Social Intelligence: The New Science of Human Relationships

by Daniel Goleman

  And so the “social brain”—those extensive neural modules that orchestrate our activities as we relate to other people—consists of circuitry that extends far and wide. There is no single site controlling social interaction anywhere within the brain. Rather, the social brain is a set of distinct but fluid and wide-ranging neural networks that synchronize around relating to others. It operates at the systems level, where far-flung neural networks are coordinated to serve a unifying purpose.

  As yet neuroscience has no generally-agreed-upon specific map for the social brain, though converging studies are starting to zero in on areas most often active during social interactions. An early proposal identified structures in the prefrontal area, particularly the orbitofrontal and anterior cingulate cortices, in connection with areas in the subcortex, especially the amygdala.1 More recent studies show that that proposal remains largely on target, while adding other details.2

  Given the widely dispersed circuitry of the social brain, precisely which neural networks are involved depends to a great extent on what social activity we engage in. Thus during a simple conversation an array of sites keeps us in synch, while a different (though overlapping) system may activate while we ponder whether we like someone. Here’s a quick survey of some findings to date on what circuitry activates during which activity.

  Mirror neurons pepper the brain. Those in the prefrontal cortex or parietal areas (and likely elsewhere) handle shared representations—the mental images that spring to mind when we talk with someone about something we are both familiar with. Other mirror neurons involved in movement activate when we simply observe someone else’s actions—including the intricate dance of gestures and body shifts that are part of any conversation. Cells in the right parietal operculum that encode kinesthetic and sensory feedback go to work as we orchestrate our own movements in response to our conversational partner.

  When it comes to reading and responding to the emotional messages in another’s tone of voice, mirror neurons prime circuitry that connects the insula and premotor cortex with the limbic system, particularly the amygdala. As the conversation continues, connections from the amygdala to the brain stem control our autonomic responses, heightening our heart rate should matters heat up.

  Neurons in the fusiform area of the temporal lobe are dedicated to recognizing and reading emotions in faces as well as monitoring where a person’s gaze has drifted. Somatosensory areas kick in as we sense the other person’s state—and as we notice our own in response. And as we send our own emotional messages back, brain stem nuclei projections to our facial nerves create the appropriate frown, smile, or raised eyebrows.

  While we attune to the other person, the brain undergoes two varieties of empathy: a fast low-road flow via connections between the sensory cortices, thalamus and amygdala, and on to our response; and a slower high-road flow that runs from the thalamus up to the neocortex and then down to the amygdala and on to our more thoughtful response. Emotional contagion runs through that first pathway, allowing our automatic neural mimicking of the feelings of the other person. But that second pathway, which loops up to the thinking brain, offers a more considered empathy, one that holds the possibility of shutting down our attunement if we choose to.

  Here the connection from the limbic circuitry to the OFC and ACC comes into play. These areas are active in perceiving another person’s emotion and in fine-tuning our own emotional reaction. The prefrontal cortex in general has the task of modulating our emotions in ways that are appropriate and effective; if what the other person says troubles us, the prefrontal area allows us to continue the conversation and remain focused despite our own upset.

  If we have to think over what to make of the other person’s emotional message, the dorsolateral and ventromedial prefrontal regions help us ponder what it all means and consider our alternatives. What response, for instance, will work both in the immediate situation and yet be in keeping with our long-term goals?

  Beneath all this interpersonal dance, the cerebellum down at the base of the brain has been keeping our attention well targeted so that we can monitor the other person, picking up even subtle cues of fleeting facial expressions. Nonverbal, unconscious synchrony—say, the intricate choreography of a conversation—requires us to pick up an ongoing cascade of social cues. And that in turn depends on ancient structures in the brain stem, particularly the cerebellum and the basal ganglia. Their role in smooth interactions gives these lower-brain areas an ancillary role in the circuitry of the social brain.3

  All these areas join in the orchestration of social interactions (even imagined ones), and damage to any of them impairs our ability to attune. The more complicated a social interaction, the more complex the interconnected networks of neurons activated. In short, numerous circuits and sites play their role in the social brain—a neural territory that we have barely begun to map in detail.

  One way to begin to identify the core circuitry of the social brain might be to outline the minimal neural networks that are engaged during a given social act.4 For instance, for the bare act of perceiving and imitating the emotions of another person, neuroscientists at UCLA have proposed the following sets of interlocking neural circuitry. The superior temporal cortex allows an initial visual perception of the other person, sending that description to neurons in those parietal areas that can match an observed act with the execution of that act. Then the matching neurons add more sensory and somatic information to the description. This more complex set of data goes to the inferior frontal cortex, which then encodes the goal of the action to be imitated. And then the sensory copies of the actions are sent back to the superior temporal cortex, which monitors the ensuing action.

  When it comes to empathy, “hot” affective circuitry must tie in to these “cold” sensory and motor circuits—that is, the emotionally dry sensorimotor system must communicate with the affective center in the limbic system. The UCLA team proposes that the most likely candidate for this connector anatomically seems to be a region of the insula, which ties together limbic areas with parts of the frontal cortex.5

  Scientists at the National Institute of Mental Health (NIMH) argue that in seeking to map the social brain, we are not talking about a single, unitary neural system but rather interlocking circuits that can work together for some tasks, and on their own for others.6 For instance, for primal empathy—the direct person-to-person contagion of a feeling—neuroscientists nominate pathways connecting the sensory cortices with the thalamus and the amygdala, and from there to whatever circuits the appropriate response requires. But for cognitive empathy, as we sense the other person’s thoughts, the circuits run from thalamus to cortex to amygdala, and then to the circuitry for the response.

  Then when it comes to empathizing with specific emotions, the NIMH researchers suggest that further distinctions are possible. Some fMRI data suggest, for instance, that there are different pathways for reading another person’s fear versus anger. Fearful expressions seem to light up the amygdala but rarely the orbitofrontal cortex, while angry ones activate the OFC and not the amygdala. That difference may relate to the differing function of each emotion: with fear, our attention goes to what has caused the fear, while with anger we focus on what to do to reverse that person’s reaction. And when it comes to disgust, the amygdala stays out of the picture; the action instead involves structures in the basal ganglia and anterior insula.7 All of these emotion-specific circuits activate both when we ourselves experience the given emotion, and when we witness someone else feeling it.

  The NIMH scientists propose still other circuitry for one variety of cognitive empathy, not just getting an idea of what the other person’s mindset might be but also deciding on what we should do in return. Here the key circuits seem to involve the medial frontal cortex, the superior temporal sulcus, and the temporal lobe.

  The link between empathy and our sense of right and wrong has support at the neural level. Studies from patients who have had brain lesions that led them to abandon thei
r previous moral standards, or to be confused when facing a question of right or wrong, suggest that these ethical acts require that the brain areas for evoking and interpreting visceral states be intact.8 Those brain areas active during moral judgments—a string of circuitry running from parts of the brain stem (particularly the cerebellum) up to areas of the cortex—include the amygdala, thalamus, insula, and upper brain stem. All these areas are involved, too, in perceiving someone else’s feelings, as well as our own. An interconnected circuit running between the frontal lobe and the anterior temporal lobe (including the amygdala and the insular cortex) has been proposed as crucial for empathy.

  Brain function can be mapped by studying what other abilities are hampered in patients with other neural lesions. For instance, neurological patients with damage to various emotional circuits in the social brain were compared with patients whose lesions were in other areas of the brain.9 While both groups were equally capable when it came to cognitive tasks, like answering questions on an IQ test, only the patients with compromised emotional areas had poor functioning in their relationships: they made bad interpersonal decisions, misjudged how other people felt, and were incapable of coping with life’s social demands.

  The patients with these social deficits all had lesions at points within a neural array called the “somatic marker” system by University of Southern California neurologist Antonio Damasio, in whose laboratory the study of the impaired patients was done. Linking areas in the ventromedial prefrontal, parietal, and cingulate areas, as well as the right amygdala and insula, somatic markers operate whenever we make a decision, particularly in our personal and social life.10 The social abilities fostered by this key part of the social brain are essential for smooth relationships. For example, neurological patients with lesions in the somatic marker circuitry are poor at reading or sending emotional signals and so can readily make disastrous decisions in their relationships.

  Damasio’s somatic markers strongly overlap with the neural systems cited by Stephanie Preston and Frans de Waal in their perception-action model. Both models propose that when we perceive an emotion in someone else, mirror neurons activate the same neural pathways for that feeling in us, as well as circuitry for the related mental images and actions (or impulse to action). Separate fMRI studies suggest that the insula links the mirroring systems to the limbic area, generating the emotional component of the neural loop.11

  The specifics of an interaction will, of course, determine which brain areas operate as we respond, as fMRI studies of differing social moments are revealing. For example, brain imaging while volunteers listened to stories about socially embarrassing situations (one told of someone spitting food into a plate at a formal dinner) revealed greater activity in the medial prefrontal cortex and the temporal areas (both activate when we empathize with the mental state of someone else) as well as in the lateral OFC and the medial prefrontal cortex.12 These same areas become active when the story made the food-spitting involuntary (the person was choking). This neural network appears to handle the more general case of deciding whether a particular action will be socially appropriate, one of the endless small decisions we continually face in interpersonal life.

  Clinical studies of neurological patients who fail to make that decision well—and so routinely commit faux pas or otherwise faulty interpersonal activities—show damage to the ventromedial region of the prefrontal cortex. Antoine Bechara, an associate of Damasio, observes that this region plays a crucial role in integrating brain systems for memory, emotion, and feeling; damage here compromises social decision-making. In the study of embarrassing moments, the most active systems suggested an alternative network in a dorsal region of the nearby medial prefrontal cortex—an area that includes the anterior cingulate.13 This region, Damasio has found, forms a bottleneck interconnecting networks that handle motor planning, movement, emotion, attention, and working memory.

  For the neuroscientist, these are all tantalizing clues, and far more needs to be known to untangle the web of the neurology of social life.

  APPENDIX C

  Rethinking Social Intelligence

  The social brain became most highly developed in those species of mammals that live in groups, evolving as a mechanism for survival.1 The brain systems that mark humans as different from other mammals grew in direct proportion to the size of the primal human bond.2 Some scientists speculate social prowess—not cognitive superiority or physical advantage—may be what allowed Homo sapiens to eclipse other humanoids.3

  Evolutionary psychologists argue that the social brain—and hence social intelligence—evolved to meet the challenge of navigating the social currents in a primate group: it equips one to determine who is the alpha male, who one can count on for defense, whom one must please and how (grooming is the usual answer here). In humans, our need to engage in social reasoning—particularly coordination and cooperation as well as competition—drove the evolution of our larger brain size and of intelligence generally.4

  The major functions of the social brain—interaction synchrony, the types of empathy, social cognition, interaction skills, and concern for others—all suggest strands of social intelligence. The evolutionary perspective challenges us to think afresh about the place of social intelligence in the taxonomy of human abilities—and recognize that “intelligence” can include noncognitive abilities. (Howard Gardner notably made this case in his groundbreaking work on multiple intelligences.)

  The new neuroscientific findings on social life have the potential to reinvigorate the social and behavioral sciences. The basic assumptions of economics, for example, have been challenged by the emerging “neuro-economics,” which studies the brain during decision-making.5 Its findings have shaken standard thinking in economics, particularly the notion that people make rational decisions about money that can be modeled by decision-tree–type analyses. Low-road systems, economists now realize, are far more powerful in such decision-making than the purely rational models can predict. Likewise, the field of intelligence theory and testing seems ripe for a rethinking of its basic assumptions.

  In recent years social intelligence has been a scientific backwater, largely ignored by social psychologists and students of intelligence alike. One exception has been the boomlet in research on emotional intelligence sparked by the seminal work of John Mayer and Peter Salovey in 1990.6

  As Mayer pointed out to me, Thorndike’s original view saw a triad of mechanical, abstract, and social intelligence, but he subsequently failed to find a way to measure the social. In the 1990s, as the localization of emotions in the brain became better understood, Mayer noted, “Emotional intelligence could be groomed as the replacement member of the triumvirate where social intelligence failed.”

  The more recent emergence of social neuroscience means the time is ripe for a revival of social intelligence on a par with its sister, emotional intelligence. A rethinking of social intelligence should more fully reflect the operation of the social brain, so adding often-ignored capacities that nonetheless matter immensely for our relationships.

  The model of social intelligence I offer in this book is merely suggestive, not definitive, of what that expanded concept might look like. Others may reshuffle its aspects differently or suggest their own; mine is but one of many ways to categorize. More robust and valid models of social intelligence will emerge gradually from cumulative research. My goal is simply to catalyze such fresh thinking.

  Some psychologists may complain that the defining capacities of social intelligence I propose add to standard definitions of “intelligence” aptitudes from noncognitive domains. But that is precisely my point: when it comes to intelligence in social life, the brain itself mixes capacities. Noncognitive abilities like primal empathy, synchrony, and concern are immensely adaptive aspects of the human social repertoire for survival. And these capacities certainly allow us better to follow Thorndike’s mandate to “act wisely” in our relationships.

  The old concept of social intelligence
as purely cognitive assumes, as many early intelligence theorists claimed, that social intelligence may be no different from general intelligence itself. Some cognitive scientists would no doubt argue that the two abilities are identical. After all, their discipline models mental life on the computer, and modules for processing information run along purely rational lines, following computational logic.

  But an exclusive focus on mental abilities in social intelligence ignores the invaluable roles of both affect and the low road. I suggest a perspective shift, one that looks beyond mere knowing about social life to include the automatic abilities that matter so much as we engage, both high road and low. The various theories of social intelligence currently in vogue detail these intertwined capacities only spottily and to quirkily varying degrees.

  Intelligence theorists’ views on the social aptitudes for life can be better understood in light of their field’s history. In 1920, when Edward Thorndike first proposed the concept of social intelligence, the newfangled concept of “IQ” was still shaping the thinking of an equally novel field, psychometrics, that aimed to find ways to measure human abilities. In those heady days psychology’s recent successes in sorting out the millions of U.S. soldiers by IQ during the First World War, and so assigning them to tasks and posts they could handle effectively, aroused understandable excitement.