120 Urban Danger

instru�ental relationships come to predominate. Cities are characterized by anomie: by a sense of normlessness, both in the sense of the individual's lack of attachment to a moral code and of a collective loss of moral consensus.

In the years since Wirth's article was published, numerous studies of the social life of cities have challenged his theory. Many researchers have described neighborhoods, workplaces, religious communities, and other urban settings in which people know one another and treat each other in terms of intimacy and interdependence (e.g., Lewis 1952; Whyte 1955· Young and Willmott 1957; Gans 1962; Hannerz 1969; Stack 1974; Hanner; 1980). In these places, the close community life which Wirth saw as charac­ terist�c . of vill�ges flouri�hed. Other critics argued that Wirth was really descnbmg the impact of mdustrialization, not urbanism. Preindustrial cities lacked the anonymity and disorder of large industrial cities (Sjoberg 1955; �apf-Askan 19�9; Fox 1977). Still others claimed that, despite the apparent disorder of the city, particularly its lack of cohesive neighborhood communi­ ties, new forms of community have emerged in modern industrial cities based on social networks, voluntary associations, and other more ephemeral, yet important, quasi-groups that stretch across neighborhoods and regions (Mitch- ell 1969; Boissevain 1974).

Despite these serious and legitimate criticisms of Wirth's vision of the city, however, sociologists and anthropologists still return to his article because they have a lingering sense that he has special insight into the issues; that there are some features of the city life he describes that do resonate with the experience of living in the modern industrial city. Such cities do have many strangers, and strangers are particularly taxing and troubling for urban­ ites. They lie behind the problems of urban crime, the fear of crime, and even the fear of immigrants and newcomers, of people who are culturally strange and different.

Although Wirth's theory of urbanism does not describe the totality of urban social life, it does apply to particular aspects of city life. It describes the bou�daries �etwee� sep

_ arate social worlds. Neither Wirth nor Park, an early

and mfluentrnl soc10logist at the University of Chicago, thought that the city lacked close-knit urban villages, but saw the city as a collection of these small communities. Park described the city as "a mosaic of little worlds which touch but do not interpenetrate" (1952:47). It is at these boundaries that the charac­ teri�tic features of urban social life, as Wirth described them, appear. Such social boundaries within a single housing project-Dover Square-and how they generate both crime and a sense of danger will be explored in this chapter.


Inspired by Wirth's vision of urban social life, I decided to conduct a cul­ tu:al study of urban _danger. I wanted to learn how people who live in high­ cnme urban areas thmk about danger, how they deal with it on a daily basis, and how anonymity fosters crime. After perusing crime statistics and neigh-

Sally Engle Merry 121

borhood descriptions, I selected a subsidized housing project located deep in neighborhood with one of the highest crime rates in a major northeastern

:ity. Since the intent was also to investigate the impact of s _ everal ethni

_ c

groups living side-by-side, a neighborhood that had a broad ffilXture. of resi­ dents was chosen. After locating an ideal small development housmg 300 white, black, Chinese, and Hispanic families, I moved into an apartment three blocks from the project and spent a year and a half carrying out anthro­ pological field research. The residents were questioned regarding how they thought about danger and about how they handled the high rate of crime sur­ rounding them. Project residents who committed crimes against their neigh­ bors were also interviewed concerning their views of crime and danger.

In the mid- l 970s, when the study was conducted, the neighborhood was slowly changing from a depressed area populated by homeless alcoholics and characterized by deteriorating housing projects to a trendy neighborhood attractive to wealthy professionals. More daring members of this group bought the old buildings, some of which had a tarnished elegance about them, and renovated them into attractive townhouses. In the 10 years siI.1.ce the completion of the study, this process of gentrification in the surroundmg neighborhood has continued, but the project has remained much the same. The ethnographic present will be employed to describe the project.

The residents of the Dover Square project are 55 percent Chinese, 14 per­ cent black, 9 percent white, and 9 percent Hispanic, by population, although the number of families is more evenly distributed. The Chinese residents, most of whom are recent immigrants from Hong Kong, speak little or no English, although their children are typically fluent in English. They plan to stay in the United States. Many of the blacks arrived in the city in the 1950s during the massive black migration from the South and have lived in less attractive parts of the city for 10 years or more. A substantial number of the whites are connected to an established Syrian-Lebanese community nearby dating from the early 1900s. Some of this community was razed to build the project. The Hispanics are recent arrivals from Puerto Rico and many do not speak fluent English. Many plan to return to their homeland and consider their stay in the city temporary.

The project is federally subsidized, designed to house both low- and moderate-income families. It opened in the mid- l 960s and, 10 years later, had a remarkably stable population for a housing project: well over half the families had lived in the development since it opened, and the rate of turn­ over was under 5 percent per year. Yet, despite this stability, the neighbor­ hood has not become a community. It is not a cohesive, integrated social system, but rather a series of distinct, unconnected social networks occupying the same geographical space. Each ethnic group is scattered throughout the development, yet residents maintain virtually all of their close social relation­ ships with neighbors of the same ethnicity. Consequently, neighbors who belong to different ethnic groups often remain strangers, despite years of shar­ ing the same stairwells, porches, and walkways.

122 Urban Danger

A social network is a way of conceptualizing those parts of social life that do not form bounded, enduring social groups (see Barnes 1954; Bott 1957; Epstein 1969; Mitchell 1969; Boissevain 1974). Each person is the center of a group of friends and relatives, the central point from which radiates a series of links to other people. This constellation forms an egocentric social net­ work. Members of this network also know others, some of whom the first person does not. These are second-order links, friends of friends. By exten­ sion, each second-order link also has. social contacts, so that one can imagine a network of social relationships extending outward from any individual to first, second, and further orders of contact. Since these networks of relation­ ship are also potential communication channels, mapping their structure and boundaries provides important clues to the flow of gossip and information through a social system (Merry 1984). Boundaries in social networks tend to be gaps in the flow of information.

In Dover Square, intimate networks-links to close friends and rela­ tives-are alr�1ost always restricted to a single ethnic group. An extended net­ work of acquaintances crosscuts ethnic Lines at a few points, but also tends to remain within a single ethnic group. A few individuals have social networks consisting almost entirely of a different ethnic group, such as the white youths who regularly hang out with a gang of young black men and women. Since each ethnic group is scattered throughout the area, the social organiza­ tion of the project consists of a series of discrete, overarching social networks. One can imagine this social composition as several layers of fishnet strung over the same space with a few threads running between the layers.

The social boundaries between ethnic groups persist because each group is encapsulated within a network of social relationships and a set of institu­ tions that stretches to nearby black, white, Chinese, and Hispanic communi­ ties. The majority of families in the project regularly visit relatives, friends, religious groups, and social organizations in their nearby ethnic communi­ ties. Chinatown lies on one side of the project, an established Syrian-Leba­ nese community on another, the black community is close by on another side, and a substantial Hispanic settlement is in the middle of the slowly gen­ trifying neighborhood nearby. Jobs, friends, marriage partners, churches, social services, and recreational opportunities are all primarily available within these communities. Consequently, relations with members of the same ethnic group carry an expectation of continuity that is not characteristic of relations with neighbors in Dover Square. Neighbors are only temporary associates, here today but gone whenever they move away, while people in the same ethnic group are connected by enduring ties. The denser mesh of personal ties and group affiliations within ethnic groups means that Dover Square residents are far more accountable to their fellow ethnics than they are to their neighbors of different ethnicity.

Because of the boundaries between ethnic groups, neighbors are often anonymous. This anonymity provides opportunities for crime, since crimi­ nals can rob their neighbors with little fear of apprehension. Many project

Sally Engle Merry 123

residents observed that, in general, criminals prefer not to work close to

home where they can too easily be identified by their neighbors, but here, where neighbors are often strangers, _a resident can rob or burglarize people close to home without fear of identification. This means that a project resi­ dent can commit crimes on his home territory, which is relatively safe, pre­ dictable, and familiar, while appearing to victims as a stranger from a distant area. The same people can be robbed whose daily habits and material posses­

sions are easily visible. At the same time, the widespread fear and distrust of neighbors under-

mines community efforts at controlling crime. As one of the leaders of the youth group active in committing crimes in the area put it:

The people who are being affected by crime don't understand that they

are the cause of crime. I think a lot of people around here don't want

other people's houses to be safe. People are beginning to be cold-hearted,

not caring enough, because, if people cared enough about other people,

they would care about theirs. In order to protect your house, you have to

protect your neighbor's.

Social control is undermined by this structure of social networks since the implementation of sanctions, of punishments for rule-breaking, is unlikely across the social boundaries. For sanctions to be effective in discouraging rule-breaking, they must be both powerful and certain of implementation. A sanction that is severe will have little deterrent effect if the offender feels that there is little chance that he or she will feel its weight. When the offender is anonymous to his victim, it is obviously difficult to catch him or to impose any penalty on him. If the person who observes a crime knows who the per­ petrator is or even where he lives or who his friends are, she might be able to impose some kind of pressure on him. When the perpetrator is a stranger, however, even observing the crime act leaves the observer powerless. The strategies Dover Square residents develop to cope with living in a hazardous environment reveal the critical role played by this knowledge of who the dan­ gerous people are, both in protecting the individual from victimization and in reducing the sense of danger.


As I became further acquainted with the residents who lived in the project, I discovered that they differed greatly in their perceptions of and approaches to the dangers in their environment. One young Chinese woman, for example, never returned home alone on foot after dark. When she arrived by car, she honked the horn to alert her parents and then dashed the 20 feet to her door. A white man cautiously packed his suitcases into his car under cover of darkness before leaving for a trip to escape being noticed by potential burglars. A middle-aged black woman sneaked surreptitiously from her home at 6:00 AM to do her laundry before the neighborhood youths gathered in the

124 Urban Danger

laundromat to visit and smoke. She was anxious not to leave her home vacant, even for a few minutes, lest the burglars she constantly feared notice that her house was unguarded.

Yet, in the same neighborhood, a young black woman moved freely, vis­ iting neighbors late at night with no thought of danger. Young men would r�ndezvo�s in dark seclude� hallways even though they were aware that they nsked bemg mugged. A Chmese man, reputed to possess marvelous skill in the martial arts, was studiously avoided by youths seeking safe and profitable robbery victims. He walked through the project without fear. Lastly, an adult black man declared that his neighborhood was very safe because he knows everyone, and everyone knows him.

All of these people face the same hazards, yet their attitudes, fears, and modes of coping vary enormously. Why do they respond so differently to the same risks? Urbanites, in general, continually make decisions about which situations and persons they consider dangerous, but these judgments are rarely based op detailed statistics about where and when crimes occur. Nor are such people's attitudes proportional to the statistical risks of victimiza­ tion, either in this project or nationwide. A survey of the victimization experi­ ences of two-thirds of the 300 households suggested that the rates of victimization of black and Chinese households are roughly the same: about half of each had experienced a robbery, burglary, or assault since moving into the development IO years earlier. Yet, when asked how dangerous they found their environment, the Chinese residents reported much higher levels of fear: 30 percent of those interviewed said the project was dangerous in contrast to only 13 percent of the black respondents, and 18 percent said it was not at all dangerous, while 65 percent of the blacks interviewed felt this way. National statistics show a similar discrepancy between fear and the chance of victim­ ization: although fear focuses on the random, unpredictable attack of the stranger, the risk of assault and murder by friends is far greater.

Clearly, danger cannot be equated with the statistical probability of being the victim of a crime. Instead, it is the individual's interpretation of the sur­ rounding environment. The process of forming attitudes about which kinds of people, places, and times of day and night are safe or dangerous, and the cues which identify them, is one facet of the elaborate process by which an individual comes to know her world. Information from the mass media from friends and neighbors, and from the urbanite's own experience is const;ucted into a mental map of the city which guides behavior and creates a sense of safety in the midst of danger. What the individual considers harmful is itself a �ltural pro�uct. For some, danger is the risk of losing property; for others, it 1s name callmg and personal humiliation; and for others, the degradation of abusive police and social service workers.

The term risk is used to refer to the likelihood of experiencing a crime or some other harm. It is thus a concept that refers to the external world and to the hazards it contains. Danger is a cultural construct that describes the way an individual conceptualizes the hazards and risks in his or her world and assesses

Sally Engle Merry 125

what they mean to him or her. Fear refers to ·the inner emotional state an indi­ vidual experiences as he or she contemplates the danger he or she believes to exist. Thus, both danger and fear are subjective in a way in which risk is not. On the other hand, they are not inevitably connected. Some may see a situation as dangerous but not regard it with fear, while others may see the same situa­ tion as dangerous and feel fearful about it. Ideas about danger are a component of culture: they are learned, shared within groups, and influence the way the world is interpreted and understood. Fear is more individualized, depending on each person's experience with harm, sense of competence and control, sense of vulnerability in general, and other psychological characteristics.

Other cultures similarly define danger in terms of more general belief systems about their world. For example, a group of Indians living in remote forest settlements in Canada, the Salteaux, do not fear wolves or bears but consider snakes, toads, and frogs dangerous (Hallowell 1938). Although these are among the most harmless of the inhabitants of the forest, they are believed to be emanations of powerful supernatural forces, capable of acting as emissaries for sorcerers, of exuding malevolent magic, and of serving as omens of ill fortune. Monster toads and frogs roam the forest. On the other hand, the Azande farmers of East Africa consider dangerous the man who is quarrelsome, spiteful, and dirty, as well as the person who defecates in oth­ ers' gardens and eats house rats, since he is believed to possess witchcraft, the power to inflict misfortune, wasting disease, and death on others (Evans­ Pritchard 1937).

Notions of danger in American cities, as in other cultural settings, draw on more general social understandings of who and what is dangerous, the kinds of persons who are believed to be violent and immoral, and the charac­ teristics of people who are believed likely to commit crimes. In both simple and complex societies, those who suffer misfortune do not always know exactly where the final responsibility lies. In small-scale societies, the witch or supernatural forces are blamed; whereas in American cities, the faceless crim­ inal receives the blame. Yet, in both, it is the outsider, the stranger, who is held responsible. Such persons are not full members of the observer's social and cultural world, but are people whose behavior appears strange and irrational.

The attitudes of the project residents I talked to vividly illustrate these images of danger. Danger has a variety of meanings for project residents. It means encounters with muggers on deserted streets, invasion by culturally alien neighbors, or the nuisance of disheveled drunks asleep on the sidewalk. Essen­ tially, danger is fear of the stranger, the person who is potentially harmful and whose behavior seems unpredictable and beyond control. Those residents who are the most convinced that their environment is dangerous tend not to be those most victimized, but rather those who lack any social connection to street youths. Such people see themselves awash in a sea of dangerous strangers.

To Chinese residents, the blacks, all of whom seem to look alike and are thought to be robbers whose favorite victims are Chinese, appear most dan­ gerous. Whites seem dangerous because they are members of a dominant

126 Urban Danger

group which has long excluded the Chinese from full membership in its social institutions and has treated them with disdain and indifference. Yet, to blacks, who recognize that only a small proportion of the project youths are actually involved in crime ( only 10 percent of the black families have children who commit crimes), the project appears as a safe place in which they know almost everyone and can anticipate which few youths might be inclined to rob them. As one young black woman said to a young Chinese woman, also a resident of the project:

To you, all the blacks are dangerous because you can't tell them apart, but to me they are George, Johnny, and Jamesy, and I know who to look out for and who will not bother me.

Black adults who endeavor to guide their children into a life of steady jobs and stable family ties find the project youths who pursue a glamorous life of hustling, easy money, and the fast lane dangerous in that they threaten to tempt their children away from their values. Some whites find the blacks dan­ gerous, but those who know the black families and have watched their chil­ dren grow up know which youths are active in crime and hustling and which are not and take comfort in the belief that, because they know them, they will not bother them.

Places that seem dangerous are not those where crimes occur most often, but are those that are unfamiliar or are favorite hangouts of tough-looking youths. Nor are places thought to be safe free from crime. A playground in the center of the project, a favorite hangout for a group of youths blamed for local crime, was generally seen as very dangerous, although few crimes actu­ ally took place there. On the other hand, the area seen most often as safe was the street in front of each person's house. Yet this was also one of the most common places for robberies. Most people said that their side of the project was safe, but they feared to venture to the other, more dangerous side. Those who lived in the center of the project avoided the edges and those .who lived on the periphery regarded the center of the project as a dangerous place to be carefully avoided. The victimization survey revealed no differences in the rate of robberies on any one side of the project. Thus, notions of safe and danger­ ous places do not simply reflect crime rates, but take into account ideas about territory, ethnic hostilities and conflicts, the presence of hostile strangers, familiarity, the availability of allies, and the design of spaces.

Not all crimes are dangerous, nor are all dangerous events crimes. Some crimes that are technically serious are not so perceived by their victims, while others that are not considered serious by the police loom large as dangerous experiences. For example, crimes of violence or threatened violence commit­ ted by strangers seem dangerous even if little or no property is taken, such as unsuccessful robbery attempts or attempted burglaries. These incidents are reported when residents are asked if they have been the victims of crimes but rarely elicit a phone call to the police. On the other hand, assaults by people who know each other are not perceived as crimes, even though they are tech-

Sally Engle Merry 127

nically defined as crimes by the police. Those �h? were assaulted i� vicious, ·nterpersonal battles never mentioned these mc1dents when quened about �eir experiences with victimization. Assaults by sttang�rs, on the other hand, are regarded as crimes and engender fear, not _knife fights betw�en rivals for a woman's affection or punches between neighbors over barking dogs or damaging gossip.

Residents of this high-crime environment respond to the dangers that rround them by constructing mental maps of the kinds of people and loca-m . . .

tions that are dangerous and safe. These maps a�e subJective r:presentat10_ns imposed on the physical realities of spa�e and time, construct10ns of rea�ity that reflect the individual's past expenence and knowledge. They guide movements through the project and channel behavior toward strangers. Yet not all mental maps are equally accurate or helpful. The process of construct­ ing these maps involves drawing distinctions and making generalizations. Maps of areas that are well known are more fmely differentiated, while maps of unfamiliar areas are blank or vague. Those with greater knowledge of the potentially hazardous people around them and o� the pa_rticular uses of the surrounding spaces develop more accurate and differentiated mental maps. Such people also find their environment far less dangerous. .

Chinese residents, for example, were generally unable to tell the black resi­ dents apart, lumping them all into a criminal and predatory population, despite the fact that 90 percent of the families had no connection to crime. They also failed to make fine distinctions between the relative danger of different parts of the development. On the other hand, the blacks were far more sophisticated in drawing distinctions between black residents and were equipped with highly differentiated locational maps of the project. Yet, they lumped all Chinese together into the category of rich restaurant owners, despite the fact that the vast majority were poor cooks and waiters in Chinese restaurants. The use of these unsophisticated mental maps thus exacerbates residents' feelings that they are surrounded by a faceless mass of dangerous strangers.

How do these people cope with their dangerous environment? Some resi­ dents adopt a defensive strategy, turning their homes into fortresses barri­ caded with expensive locks and elaborate window bars, stockpiling guns, learning to live with large guard dogs in small apartments, and calling the police to report every incident. They are always cautious, staying at home at night and avoiding social contact with anyone but close friends and relat�ves. These are the people whose lives are most constricted by the fear of cnme: the elderly, residents who speak only Chinese, and social isolates. Their mo�e of defense is escape and retreat, but, if the fragile shell of safety around th err homes is violated by a crowbar mark on the door or an attempted purse snatching on the porch, the loss of a sense of security is devastating.

Others adopt offensive strategies, developing reputations as dangerous, tough people who are willing to fight back if abused, either by violence, by calling the police, or by going to court. These people are still vulnerable to vic­ timization by outsiders who do not know their reputations or by insiders who

128 Urban Danger

are angry at them, but they do not feel the same sense of helplessness in the face of anonymous dangers. Unlike the defensive residents, who are vulnera­ ble every time they leave their homes, those who adopt offensive strategies carry their protective armor around with them. Thus, the residents of this project range from those who cower in fear in a barred haven of safety to those who traverse the city at any time of day or night with a sense of confidence and ease springing from their mastery of the urban environment and their extensive knowledge of its locations, its residents, and its cultural patterns.


This analysis suggests a more general theory: that the sense of danger is rooted in feelings of uncertainty, helplessness, and vulnerability triggered by encounters with strangers who belong to unfamiliar, hostile, and potentially harmful groups. A stranger is not perceived as a unique individual having a personal his!,ory, reputation, and location in social space. Instead, visible cues such as age, sex, dress, demeanor, ethnicity, location, and mode of speaking are used to place a stranger in a social category associated with certain expected behaviors. Mitchell terms this a categorical relationship, one that arises in situations in which interactions are superficial and perfunctory (1966:52). Such categories codify and order social interaction in otherwise uncertain, unstructured urban situations. These categories are constructed through experience and shared cultural conceptions, but the process of con­ struction is rarely conscious or deliberate; rather, it proceeds through the cre­ ation of implicit categories that feel like instinctive descriptions of the world.

Categorical identities inevitably ignore individual variation and are likely to lump very different individuals together. Because finely honed categories develop through familiarity and contact, socially distant and unfamiliar strangers will be assigned to grosser and less refined categories than those who inhabit more familiar social worlds. The less contact an individual has with members of other groups, the less accurately will she categorize these groups. Entire ethnic or age groups can be lumped into the dangerous, immoral or threatening camp. The dangerous group generally differs in ethnic back­ ground, but suspicions may also arise due to differences in class and lifestyle.

Predictions of behavior based on categorical identities are far less certain than predictions based on knowledge of the particular habits and propensities of a specific person. The stranger's behavior is likely to appear unexpected and unprovoked, leaving the observer with the feeling that there is little she can do to avoid attack. Psychological research suggests that fear comes from the …