Behave: The Biology of Humans at Our Best and Worst

by Robert M. Sapolsky

  The second study comes from Greene and colleague Joseph Paxton.44 Subjects in a scanner would predict the outcome of coin tosses, earning money for correct guesses. The study’s design contained an extra layer of distracting nonsense. Subjects were told the study was about paranormal mental abilities, and for some of the coin tosses, for this concocted reason, rather than state their prediction beforehand, subjects would just think about their choice and then tell afterward if they were right. In other words, amid a financial incentive to guess correctly, there were intermittent opportunities to cheat. Crucially, this was detectable—during the periods of forced honesty, subjects averaged a 50 percent success rate. And if accuracy jumped a lot higher during opportunities to cheat, subjects were probably cheating.

  The results were pretty depressing. Using this form of statistical detection, about a third of the subjects appeared to be big-time cheaters, with another sixth on the statistical border. When cheaters cheated, there was activation of the dlPFC, as we’d expect. Were they struggling with the combination of moral and cognitive conflict? Not particularly—there wasn’t activation of the ACC, nor was there the slight lag time in response. Cheaters typically didn’t cheat at every opportunity; what did things look like when they resisted? Here’s where you saw the struggling—even greater activation of the dlPFC (along with the vlPFC), the ACC roaring into action, and a significant delay in response time. In other words, for people capable of cheating, the occasional resistance seems to be the outcome of major neurobiological Sturm und Drang.

  —

  And now for probably the most important finding in this chapter. What about subjects who never cheated? There are two very different scenarios, as framed by Greene and Paxton: Is resisting temptation at every turn an outcome of “will,” of having a stoked dlPFC putting Satan into a hammerlock of submission? Or is it an act of “grace,” where there’s no struggle, because it’s simple; you don’t cheat?

  It was grace. In those who were always honest, the dlPFC, vlPFC, and ACC were in veritable comas when the chance to cheat arose. There’s no conflict. There’s no working hard to do the right thing. You simply don’t cheat.

  Resisting temptation is as implicit as walking up stairs, or thinking “Wednesday” after hearing “Monday, Tuesday,” or as that first piece of regulation we mastered way back when, being potty trained. As we saw in chapter 7, it’s not a function of what Kohlbergian stage you’re at; it’s what moral imperatives have been hammered into you with such urgency and consistency that doing the right thing has virtually become a spinal reflex.

  This is not to suggest that honesty, even impeccable honesty that resists all temptation, can only be the outcome of implicit automaticity.45 We can think and struggle and employ cognitive control to produce similar stainless records, as shown in some subsequent work. But in circumstances like the Greene and Paxton study, with repeated opportunities to cheat in rapid succession, it’s not going to be a case of successfully arm wrestling the devil over and over. Instead, automaticity is required.

  We’ve seen something equivalent with the brave act, the person who, amid the paralyzed crowd, runs into the burning building to save the child. “What were you thinking when you decided to go into the house”? (Were you thinking about the evolution of cooperation, of reciprocal altruism, of game theory and reputation?) And the answer is always “I wasn’t thinking anything. Before I knew it, I had run in.” Interviews of Carnegie Medal recipients about that moment shows precisely that—a first, intuitive thought of needing to help, resulting in the risking of life without a second thought. “Heroism feels and never reasons,” to quote Emerson.46

  It’s the same thing here: “Why did you never cheat? Is it because of your ability to see the long-term consequences of cheating becoming normalized, or your respect for the Golden Rule, or . . . ?” The answer is “I don’t know [shrug]. I just don’t cheat.” This isn’t a deontological or a consequentialist moment. It’s virtue ethics sneaking in the back door in that moment—“I don’t cheat; that’s not who I am.” Doing the right thing is the easier thing.

  Fourteen

  Feeling Someone’s Pain, Understanding Someone’s Pain, Alleviating Someone’s Pain

  A person is in pain, frightened, or crushed with a malignant sadness. And another human, knowing that, is likely to experience something absolutely remarkable—an aversive state that is approximated by the word “empathy.” As we’ll see in this chapter, it is a state on a continuum with what occurs in a baby or in another species. The state takes varied forms, with varied underlying biology, reflecting its sensorimotor, emotional, and cognitive building blocks. Various logical influences sharpen or dull the state. All leading to this chapter’s two key questions: When does empathy lead us to actually do something helpful? When we do act, whose benefit is it for?

  “FOR” VERSUS “AS IF” AND OTHER DISTINCTIONS

  Empathy, sympathy, compassion, mimicry, emotional contagion, sensorimotor contagion, perspective taking, concern, pity. Let the terminology and squabbles begin over definitions of ways in which we resonate with someone else’s adversity (along with the question of whether the opposite of such resonance is gloating pleasure or indifference).

  We start with, for want of a better word, primitive versions of resonating with someone’s pain. There’s sensorimotor contagion—you see a hand poked with a needle, and the part of your sensory cortex that maps onto your hand activates, sensitizing you to the imagined sensation. Perhaps your motor cortex will also activate, causing you to compress your own hand. Or you watch a tightrope walker and involuntarily put your arms out for balance. Or someone has a coughing fit, and your throat constricts.

  Even more explicitly motoric is the act of matching movements with simple mimicry. Or there’s emotional contagion, the automatic transfer of strong emotive states—such as one baby crying because another is, or someone catching the fever of a mob plunging into a riot.

  Your resonance with someone’s plight can carry an implicit power differential. You can pity someone in pain—recalling Fiske’s categories of Thems in chapter 11, this belittling pity means you view the person as high in warmth and low in competence and agency. And we all know the everyday meaning of “sympathy” (“Look, I sympathize with your situation, but . . .”); you have the power to alleviate their distress but choose not to.

  Then there are terms reflecting how much your resonance is about emotion versus cognition. In that sense “sympathy” means you feel sorry for someone else’s pain without understanding it. In contrast, “empathy” contains the cognitive component of understanding the cause of someone’s pain, taking his perspective, walking in his shoes.

  And then there are distinctions introduced in chapter 6, describing how much you and your own feelings play into resonating with someone else’s distress. There’s the emotionally distanced sense of sympathy, of feeling for someone. There’s the rawer, vicarious state of feeling their pain as if it were happening to you. And then there is the more cognitively distanced state of perspective taking, of imagining what this must be like for her, not you. As we’ll see, an as-if state carries the danger that you experience her pain so intensely that your primary concern becomes alleviating your own distress.

  Which raises a different word—“compassion,” where your resonance with someone’s distress leads you to actually help.1

  Perhaps most important, these words are generally about inwardly motivated states—you can’t force someone to truly feel empathy, can’t induce it in them with guilt or a sense of obligation. You can generate ersatz versions of it those ways, but not the real thing. Consistent with that, some recent work shows that when you help someone out of empathy, there is a very different profile of brain activation from when you do so out of an obliged sense of reciprocity.2

  As usual, we gain insights into the nature and biology of these states by looking at their rudiments in other species, their development i
n children, and their pathological manifestations.

  EMOTIONALLY CONTAGIOUS, COMPASSIONATE ANIMALS

  Lots of animals display building blocks of empathic states (I use “empathic state” throughout the chapter when referring to the collectivity of sympathy, empathy, compassion, etc.). There’s mimicry, a cornerstone of social learning in many species—think young chimps watching Mom to learn to use tools. Ironically, humans’ strong proclivity for imitation can have a downside. In one study chimps and children observed an adult human repeatedly accessing a treat inside a puzzle box; crucially, the person added various extraneous movements. When exploring the box themselves afterward, chimps imitated only the steps needed to open it, whereas kids “overimitated,” copying the superfluous gestures as well.*3

  Social animals are also constantly buffeted with emotional contagion—shared states of arousal in a pack of dogs or male chimps going on a border patrol. These are not terribly precise states, often spilling over into other behaviors. For example, say some baboons flush out something good to eat—say, a young gazelle. The gazelle is running like hell, with these baboons in pursuit. And then the male in front seems to think something along the lines of “Well, here I am running fast and—WHAT? There’s my hated rival running right behind me! Why’s that jerk chasing me?” He spins around for a head-on collision and fight with the baboon behind him, gazelle forgotten.

  Mimicry and emotional contagion are baby steps. Do other animals feel one another’s pain? Sort of. Mice can learn a specific fear association vicariously by observing another mouse experiencing the fear conditioning. Moreover, this is a social process—learning is enhanced if the mice are related or have mated.4

  In another study a mouse would be exposed to an aggressive intruder placed in its cage.5 As shown previously, this produces persistent adverse consequences—a month later, such mice still had elevated glucocorticoid levels and were more anxious and more vulnerable to a mouse model of depression.* Importantly, the same persistent effects would be induced in a mouse merely observing another mouse experiencing that stressful intruder paradigm.

  An even more striking demonstration of “your pain is my pain” in another species came in a 2006 Science paper from Jeff Mogil of McGill University.6 A mouse would observe another mouse (separated from it by Plexiglas) in pain, and, as a result, its own pain sensitivity increased.* In another part of the study, an irritant would be injected in a mouse’s paw; mice typically lick their paw at that point, with the amount of licking indicating the amount of discomfort. Thus, X amount of the irritant would produce Z amount of licking. However, if the mouse was simultaneously observing a mouse who had been exposed to more than X amount of irritant and who thus was licking more than Z amount, the subject mouse would lick more than usual. Conversely, if the subject observed a mouse licking less (having been exposed to less than X amount of irritant), it would also lick less. Thus the amount of pain a mouse was feeling was modulated by the amount of pain a nearby mouse was. Importantly, this was a social phenomenon—this shared pain only occurred between mice that were cagemates.*

  Obviously we can’t know the internal state of these animals. Were they feeling bad for the other mouse in pain, feeling “for” or “as if,” taking the other mouse’s perspective? Pretty unlikely, making the use of the word “empathy” in this literature controversial.7

  However, we can observe overt behavior. Do other species proactively lessen the distress of another individual? Yes.

  As we will see in the final chapter, numerous species show “reconciliative” behavior, where two individuals, soon after a negative interaction, show higher-than-chance levels of affiliative behaviors (grooming, sitting in contact) between them, and this decreases the odds of subsequent tensions between them. As shown by de Waal and colleagues, chimps also show third-party “consolation” behavior. This is not when, after two individuals fight, some bleeding-heart chimp indiscriminately nices both of them. Instead the consoler is preferentially affiliative to the victim over the initiator of the fight. This reflects both a cognitive component of tracking who started a tension and an affective desire to comfort. Similar consolation, focused on fight victims, also occurs in wolves, dogs, elephants, and corvids (who preen the feathers of victims). Ditto for bonobos—with some bonoboesque sex thrown in for victims along with all that platonic grooming. In contrast, such consolation doesn’t occur in monkeys.8

  Consolation is also shown among those heartwarming pair-bonding prairie voles, as shown in a 2016 Nature paper from Larry Young of Emory University, a pioneer of the vole/monogamy/vasopressin story, along with de Waal.9 Members of a vole pair would be placed in separate rooms. One of the pair would be either stressed (with a mild shock) or left undisturbed; pairs were then reunited. As compared with unstressed individuals, stressed ones would be licked and groomed more by their partner. Partners would also match the anxiety behaviors and glucocorticoid levels of their stressed pairmate. This didn’t occur for a stressed stranger, nor among polygamous meadow voles. As we’ll see, the neurobiology of this effect is all about oxytocin and the anterior cingulate cortex.

  Animals will intervene even more proactively. In one study rats worked more (pressing a lever) to lower a distressed rat, dangling in the air in a harness, than a suspended block. In another study rats proactively worked to release a cagemate from a stressful restrainer. Subjects were as motivated to do this as to get chocolate (nirvana for rats). Moreover, when a rat could both release the cagemate and get chocolate, they’d share it more than half the time.10

  This prosociality had an Us/Them component. The authors subsequently showed that rats would work to release even a strange rat—so long as it was of the same strain and thus nearly genetically identical.11 Is this automatic Us/Them-ing built on the genetics of shared pheromone signatures (back to chapter 10)? No—if a rat is housed with a cagemate of another strain, it will help individuals of that other strain. And if a rat is switched at birth and raised by a female of another strain, it helps members of its adopted but not its biological strain. “Us” is malleable by experience, even among rodents.

  Why do all these animals labor away consoling another individual in distress, or even helping them? It’s probably not conscious application of the Golden Rule, and it’s not necessarily for the social benefits—rats were just as likely to release cagemates from restrainers even if they didn’t get to interact afterward. Maybe it’s something resembling compassion. On the other hand, maybe it’s just self-interest—“That dangling rat’s incessant alarm calls are getting on my nerves. I’m going to work to lower him so he’ll shut up.” Scratch an altruistic rat and a hypocrite bleeds.

  EMOTIONALLY CONTAGIOUS, COMPASSIONATE CHILDREN

  A recap of material covered in chapters 6 and 7:

  As we saw, a developmental landmark is attaining Theory of Mind, something necessary but not sufficient for empathy, which paves the way for increasing abstraction. The capacity for simple sensorimotor contagion matures into empathic states for someone’s physical pain and, later, for someone’s emotional pain. There’s the progression from feeling sorry for an individual (e.g., someone homeless) to feeling sorry for a category (e.g., “homeless people”). There is increasing cognitive sophistication, as kids first distinguish between harming an object and harming a person. Likewise for distinguishing between intentional and unintentional harm, along with a capacity for moral indignation that is more readily evoked by the former. Along with this comes a capacity to express empathy and a sense of responsibility to act upon it, to be proactively compassionate. Perspective taking matures as well, as the child transitions from solely being capable of feeling “for” to also feeling “as if.”

  As we saw, the neurobiology of this developmental arc makes sense. At the age where an empathic state is evoked only by someone’s physical pain, brain activation centers on the periaqueductal gray (PAG), a fairly low-level way station in the brain’s pain circuitry. On
ce emotional pain can evoke an empathic state, the profile is mostly about coupled activation between the (emotional) vmPFC and limbic structures. As the capacity for moral indignation matures, coupling among the vmPFC, the insula, and amygdala emerges. And as perspective taking comes into play, the vmPFC is increasingly coupled to regions associated with Theory of Mind (like the temporoparietal junction).

  This was our picture of empathic states in kids being built upon the cognitive foundation of Theory of Mind and perspective taking. But as we also saw, there are empathic states earlier on—infants showing emotional contagion, a toddler trying to comfort a crying adult by offering her stuffie, long before textbook Theory of Mind occurs. And just as with empathic states in other animals, one must ask whether compassion in kids is mostly about ending the sufferer’s distress or ending their own.

  AFFECT AND/OR COGNITION?

  This again. We can predict the major punch lines, thanks to the previous three chapters: both cognitive and affective components contribute to healthy empathic states; it’s silly to debate which is more important; what’s interesting is seeing when one predominates over the other. Even more interesting is to look at the neurobiology of how those components interact.

  The Affective Side of Things

  When it comes to empathy, all neurobiological roads pass through the anterior cingulate cortex (ACC). As introduced in chapter 2, this frontal cortical structure has starred in empathy neuroscience ever since people felt someone else’s pain while inside a brain scanner.12