With only minimal cues, we are able to form vivid and detailed impressions of other people. In person, we see another’s face, clothing, perhaps hear him speak—and after less than a minute, will have formed a strong opinion about his personality and his role and status in society. This impression helps us decide whether to trust him, and shapes our expectations of what he will do and our understanding of how we should act toward him. These impressions may, of course, be wrong—they are subjective and stereotypical, based on our previous experiences and influenced by our culture’s mores and prejudices. But without this ability, we would be unable to make sense of society: each new person or situation we encountered would be completely novel, and we would be reduced to the social illiteracy of the tourist in a completely foreign culture, unable to understand how to act toward others or what to expect from them, unable to judge if others were acting appropriately or not.

Online, the process is the same, but the cues are different—and sparser. In many settings we do not see the other’s face. There is no voice to hear, no clothing to evaluate. We may have only a screen name and a few lines of commentary to go on. With too little data, the result is a cipher. But with just a bit more information—sometimes even just a few words or an evocative name—we again form vivid, if not necessarily accurate, assessments.

With its different identity cues, the online world offers the intriguing possibility of being a place in which we can make sense of other people in fundamentally different ways. In particular, it raises the promise of eliminating the stereotypical impressions we form based on physical characteristics. In person, physical characteristics dominate our initial impressions. Although we do make some insightful judgments, physical cues also lead us to make erroneous, though pervasive and strongly held, assumptions about others.

Online, we can create spaces where identity is firmly tethered to one’s real-world self, and where the impression you make online is as close as possible to the way you appear in person. We can create spaces where identity floats freely and fantastically, and you can claim to be whoever or whatever you please. We can create spaces where your history anchors your identity, but where many of the physical features people would see immediately in person are unknown.

In this chapter, we will take a close look at how we form impressions of others, with these questions in mind: Do we need the context of real-world identity to make sense of each other? Can we design social spaces in which people make better sense of each other than they do face to face—that ameliorate the prejudices stemming from physical trait–based impressions, while also avoiding the confusion and deceptions of easy, ungrounded identity claims?

Cyber Utopia

In the late 1980s and early 1990s, online social interaction—and identity experimentation—was becoming a mainstream experience. Colleges had begun routinely providing online access to incoming students, away from home for the first time and primed for self-reinvention. America Online and other consumer services offered easy access to disposable identities and anonymous chat-rooms. In 1993, the New Yorker published what would become its most reprinted cartoon ever, a picture of a dog sitting at a computer, telling his canine companion, “On the Internet, nobody knows you’re a dog.”

Many greeted this new, potentially post-gender, post-racial, and disembodied world with enthusiasm. “Cyber-utopian” writers predicted a future in which people judged each other on the merits of their words rather than their skin color, and where anyone could experience what it was like to be a man, or black, or fantastically beautiful—transformative experiences that might even carry over to reforming face-to-face society.1

In Life on the Screen (1995), psychologist and MIT professor Sherry Turkle wrote about the therapeutic value of online identity play:

When people adopt an online persona … some sense the possibilities for self-discovery, even self-transformation. A more fluid sense of self allows a greater capacity for acknowledging diversity. … We do not feel compelled to rank or judge the elements of our multiplicity. We do not feel compelled to exclude what does not fit. (Turkle 1995, 260)

Turkle was writing about MUDs,2 online fantasy worlds that Howard Rheingold, writing at around the same time, described as

the wild side of cyberspace culture, where magic is real and identity is a fluid. … MUDs are part of the latest phase in a long sequence of mental changes brought about by the invention and widespread use of symbolic tools. … Similar to the way previous media dissolved social boundaries related to time and space, the latest computer-mediated communications media seem to dissolve boundaries of identity as well. (Rheingold 1993, 149)

In this view, “real life” is constraining, with limits imposed by social conventions, by accidents of birth, by lack of money. Online, one is free to imagine anything. If you are curious about what it is like to be the opposite sex, you can simply create a character of that gender and experience it. If you have always wanted to be a darkly mysterious foreigner, or a blond cheerleader, or a flying elf, online you could find a community where you could identify yourself as such, and others would go along with the fantasy. And if it was not too farfetched—that is, if you claimed to be a gifted male graduate student (though in fact were a forty-two-year old mother of three) rather than, say, a talking, clairvoyant frog—there was a good chance that the others would not just humor you, but would actually believe you, since any evidence to the contrary was sparse. Turkle quoted one MUD user:

You can be whoever you want to be. You can completely redefine yourself if you want. You can be the opposite sex. You can be more talkative. You can be less talkative. Whatever. You can just be whoever you want, really, whoever you have the capacity to be. You don’t have to worry about the slots other people put you in as much. It’s easier to change the way people perceive you, because all they’ve got is what you show them. They don’t look at your body and make assumptions. They don’t hear your accent and make assumptions. All they see is your words. (Turkle 1995, 184)

It was not only the fantastical worlds of MUDs and role-playing games that provided opportunities to explore new identities. All over the Net, people were engaging in discussions with others whom they had never seen face to face, people whose race, attractiveness, and habitual expressions were unknown. Even if their names were known, in this disembodied world, identity was still markedly freer than in face-to-face experience. Discussing the WELL, an early online text-based community where people gathered to discuss anything from headache remedies to world politics, Howard Rheingold observed:

Because we cannot see one another in cyberspace, gender, age, national origin, and physical appearance are not apparent unless a person wants to make such characteristics public. People whose physical handicaps make it difficult to form new friendships find that virtual communities treat them as they always wanted to be treated—as thinkers and transmitters of ideas and feeling beings, not carnal vessels with a certain appearance and way of walking and talking (or not walking and not talking). (Rheingold 1993, 11)

Today, elements of this ideal persist in various forms. News articles spark lively discussions among readers identified only by screen name on topics ranging from foreign policy to health care to vacation tips. Virtual worlds are lauded as places where the disabled can dance and autistic children learn to socialize. Millions of people worldwide play online role-playing games, in which they assume the guise of mythical figures, join guilds with other players, seek treasures, battle enemies, and live out an alternative existence.

But, as is often the case with idealistic visions, reality has proven to be more complex. Many people have found the fluidity of online identity to be deceptive rather than empowering; they do not enjoy finding out that the beautiful girl they had been flirting with is really a man, or that the expert advisor is still in junior high school.

Moreover, the ease of creating and abandoning identities has allowed people to misbehave without repercussions. In the earliest days of online communication, one’s account was attached to one’s position in a university or corporation, a connection that helped ensure responsible behavior. But by the early 1990s—the time of much of the most enthusiastic cyber-utopian writing—consumer Internet access was available for a monthly subscription, and Web-based sites, which are often anonymous, were rapidly gaining popularity.3 These online identities are not only mutable, but disposable. And disposable identity has proved to be incompatible with community. Sites allowing anonymous postings are almost instantly flooded with pornography, commercial spam, and vicious hate speech. It appears that many people, freed from the responsibility of being their real-world self, behave rather badly.

Partly in reaction to the slipperiness of much online identity, social networking sites such as Facebook, which ground identity with solid ties to one’s real-world self, became prominent in the early 2000s. These sites are the antithesis of the “wild side of cyberspace culture,” for they make identity less fluid than it is even in face-to-face experience. In everyday life we typically maintain separate facets of identity: we play different roles in different social contexts, and may consciously try to reinvent ourselves at times of major personal disruption, such as moving away from home to college, and so on. Social network sites collapse these contexts, thus compelling their users to attempt to present a single identity that is appropriate across all situations.4

One can argue that reality has deflated the cyber-utopian dream and that the explosive growth of networking sites featuring participants’ real-world identity shows where the future lies. Online discussions increasingly require participants to register with a verified and sometimes network-based identity; this not only reveals name and gender, but clearly situates the person in his or her real-world social position (Donath 2007; Ellison, Steinfield, and Lampe 2007; Jernigan and Mistree 2009; Utz 2010; Walther et al. 2008). Using a pseudonym and wanting not to connect your online actions with your real-world self is seen as suspicious: why, if your intentions are good, would you want to appear under a fake name? Anonymity is associated with trolls, scammers, and fringe activist groups.

It is easy, in this atmosphere, to forget the excitement and idealism that identity experimentation and anonymous communities sparked a couple of decades ago. Underlying the cyber-utopian ideal was a critique of the physical world as one in which people are tethered to identities based on immediately visible facial and bodily features—many of which, such as age, gender, and race, trigger deep-seated and often prejudicial stereotypes. Inherent to the ideal was the belief that these embodied social identities are not innate but culturally imposed, and that their markers are features only of the body. Online, where there is no body, the identities could vanish.

It is a controversial ideal. Does the absence of the body and of the visibly prominent and embodied features of race, gender, age, attractiveness, and so on lead to an absence of prejudice based on physical cues? Or are the cues to our real-world identities more than skin deep? Is identity volitional—a performance rather than an inherent state of being—and how does technology change this? What, if anything, is fundamentally different about how we make sense of other people online compared with face to face? To answer these questions, we need to look more closely at how we construct and perceive identity.

How We Make Sense of People

To start, we need to clarify what we mean by identity. The word “identity” refers to two different but intertwined phenomena. One is individual identity: who you are as opposed to any other person. This is the identity of social security numbers, fingerprints, passports, DNA testing, face recognition, and the like. The other is social identity: the type of person you are and your role in society.

Individual identity is fundamental for forming relationships: if you are my friend or lover, you are not replaceable with a similar other, no matter how strong the resemblance. We develop a shared history with individuals; as time passes, each relationship becomes more detailed and distinct. We base our decisions about future interactions with others on our knowledge of their past actions. In addition to immediate experience, this knowledge can come via reputation (the experiences of others) or history (records of the person’s past actions), both of which require individual identity, the ability to connect memories and stories to a specific individual. We look more closely at issues around individual identity and identification online in chapter 11, “Privacy and Public Space.”

This chapter focuses on the other phenomenon that the word “identity” refers to, social identity, a loosely defined concept that includes personality, interests, and social role. If you ask about a person “What is he like?” you are asking about social identity. Social identity is how people make sense of you—it is how they understand what sort of person you are, what type of relationship they might have with you, what behaviors they can expect from you, and how they should act toward you (Holland and Skinner 1987). We want to shape the impression others have of us so that they will act in desirable ways toward us (Goffman 1959).

These two aspects of identity are intertwined. The identifiers that we use to recognize and refer to an individual also reveal varying bits of social information. Numeric identifiers such as social security numbers reveal relatively little (though some encode social and demographic data), whereas names may indicate gender, ethnicity, social class, and so on.

Perceiving Social Identity

In a 1908 essay entitled “How Is Society Possible?” Georg Simmel, a founder of the field of sociology, presented the question of how we make sense of other people:

The picture of another man that a man gains through personal contact with him is based on certain distortions. … We see the other person generalized, in some measure. This is so, perhaps, because we cannot fully represent to ourselves an individuality which deviates from our own. … We cannot know completely the individuality of another. …

We conceive of each man—and this is a fact which has a specific effect upon our practical behavior toward him—as being the human type which is suggested by his individuality. We think of him in terms not only of his singularity but also in terms of a general category. This category, of course, does not fully cover him, nor does he fully cover it. …

… Just as we compensate for a blind spot in our field of vision so that we are no longer aware of it, so a fragmentary structure is transformed by another’s view into the completeness of an individuality. (Simmel 1959, 10)

Imagine being introduced to someone at a party. You observe that he is a man, white, middle-aged. You see the clothes he wears; you know the setting in which you are meeting; perhaps you see some of his acquaintances or exchange a few words. After even a brief meeting, you are likely to feel able to make a good guess about many unobserved characteristics—is he a sports fan? Would he like this joke? Whom did he vote for in the last election?

We can make these guesses because we have in our minds types of people. These types provide information about many features of their members; once we have classified someone as being of a particular type, we infer a much broader set of beliefs about him of her than our observations alone would bring.

These inferences may be wrong. Indeed, as Simmel pointed out, they are inevitably distorted. We develop types based on our experience. Since each person’s experience is different from every other person’s, each person’s types are different, and thus the inferences they make about another, given the same observations, are different. For example, upon returning from a dinner party, a couple may discuss the people they had met that evening. Although both participated in the same experience, their impressions may be quite different.

“What did you think of the woman that Danny brought with him?” “She seemed very interesting. So artistic and creative.” “You think so? She was certainly weirdly dressed, with everything mismatched and asymmetrical. But I didn’t hear her say anything particularly interesting.” Is Danny’s date imaginatively talented or just an eccentric dresser? She may be both—or neither. The two people who are discussing her have formed different impressions based on their individual experiences and personal categories. Such conversations highlight the subjectivity of the participants’ interpretations of the same social information.

Such conversations also help people develop shared cultural models, as they exchange and refine their viewpoints (Holland and Skinner 1987). One’s own personality and experiences affect individual perceptions, but the basic types that we perceive, as well as our vocabulary for thinking about them and discussing them with others, we share across our culture. Our culture shapes our impressions of other people.

Prototype Theory

In the hundred years since Simmel wrote about our perception of people as different social types, anthropologists and cognitive scientists have learned more about how our social cognition works. To understand how we categorize people, it is useful to first look at how we comprehend everyday objects.

Classification is at the heart of how we make sense of the world—it is how we can see a new piece of furniture and know that it is a chair and can be sat upon, or meet a new person and know that she is a sweet old lady and should be deferred to. How do we do this? One might think we have definitions, for example, chairs are the set of objects with four legs and a platform big enough to sit on and a back on one side. But, while formal systems such as mathematics and biological classification are based on definitions (even numbers are whole numbers that are divisible by 2; mammals are warm-blooded vertebrates that have hair or fur and nurse their young), that is not how we make sense of our everyday world. Instead, we understand the world through prototypes (Lakoff 1987; Rosch 1975; Rosch et al. 1976).

Prototypes are the most central member of a category, its clearest example. Think of a bird. Chances are the bird you imagined resembles something like a robin or a sparrow and not a chicken or a penguin, even though those are birds, too. The bird in your imagination—your prototype of the category bird—might not be any specific species of bird, but it is distinct enough so you can note that some birds (robins, sparrows, finches) resemble it quite closely whereas others, such as owls, are less similar, and others, like ostriches, are more distant still. Robins and finches are termed prototypical members of the category bird: they closely resemble our conception of the typical member of the category.

These categories are learned and subjective. If you grew up in a world where penguins were the dominant avian species, your idea of the prototypical bird would be quite different. Culture also influences categories—even if you grew up surrounded by penguins, but your schoolbooks featured illustrations of typical English garden birds, you might still think of robin-like creatures as prototypical birds, although they were outside of your personal experience. Categories do not exist as independent external entities; we construct them through our interactions with the world and with each other.

Although the categories into which we classify physical things—birds, chairs, cars, trees, shades of red—are socially constructed, they are not arbitrary; they reflect the structure of the objects themselves. We also have many categories into which we classify social things—family, mother, geek, spinster, untouchable, hippie, hipster, and so on. These too are not arbitrary; they reflect the larger framework of how a culture structures the world. To understand why some unmarried men are “bachelors” and others are not (gay men, Catholic priests, men confined to mental institutions or in long-term, cohabiting relationships) requires understanding the culture notions about eligibility for marriage (Lakoff 1987).

Much can go unsaid in conversations with friends who share your cultural models, for you can simply refer to a shared prototype to conjure up a detailed picture. Discussions with people with different beliefs or from different cultures can be more difficult, for disparate types may populate your social worlds. Even where you share the basic types, your assessments of them may differ. A mother offers to introduce her twenty-year-old daughter to her friend’s son, a nice boy who has a steady job in a bank and volunteers at church on the weekend; mother and daughter may share enough cultural foundation to agree upon the type of person he is, but may differ significantly on their assessment of whether it makes him desirable.

Some cultural categories exist only within a small community. In To Kill a Mockingbird, Miss Caroline, a new first-grade teacher in town, tries to lend a quarter to a boy who did not bring any lunch to school. Jean, another first-grader, attempts to explain to her why he will not accept it (Lee 1962).

I rose graciously on Walter’s behalf: “Ah—Miss Caroline?”

“What is it, Jean Louise?”

“Miss Caroline, he’s a Cunningham.”

I sat back down.

“What, Jean Louise?”

I thought I had made things sufficiently clear. It was clear enough to the rest of us: Walter Cunningham was sitting there lying his head off. He didn’t forget his lunch, he didn’t have any. He had none today nor would he have any tomorrow or the next day. He had probably never seen three quarters together at the same time in his life.

The students in the class, as inhabitants of the town, shared a common knowledge of the different clans and neighborhoods; they knew what to expect of each and how to act toward them. Ordinarily, they would never state this knowledge out loud. However, the presence of a newcomer, who did not understand any of this and yet was in a position of authority, required that they explicitly describe their cognitive model of the local society.5

Online, the cues to identity are different and sparse, but the cognitive process of impression formation is the same,6 and we can still form vivid impressions of others. In one study of this process, a participant, given simply a screen name, filled in the details as follows:

JoshSamBob’s name sounds like a white southerner. He fits my idea of a stereotypical frat boy. He’s just under 6 feet tall, medium build, blonde hair, blue eyes, his lips are a bit thin. He’s particular about what he wears and is in general very neat. He tends to fidget a lot and would much rather be out playing sports than sitting in a lecture hall or library. He enjoys horseplay and being where the action is. (Jacobson 1999)

Sparse cues can have the paradoxical effect of producing more intense impressions: absent specific individual details, we rely more heavily on our internal prototypes, including exaggerated stereotypes,7 to fill in the blanks (Hancock and Dunham 2001).

Not all cues are evocative. A generic name, with no clear cultural meaning, such as gss93, will conjure up very little. But in the absence of information, we do not perceive others to be complete ciphers; instead, we bestow upon them what we imagine to be the typical identity in that setting. If you see a generic name in an online forum that you imagine to be populated by white men in their thirties, you will ascribe the qualities you associate with that type to an otherwise unmarked person.

Online discussions bring together people from around the world, with greatly varying cultural models. In the example above, you can imagine that the name “JoshSamBob” would evoke a different impression (or none at all) in a participant unfamiliar with the college culture of the American South. A name or behavior that makes a vivid impression on one person will imply something quite different to another. One challenge in making cross-cultural discussions work is to help people recognize these underlying differences and work toward creating common ground.

Mistaken Impressions

It is remarkable how much we can sometimes accurately infer from very little data. Studies estimate that people accurately assess personality about 70 percent of the time after observations as brief as thirty seconds (Ambady and Rosenthal 1992). Certain traits are more readily observable: people quickly recognize whether another is outgoing or introverted, dominant or passive, anxious or confident, warm or chilly. Other traits, such as creativity and intelligence, are harder to detect. Too, some people are more legible than others—unsurprisingly, those who are outgoing and confident provide more cues about themselves (Ambady, Hallahan, and Rosenthal 1995; Ambady and Rosenthal 1992; Funder 1995). And, the best judges are the “socially vulnerable”: women, shy people, and people with lower self-esteem; they need to be more alert to the moods and intentions of others. Confident, outgoing people—less dependent on the whims and wishes of others—tend to be less accurate judges.

Overall, we are “good enough perceivers” of social traits (Fiske 1993): although we make many errors in our assessments of other people, we generally do well enough to make sense of and function in the world.

However, we do make pervasive errors in assessing others. The leaps of categorization that allow us to quickly make sense of a diverse social world also lead us to the distorted and harmful assumptions of racial prejudice and gender stereotypes. Studies have shown that people judge essays, job applications, musical performances, and the like lower when they think they were produced by a woman rather than a man (Bertrand and Mullainathan 2003; Goldin and Rouse 1997; Steinpreis, Anders, and Ritzke 1999).

Furthermore, we overgeneralize, too broadly assuming that a feature implies that a person has certain character and personality traits (Zebrowitz 1997; Zebrowitz and Montepare 2008). For example, Leslie Zebrowitz and colleagues have found that people ascribe characteristics such as naiveté and trustfulness to strangers who have small chins, upturned noses, big eyes, and other childlike features—although having those facial features does not correlate to actual personality differences. Zebrowitz hypothesized that our evolved response to babies and children—to nurture them, to expect naive behavior from them—was also triggered by individuals who happened to have those sorts of features as adults.

Performing Identity

One of the key visions the cyber-utopian writers set forth for online society was to have people’s words and actions assessed on their merits rather than filtered through someone’s belief that, say, black people or young mothers or people who speak with an accent are less intelligent and able. Merit-based assessment is possible online because embodied cues such as facial features that reveal gender and race or a receding chin that is equated with weakness can be unseen.

The desirability of masking an identity cue online depends on one’s assessment of the validity of the prototypes it triggers. Let’s say that the cue “female” tends to make people think of someone nurturing, passive, risk-averse, emotional, not intellectually rigorous. If you think that this is an accurate model for the prototypical woman, then you will see the cue of gender as being a very useful guide for understanding the other. If you think this prototype distorts the way people see women, then you may believe that eliminating the cues to gender can help women be perceived more accurately. These controversial questions arise about all aspects of identity: they are at the center of debates about racial and ethnic profiling, women’s role in combat, men’s parenting instincts, and so on.

In a society where roles are immutable, no such controversy arises. It is when roles are changing that claims about what traits are cultural or innate become hotly contested. The Internet, with its novel opportunity for identity reinvention, arrived at a historical moment that was already wrestling with these questions.

In medieval Europe, rulers and religious leaders exhorted their subjects to be content with their station in life, for it was divine will that made unequal positions and placed people in them (Herlihy 1973). Since then, rigid ideas about innate identity have been steadily loosening. Five hundred years ago, the appearance of guilds and a growing merchant class shook up the highly codified social structure of medieval feudalism. The eighteenth century saw the rise of democracy over inherited monarchy. In nineteenth and early twentieth-century America, slavery was abolished and voting rights were extended to all women. Today, although we still have many entrenched stereotypes, we accept that many aspects of someone’s identity, including social status, religious beliefs, and the like, can change over time.

By the mid-twentieth century, the notion of identity as performance arose in both popular culture and intellectual writing. Popular fashion reflected a growing ambiguity in gender roles with the acceptance of women dressing in pants and other traditionally male garb. The world of ideas included works such as Erving Goffman’s influential book The Presentation of Self in Everyday Life (1959), which used the metaphor of theater to describe society: daily life is a performance in which people attempt to create a desirable impression, within the constraints of their role in a particular context and their social (acting) abilities.

This dramaturgical model of society resonated strongly with growing social movements that questioned traditional roles. For generations, women had not been allowed to vote, own property, or pursue most careers—constraints based on the argument that women were biologically unfit owing to their smaller brains, excitable hormones, and destiny as child-bearers. As the women’s movement gained force and these constraints were lifted, it became clear that women’s lesser abilities were mostly culturally imposed. Girls acted silly not because they were silly, but because they were taught to do so.

In her book Gender Trouble (1990), radical feminist Judith Butler took this argument further, positing that all of masculinity and femininity is performance, separable from the physical shape of male or female bodies. Radical as her writing might be, she noted that the response to it had been even more extreme than she intended: “One of the interpretations that has been made of Gender Trouble is that there is no sex, there is only gender, and gender is performative. People then go on to think that if gender is performative it must be radically free. … What’s interesting is that this voluntarist interpretation, this desire for a kind of radical theatrical remaking of the body, is obviously out there in the public sphere. There’s a desire for a fully phantasmatic transfiguration of the body” (Osborne and Segal 1994, 110–111).

By the late twentieth century, traditional stereotypes based on visible characteristics were under attack, in realms ranging from academic cultural theory to everyday life. A growing body of laws, especially in Europe and the United States, forbade discrimination based on gender, race, or age. These traits, which people a generation earlier had assumed were legitimate, commonsense cues about a person’s character and suitability, have now become legally invisible. Although the goals of such legislation are less extreme than those of race- and gender-bending activists, they are evidence of a worldview in which identity is seen as personally constructed rather than anatomically destined.

It is into this cultural atmosphere that online communication arrived. It is thus not surprising that many idealistic visions for the technology focused on the opportunity to create social spaces in which one’s physical traits were unknown.

In the face-to-face world, even if it can be shown that most of the personality and cognitive traits associated with categories such as race and gender are culturally produced, many of the cues that place a person in these categories are inborn and immediately visible in one’s appearance, making it difficult to escape from them as basic ways of organizing the physical world.

But the online world seemed to present no such difficulty. With no body, you could just claim to be, well, whatever you wanted. Or could you?

Identity cues are given off in many forms, not just appearance. Writing—the medium of online communication—also provides cues to gender. Susan Herring, a linguist who has carried out numerous studies of gender online, notes that, in general,

males are more likely to post longer messages, begin and close discussions in mixed-sex groups, assert opinions strongly as “facts,” use crude language (including insults and profanity), and in general, manifest an adversarial orientation towards their interlocutors. … In contrast, females tend to post relatively short messages, and are more likely to qualify and justify their assertions, apologize, express support of others, and in general, manifest an “aligned” orientation towards their interlocutors. (Herring 2000)

In the absence of visible cues to gender, verbal cues can reveal it.

Misclassifying someone—whether because he or she claimed to be a different identity or had provided no cues and we had made an erroneous assumption—distorts our interpretation of his or her behavior. The norm for women is to be more polite than men. Thus, if I correctly assume that a particular writer is a woman, and she writes in a typical female fashion, I will mark her neither as unusually nice nor aggressive. But let’s say her screen name gives no indication of her gender and I initially assume she is male. I will then read “his” words differently—they will sound quite deferential, even meek. My impression of this person—a woman writing in a normal tone for a woman—would then be one of a rather submissive and self-effacing man.

This distortion in social perception happens as well in the face-to-face world when a person’s behavior does not match the expected cultural norm. One of the problems facing women who seek leadership positions is that candidates for these roles are expected to be outspoken and assertive. A man with these qualities is perceived to be authoritative; a woman with them is seen as bossy, shrill, and aggressive.

In an online discussion, a woman with strong, clear opinions might find her ideas received more favorably with her gender masked—those who would reject the suggestions of a “strident woman” might welcome them when they thought they came from an authoritative man. Yet this is not an ideal solution; at best, it effects improving perception by guiding others to a better-fitting, but still quite inaccurate, stereotype.

Negative stereotypes, which unfairly denigrate their subject, are clearly harmful. But even positive ones can cause harm, if they make someone seem to be more trustworthy or well-meaning than he or she actually is. One goal for an ideal society is thus not to eliminate social categorization—it’s essential for helping us make sense of the world—but to make the prototypes and their triggering cues more closely aligned with actual behavior.

To truly effect change in how people are understood, the cultural models themselves need to change.

Changing Cultural Models

Changing cultural models is a slow process. It is often easier simply to miscategorize the other rather than to adapt to a new worldview. In the famous husband-and-wife architect team, Robert Venturi and Denise Scott Brown were equal collaborators, but Venturi gained far more acclaim: the prototypical architect was male. Scott Brown said that no matter how forcefully she and Venturi explained that that were equal partners, people assumed she did not have a creative role: “Whatever you say to them, they say, ‘Well, she must be something else. Maybe a planner, maybe a typist, maybe she takes photographs. It has to be something else!’” (Cook 2013). People found it easier to distort their own observations than to change their assumptions about women in architecture.

Pervasive cultural stereotypes are especially hard to change because one’s distorted perception will be reinforced by the similar beliefs of one’s peers. If a population of observers thinks that all white people are greedy or all Asians are treacherous or all black people are violent, their shared stereotype allows them to be in agreement without being accurate (Allport 1979; Lee, Albright, and Malloy 2001).

Yet cultural models do change and new types arise.8 By the end of the 1960s, most Americans could identify a person with long wild hair and tie-dyed clothing as a hippie, a type that had not existed a decade earlier. They could make numerous assumptions about the latter’s politics and lifestyle and would have beliefs about how to act toward him or her—whether as the person they would gravitate to at a party or as the one they would forbid their daughters from befriending.

Online, new social types have emerged in many communities. People discuss what they are, how to recognize them, and most importantly, how to act toward them. Disruptive types get much attention, because figuring out how to discourage them is necessary for the community to thrive: “newbie,” “bot,” and “troll” were among the most widely recognized and discussed in early text-based discussion groups (Golder and Donath 2004). “Trolls” initially pretend to be legitimate participants but become increasingly disruptive, often by making offensive statements designed to derail the discussion (Donath 1998). Communities developed detailed cultural models of this type, as they analyzed whether an offensive user was actually a troll or simply an abrasive individual and then debated how to respond. Understanding the trolls’ motivations was essential for figuring out how to discourage them, and once they were seen as pranksters poking fun at the infuriated response they drew, it became clear that ignoring them would be most effective response. “Newbie” is a somewhat derogatory term for newcomers to the online community; they are people who do not yet know the rules of behavior. One description of newbie says that a “well-constructed” troll’s post “induces lots of newbies and flamers to make themselves look even more clueless than they already do, while subtly conveying to the more savvy and experienced that it is in fact [deliberate]” (Raymond 2003). Newbies are those who have not yet learned about cultural categories such as trolls.

Interface design can help shape culture by making cues to the qualities that are important to the community prominent. The data portraits we discussed in an earlier chapter are one way to do this. Seeing others via a representation that highlights actions such as responsiveness, the things they talk about, and the rhythm of the interactions, as opposed to one that provides cues to features such as gender, race, and age, could help create a society in which those traits form the basis of new prototypes—and become the primary models for a community’s structure.

How big a change can design make? One answer comes from evolutionary psychologists who suggest that the categories associated with race would be relatively easy to transform (Cosmides, Tooby, and Kurzban 2003; Kurzban, Tooby, and Cosmides 2001). They argue that, whereas people immediately and automatically encode new individuals by age, gender and race, age and gender are more fundamental, while race is a “by-product” of adaptations for detecting coalitions and alliances.

Our subjects had experienced a lifetime in which ethnicity (including race) was an ecologically valid predictor of people’s social alliances and coalitional affiliations. Yet less than 4 min of exposure to an alternative social world in which race was irrelevant to the prevailing system of alliance caused a dramatic decrease in the extent to which they categorized others by race. This implies that coalition, and hence race, is a volatile, dynamically updated cognitive variable, easily overwritten by new circumstances. If the same processes govern categorization outside the laboratory, then the prospects for reducing or even eliminating the widespread tendency to categorize persons by race may be very good indeed. (Kurzban, Tooby, and Cosmides 2001, 15391)

Human society organizes around in-group and out-group clustering. We may not be able to change this tendency, but we can change the features that define the clusters.9

As for gender and age, although we may never be able to fully tease apart the innate from the cultural, we can design spaces that make these features more or less prominent, depending on their relevance in the context. Your gender should not be relevant in a forum on math, but may be in a discussion on sexism—or rating shoes.

The anonymous comments on New York Times stories provide an interesting example of people choosing to selectively reveal different facets of their identity. Depending on the story, people will note relevant aspects of their identity, to provide context for their opinions and establish their authority on the subject. An article about Trayvon Martin (Barry et al. 2013), a black teenager shot and killed while visiting his aunt by a white vigilante who though he was a trespassing criminal, had comments with identifiers such as: “I’m an old white guy and I formed the opinion that when the dispatcher said ‘we don’t need you to do that’ Zimmerman should have stopped and met the police as he was instructed to do,” and “Even though my husband and I are white, we could never live in a state where the law enforcement fails its citizens and stands behind such an ill conceived law as ‘Stand Your Ground.’” People commenting on an opinion piece by Sherry Turkle (2012) bemoaning the growing use of social technologies at the expense of face-to-face interaction often mentioned their age: “Although I’m in my late 60’s and lived most of my life without anything like email, even I now find that emails are an important part of my life”; “I’m twenty-five & rather prone to anxiety. I don’t use Twitter or Facebook, I don’t have a smartphone or tablet computer, I don’t watch TV. I own an iPod but no longer use it.” These writers all felt that others needed to be aware of these aspects of their identity in order to understand their opinions.

Identity Play versus Deception

In the examples above, nothing prevents the anonymous writers from falsely claiming an identity (if the accounts were pseudonymous, one could still make false claims, but would need to stick with them. One could not be a twenty-year-old one week, and fifty the next). The comments on a news article may not be the weightiest of texts, but the reader is still arguably harmed by such deceptions. I might feel that, for example, a stand against vigilante patrols in the Trayvon Martin case would be more powerful coming from a middle-aged, white Southern male—the type many people would imagine support such patrols—but it is deceptive and misleading to falsely claim this identity to give my comments a credibility that they have not earned. Even worse would be using a false identity to make trouble; this is the behavior of an Internet troll.

Yet, you might ask, in the last section we talked about masking identities to be more accurately perceived—is that not deceptive also? Not generally, in cases where identity information was simply absent, and the perceiver made the wrong assumption; we cannot be responsible for all the mistaken assumptions people make (though one could argue it is deceptive if the mistake was very likely and the person counted on it being made). A more ambiguous case is the one in which the person deliberately provides misleading identity information—but with the intention of avoiding being inaccurately stereotyped. We can think of examples that fit this model that we would call deception and others that we would find acceptable. Here we want to assess whether the motivation was to avoid misjudgment by an erroneous stereotype and whether the resulting assessment proved beneficial to the perceiver.

In 2011, during the “Arab Spring,” a time of great political unrest in the Middle East, a blog appeared with the title “A Gay Girl in Damascus.” The author, Amina Abdallah Arraf, wrote about being gay and a woman in an oppressive society. People were fascinated by her story and the insight it provided into Syrian culture; she had many online supporters and journalists covered her story. In June, she was abducted. Activists lobbied for her release, even getting the attention of the US State Department. Soon, however, the hoax was uncovered. Amina turned out to be neither gay, nor a girl, nor in Damascus, but instead was the creation of a white male graduate student in Edinburgh, invented in order to attract attention and garner credibility for his views on the Middle East. In his apology, Tom McMaster, Amina’s creator, defended the righteousness of his actions: “I feel that I have created an important voice for issues that I feel strongly about. … I do not believe I have harmed anyone” (quoted in Whitaker 2011). Yet his creation was harmful: besides misleading people into believing they were gaining real insight into the world and thoughts of a gay woman in Syria, the journalists who had written about Amina lost credibility, legitimate Syrian bloggers faced increased scrutiny and doubt about their veracity, and the people in Syria who attempted to rescue her took serious risks trying to save a fictional character (Addley 2011; Hajratwala 2011). Here, real social identity matters a great deal—not individual identification; if Amina were a real person, revealing her name would put her in danger from a repressive government (a fact that helped sustain the deception)—but the writer’s actual cultural identity. For many stories, authenticity matters.

Had McMaster written an openly fictional account, these harms would not have occurred.10 Many acclaimed novels show the world through the eyes of a character far removed from the author’s identity. The make-believe of fiction takes place with the reader’s consent.

Online, there are numerous sites where fictional identities exist by common consent. Yet even in social spaces that encourage creative identity, such as the graphical worlds (The Palace, There, Second Life) that come in and out of online fashion every few years, most people create an online persona that strongly resembles their real-world self—or quickly devolves back to it. Maintaining a fiction is hard. For the most part, what people want to do is shape the impression they make. The online environment changes what elements we can play up or tone down, but in the theater of social life, we mostly want to be performing as ourselves.

The extremely popular massively multiplayer online role-playing games (MMORPGs) are an interesting example. The roles are predesigned by the game makers: the players inhabit an elaborate costume, typically including a fictional racial type (role-playing games neatly illustrate the hypothesis that people can quickly and fervently adapt to arbitrary ethnic divisions) and clearly defined characteristics. At first glance, they appear to be worlds in which everyone plays a character far removed from their mundane selves. Yet while the participants interact as fictional characters, many build close friendships that carry over outside of the game.11 Nicholas Yee, a social scientist who studies behavior in online games, argues that these games provide an advantageous environment for developing relationships. The real physical identity cues are all hidden (age, gender, attractiveness, etc.) and the resulting pseudonymity encourages self-disclosure: people are more open than they might be face to face. At the same time, the vivid mask of the role makes each player memorable. Unlike many other forums, where interactions are fleeting, players in these games work as teams and spend considerable amounts of time together online. The gameplay itself provides a proving ground for character: players get to see who takes risks to help others, who acts selfishly, and so on. Participants thus form impressions of each other based on many and various shared experiences, before they learn what the other looks like (Yee 2009).

Popular as they are, MMORPGs are not to everyone’s taste. But they demonstrate the feasibility of one cyber-utopian goal—to have a place where people can get to know each other without immediately resorting to preconceived opinions based on real-world physical appearance. They show that to achieve this in a meaningful way, simply making up an identity is insufficient. The identity needs to be persistent and needs to exist within a community with shared goals and activities.

Another cyber-utopian goal—that playing another ethnicity or gender online would give someone special insight into what it is like to have that identity in the real world—has proven less promising. Much online identity play has ended up being rather shallow. People create a character, claim a gender, but soon forget about the role, like a masquerade party when, after the first excitement of seeing the costumes, the Supermen and witches and ghosts sit around eating dip and talking about sports and the PTA. Social scientist Lisa Nakamura used the phrase “identity tourist” to critique the notion that radically free identity choice is a transformative experience: “It gave users a false notion of cultural and racial understanding based on an episodic, highly mediated experience” (Nakamura 2008, 1675). Being able to painlessly assume the appearance of a different identity does not provide an understanding of the social consequences to being seen as, raised as, and living as male or female, or black, white, or Asian, and so on. At most, it gave people the opportunity of experiencing how others act toward the claimed identity.

Multiple Identity Spaces

Identity is not monolithic: there is not a singular, unchanging self that is you. But neither is it a constantly shape-shifting shroud. We have inner qualities—our beliefs, personality, knowledge, and experience—that make up who we are. We have roles in society that shift over time and vary by context, ranging from our individual relationships to our position in a large-scale community. We try to make a desired impression on others. We may wish to appear tough, kind, cool, or loving. Sometimes the impression we wish to give to others is a great match with our inner self and social role, and sometimes it is quite divergent. Whether this divergence is deception, ambition, role-playing, or self-expression depends on the person’s intentions. A man who presents himself as a woman online may be doing so for a range of reasons. He may enjoy fooling people (e.g., a troll on a dating site), he may wish to understand better the experience of being a woman, or he may feel that he is in fact more female than male and it is his body that is at odds with reality. The level of reinvention people find acceptable depends on the context. There can be great freedom in games and role-playing spaces, whereas in other situations we want to maintain a level of honesty, ensuring that the impression and information we get from each other is reliable. Almost anyone who has posted anything online—which is now a sizeable percentage of the world’s population—has had to think about identity. What name to use? What tone to take?

What we want to know (or ought to know) about the identity of another depends on context. In a technical support forum, it may be better if the social identities of the customers and technicians remain hidden, the better to avoid immediate, perhaps even subconscious assumptions that, say, a non-white non-male is technically illiterate. On a dating site, gender is quite relevant. Age, too, can sometimes be important. Numerous sites on the Web today explicitly forbid minors or children under a certain age from entering, because the content and activity is not appropriate for them; at the same time, other sites are set aside to be places for children to interact with other children (and hopefully not with forty-year-olds pretending to be kids).

The design of an online space can make it easier for people to fabricate claims about themselves, or it can make it more likely that one will see, over time, sufficient detail about another to form a well-grounded sense of them. Online, it is technologically easy to create a space where little or no information is given about the effectively anonymous participants; it is only a bit harder to create one that insists on a real name for identity. The key challenge is in creating online spaces that are in between, where we are known but not entirely transparent. The design goal is to create a space in which the most obvious and salient properties that we see of the other are the qualities around which we want society to structure itself.