Lifting the Veil: How the Brain Decides Blame and Punishment

During the early 20th century, legal scholars were gripped in fierce debate: legal formalists, led by Harvard Law School dean C.C. Langdell, argued that judges should be impervious to emotion and guided by reason alone; in contrast, legal realists maintained that judges are people who have the same foibles and idiosyncratic biases as everyone else. The realists were building on ideas expressed by future U.S. Supreme Court Justice Benjamin Cardozo, who wrote that forces “deep below consciousness”—“the likes and the dislikes, the predilections and the prejudices, the complex of instincts and emotions and habits and convictions”—shape judges and litigants alike (1).

Cardozo’s ideas were well ahead of their time. In the 1940s, as B.F. Skinner and other behaviorists’ work was on the rise, an emotionless approach took hold. Theorists sought to explain the law in terms of observable inputs and outputs. Mind and emotion took a backseat for more than a generation.

The road back to realism was slow. Not until the 1980s and 1990s did feminist scholars, critical legal theorists, and others shatter the presumption that law is a neutral, value-free enterprise. There were far too many examples of personal influence affecting cases. Studies showed judges to be influenced by their own demographics (including race), the physical attractiveness of litigants (2), and, amazingly, something as mundane as whether their college football team won or lost (3). Troublingly, though unsurprisingly, judges were found to harbor the same kinds of implicit biases as most people (4). Legal decision making had finally become a topic of serious inquiry.

One helpful framework for understanding how decisions can go awry is dual-process theory (a version of which was eventually popularized by Daniel Kahneman in Thinking, Fast and Slow). According to this model, when people are learning new skills, practice is slow and effortful, requiring a high degree of executive function. Once people become experts, however, they rely on intuitive (“fast”) processes as a default problem-solving tool. The key idea: intuition works most of the time but can lead to mistakes.

While dual-process theory offers a compelling higher-level model for understanding decision making, it doesn’t speak to the underlying biology. Moreover, legal decision making may be especially complicated, as it is deeply rooted in interpersonal dynamics and requires a core skill—a sort of cognitive super-power—that most humans develop by age 5: the ability to attribute mental states, such as beliefs, intentions, and desires, to other people. Psychologists call this “theory of mind” (5). Without it, the social world would be a deeply confusing place.

As a graduate student at MIT, Rebecca Saxe wanted to understand which brain systems underpin theory of mind. Using functional neuroimaging, she found that a particular region located above and behind the right ear—the right temporoparietal junction (rTPJ)—is preferentially activated when making inferences about other people’s mental states (e.g., when considering how others’ beliefs may differ from their own). And if the rTPJ is involved in thinking about other people’s thoughts, Saxe wondered, could it also play a role in judging people’s behavior? To answer this question, Saxe designed an experiment in which she recruited healthy adults and read them two stories. In the first story, the protagonist (Grace) poisoned her friend’s coffee by complete accident; in the second, she intentionally tried (but failed) to do the same thing (6). After hearing both stories, participants used a scale to rate Grace’s blameworthiness in each scenario. While there was some variability in responses, everyone agreed that Grace was more blameworthy for intentionally trying to poison her friend. Of note, participants who blamed Grace more for the accidental poisoning showed less activity in the rTPJ than those who were willing to overlook her innocent mistake. The rTPJ, in other words, seemed to help participants judge Grace by her best intentions.

Saxe then took the experiment a step further and asked whether moral judgments could be changed by modulating rTPJ functioning. To do this, she used the then-novel interventional tool repetitive transcranial magnetic stimulation (TMS), which can temporarily block brain activity in the targeted region. The effect was striking: after TMS was applied to the rTPJ, participants found Grace less blameworthy for the attempted (but failed) poisoning and more blameworthy for the accidental poisoning. Having lost some of their ability to consider Grace’s intentions, they were left to judge her based solely on the outcomes.

But intentionality is only the first part of this complex story: once someone attributes blame, how does the brain adjudicate punishment? In a procedure similar to Saxe’s, Joshua Buckholtz (then a graduate student at Vanderbilt University) scanned the brains of participants while they judged blameworthiness in hypothetical criminal cases (a task highly dependent on theory of mind) (7). Consistent with evidence from previous studies, Buckholtz found that participants who deemed defendants more blameworthy had higher levels of activity in the right dorsolateral prefrontal cortex (rDLPFC), a structure involved in a host of cognitive functions that works in concert with the rTPJ. These findings underscored that adjudicating blame is largely a cognitive process.

However, the study didn’t end there. Buckholtz and his team also examined how participants meted out punishment. Surprisingly, the severity of their punishment was not correlated with rDLPFC activity; instead, it was correlated with increased activity in the amygdala and other limbic regions. In other words, severity of punishment seemed to be driven by emotion rather than by rational deliberation.

But how much can someone change the way they feel by trying to see something from another person’s perspective? To answer this question, Buckholtz’s team gave participants descriptions of crimes that varied based on emotional valence (graphic vs. neutral) and intentionality (did they mean to do this or was it an accident?). Participants who were provided graphic descriptions meted out harsher punishments (and showed stronger functional connectivity between the amygdala and DLPFC) than those who were given neutral descriptions. However, this held true only when individuals were asked to judge intentional harms (8). When judging accidental harms, medial-prefrontal structures remained firmly in control. Moreover, as in Saxe’s work, the TPJ¹ was more active when participants were using theory of mind in an effort to reconcile the offender’s (innocent) mental state with the resulting harm (Figure 1).

Figure 1.

Cognitive model (A) and neural circuitry (B) of blame and punishment. Punishment by third parties (i.e., those not directly affected by a transgression) is driven by blame, which is a function of two factors: the moral responsibility of the perpetrator and the harm caused to the victim. First, the salience network (red) detects a social norm violation (dorsal anterior cingulate cortex [dACC]) and provides an emotional signal based on the severity of the harm (amygdala). Next, the medial prefrontal cortex (mPFC) and associated regions (gray) integrate harm and intent, including from the temporoparietal junction (TPJ) (pink), into an assessment of blame. Finally, the central executive network (green) receives these signals and selects a specific punishment response (dorsolateral prefrontal cortex [dlPFC]). Adapted from Krueger and Hoffman (10).

The lesson from both studies? Our “fast” decisions about how much to punish may be largely emotion driven, intuitive, and sometimes erratic—and, at the same time, it’s possible to overturn these decisions using slow, deliberative processes.

These studies raise a number of intriguing questions about our legal system. For instance, what is the significance of the fact that some laypeople may have “more” or “better” theory of mind than others? Jurors are frequently asked to make finegrained distinctions between different, legally relevant mental states (9). It is deeply problematic if they are unable to make these distinctions, especially when defendants’ lives and liberty are at stake. Judges, for their part, would be wise to heed Cardozo’s counsel. Their emotional lives, both on and off the bench, may significantly influence their behavior, especially during sentencing. By becoming aware of the risk of bias, judges may be able to take steps to mitigate it (e.g., through mindful deliberation or peer consultation) (10).

And what about the rest of us—those who may never set foot in a courtroom? Even if we’re not immersed in legal decisions, our choices are still important. We each occupy positions of power over others at least some of the time, whether as teachers, parents, supervisors, or doctors. In medicine, while a “formalist” approach to complex clinical cases (i.e., that clinicians should be impervious to emotion and guided by reason alone) would be theoretically laudable, it is as unrealistic in a hospital as it is in a courtroom. Take, for example, the robust literature on “hate in the countertransference”—the solution to difficult encounters lies not in equanimity but in acknowledging intense affect, understanding its origin, and constructively leveraging this information.

As psychiatrists, we also see this problem from the other side: many of our patients present to us because they have difficulty with mentalization (with resultant interpersonal problems) or with balancing cognitive and affective processes in decision making. Understanding the role of these basic circuits (in a transdiagnostic way) may allow us to better understand the problems they face—and perhaps explain them in a less stigmatizing way. It may also point to effective biopsychosocial interventions, including different forms of psychotherapy (or mindfulness). Moreover, Saxe’s proof of concept work illustrates how interventional tools could (at least theoretically) be used to modify these circuits in potentially therapeutic ways.

Thanks to new scientific discoveries, we are in a better position than ever to appreciate Cardozo’s insights. As society seeks to resolve all manner of conflict, it is useful to bear in mind that we each harbor our own “complex of instincts and emotions and habits and convictions.” We would be wise to manage them—and also to accept that these imperfect decision-making processes are a part of what make us so inescapably human.