Behave: The Biology of Humans at Our Best and Worst

by Robert M. Sapolsky

  Human infants also recognize status differences, as shown in a truly clever study. Show an infant a computer screen displaying a big square and little square; each has eyes and a mouth.10 The two squares are at opposite ends of the screen and repeatedly move to the other side, passing each other in the process. Then show a version where the two bump into each other—conflict. The squares bump repeatedly until one of them “gives in” by lying down, letting the other one pass. Toddlers look at the interaction longer when it’s the big square that gives in, rather than the little one. The first scenario is more interesting because it violates expectations—“Hey, I thought big squares dominated little squares.” Just like monkeys and corvids.

  But wait, this may just reflect folk physics, not attunement to hierarchy—big things knock over little things, not the other way around. This confound was eliminated. First, the adversarial squares were not touching when one gave in. Second, the subordinating one would fall in the opposite direction from that predicted by physics—rather than being knocked backward, it prostrates itself before the alpha square.

  Along with this expertise comes intense interest in hierarchy—as emphasized in chapter 9, gossip is mostly about the status of status: Are there any fallen mighty? Have the meek inherited anything lately? Regardless of which square wins, infants look longer at the conflict situation than when the squares peacefully glide past each other.

  This is logical self-interest. Knowing the hierarchical lay of the land helps you navigate it better. But there’s more to that than just self-interest. Those monkeys and corvids not only pay attention when there are rank reversals in their group; they do the same when eavesdropping on the neighbors. Same with us.11

  What’s happening in our brains when we contemplate rank?12 Naturally, the prefrontal cortex weighs in. Frontal damage impairs skill at recognizing dominance relations (along with recognizing kinship, deception, or intimacy in faces). The vlPFC and dlPFC activate and become coupled when we figure out dominance relations or look at a dominant face, reflecting the combined affective and cognitive components to the process. These responses are most pronounced when considering someone of the opposite sex (which may reflect mating goals more than mere academic interest about hierarchy).

  Seeing a dominant face also activates the superior temporal gyrus (STG, with its role in Theory of Mind) and increases its coupling to the PFC—we’re more interested in what dominant individuals are thinking.13 Moreover, individual “social status” neurons occur in the monkeys. And as noted in chapter 2, contemplating an unstable hierarchy does all of the above, plus activates the amygdala, reflecting the unsettling effects of instability. Of course, though, none of this tells us what we are contemplating at these times.

  Your Brain and Your Own Status

  Your own rank does logical things to your brain. In macaque monkeys an increase in rank increases mesolimbic dopamine signaling. And back to that rhesus monkey study showing that being in a larger social group causes expansion and functional coupling of the STG and PFC. The study also showed that the higher the rank attained within each group, the greater the expansion and coupling. Consistent with that, a study of mice showed that higher-ranking animals had stronger excitatory inputs into the mouse equivalent of the (cognitive) dlPFC.14

  I love these findings. As I said, in lots of social species, attaining high rank is about sharp teeth and good fighting skills. But maintaining the high rank is about social intelligence and impulse control: knowing which provocations to ignore and which coalitions to form, understanding other individuals’ actions.

  Does the monkey make history, or does history make the monkey? Once groups were formed, did individuals who became dominant respond with the biggest expansions of those brain regions? Or, prior to group formation, were the individuals destined to become dominant already endowed in those regions?

  Unfortunately, animals weren’t imaged before and after group formation in the study. However, subsequent work showed that the larger the size of the group, the larger the association between dominance and those brain changes, suggesting that attaining high rank drives the enlargement.* In contrast, the mouse study showed that when synaptic excitability was increased or decreased in the dlPFC, rank rose or declined, respectively, suggesting that enlargement drives attainment of high rank. The brain can shape behavior can shape the brain can shape . . .15

  Your Body and Your Own Status

  What about biological differences outside the brain as a function of rank? For example, do high- and low-ranking males differ in their testosterone profiles and, if there are differences, are they causes, consequences, or mere correlates of the rank differences?

  Folk endocrinology has always held that high rank (in any species) and elevated testosterone levels go hand in hand, with the latter powering the former. But as covered at length in chapter 4, neither is the case in primates. As a reminder:

  In stable hierarchies high-ranking males typically don’t have the highest testosterone concentrations. Instead it’s usually low-ranking adolescent males, starting fights they can’t finish. When there is an association between high rank and high testosterone, it generally reflects the higher rates of sexual behavior among dominant individuals driving secretion.

  An exception to the above is during unstable times. For example, among a number of primate species, high-ranking males have the highest testosterone levels for the first months but not years after group formation. During unstable times, the high-testosterone/high-rank relationship is more a consequence of the high rates of fighting among the high-ranking cohort than of rank itself.16

  Reiterating the “challenge hypothesis,” the elevation in testosterone levels caused by fighting is not so much about aggression as about challenge. If status is maintained aggressively, testosterone fosters aggression; if status were maintained by writing beautiful, delicate haikus, testosterone would foster that.

  Next we consider the relationship between rank and stress. Are different ranks associated with different levels of stress hormones, styles of coping, and incidences of stress-related disease? Is it more stressful to be dominant or subordinate?

  An extensive literature shows that a sense of control and predictability reduces stress. Yet monkey research conducted by Joseph Brady in 1958 produced a different view. Half the animals could press a bar to delay shocks (“executive” monkeys); the passive other half received a shock whenever the executive did. And the executive monkeys, with their control and predictability, were more likely to ulcerate. This birthed the “executive stress syndrome”—those on top are burdened with the stressors of control, leadership, and responsibility.17

  Executive stress syndrome became a meme. But a huge problem was that monkeys were not randomly assigned to be “executives” and “nonexecutives.” Instead, those that pressed the bar soonest in pilot studies were made executives.* Such monkeys were subsequently shown to be more emotionally reactive, so Brady had inadvertently stacked the executive side with the ulcer-prone neurotics.

  So much for ulcerating executives; contemporary studies show that the worst stress-related health typically occurs in middle management, with its killer combo of high work demands but little autonomy—responsibility without control.

  By the 1970s dogma held that subordinate organisms are the most stressed and unhealthy. This was first shown with lab rodents, where subordinate animals typically had elevated resting levels of glucocorticoids. In other words, even in the absence of stress, they showed signs of chronically activating the stress response. The same is observed in primates ranging from rhesus monkeys to lemurs. Same for hamsters, guinea pigs, wolves, rabbits, and pigs. Even fish. Even sugar gliders, whatever they are. In a pair of unintentional studies of captive monkeys in which subordinate individuals were basically subordinated to death, such animals had extensive damage to the hippocampus, a brain region very sensitive to the damaging effects of glucocorticoid excess.18

>   My own work with baboons in Africa showed the same (being the first such studies of wild primates). In general, low-ranking male baboons had elevated basal glucocorticoid levels. When something stressful did occur, their glucocorticoid stress response was relatively sluggish. When the stressor was over, their levels returned to that elevated baseline more slowly. In other words, too much of the stuff in the bloodstream when you don’t need it, and too little when you do. Remarkably, at the nuts-and-bolts level of brain, pituitary, and adrenals, the elevated basal glucocorticoid levels of a subordinate occurred for the same reasons as the elevated levels in humans with major depression. For a baboon, social subordination resembles the learned helplessness of depression.

  Excessive glucocorticoids get you into trouble in various ways, helping explain why chronic stress makes you sick. Subordinate baboons paid a price in other realms as well. They had (a) elevated blood pressure and a sluggish cardiovascular response to a stressor; (b) lower levels of “good” HDL cholesterol; (c) subtle immune impairments, a higher frequency of getting sick and slower wound healing; (d) a testicular system more easily disrupted by stress than that of dominant males; and (e) lower circulating levels of a key growth factor. Try not to be a subordinate baboon.

  Chickens and eggs reappear—does a particular physiological attribute contribute to rank, or the reverse? This is impossible to determine in wild animals, but in captive primate populations the distinctive physiological features of a rank generally follow, rather than precede, the establishment of the rank.19

  At that point I’d happily proclaim that these findings reflected the nature of Hierarchy, with a capital H, and of the stressfulness of social subordination. Which turned out to be totally wrong.

  A first wrinkle was provided by Jeanne Altmann of Princeton and Susan Alberts of Duke, studying wild baboons with stable hierarchies. They found the familiar picture, namely of subordination associated with elevated basal glucocorticoid levels. However, unexpectedly, levels in alphas were elevated into the range seen in lowest-ranking males. Why is life more stressful for alpha than beta males? The two ranks had similar rates of being challenged by lower-ranking males (a source of stress) and being groomed by females (a source of coping). However, alpha males fight more often and spend more time in sexual consortships with females (which is majorly stressful, as the male has to defend his consortship from harassing males). Ironically, a chief benefit of alphadom—sexual consortships—can be a major stressor. Be careful what you wish for.20

  Modified from R. Sapolsky, “Sympathy for the CEO,” Sci 333 (2011): 293.

  Visit bit.ly/2ngWthp for a larger version of this graph.

  Okay, so except for the curse of being alpha, social subordination is generally stressful. But this is also wrong. It’s not just the rank that matters but what the rank means.

  Consider the primate species in which a relationship has been found between rank and glucocorticoid levels. Across these species, basal glucocorticoid levels are relatively elevated in subordinate animals if: (a) dominant individuals in bad moods frequently displace aggression onto subordinates; (b) subordinates lack coping outlets (such as a grooming partner); and/or (c) the social structure is such that subordinate animals have no relatives present. And when the profile was the opposite, it was dominant animals with the highest glucocorticoid levels.21

  The “meaning” of rank and its physiological correlates also vary between different groups of the same species. For example, while health of subordinate baboons fared particularly badly in a troop with high rates of dominant males displacing aggression, the health of dominant males fared badly in a troop during a period of instability centered around the top of the hierarchy.

  And superimposed on all this, personality shapes the perception of the reality of rank. Using the word “personality” about other species used to cost you tenure, but it’s now a hot topic in primatology. Individuals in other species have stable differences in temperament—how likely someone is to displace aggression when frustrated, how socially affiliative they are, how rattled they get by novelty, and so on. Primates differ as to whether they see watering holes as half empty or full; in the context of hierarchy, some individuals who are number two care only that they’re not number one, and some individuals who are number nine gain comfort from at least not being number ten.

  Not surprisingly, personality influences the rank/health relationship. For the same high rank, an individual is likely to be less healthy if he (a) is particularly reactive to novelty; (b) sees threats in benign circumstances (e.g., his rival showing up and merely taking a nap nearby); (c) doesn’t take advantage of social control (e.g., letting a rival determine the start of an obvious showdown); (d) doesn’t differentiate between good and bad news (e.g., distinguishing behaviorally between winning and losing a fight); and/or (e) doesn’t have social outlets when frustrated. You could make a living giving baboons “how to succeed in business” seminars built around these factors.22

  Meanwhile, on the flip side, for the same low rank, an individual tends to be healthier if (a) he has lots of grooming relationships; and/or (b) there’s someone even lower ranking than him to serve as a target for displaced aggression.

  Thus, in other species, how does rank affect the body? It depends on what it’s like to have a certain rank in that species and particular social group, and the personality traits that filter the perception of those variables. What about humans?

  And Us

  A smidgen of neurobiology research has examined differences in how people feel about hierarchy. Back to a concept from the last chapter, social-dominance orientation (SDO), the measure of how much people value power and prestige. In one study subjects viewed someone in emotional pain. As reviewed in chapter 2, this activates the anterior cingulate cortex and insular cortex—empathy, and disgust at the circumstance that evoked the pain. The higher someone’s SDO score, the less activation of those two regions. Those with the most interest in prestige and power seem least likely to feel for those less fortunate.23

  What about the biological correlates of a human having a particular rank? In some ways we’re more subtle than other primates; in others, far less so.

  Two studies examined high-status individuals in government or the military (in the latter case, officers up to the level of colonel). As compared with low-status controls, these folks had lower basal glucocorticoid levels, less self-reported anxiety, and an enhanced sense of control (this telling us nothing, however, as to which came first—the rank or the unstressed profile).24

  Baboons redux. But something subtler was happening. The authors deconstructed high rank with three questions: (a) How many people ranked lower than the subject in his organization? (b) How much autonomy did he have (e.g., to hire and fire)? (c) How many people did he directly supervise? And high rank came with low glucocorticoids and anxiety only insofar as the position was about the first two variables—lots of subordinates, lots of authority. In contrast, having to directly supervise lots of subordinates did not predict those good outcomes.

  This gives credence to executives’ bellyaching about how they aren’t supervising eleventy people; instead they have eleventy bosses. To accrue the full physiological benefits of high status, don’t supervise people; instead, glide through the workplace like a master of the universe while minions whom you never interact with smile obsequiously. It’s not just rank; it’s what rank means and entails.

  In what sense is the status/health relationship in humans less subtle than in other primates?25 In that it reflects the most permeating form of status any primate has invented, namely socioeconomic status (SES). Numerous studies examine the “health/SES” gradient, the fact that life expectancy and the incidence and morbidity of numerous diseases are worse in poor people.

  To summarize this sprawling topic that was reviewed in chapter 9:

  Which comes first—poverty or poor health? Overwhelmingly the former. Recall that d
eveloping in a low-SES womb makes poorer health as an adult more likely.

  It’s not that the poor have poor health and everyone else is equally healthy. For every step down the SES ladder, health is worse.

  The problem isn’t that poor people have less health-care access. The gradient occurs in countries with socialized medicine and universal health care and for diseases whose incidence is independent of health-care access.

  Only about a third of the variability is explained by the poor being exposed to more health risk factors (e.g., pollution) and fewer protective factors (e.g., health club memberships).

  The gradient seems to be about the psychological baggage of SES. (a) Subjective SES predicts health at least as accurately as objective SES, meaning that it’s not about being poor. It’s about feeling poor. (b) Independent of absolute levels of income, the more income inequality in a community—meaning the more frequently the poor have their noses rubbed in their low status—the steeper the health gradient. (c) Lots of inequality in a community makes for low social capital (trust and a sense of efficacy), and that’s the most direct cause of the poor health. Collectively these studies show that the psychological stress of low SES is what decreases health. Consistent with that, it is diseases that are most sensitive to stress (cardiovascular, gastrointestinal, and psychiatric disorders) that show the steepest SES/health gradients.

  The SES/health gradient is ubiquitous. Regardless of gender, age, or race. With or without universal health care. In societies that are ethnically homogeneous and those rife with ethnic tensions. In societies in which the central mythology is a capitalist credo of “Living well is the best revenge” and those in which it is a socialist anthem of “From each according to his ability, to each according to his need.” When humans invented material inequality, they came up with a way of subjugating the low ranking like nothing ever before seen in the primate world.