributions of responsibility. In Path 4(b), relevant features of the case may provoke emotional responses, which in turn affect attributions of responsibility or blame. This path is established by research showing that jurors’ emotions mediate [4] the effects of case features, such as the severity of an accident or a party's blameworthiness, on attributions of responsibility and damage awards. For instance, Bornstein [11] has found that sympathy mediates the effect of outcome severity on mock jurors’ responsibility judgments. In one set of experiments, a product liability lawsuit against the manufacturer of a birth control pill, mock jurors were more sympathetic to the more seriously injured plaintiff, and this greater sympathy made them more likely to find the defendant liable. Similarly, Feigenson et al. [33] found that anger mediated the effect of the parties’ blameworthiness and the severity of the outcome on their apportionments of fault (but not their damage awards) in comparative negligence cases. Increasing the severity of the accident made participants angrier at the defendant, which led them to apportion more fault to the defendant; increasing the plaintiff's blameworthiness made them angrier at the plaintiff, which led them to apportion more fault to the plaintiff. Finally, Douglas et al. [20] found that mock jurors in a murder case who viewed autopsy photographs were more likely to report feeling anxious, anguished, disturbed, and shocked than those who did not view the photographs, and that the more anxious and shocked the mock jurors were, the more they believed that the defendant was guilty. These emotions, therefore, mediated the effect of the independent variable (autopsy photograph vs. no photograph) on verdicts.

The most plausible explanation for the emotion effects in Path 4(b) is that, as in Path 4(a), people are using their current emotional state as an informational cue regarding the judgment target. For example, because the cognitive or appraisal structure of anger is “disapproving of someone else's blameworthy action and being displeased about the related event” ([62], p. 148), being angry sends a signal [17] to the person that the target of judgment has behaved in a blameworthy fashion and, therefore, deserves to be blamed. The difference between Path 4(b) and Path 4(a) is that the emotion in Path 4(b) is integral to the judgment task.

In Path 5, relevant features of the case can affect attributions of responsibility and blame, which in turn affect emotional responses and associated action tendencies. The research design takes some stimulus of interest—say, how blameworthy the victim is—as the independent variable, and measures emotional response and inclination to act on it as the dependent variable. For instance, in a series of studies spanning a generation, Weiner and his associates have found that emotional responses to suffering depend on attributions of responsibility [87]. When an observer perceives a person in need of aid (including a victim of accident, disease, or natural disaster), the observer attempts to discern the cause of the need. If the cause is perceived to be outside the sufferer's control, the observer reacts with sympathy and is inclined to help. If the cause is perceived to be within the sufferer's control, the observer reacts with anger and is inclined to ignore the sufferer. Thus, emotion figures as an output of the attribution of responsibility or blame.

Footnote 7

It seems likely that the paths of affective influence on attributions of responsibility or blame in any given case may be recursive rather than linear. That is to say, the decision maker may respond emotionally to the facts of the case (features of the judgment target) or to his or her own attribution of responsibility, and these emotions may in turn influence further consideration of the facts or further rumination toward the attribution. In this way, emotions and legal judgments can form feedback loops, identified by the several Paths 6 in Fig. 1. For instance, the construal of target features can generate anger (Path 4(b)), as can the attribution of responsibility to the victim for the harm he or she has suffered (Path 5). That anger then makes salient the role of dispositional factors of other people as causes of harm, engendering blame (Path 3) [52]. Thus, anger and attributions of blame comprise a reciprocal relationship in which each can increase the other (see also [71, 82]. In addition, given research showing that jurors subconsciously adjust their ultimate judgments and their evaluations of the evidence and arguments on which those judgments are based to achieve cognitive coherence [78, 79], jurors may take the emotions they experience during this judgment process as a cue to whether they have completed this process satisfactorily (see [29, 33]).

The most explicit, albeit indirect, empirical support for feedback loops is provided by Quigley and Tedeschi [71], who found that participants’ anger mediated the effects of participants’ perceptions of the amount of harm, the target's intent to harm, and the target's justification for inflicting harm on their judgments of blame, and that their judgments of blame mediated the effects of these same variables on their anger. Their structural model is consistent with the feedback loops in our model but, because of differences in design, it does not directly support them. Nevertheless, it seems reasonable to suppose that such feedback loops are a common feature of actual legal decision making.

Footnote 8

Finally, it is important to point out that even the multiple paths of affective influence indicated by Fig. 1 understate the complexity of emotion effects on judgments of legal responsibility and blame. For instance, the same emotion relating to the same judgment target might incline decision makers in conflicting directions: Angry jurors might in general be more inclined to hold a criminal defendant responsible for engaging in the charged behavior, but if the defendant belongs to a group stereotypically thought to possess positive rather than negative criminal traits, jurors’ anger might incline them to rely more on those stereotypes and find the defendant less blameworthy. Jurors may also experience multiple (integral) emotions toward a given party that conflict with one another. For instance, in a comparative negligence case, jurors may feel sympathy for an accident victim, which inclines them to hold the defendant more responsible [10]; yet they may also feel anger toward that same victim, which would incline them to hold the victim more responsible and the defendant less [32].

The fact that legal cases present multiple judgment targets creates further complex relations among jurors’ emotional reactions and between those reactions and their ultimate decisions. Jurors’ emotional responses to multiple parties may, for instance, complement each other: In a comparative negligence case, for example, the angrier jurors are at the defendant, the more sympathy they may feel for the plaintiff [33], and the more sympathy they feel for the plaintiff, the less sympathy they may feel for the defendant [10]. Jurors’ emotional responses to multiple targets may also pull in opposite directions. Jurors in a comparative negligence case may feel increased anger at a defendant who causes a more severe injury, which leads them to apportion a greater percentage of negligence to the defendant, but this may be counterbalanced by their increased anger at a more blameworthy plaintiff, which may lead them to attribute more negligence to the plaintiff (and thus less to the defendant) [33]. And Horowitz et al. [49] found that decision making in a criminal case was influenced not only by jurors’ emotional response to the victim (as distinguished from the defendant) but also by aspects of the decision making context: Increased sympathy for a sympathetic crime victim made jurors who received nullification instructions more upset about the crime, and this feeling of upset made them more likely to convict the defendant. Still more complex combinations of emotional influence on legal judgments can be imagined. In short, the model of mood and emotion effects we present here does not encompass every permutation of affective influence on judgments of legal responsibility.

Do the differences among paths of affective influence matter?

On the basis of the relatively few studies reported in the literature, it is not yet possible to decide whether the apparent complexity of the emotion effects included in the model we propose indicates a corresponding complexity in underlying processes, or whether the apparent differences among paths of emotional influence may be at least partly epiphenomenal. We have already noted, for instance, that both appraisal tendency and affect-as-information are types of informational effects, which is consistent with the generally adaptive nature of emotions as perceiver-relevant signals about the environment [17]. Our model does make clearer precisely how the different paths of affective influence on judgments of legal responsibility identified in the literature may to some extent be the result of various experimenters’ choices of different independent and dependent variables, rather than being truly indicative of different underlying processes.

The model also suggests, however, that the different paths of affective influence may have both basic psychological and practical legal significance. For instance, the distinction between integral and incidental affect may be important. Integral emotion would be expected to exercise a greater influence on ultimate judgments of responsibility than would incidental emotion, for at least two reasons. First, a person experiencing integral emotion (e.g., as a mediator of the effect of target features on attributions, Path 4(b)) is less likely than one experiencing incidental emotion (e.g., Path 4(a)) to be persuaded that the emotion is in fact incidental, and thus less likely to treat it as irrelevant to the judgment task [75] and possibly to undertake corrective or debiasing measures (discussed later). Second, integral emotion can feed back into reconstruals of the judgment target, further shaping interpretation of the evidence and consequent judgments. On the other hand, it could be argued that because incidental emotion is more likely than integral emotion to be implicit, decision makers are less likely to be aware of incidental emotion in the first place and thus less likely to undertake corrective measures in response; hence, incidental emotion may exercise a greater influence on judgments of responsibility than would integral emotions.

The distinction between indirect (i.e., mediated) and direct (i.e., non-mediated) paths of affective influence may also be significant. For instance, because appraisal tendency triggered by incidental emotion affects attributions of responsibility indirectly, through construal of judgment target features, whereas affect-as-information triggered by incidental emotion affects those same attributions directly, the nature of particular target features or particular items of evidence would in some circumstances be predicted to moderate appraisal tendency effects but not affect-as-information effects. To test this, participants in angry or neutral moods because of incidental stimuli could be exposed to cases of accidental injury in which the victim's responsibility (as opposed to situational factors) was clearly high, ambiguous, or clearly low. Consider, for example, the driver of a car who approaches an intersection at which, according to traffic signs, the driver should yield right of way to vehicles traveling on the other road; the driver proceeds into the intersection and is struck by one of those vehicles. In the high responsibility condition, the driver has a clear view of crossing traffic but barely pauses before entering the intersection; in the ambiguous responsibility condition, the angle of intersection and overhanging foliage partially obscure the driver's view of other traffic, and the driver slows before easing into the intersection; in the low responsibility condition, the driver's view is even more occluded and the driver comes to a complete stop before proceeding. Appraisal tendency predicts interactions of incidental affect and judgment target features, such that angry participants should attribute greater blame to the victim than neutral participants in the ambiguous responsibility condition but not the others. (This is because incidental affect should influence target feature construals by making the victim's responsibility, as opposed to situational factors [the angle of intersection and the blocking of the driver's view by foliage] more salient in the case of anger, but not where the target features already make highly salient the victim's responsibility or situational factors, respectively; see [52].) Affect-as-information, by contrast, predicts main effects for incidental emotion: Relative to neutral participants, angry ones should attribute more blame to the victim regardless of target responsibility condition (assuming that the high and low responsibility conditions do not yield ceiling and floor effects, respectively). Although emotion might not be needed as an informational cue through either process if the evidence going to victim responsibility is clear enough, greater sensitivity to variations in that evidence would seem to indicate appraisal tendency rather than affect-as-information.

Another potentially important difference between the appraisal tendency and affect-as-information mechanisms, despite their basic functional commonality as informational cues to judgment (see [21]), is that decision makers who recognize an emotion source as incidental would not be likely to take their emotion as informative of the judgment target; therefore, affect-as-information predicts no emotion effect on judgment in this situation, whereas appraisal tendency would still allow for some residual emotional influence on judgment [55]. This difference might become manifest in a case in which a powerful emotion such as anger could plausibly be attributed to either an incidental or an integral source, but not both. For example, in a civil suit brought on behalf of victims of the terrorist attacks on the World Trade Center against the airlines whose planes were hijacked and other non-terrorist defendants, jurors might approach the case with an anger that stems from their response to the terrorists but that could spill over to the defendants [30]. Stimulus materials could be designed so that the defendants are not legally blameworthy and not deserving of anger or blame; participants in one experimental condition could be instructed to recognize the terrorists as the real source of their anger, and thus to cease misattributing that anger to the judgment target. If anger continued to affect participants’ attributions of responsibility and blame to the defendants, that would be evidence for the operation of the appraisal tendency process.

There are, therefore, good reasons to try to untangle the effects of appraisal tendency from those of affect-as-information within legal decision making. Which would be expected to be more influential? Forgas's [37] “Affect Infusion Model” posits that when judgments are made through heuristic processing, as they are when the judgment target is unfamiliar, the judgment task is relatively simple, and the decision maker has limited processing resources and is not strongly motivated toward a particular outcome, the affect-as-information mechanism is triggered. When, in contrast, the judgment task is complex or unusual and the situation demands accuracy or accountability, then substantive as opposed to heuristic processing occurs, and affect plays an even greater role in judgment, especially via affect priming [37]; appraisal tendency is (as noted) a kind of priming mechanism. Persons who must determine legal responsibility at trial face a task that has features indicative of both heuristic and substantive processing. One way to define more precisely the factors that moderate decision makers’ inclination to utilize affect-as-information more than appraisal tendency (or vice versa) would be to conduct either of the experiments suggested earlier while also manipulating one or more of these features of the judgmental target and situation. For instance, simplifying the facts and reducing the available time for deciding would be predicted to increase affect-as-information effects.

Correcting for the influences of emotion on legal judgments

To the extent that any emotion effects on judgments of legal responsibility are considered undesirable, what does the research show about decision makers’ ability to reduce or eliminate those effects? Generally speaking, the research on debiasing (or correction) indicates that in order to purge judgments of unwanted bias, the decision maker must be (i) aware of the unwanted influence; (ii) motivated to correct the bias; (iii) aware of the magnitude and direction of the bias; and (iv) able to adjust the response appropriately [88, 89].

Legal decision makers may fail to satisfy any or all of these criteria. First, decision makers may perceive no need to correct for bias; they are likely to remain unaware of many sources of unwanted influence on or “mental contamination” of their decision making, if only because people usually believe that their own thinking and judgments are unbiased [23, 89]. Second, some legal decision makers may not be sufficiently motivated to correct for any emotional influence, believing that taking at least certain emotional responses into account—such as their sympathy for accident victims—is proper, notwithstanding the judge's instructions to the contrary. Third, judicial instructions to disregard emotional influence may even lead to a “paradoxical” effect in which that influence is enhanced, not diminished, because of the increased availability of the proscribed influence [22] or reactance on the part of jurors [59]. Fourth, decision makers may not know how to adjust appropriately even if they perceive the need to debias and are motivated to try. For instance, decision makers who are made highly aware of their feelings and are highly motivated to reach a fair and accurate decision may overcorrect for any emotional influence [6].

On the other hand, there is some evidence that legal decision makers can follow instructions not to be improperly influenced by their emotions. DeSteno and his colleagues found that participants with high need for cognition could avoid the biasing effects of emotion on risk perception when instructed to be careful and accurate (and where the emotion manipulation was very salient) [19]. Moreover, knowing that one will be accountable for one's decision has been shown to attenuate the effect of incidental emotional influence on that decision, specifically, anger leading to punitiveness [55, 58]. Thus, although the research suggests that the affective influences on judgments of legal responsibility and blame are likely to persist in real legal settings, the moderating factors identified by DeSteno et al. [19] and Lerner et al. [55, 58] add support to the position that affective influences may not be as intractable and pervasive as Haidt [45, 46] and Alicke [1] contend.

Emotions and legal judgments: normative issues

To this point our goal has been purely descriptive: We have summarized the relevant research on the role of affect in judgments of legal responsibility and blame, identified various causal relationships in the judgment process, and suggested directions for further research to test aspects of the proposed model. We have remained agnostic concerning the roles, if any, that emotions should play in legal judgment. In this concluding section, we briefly address this topic and discuss how our descriptive work and subsequent research may shed light on it.

It could be argued that legal doctrine and the formal legal system display ambivalence toward emotion as a component of law making and adjudication. On the one hand, norms of legal decision making have traditionally stressed rationality and dispassion [28]. Standard jury instructions discourage decision makers from using their emotions to decide cases (e.g. [90]), and trial judges often exclude evidence precisely to avoid provoking jurors’ emotional responses, fearing that those emotions will “prejudice” jurors’ decisions [27]. On the other hand, many aspects of evidence law and trial procedure acknowledge jurors’ emotions and perhaps even enhance their salience, for example, through dramatic concentration (see [14]). Certainly trial lawyers may seek to activate jurors’ emotional as well as their cognitive processing, whether through captivating storytelling (e.g. [81]), witness examinations that seek to bring out “visceral” case themes (e.g. [2]), or vivid demonstrative evidence (e.g. [20, 31]), although emotional appeals that jurors not only recognize but perceive to be excessive can be ineffective [47]. Indeed, some legal scholars have promoted a greater role for emotions in the framing of legal rules and the deciding of particular cases (see, e.g. [3]). In any event, the widespread belief that legal decision makers’ (perhaps especially but not exclusively jurors’) emotions do influence their judgments makes considering the roles of affect in legal judgment a matter of some significance to judges, advocates, litigants, and members of the public who are interested in how the legal system works.

How might an analysis of appropriate and inappropriate emotion in judgments of legal responsibility and blame be informed by the model of affective influences we offer here and further empirical research guided by that model? We begin with the distinction between incidental and integral emotion sources. It may seem intuitively obvious that incidental emotional influence cannot be justified within any conceivable model of appropriate legal (or moral) judgment [48]; incidental emotion sources, like nonemotional information sources, would seem to be simply irrelevant to the judgment task. Some have argued to the contrary, that certain kinds of emotionality are desirable traits for legal decision makers to possess (e.g. [68]), even though the origins of those traits are necessarily incidental to the case at hand. In any event, the model we present makes clear, on the one hand, just how difficult it may be to disentangle incidental from integral emotion in legal decision making, and on the other, that the two kinds of sources of emotion may yield different effects in certain circumstances, as indicated by the study we propose comparing the appraisal tendency and affect-as-information mechanisms.

Not all integral emotions are normatively justifiable in legal decision making, even for those who believe that affective influences are not improper per se. Among integral emotion sources, it is necessary to distinguish between “legal” (or “evidential”; [1]) and “extralegal” (or “extraevidential”) sources. Legal integral emotion sources would include, for instance, the nature of the case-relevant behavior of relevant persons and the outcomes of those behaviors (cf. [1]). Extralegal integral emotion sources would include a party's race or gender (where not relevant to a legal claim or defense), information learned about the case through pretrial publicity, and so on. Emotional responses triggered by extralegal factors, like incidental emotion sources, would appear to be normatively unjustifiable [48].

A closer question is posed by the difficulty of classifying certain integral emotion sources as legal or extralegal. For instance, under Federal Rule of Evidence 404, whether information regarding a party's character (and hence any emotional responses that information elicits) should be considered legal or extralegal may depend on the way in which that information is introduced at trial [27]. To choose one example, the defendant in a criminal trial may introduce evidence about his own relevant, positive character trait (e.g., for peacefulness) to prove that he acted in conformity with that trait in the events at issue, but (with certain exceptions) the prosecution may not introduce evidence about a corresponding negative trait (e.g., for violence) unless the defendant offers the positive character evidence first. And quite apart from what the law provides, it may be that only a prior moral theory can tell us whether decision makers ought to take account of their emotional reactions to a party's character or values in attributing responsibility or blame. It could be argued that in order to attribute blame properly, decision makers need to determine why the party has acted as he or she did, and that such “extralegal” emotion sources as the party's character and/or values are important to that determination.

Footnote 9

Nor does it follow that all emotional responses even to integral legal emotion sources are desirable. First, the emotional response has to be an “appropriate” one, based on an accurate perception of the relevant facts [51]. Second, although appropriate emotional responses to relevant features of the case may assist good decision making by providing the decision maker with information about the case otherwise less likely to be obtained and by motivating the decision maker to act in accordance with his or her (proper) judgment, those emotional responses can also bias the interpretation of the facts, leading to normatively incorrect decisions [28]. The law then confronts the daunting challenge of accommodating decision makers’ emotional responses to integral legal emotion sources but then “educating” those emotions (presumably by means of appropriate judicial instructions) to reduce the likelihood of incorrect judgments (see generally [3, 28]).

Empirical research can shed light on how and to what extent particular emotions and moods variously inform and/or bias particular kinds of judgments. And where some adjustment of emotional influence on decision making is deemed appropriate, our model suggests that some influences on judgments of legal responsibility or blame may be more or less amenable to modification than others. For instance, as noted earlier, when people are made aware that the source of their emotion is actually incidental to the target, they should regard the emotion as irrelevant to their judgment task; the emotion should, therefore, cease to have any directly informational effect. Relatively, jurors are more likely to follow judicial instructions to disregard emotional influences that the jurors themselves perceive to be incidental, and therefore irrelevant to their judgment task (cf. [43]). On the other hand, legal decision makers will frequently need to confront integral emotion sources and make the difficult determination whether they should take their corresponding emotional responses into account in reaching their judgments (assuming that they do not regard emotional judgment as per se improper). This is because lawyers, in the ordinary course of appropriately presenting evidence and making arguments regarding the severity of the harm that plaintiffs or crime victims suffered and the blameworthiness of defendants’ conduct, are likely to trigger the integral emotional effects described by Paths 4(b) and 5 (and thence the feedback loops indicated by Path 6). Further research that tests the comparative effects of these various affective influences in a variety of legal contexts could help inform judges, lawyers, and others interested in the performance of the litigation system about how large a role emotional influences actually play, and whether the legal system can or should take steps to modify that role.