No Two Alike

by Judith Rich Harris

  In the Paleolithic, your life or the lives of your children might have hinged on the answer to such a question.

  For hundreds of years, the writers of plays, operas, and novels have toyed with the idea of identity, particularly mistaken identity. A woman’s husband disguises himself and pretends to be a stranger. She allows the stranger to seduce her. Has she betrayed her husband? Another scenario, this one from science fiction: a machine makes a replica of the husband that is accurate down to the last little molecule—it has his exact brain with all his memories and so on. If the woman has sex with the replica, has she betrayed her husband? The answer to the first question is no but the woman is nonetheless guilty; to the second it is yes but the woman cannot be blamed. There are two separate issues here: the identity of the seducer and what the woman thinks is the identity of the seducer. Put a mustache on a man or cut it off, dress him in mufti or in motley, have him speak with a British or Italian accent, and he will still be the same man. But a replica, however accurate and convincing, is a different man. We don’t want Shawn sleeping with Jonathan’s wife any more than we want him convicted of Jonathan’s crimes.

  But Jonathan’s wife, if he has one, wouldn’t want to sleep with Shawn. According to a twin study carried out by the behavioral geneticists David Lykken and Auke Tellegen, people who are married to one of a pair of identical twins feel, on average, only so-so about the other twin; only 13 percent of the men and 7 percent of the women think they could have fallen for their spouse’s twin. The feeling, or lack thereof, is mutual: twins are generally not attracted to their co-twin’s wife or husband.13

  This, as Lykken and Tellegen pointed out, requires some explaining. Evolutionary psychologists sometimes give the impression that mate selection is pretty much cut and dried. First, eliminate Mom and Dad and other close relatives. Second, look for someone with youth and beauty (if you’re a male) or high status (if you’re a female) and rank your possible choices in order of desirability. Third, decide how high you can go on the desirability index—a judgment that entails some understanding of your own desirability. Shooting too high is a waste of energy and could prove embarrassing, so cross out the top three names on your list. Now flirt with the fourth, and if he or she is unreceptive, try the fifth.

  Hold on, I’ve forgotten about similarity. At least in modern times, people tend to choose mates who are reasonably similar to themselves in characteristics such as physical attractiveness, height, education, intelligence, attitudes toward religion and politics, and (to a lesser extent) personality.14 But if mate selection were simply a question of finding the person of the desired sex who is most like you, then the mates of identical twins would be more alike than the mates of fraternal twins, and they are not, Lykken and Tellegen discovered. The researchers concluded that the following model of mate selection would fit their data: eliminate half of the eligible population because they are too different from you or you have no hope of winning them, randomly select one from the remaining half, and fall madly in love with him or her. “The heart has its reasons that the reason knows not of,” sighed Pascal.15

  The point is that you don’t fall madly in love with a class or group of individuals, or with everyone who meets your basic criteria. You fall in love (at least for a time) with a particular individual: your “one and only.” It’s called pair-bonding. Humans are not the only species that practice it, of course. Monogamy—or monogamy that lasts but a season or that admits a little playing around on the side—is fairly common in the animal kingdom. Its evolutionary purpose is clear: to motivate the male to help rear the young produced by the female. It’s worth his trouble only if he has a reasonably good chance of being the biological father of her offspring.

  The members of monogamous species, or quasi-monogamous species like ours, have to be able to identify their mates and not make errors of the “Oops, I thought you were my husband” sort. Pair-bonding is possible only if the members of the pair can reliably recognize each other. Birds that pair-bond (many birds do) have to be able to tell their male robin from all the other male robins, or their female goose from all the other female geese.

  The same problem faces the members of species that rely on reciprocal altruism to get them through hard times—vampire bats, for example. These bats live as long as eighteen years, and the ones that share a nesting site (not necessarily close relatives) get to know one another. They remember which of their buddies are indebted to them and which they are indebted to. After an unsuccessful night of hunting, they’ll go for a handout to the bat that owes them a favor.16

  The methods used to distinguish individuals vary from one species to another. Dogs are famous for their ability to tell humans apart on the basis of smell. But in case you wondered, even a dog probably couldn’t have solved the problem of Shawn versus Jonathan. Identical twins, if they live in the same house and eat the same diet, smell alike to a dog.17

  Although there is some evidence that human babies and their mothers can identify each other by smell,18 and no question that we recognize familiar voices, humans rely principally on vision to distinguish one individual from another. Hence the importance of the face-recognition module. The neuroscientist Martha Farah and her colleagues have explored some of the ways that face recognition differs from ordinary vision—the kind of vision we use to identify an object. For example, recognizing an object is somewhat harder when it is upside down than when it is right side up, but recognizing a face is dramatically more difficult if it is upside down.19 You can try it yourself: just turn the newspaper upside down and look at the photos.

  Mechanisms of the mind do not necessarily correspond to specific areas in the brain; some are localized, others are spread around. The face-recognition module is localized. Brain-imaging studies have pinpointed a place in the brain that responds selectively to faces; it’s called the fusiform face area and is located near the junction of the occipital and temporal lobes. Damage to this area, particularly on the right side of the brain, can produce the neurological disorder called prosopagnosia. People with this disorder don’t recognize familiar faces—even their own face, seen in a mirror, looks like a stranger—and they can’t learn to identify new ones.20

  People with autism, too, are notoriously poor at recognizing faces. Brain-imaging evidence tells us why: their fusiform face area isn’t working properly. Consequently they have to process faces as though they were objects. Because autistic people find it no more difficult to identify an upside-down face than an upside-down object, they are actually better than neurologically normal subjects at identifying faces shown upside down, though not nearly as good as the normal subjects when the faces are shown in their usual orientation.21

  Another sign of malfunction of the face-recognition module is a failure to discriminate between familiar and unfamiliar faces, and this too is found in people with autism. Brain-imaging research has shown that a neurologically normal brain lights up in one way to a familiar face, in a different way to the face of a stranger. This difference in neural response is absent in four-year-old autistic children, but is present in normally developing babies as young as six months of age.22 By six months, neurologically normal babies are also beginning to be wary of strangers, so this makes sense: their brain is discriminating between known and unknown. But oddly enough, the distinction is not a simply a matter of identifying a face as someone you know or failing to do so: there’s apparently a brain device that signals (or fails to signal) familiarity and that operates independently, to some extent, of the one responsible for identification.

  Malfunction of the familiarity device can have strange effects. A brain-injured patient may say that, yes, this person looks exactly like my wife, but she’s not my wife—she’s an impostor. The identification mechanism is working okay but the familiarity signal is missing. In other cases, the familiarity signal lights up when it shouldn’t and patients become convinced that a stranger is a familiar person—that the patient in the next bed is really their mother masquerading
as a stranger or inhabiting the stranger’s body. In case you still enjoy learning new words, the first disorder is called the Capgras syndrome, the second, the Frégoli delusion.23

  These are rare disorders. The familiarity signal usually works flawlessly, as parents discover when they put their young children on the lap of the department store Santa Claus. The children are not deceived by the name or the outfit; they know full well that this is a stranger.24

  Even in the Paleolithic, face recognition alone wouldn’t have been enough, because humans also need a way of identifying individuals to others. If you want to tell A what B is doing with C—and you do, you do!—it’s awkward to have to identify B as “the guy with a big nose and a scar on his cheek.” If you want to keep giving A updates on B’s adventures, the description can be shortened to Bignose or Scarface, but what if B is just an ordinary-looking guy? If the exchange of social information is one of the purposes for which humans developed language, then people’s names—some quick and reliable way of verbally identifying them—must have been among the very first items in the mental word lexicon.25

  Perhaps they were also the first to be forgotten. You don’t have to have anomia to experience the feeling of not being able to think of someone’s name when it’s right on the tip of your tongue: it happens to all of us, more frequently as we get older. Alzheimer’s patients tend to forget proper nouns—names of people and places—before they forget common nouns. Injuries to the brain can sometimes produce the same effect: the loss of people and place names, while common nouns are spared. But less extensive injuries can sometimes lead to difficulties retrieving people’s names without difficulties in retrieving names of places,26 which suggests to me that the storage of people’s names may be handled by a specialized mechanism—a name-storage device that works in a similar fashion to the face-recognition module. Perhaps the reason why names seem to be easier to forget is that the name-storage device has a much shorter evolutionary history. Natural selection hasn’t had time to work out the bugs in the program.

  On the other hand, it isn’t only names we forget. Other kinds of failures in identifying people also occur. We see someone who looks familiar but we can’t remember where we saw them before. Or we see someone at a distance and think it’s someone we know, but when they get closer we discover we were mistaken. To account for such experiences, Andrew Young and his colleagues at Lancaster University proposed a model of memory storage for people information. They hypothesized that our minds contain a “person identity node” for each individual we know. Recognizing someone depends on the action of “recognition units” that indicate how closely the person’s physical characteristics—face, hair, body build, voice, maybe even clothing—resemble those of a known individual. If the match is good, the person identity node for that individual is activated, which then gives us access to other information about him or her: name, occupation, and so on. But the links to the other information can fail to work. The person identity node, then, is a sort of junction where various kinds of information about a given individual are joined together, some of the joins being weaker (or more easily blocked) than others.27

  Young’s model of a device for storing people-information is based on nodes and links. I’ve used, and will continue to use, the more vivid image of a lexicon, or Rolodex, of people-information. But you shouldn’t take this image literally: it’s meant as a metaphor, not as a model. The node model is no doubt closer to the way the brain actually works; moreover, it correctly conveys the idea that there are many ways of getting to the information stored for a particular person: we can identify someone by name, by face, or by profession, and once we get to the right node we have access to other information about that person. The cards in a Rolodex, in contrast, are kept in alphabetical order, and if you don’t happen to remember the person’s name you’re out of luck.

  On the other hand, the lexicon image does a better job of conveying the idea of a vast collection of entries. We have thousands and thousands of mental storage sites for people-information. Each is associated with a particular individual; each contains (or is linked to) other information we have about that individual. Picture a mental lexicon with a page for each individual you know, with slots for the face, name, and whether he or she is a close relative; plus other information such as occupation; plus memories of the experiences you’ve had with him or her. There may also be an emotional marker, indicating how you feel about this person. The contents of some slots may be hard to read; other slots may never have been filled in. A page can be set up in the lexicon even if you’ve never set eyes (or ears) on the individual it refers to. You collect and store information on characters you read about in novels or hear about from other people. Folks you’ve never met may have a page in their lexicon for you! And they don’t hesitate to tell you so, when you’re finally introduced. “I’ve heard a lot about you,” they say.

  One more thing about the lexicon: each page also has slots for the person’s group memberships and social categories. The first things we notice about people are their gender, age group, and race,28 and these are duly listed on their page in the lexicon. The category information is analogous to the information about parts of speech in Pinker’s word lexicon. Sometimes we treat words or people as unique individuals, as when we retrieve an irregular past tense; sometimes we treat them as members of a category, as when we add -ed to a verb. This chapter is about relationships; in relationships, people are treated as unique individuals. We will treat people as members of categories in the next chapter.

  A word, according to Pinker, is an arbitrary sign that “works because a speaker and a listener can call upon identical entries in their mental dictionaries.”29 A name is an arbitrary word that refers to a particular individual: the speaker and the listener can use it to call upon corresponding entries in their mental lexicons of people. The virtue of a word is that its arbitrary association with a meaning permits us to exchange information with our conversational partners. The virtue of a name is that its arbitrary association with a person permits us to transmit information about that particular person.

  Once we can do that, gossip becomes a commodity—a medium of exchange. We can do a favor for A by conveying information about what B and C are doing. Watch out for D; he’s stronger than he looks. Don’t trust E; she tells lies. F has her eye on G. If J gets mad at you, you’ll have K to contend with too. One of the reasons why gossip is a pejorative term is that much of the information passed along this way is unfavorable, which is why it is always a little unsettling to be told, “I’ve heard a lot about you.”

  The evolutionary purpose of the people lexicon is as clear as the evolutionary purpose of pair-bonding: to enable us to behave appropriately toward different individuals, depending on what we have learned about them. To enable us to tailor our behavior to the nature of the relationship we have with each. The baby lifts up its arms to its mother but not to the stranger, even if the stranger is the right age and sex. The child learns to avoid the bully but to seek out other kids in the neighborhood. People stop doing favors for people who never pay them back, unless they are close relatives.

  “There is not just one ‘social intelligence,’” explained the evolutionary developmental psychologists David Bjorklund and Anthony Pellegrini, “but a set of hierarchically arranged, relatively specific abilities that evolved to deal with the variety of social problems faced by our ancestors.” Another evolutionary developmental psychologist, Daphne Bugental, has divided the problems of social life into several “domains” and proposed that humans are provided with specialized “regulatory mechanisms” to guide behavior in each domain.30 These mechanisms may make use of the same kind of information and yet follow different rules. For example, what do we do with the information that someone is a close relative? The information is used in one way if it’s a question of deciding whether or not to help them, in quite a different way if it’s a question of whether or not to have sex with them.

  So the relationship s
ystem contains many intricately connected parts. There’s a people-information acquisition device that constructs and stores a lexicon of people and provides the motivation to collect the information. There are regulatory mechanisms that make use of the information stored in the lexicon to guide behavior in different domains of social life and that provide their own motivations, the sex drive being an obvious example. Other specialized modules deliver input to the relationship system; they include the face-recognition module, a device that assesses kinship, and the mindreading mechanism I described in the previous chapter. Whatever you are considering doing with another person—help them, mate with them, engage in trade with them, pick a fight with them—it is extremely useful to have an idea of what their intentions are and what they are thinking about you. Information provided by the mindreading mechanism is specific to a particular person (we understand that Anne may know where the ball is but that Sally may be ignorant of its whereabouts) and is stored on the appropriate page of the people lexicon. Or, if you prefer, linked to the appropriate person identity node.

  The purpose of the relationship system is to guide the way we behave toward specific individuals. To show how it works I’ll use as my example one particular kind of relationship: what Bugental calls the “hierarchical power domain.”31 These are relationships in which one individual dominates the other. What particularly interests me about dominance relationships is that they are an example of a process in which pairwise interactions between individuals give rise to a larger social structure—in this case, the dominance hierarchy or “pecking order.”