David Snow and Leon Anderson show us the wretched face of homelessness in late twentieth-century America in countless cities across the nation. Through hundreds of hours of interviews, participant observation, and random tracking of homeless people through social service agencies in Austin, Texas. Snow and Anderson reveal who the homeless are, how they live, and why they have ended up on the streets. Debunking current stereotypes of the homeless. Down on Their Luck sketches a portrait of men and women who are highly adaptive, resourceful, and pragmatic. Their survival is a tale of human resilience and determination, not one of frailty and disability.

Down on Their Luck: A Study of Homeless Street People

By David Snow and Leon Anderson

University of California Press

Copyright 1993 David Snow and Leon Anderson
All right reserved.
ISBN: 0520079892
Chapter One
Studying the Homeless

It is a mid-January afternoon. A chilling "Blue Norther" has blown down from Canada into central Texas. A raw wind blows in the alleys and near-freezing rain pelts the streets. Not a day to spend outside. No matter, a twenty-five-year-old man in cowboy boots, a grimy denim shirt, and an oil-stained vest is sprawled on a rain-soaked piece of cardboard beneath his rusted-out black '65 Cadillac parked down the street from the Salvation Army. Aside from two short breaks to get warm, he has been lying under the car since breakfast, trying to install a secondhand starter. If he had had the right tools, he might have gotten the starter in by 10:00 A.M. Now time is getting short for making it to the plasma center before it closes. "If I don't get this car running in the next half hour," he mutters, "I'm going to have to walk the whole three miles to the plasma center in this rain."

Up the block from where the young cowboy is struggling with his car, nearly two hundred people have taken refuge from the cold, driving rain in the Salvation Army's drafty, run-down transient lodge. The Sally, as people on the streets call it, is an anomaly in the renovated downtown area, a diminutive tattered structure dwarfed by glittering highrises. Inside the Sally's Big Room, some of Austin's most destitute citizens are waiting out the storm. With 200 people, mostly male, the so-called Big Room is bursting at its seams. At the front it is standingroom only.

By the door, several young men with rumpled clothes and unkempt hair pass a cigarette around and peer through the steamed-up windows.

They discuss the dismal employment scene. "People keep saying this is the workingest city in the country, but you couldn't prove it by me," one of them laments. "I haven't been able to get a day's work since I been here."

"And you can write this week off," replies one of his companions, a middle-aged man with a wandering eye and a week's stubble on his face. "You ain't getting nothing in this weather."

Their conversation is interrupted by shouts from the audience of a game show that is playing on a small black-and-white television perched on a card table. Two Sally workers and several of their friends sit by the TV on the few chairs in the room. Behind them stretches a sea of disheveled bodies, discarded newspapers, cigarette butts, and wet paper bags filled with mildewed clothes. Many lie asleep. An acrid blanket of cigarette smoke hangs about four feet off the floor.

A few pockets of animated conversation punctuate the somber mass. Four men in the middle of the room are playing a lively game of hearts. Periodically their banter is halted when their eldest member, a sweettempered black man who reeks of stale urine, has such a violent coughing fit that they fear he will pass out.

Nearby, a couple of long-haired young men with two sets of earphones connected to one Walkman pretend to play guitars as they sing along loudly to a rock song no one else can hear. Behind them stands a heavily bearded young man with wild hair and a crude tattoo of a twoheaded snake stretching the full length of his right arm. He is throwing karate punches and arguing with an unseen foe. All the while a blond boy who looks to be about twelve years old darts around the room, stumbling over sleepers and leaving a muffled chorus of curses in his wake.

In a back corner of the room, half a dozen men and two women surreptitiously pass around a couple of joints. As the twelve-year-old sails into the back of the room, he sees what they are doing and begins to chant, "People back here are smoking joints!" Suddenly he is pulled to the floor and punched in the side. "Shut up, you little shit!" yells the man who pulled him down. "You're gonna end up dead if you don't shut up now. What the hell're you doing here anyway? Don't you have a mother or something?"

Squirming and whimpering, the kid slips out of the man's grip and shoots off through the crowd. The man shakes his head disgustedly, then gets up and worms his way through the crowd to the restroom. There he finds a long line of men waiting to use the single toilet. "What's

the problem?" he asks, after five minutes during which the line hasn't moved.

"Some guy in there's puking his guts up," a man at the front of the line informs him.

"He better finish pretty soon," the man behind the first says, pounding on the door. "Hurry up in there or I'm gonna go in my pants!" After another minute he steps out of line and heads for the door. "To hell with it," he grumbles. "I'll go outside." He winces as his bare feet hit the wet, cold pavement, and he vows when he goes back in that he'll "find the bastard" who stole his shoes.

As the barefoot man shuffles off to the side of the building, the young cowboy-mechanic sloshes into the Big Room. His clothes are soaked, he's shivering, and his right hand is wrapped in a bloody rag. In his left hand he holds the ratchet set, pliers, and screwdriver he's been using all day. Slowly he weaves his way to the back wall, where several older, crippled men are lying on cots.

"You look like a drowned rat," a little man with a dirty cast on his left ankle tells him. "What happened to your hand?"

The young cowboy drops the tools on the little man's cot. "I sliced it up when the ratchet slipped," he answers wearily. "Still haven't got the starter in. Can I bum a smoke?"

The little man reaches under his blanket, pulls out a package of Bugler tobacco, and rolls them each a cigarette. The man on the next cot watches them enviously. Finally he gets up the nerve to ask, "Can I have a couple of drags?"

"Here, have the rest of it," the little man says, passing him what's left of his cigarette.

The young cowboy slumps down by the cot. His mind is still focused on heading up to the plasma center to sell some blood, but his body is too weary to follow through on it. He's starting to warm up now, but his cut hand throbs and his throat feels scratchy. "Please don't let me catch a cold," he mumbles to himself. "It's for doggone sure I don't need that." He closes his eyes and imagines working a roofing job on a spring day: the smell of hot tar and the sun beating down on his back. Just as he is about to fall into a peaceful dream, the rambunctious blond boy stumbles and falls into his lap. Before the cowboy quite knows what hit him, the kid is up and running through the crowd again.

"No use going to sleep anyway," he muses. "Any time now they'll want us to get in line for dinner." He slouches against the cot, staring blankly at a large, gray-haired woman in a tattered dress. He's been

watching her for several minutes before he realizes with disgust that she is tearing scabs off large sores on her arms. The sight disturbs him so much that he gets up and wanders off to the front of the room. He's about halfway to the front when a commotion erupts by the bathroom. "Call an ambulance!" someone shouts. "I think the guy in the bathroom's had a heart attack!"

One of the Sally workers heads over to the bathroom; another goes to call 911. In a few minutes the Emergency Medical Service workers arrive and put the man on a stretcher. One EMS worker keeps asking whether anyone knows the unconscious man's name, but nobody does. They clear a path to the door and rush him out to the ambulance, which speeds away with its siren blaring in the rain.

The cowboy bites his lip as the ambulance disappears. The room hums with nervous excitement, and the blond kid runs around the room shouting, "That man's gonna die!"

One of the Sally workers tries to calm things down by collaring the boy and putting a hand over his mouth. Another worker yells out, "Okay, let's get in line for dinner!"

Slowly the crowd reorganizes itself into a long line snaking back and forth across the room. A few determined sleepers lie scattered on the floor. The cowboy feels a fever and a headache coming on. "But at least I'm near the head of the line," he tells himself. "At this rate I'll be able to get a mat by the television after dinner." That provides some consolation for his day of disappointment. After all, he knows, it's going to be a long night.

The preceding scene, based on participant observation in Austin in 1985, could have occurred in any of America's urban centers during the 1980s, when they were besieged by increasing numbers of homeless people. The question of just how many Americans were homeless during the 1980s has been hotly debated, with numbers ranging from around 250,000 to over 3,000,000.1 But a host of other questions spring to mind as well. Who are these people? Where do they come from? What are their lives like? How do they manage to survive physically, socially, and psychologically in this netherworld of the streets that is so alien to most Americans? How do they manage to make sense of lives that strike most of us as waking nightmares?

These questions are relevant not only to the homeless who took refuge in Austin's Sally on that blustery winter day in 1985 but to hundreds

of thousands of homeless across the United States. They are also the questions that anchor this book. Our goal is to provide a detailed description and analysis of street life as it was lived by the homeless in Austin, Texas, in the first half of the 1980s. But before turning our attention to Austin, we provide a broader analytic and historical perspective on homelessness. First, we clarify three analytically useful dimensions of homelessness and categorize several distinct varieties. Then we take a brief historical look at street homelessness, the kind of homelessness that we are concerned with in this book.

Varieties and Dimensions of Homelessness

Homelessness in one form or another has existed throughout much of human history.2 Yet even a cursory examination reveals striking differences among homeless individuals and their circumstances. Some people are rendered homeless by mass disasters such as earthquakes, floods, or hurricanes. Others are homeless because the labor they perform forces them into perpetual migration. Some face homelessness when they flee their countries for political or economic reasons. And still others confront homelessness when deep-seated economic and institutional changes push them onto the streets. Each of these groups experience some degree of homelessness, but the nature of that experience often varies considerably among them. These differences can be clarified by considering three separate dimensions of homelessness: a residential dimension; a familial-support dimension; and a role-based dignity and moral-worth dimension.

Homelessness is typically thought of in terms of the first dimension, that is, as a life-style characterized first and foremost by the absence of conventional permanent housing. This residential dimension has been the primary basis for most recent conceptualizations of homelessness.3 The term "literal homelessness," meaning "those who sleep in shelters provided for homeless persons or in places, private or public, not intended as dwellings," captures this dimension well.4 Although this dimension is a necessary starting point for conceptualizing homelessness, it is inadequate for distinguishing among the varieties of individuals who share the obdurate reality of residential dislocation, whether for a few nights or weeks on end.

The second dimension is the presence or absence of familial support. We include this dimension for three reasons. First, the concept of family, with its emphasis on social bonds, networks, and the linkage be-

tween individuals and the larger society, is a basic component of sociological theorizing more generally. Second, the inclusion of some notion of family bondedness is partly consistent with work on earlier generations of homelessness, which focused on the "absence or attenuation of affiliative bonds that link settled persons to a network of interconnected social structures."5 And, third, the family dimension resonates with traditional folk images of the home. The term has traditionally symbolized far more than simply shelter or a roof over one's head. As John Howard Payne wrote in 1823 in "Home Sweet Home":

How sweet 'tis to sit 'neath a fond father's smile,
And the caress of a mother to soothe and beguile!
Let others delight mid new pleasures to roam,
But give me, oh, give me, the pleasures of home!6

The notion of home in American poetry also includes a feeling of shared history with significant others, as in Edgar Guest's memorable "It takes a heap o' livin' in a house t' make it a home,"7 and the sense of unconditional support that was captured in Robert Frost's famous line, "Home is the place where, when you have to go there / They have to take you in."8 This image of the home approximates sociological conceptualizations of the family as a web of mutually affective and supportive relationships. Although recent research indicates that the familial home is often a far cry from the poetic "haven in a heartless world,"9 it undoubtedly still performs this function for many, especially in times of crisis. For many homeless, however, this web of relationships has been weakened or shredded.

The degree to which the homeless lack familial support varies, of course. For some categories of homeless people, such as mass-disaster victims, familial connections generally remain intact.10 For most other categories, including migrant workers, refugees, and street families, familial bonds are often weakened but not totally shattered. The unattached street homeless, though, tend to face their daily lives almost devoid of reliable familial support. Indeed, as we will see in Chapter 8, many of the homeless never had a reliable familial network to begin with, and for others it is only a distant memory, seldom functioning as a source of pleasant reminiscence, much less as a resource that can be tapped in times of crisis. We thus find it useful to incorporate the familial-support dimension as a variable that helps distinguish among different forms of homelessness.

The third distinguishing feature of homelessness is the degree of dig-

nity and moral worth associated with the various categories of homelessness. From a sociological standpoint, to be homeless is, among other things, to be the incumbent of a basic role or master status. These terms refer to statuses that are central to the way we are viewed and the way we view ourselves.11 Basic roles or master statuses are usually highly visible, are relevant to interaction in most situations, and are generally repositories of moral worth and dignity. Dignity and moral worth, then, are not primarily individual characteristics but instead flow from the roles we play.

Incumbents of different social roles can thus vary considerably in terms of perceived moral worth. Some social roles, such as the occupational roles of physicians or attorneys, tend to imbue their incumbents with considerable prestige. Stigmatized roles or statuses, by contrast, confer disrepute on their incumbents. Implicit in most such roles is an imputation of character defect. This imputation varies in degree among the categories of homeless people, based largely on public perception of the extent to which they are responsible for their plight and the threat they are seen as posing to the safety and welfare of other citizens. Massdisaster victims, for instance, are seldom seen as having caused their homeless condition or as threatening the larger community, and therefore they are not stigmatized. The situation is more ambivalent in the case of migrant workers and refugees. In contrast, homeless street people are more likely to be perceived as homeless by virtue of their own imperfections or moral failings.12 Yet even among street people distinctions tend to be drawn. Homeless families and children tend to be treated more sympathetically than homeless street adults. And it is our sense that adult street women are seen as less responsible for their plight than street men, who often tend to be depicted as improvident and lazy individuals who are threats to the property and physical safety of the domiciled.13 This general perception is reflected in the views of most agencies and individuals who interact with the homeless in a caretaker fashion and who tend to treat homeless men as less worthy or deserving than others.14 The point, then, is that a community's "span of sympathy" is not as likely to be extended to unattached homeless men as to homeless families, children, and women.15

As the preceding discussion demonstrates, homelessness can be viewed as a generic category, with homeless street people constituting a subcategory. Our focus in this book is not on homelessness in general but only on unattached, homeless street adults, mostly male, in the United States in the 1980s.16 These homeless lived largely in the public domain,

and their lives were characterized not only by the absence of conventional, permanent housing but also by an absence or attenuation of familial support and of consensually defined roles of social utility and moral worth. The combination of these three deficits is reflected in the lives and experiences of the street homeless in Austin in the mid-1980s. But these defining characteristics are hardly peculiar to the street homeless of this era. We therefore turn to an historical overview of street homelessness in order to place the street homeless of today in perspective.

An Historical Overview of Street Homelessness

Street homelessness has had a long and varied history throughout much of the world. The preindustrial city was characterized in part by "the omnipresence of beggars."17 The beggars were just the tip of the iceberg, however, as the preindustrial city contained large numbers of impoverished and organizationally unattached persons who were referred to collectively as "floating populations."18 It is uncertain just how these people managed to survive. Begging was one common means of livelihood and was sometimes combined with thievery and prostitution, but "much of their time appeared to have been spent hanging about waiting to involve themselves in whatever was happening."19

Although the unattached homeless were a common feature of the preindustrial cities of Europe, two overlapping philosophies mitigated their stigmatization. First, folk traditions emphasized the importance of offering hospitality to needy itinerants.20 Second, during the Middle Ages there was a tendency to idealize poverty. This tendency can be seen in the creed of Saint Francis, who taught "that beggars were holy, and that the holy should live as beggars."21 It was an age of considerable charity toward the destitute.

This spirit was challenged in the fourteenth century. The change in attitude resulted from multiple social forces in medieval society. Religious values that denigrated poverty emerged, in part due to the discrediting of the Franciscan ideal because of the great wealth that the order amassed, and partly because Renaissance humanists valorized worldly activity and success.22 But powerful material forces also underlay this ideological shift. The decimation of the population by the Black Death, which struck England about 1348, prompted the passage of the country's first full-fledged vagrancy statute in 1349.23 Since at that time

England's feudal economy was highly dependent on a ready supply of cheap labor, the first vagrancy statutes were designed expressly to force the dwindling pool of laborers to accept low-wage employment and to keep them from migrating in search of better opportunities. "Every ablebodied person without other means of support was required," according to one observer, "to work for wages fixed at the level preceding the Black Death; it was unlawful to accept more, or to refuse an offer to work, or to flee from one country to another to avoid offers of work or to seek higher wages, or to give alms to able-bodied beggars who refused to work."24 In essence the statutes acted as a substitute for serfdom, curtailing "mobility of workers in such a way that labor would not become a commodity for which the landowners would have to compete."25

A change in the perception of vagrancy and poverty accompanied the passage of these statutes. The sin of sloth, originally conceived as a spiritual vice, was redefined to include physical idleness. Criticism of the homeless in the fourteenth century was minor, however, in comparison to the flood of vigorous attacks unleashed in the Tudor period, beginning about 1485. These signaled a shift from concern with idleness toward a concern with what was perceived as dangerous criminality.26 A few perceptive contemporaries, such as Sir Thomas More, recognized the poverty and exploitation that were forcing so many people into homeless destitution, but most writers of the period railed against a subculture of rogues and vagabonds that they feared would destroy civilized society. Vagrants were portrayed as cut-throat thieves and sorcerers and often as being in league with the devil.

The same sentiments prevailed on the Continent, as is evidenced by Martin Luther's editing in 1528 of Liber Vagatorum, a purported confessional by "Expertus in Truffis" (Expert in Roguery), who revealed the criminal secrets of the vagrant underworld. Luther promoted the publication of the work "in order that men may see and understand how mightily the devil rules in this world."27

At the same time that this ideological shift in the public perception of the homeless took place, the number of homeless in England was growing dramatically. The Enclosure Laws and early industrialization are widely recognized as primary reasons for this growth in homelessness, but other factors operated as well. Between the mid-1500s and the mid-1600s the population of England nearly doubled, and the proportion of adolescents, the demographic group most at risk of homelessness, rose as well.28 In addition to increased competition for resources,

members of the lower classes experienced constant increases in rents and food prices during this hundred-year span, increases that were exacerbated by a drop in real wages.29

The official response to the growth in vagrancy during the Tudor years was frequently brutal. Although attempts were made to succor the local poor, itinerant vagrants were punished harshly. Military manhunts were organized periodically to round up the homeless, and once arrested they were often summarily subjected to the stocks, flogging, and sometimes even hanging. Branding was common, as was ear boring, which was introduced in a 1572 statute that required all vagabonds to be "grievously whipped and burned through the gristle of the right ear with a hot iron an inch in diameter."30 Imprisonment of vagrants was common, and they were often confined in the bridewell, an early British version of the workhouse. Conscription into the military was a frequent alternative to harsher options, with vagrants comprising a major part of the period's armies.31 The Slavery Act of 1547 placed convicted vagrants in slavery for two years, and the Vagrancy Act of 1597 permitted a sentence of transportation to the colonies.

During the seventeenth and eighteenth centuries, many of Britain's homeless, then, were sent to the American colonies as laborers and servants. Still others with few skills or possessions emigrated in the hope of establishing a better life. Life in the colonies, however, was frequently brutal and exhausting, in part because a constant shortage of workers led employers to overwork those they did employ.32

In colonial America, two separate systemsboth of which had their origins in the Elizabethan Poor Laws in Englandexisted for dealing with the problem of the poor. The first system was directed toward legally recognized members of the community. Although it was by no means extravagant, some "outdoor relief" was provided to community members who were unable to support themselves. Maximum efforts were made to keep families together, and the workhouse, so common in England, was seldom a preferred method for dealing with the communities' poor during the colonial period.33 The principle of requiring a pauper to show legal residency in order to receive support, however, had been imported from England, and this created great difficulties for new immigrants and itinerant workers. Colonial America's agrarian economy in many ways encouraged mobile labor, but its system of support was biased against such laborers.

The poor were allowed to petition communities for settlement rights, but they were often denied admittance as bad risks. "There thus arose

a kind of transient poor, shunted from community to community because in place after place they were denied settlement rights."34 The major seaboard cities of the colonial era experienced a particularly heavy influx of nonlocal poor, especially immigrants and sailors. These cities were among the first communities to develop workhouses, shelters, and soup kitchens.35

During the preCivil War years the American economy suffered several minor economic depressions that swelled the ranks of the homeless, but it was after the Civil War that homelessness rose most dramatically.36 The war itself displaced tremendous numbers of people. The country also experienced heavy waves of immigration during the later decades of the nineteenth century. And the industrial and agricultural state of the nation demanded large numbers of mobile workers. As one account of this era noted: "The drive of American industry westward opened new kinds of jobsat the railroad construction sites, in the mines, in the timberlands, on the sheep and cattle ranches, in the orchards. The call was for a special kind of labor, a labor remote from family and community life."37 The development of the railroad system across the country provided jobs as well as the cheap transportation necessary for those willing to answer this call for a large, itinerant work force. The confluence of these factors gave birth in the late 1800s and early 1900s to the homeless life-style of the American hobo.

During this period hoboes both played a central role in American labor and became a prominent feature of the urban landscape. For all their work in rural areas, the hoboes were, as one observer phrased it, "urbancentric."38 The reasons were twofold. First, from the 1870s to the 1920s, many major cities contained a district known as the Main Stem (or Hobohemia), to which these men returned when their jobs were finished in the hinterlands or to wait out the winter months. With its lodging houses, pawnshops, saloons, dance halls, and inexpensive restaurants offering "coffee an's" (coffee and donuts or biscuits) for a nickel, the Main Stem provided for all the needs of these migrant workers at reasonable prices. Second, and perhaps more important, the Main Stem contained numerous employment agencies where the men might find new work. Here they could sign on for jobs at remote railroad construction and logging sites and have their transportation paid for. Through these agencies "battalions of workers marched to the pulse of on-again, off-again employment."39

In major cities, especially those on primary rail routes, the Main Stem contained a substantial portion of the city's population and businesses.

In San Francisco, the city directory listed one-third of the city's restaurants as located in the Main Stem area.40 And according to Nels Anderson's classic sociological study, The Hobo, Chicago's Main Stem in the early 1920s was populated by "30,000 people in good times and 75,000 in hard times," with roughly half a million transient workers passing through in the course of a normal year.41

The Main Stem was politically and intellectually active. Several unions, most notably the International Workers of the World (the I.W.W., popularly known as the Wobblies), worked constantly to recruit hoboes. Chicago's Bughouse Square was a gathering spot for hobo intellectuals, vagabond poets, and revolutionaries. The bookstores of the area sold a wide array of radical literature, including some thirteen publications of the I.W.W.42

The hoboes' economic fortunes were better than those of future homeless generations, but their lives, too, had a dark side. The work was unsteady, and they were often subjected to exploitation at the hands of employment agencies and employers. Furthermore, riding the rails was a dangerous business, as over 2,500 men were killed and another 2,600 were injured on the railroads in 1919 alone.43

The hoboes occupied an ambivalent place in American culture. On the one hand, they were romanticized as "frustrated Western pioneers with too few lands to conquer."44 On the other, they were vilified as scoundrels. They were so despised in some quarters that in 1877 the Chicago Tribune advised its readers, admittedly tongue-in-cheek, that "the simplest plan, probably, where one is not a member of the humane society, is to put a little strychnine or arsenic in the meat and other supplies furnished the tramp. This produces death within a comparatively short period of time, is a warning to other tramps to keep out of the neighborhood, keeps the coroner in good humor, and saves one's chickens and other portable property from constant destruction."45

In the mid-1920s the hobo era died, from a combination of causes. Mechanization of agriculture had depleted the job market for seasonal farm workers. The western frontier had largely been settled. The American economy became strong enough to support large numbers of Americans in more stable jobs. Finally, the railroad, the matrix of the hobo life, had gradually been replaced by automobiles as the major mode of transportation, and those remaining switched from steam to diesel locomotion, a change that made it more difficult to ride the rails.46

With the coming of the Great Depression, however, the numbers of homeless people again rose quickly.47 Testifying before the U.S. Senate

in 1934, Nels Anderson estimated that the country contained at least 1.5 million homeless persons.48 The situation of the homeless had changed dramatically, however, as the Main Stem, home of self-supporting hoboes, was transformed into skid row, where men primarily dependent on charity lived.49 And inlike the more solidaristic hoboes, transients during the Depression years engaged in little collective organizing. The transition in the skid-row population and its situation is captured in one observer's lament that it was "pathetic to see beggars where rebels once shouted, sang, and whored."50

The Depression also witnessed a rise in the number of homeless families and single children on the road. John Steinbeck's The Grapes of Wrath realistically portrays the dislocation of the Great Plains refugees who lost their farms and wandered west in search of a new life. Roadside Hoovervilles, named after President Hoover, who had been notoriously unsympathetic to the plight of the poor, sprang up around the country. Police in many communities used their jails as overnight shelters. And in 1932 the U.S. Children's Bureau conducted a study that documented over 200,000 homeless children.51

Although in 1933 the federal government began efforts to alleviate the plight of the homeless, the programs were usually modest and often contradictory. The Federal Transient Bureau agreed to pay cities for the shelter and meal costs on nonlocals at established tent camps along the road. But in 1935 the Works Progress Administration took over these functions, resulting in "a significant curtailing of the federal effort on behalf of transient workers" because it strongly supported stiff residency requirements.52 Furthermore, the transient homeless were kept on the move by local officials who did not want to assume responsibility for their support. In New York state, for example, the money spent on transporting transients out of communitiesf sometimes equaled the amount that was spent on supportive services for the local homeless.53

By 1936 many of the Depression's homeless people had been reabsorbed into the work force, but it was only with the onset of the war effort in the work effort in the late 1930s that the nation's skid-row population diminished appreciably, as the homeless were recruited into the armed services and into war industry. During World War II the skid-row population almost disappeared, reaching a low in 1944, when the city shelters in New York reported an average of only 550 lodgers per day, compared with 19,000 in 1935.54

After World War II the federal government provided benefits for veterans in an efforts to assist them in the transition back to civilian life and

employment. This was the first time in American history that the end of a war did not substantially increase the homeless population.55

Skid-row populations grew only modestly in the postwar years, and the demographic composition of skid row changed. The district no longer drew a young and mobile labor force, as it had in the heyday of the Main Stem or even in the Depression. Rather, the cheap food, hotels, pawnshops, soup kitchens, and missions of the Row attracted an older, often disabled population of down-and-outers, averaging over fifty years of age.56 Some had small railroad pensions, military disability checks, or Social Security income, but the majority scraped by on a meager income from intermittent unskilled labor, frequently supplemented by the sale of blood to commercial blood banks. In 1958, for instance, the median income of the skid-row men in Minneapolis was eighty dollars per month.57 Most skid-row residents had regular or at least semiregular private accommodations in cheap, single-residence occupancy (SRO) hotels, but at any given time about 10 percent of the population of the Row was either sleeping outside or in free missions and shelters.58 The skid-row men of this period also exhibited a high degree of residential stability, in striking contrast to the hoboes of the Main Stem and the Great Depression's job-seekers.59

The postwar skid row was indeed a smaller, older, and less economically productive version of the once dynamic Main Stem. However, the differences between the Main Stem and skid row were highly exaggerated in the minds of the general public and of the academic community, both of which took a far more pejorative stance toward skid row and its habitues. Skid row was perceived primarily as the part of town that catered to down-and-out drunks. Newspaper cartoons of the period frequently portrayed the men of the Row as social misfits and alcoholic degenerates.60 Academic researchers supported this perception by focusing the bulk of skid-row research on alcohol problems, despite the fact that only a minority of the population there was actually alcoholic.61

In addition to providing inexpensive services for unattached older men at the bottom of the social order, skid row during this period served a broader social function that was consistent with the negative public and academic stereotype of skid-row men. As a geographically distinct district, skid row spatially segregated the stigmatized down-and-outers from the rest of the community. Police tended to keep a watchful eye on the district to prevent spillover into middle-class areas.

By the early 1960s, the populations of America's skid rows had, for

several reasons, declined even more dramatically.62 First, the number of transient workers cycling through the skid rows had continued to decline as the Row no longer functioned as a major labor exchange.63 Second, the availability of more generous welfare benefits and other entitlements enabled many who would otherwise have been dependent on skid row's cheap housing and services instead to live in other city neighborhoods. Many welfare agencies encouraged their clients to locate elsewhere in the belief that they would thus be saved from the negative influence of the skid-row subculture. Additionally, skid rows around the country fell prey to gentrification and urban renewal. The decline of skid-row populations, therefore, also reflected the demolition of many of the cheap hotels and other services on which the residents had relied.64

This decline in the skid-row population around the country led some observers to forecast the eventual disappearance of the skid-row homeless in the United States. Others suggested a more cautious interpretation, arguing that "this decline [was] not due to a decrease in the absolute size of the [homeless] population" but, rather, to its dispersal in American cities.65 But even those who cautioned against the assertion that the number of homeless was declining did not anticipate the surge in homelessness that was to occur in the early 1980s.66

Perhaps as startling as the sudden increase in the number of homeless in the 1980s was the growing recognition that the characteristics of the population were shifting as well. One change was in the age composition of the homeless. In stark contrast to the skid-row men's average age of over fifty, the homeless of the 1980s tended more often to be in the earlier years of adulthood, with a mean age somewhere in the midthirties.67 They were also more ethnically diverse. Although researchers found considerable variation from one community and region to another, most found among the homeless a greater proportion of ethnic minorities than had existed on skid row.68 The proportion of women had also increased, as had that of families.69

The differences between the homelessness of the skid-row era and that of the 1980s extended beyond demographics. Most significantly, it included a shift in the public perception of the problem of homelessness. Urban renewal and the gentrification of skid rows around the country had destroyed the urban niche in which many of the homeless of the previous period had existed. As a result, the homeless of the 1980s were more visible and faced more frequent contact with domiciled citizens than had their earlier counterparts. Because of this increased visibility

and contact, coupled with the dramatic growth during the 1980s in the sheer numbers of homeless, homelessness generated more public interest and debate during the decade than did almost any other domestic issue. One result was a plethora of research reports on the topic.

Indeed, it is probably not too farfetched to assert that academics or researchers in the social-service sector have generated reports of one kind or another on the homeless in almost every community throughout the country. In a monograph reviewing many of these studies, for example, the author noted that his "working bibliography on the homeless exceeds sixty single-spaced pages of entries, of which three-quarters are from 1980 or later."70 The vast majority of the more recent studies that contain primary data share two fundamental characteristics. Most are based on questionnaire surveys of the homeless or of shelter providers, and most are concerned primarily with the demographics and disabilities of the homeless.

Evidence of the first tendency can be readily gleaned from the research literature, but it is also clearly indicated by a General Accounting Office report (1985) summarizing research on the homeless. This overview identified one hundred and thirty studies or reports, seventy-five of which contained primary data. Thirty of the seventy-five were based on street and shelter surveys of the homeless; the remaining forty-five reported data retrieved from shelter providers. There was nothing resembling a case study or ethnography among these studies. Clearly, there has been a proliferation of research on the homeless since 1985, but with the exception of a few studies such as Jonathan Kozol's Rachel and Her Children and Irene Glasser's More Than Bread, most of this recent research has been of the survey variety.71

The tendency to focus on the demographics and disabilities of today's homeless is also evident in both popular and social science literature. Indeed, it is difficult to find current publications on the homeless that go much beyond enumeration of their demographic characteristics and presumed disabilities, such as mental illness, alcoholism, and poor health.72 These focal concerns are congruent both with the interests and agendas of funding agencies (such as the National Institute of Mental Health, the National Institutes of Health, and the Robert Wood Johnson Foundation) and with the kinds of data best procured by questionnaire surveys. The resultant findings certainly advance understanding of the demographics and some of the disabilities of the homeless, and they are of some utility to service agencies interested in doing something about or for the homeless, but they tell us little about actual life on the

streets. Thus, it seems reasonable to assert that such survey studies of today's homeless have tended to deflect attention from questions and issues pertinent to an understanding of the nature and texture of street life, and particularly from the perspective of the homeless themselves. They have generated, in other words, what Clifford Geertz has called "experience distant" rather than "experience near" constructions and understandings.73

Since it was the latter type of understanding we were most interested in securing, and since we were particularly interested in learning about the survival routines of the homeless and how these routines vary among them, we thought an ethnographic field study would be the appropriate research strategy. It was with this in mind that we began our field research of homeless street people in Austin, Texas, in September of 1984.

Local Context and Procedures

Nestled on the eastern edge of Texas's verdant hill country, Austin has long been perceived by both natives and visitors as the state's garden spot. As the state capital, the home of the main campus of the University of Texas, and the seat of progressive country music, Austin has been regarded as a cultural oasis as well. In the decade between 1975 and 1985, it was also one of the sunbelt's shining boomtowns, nearly doubling its population, from around 250,000 to close to 450,000. But its booming economy and growing population did not shelter it from the scourge of homelessness. On the contrary, its very prosperity and growth seemed to give rise to a number of local homeless and to attract even larger numbers from other states and cities. The fact that it is located on the interstate between Dallas and San Antonio and is connected by several highways to Houston, two and a half hours to the east, placed Austin at the intersection of the flow of homeless between three of the country's ten largest cities. For these and other reasons that will be discussed later, the ranks of Austin's homeless population swelled during the first half of the 1980s.

Estimates of the size of Austin's daily homeless population during the mid-1980s ranged from a low of around 650 to a high of 4,000, with perhaps the most reasonable estimate being 1,000 to 1,300.74 Although this and other estimates of the size of Austin's homeless population are open to debate, there is no mistaking the dramatic growth in that population during the first half of the 1980s. This growth is clearly demonstrated by the increase in services provided to the homeless by

the Salvation Army, by Caritas (a local charity agency), and by the Texas Employment Commission (TEC). The Salvation Army, for example, served 4,928 people in 1979 and 11,271 in 1984an increase of 128 percent. That the vast majority of these individuals were indeed destitute and undomiciled is suggested further by the quantum jump in lodgings and meals provided by the Salvation Army during the same time period, from 16,863 to 156,451, an increase of 828 percent. Caritas experienced an even more phenomenal increase of 1,602 percent in the services it provided during this period, and TEC witnessed a 72 percent increase in day-labor applicants between 1982 and 1984. Even though there is some duplication in services, both within and across agencies, the rate of increase in the number of homeless served by each agency is so robust that the conclusion is unmistakable: whatever the exact number of homeless living in or passing through Austin in the mid-1980s, the city had experienced a remarkable leap in the size of that population since the late 1970s.

This pool of homeless street people was the focus of the case study we conducted between 1984 and 1986. Our primary research interest was not with their demographics and disabilities but with three other considerations: the repertoire of material, interpersonal, and psychological strategies and routines the homeless fashion or appropriate to facilitate their survival; the variation in the use of these strategies among the homeless; and the array of factors that shape these survival strategies and routines. In short, we were most interested in the subculture of street life as it manifested itself in the lives of the homeless.

To respond to these interests we had to direct our research to meet four basic criteria. First, it had to be appreciative of the institutional contexts in which the routines and experiences of the homeless are embedded. Second, the research had to attend to the perspectives and voices, not just of the homeless, but also of other groups whose actions affect the daily lives of the homeless. Third, it had to use a variety of procedures in order to tap a range of data sources. And, fourth, the research had to be longitudinal, capturing events and happenings as they unfolded over time. In addition, we wanted the research to permit the discovery of unanticipated findings and unexpected data sources.

Contextualization

Concern with contextualization is predicated on the contention that social actions and events can be adequately understood only in relation

to the social contexts in which they are embedded. In practice, a contextual orientation manifests itself in a persistent commitment to understanding how actions or processes are produced and reproduced or changed by their interaction with other elements within a particular sociohistorical context.

In our research we were interested not only in the repertoire of survival strategies fashioned or appropriated by the homeless and in their corresponding daily routines but also in the factors that shaped these strategies and routines. Repertoires of survival strategies do not emerge willy-nilly. They are the product of the interplay between the resourcefulness and ingenuity of the homeless and local organizational, political, and ecological constraints. An understanding of the experience of homelessness and how it is managed thus requires consideration of the local matrix of social service and control agencies and commercial establishments that deal directly with the homeless. Accordingly, in addition to spending over four hundred hours in the field with homeless individuals, we spent another two hundred hours with agency personnel, police officers, local political officials, and neighborhood activists. This aspect of the field research was facilitated by ongoing examination of relevant agency reports, news releases, and articles and editorials in the Austin American-Statesman, the city's daily newspaper. These considerations take us to the second feature of the research.

Multiperspectival Analysis

To contextualize social activities, issues, and processes involves more than providing a descriptive overview of the encompassing context. It also requires consideration of the voices and experiences of the range of actors of focal concern, of the perspectives and actions of other relevant groups of actors, and of the interaction among all of them.

We attempted, therefore, to identify and map the social settings and organizations and the types of homeless that together constituted the subculture of street life in Austin. Relevant social settings and organizations included the major institutions or agencies (e.g., Salvation Army, the city hospital, the city police department), commercial establishments (e.g., plasma centers, labor pools, bars), and spatial or territorial niches (e.g., campsites, bridges, parks, street corners) that were central to the daily rounds, life-style, and prospects of the homeless. Social types, in contrast, connote characteristic ways of thinking and acting among individuals within a given context.75

Since we were interested in documenting the diversity of relevant settings and types of homeless, we employed a type of nonprobability, judgmental sampling technique called maximum variation sampling.76 The mode of procedure is almost Darwinesque: it is to sample as widely as possible within the specified sociocultural (ecological) context until exhaustion or redundancy is reached with respect to types of adaptation or response. We thus spent time with as many homeless as possible in the settings most relevant to their daily lives in Austin. In total, over six hundred waking hours were spent with 168 homeless men and women and with other individuals dealing with them in one capacity or another in the twenty-five street settings and organizations shown in Table 1.1.

Here it is important to emphasize that our interest in street life as it was experienced, whether from the differing standpoints of the homeless or of the agencies that dealt with them, meant that we were primarily interested in "perspectives in action" in contrast to "perspectives of action."77 Perspectives in action are accounts or patterns of talk formulated for a particular end in a naturally occurring situation that is part of some ongoing system of action, such as when a homeless street person panhandles a passerby. Perspectives of action, by contrast, are constructed in response to the queries of researchers or other outsiders, as when a street person tells a researcher about panhandling. Perspectives of action are thus produced "not to act meaningfully in the system being described, but rather to make the system meaningful to an outsider."78 Both perspectives yield useful information, but they are of different orders. Perspectives of action are ex post facto accounts that place the action in question within a larger normative framework; perspectives in action contain the cognitions and feelings that are inseparable from the sequences of action themselves. The more interested researchers are in lived experience and the management of everyday routines, the more critical to their project is the elicitation of perspectives in action.

We attempted to elicit such perspectives primarily by two means:interviewing by comment, and listening unobtrusively to conversations among the homeless that arose naturally rather than in response to the researcher's intervention.79 This listening took two basic forms: eavesdropping, which involved listening to others in a bounded interactional encounter without being part of that encounter, as could be done when waiting in meal lines or in day-labor offices; and a kind of nondirective, conversational listening that occurred when we were engaged in encounters with one or more homeless individuals. The elicitation of perspectives in action through these means enabled us, we believe, to gain

TABLE 1.1 FOCAL SETTINGS 1+ ABC Plasma Center 2+ Angels House (Austin's only soup kitchen) 3 Austin Police Department 4 Brackenridge Hospital (city hospital) 5 Bunkhaus ( a men's dormitory) 6+ Caritas (private, nonprofit welfare agency) 7 Central Assembly of God Church 8 City Planning and Zoning Commission meetings 9+ "The Drag" (street constituting western border of University of Texas campus) 10+ Labor Corner 11+ Labor Pool 12 Legal Aid Society 13+ The Longue (demolished in spring of 1985) 14 Mental Health and Mental Retardation (MHMR; state and local agencies) 15 Oak Springs De-Tox Center 16 Salvation Army Alcohol Center 17+ Salvation Army Shelter (Sally) 18 Stratford House (private alcoholism center) 19 Task Force on the Homeless (City of Austin) 20 Texaco Truck Stop 21+ Texas Employment Commission 22+ Town Lake Parks and Bridges 23+ Whataburger (on the Drag) 24 Winter Shelter 25 Wright Road Farm (for recovering alcoholics) + Plus signs indicate major stopping and/or hanging-out points that comprise the daily round of the homeless in Austin. Italics indicate agencies whose personnel were interviewed regarding the homeless and agency services.

a reasonable approximation of a multiperspectival understanding of street life as it was actually lived by the homeless.

Triangualted Research

The third feature of our research is that it was heavily triangulated. Triangulation has traditionally been associated with the use of multiple

methods in the study of the same phenomenon,80 but it can also occur with respect to data, investigators, and theories.81 Broadly conceived, triangulation entails the use of multiple data sources, methods, investigators, and theoretical perspectives in the study of some empirical phenomenon.

The logic underlying triangulation is rooted in the complexity of social reality and the limitations of every research methodology. The basic argument is that social reality is too multifaceted to be grasped adequately by any single method. Consequently, rather than debate the merits of one more or less flawed method vis-a-vis another, it is better to combine multiple strategies so that they make up for one another's weaknesses.82

With that philosophy, we pursued two basic research strategies. One entailed extensive ethnographic research among the homeless and the settings in which they found themselves; the other was to track a sample of homeless through a set of core institutions.

Ethnographic Strategy We have discussed our interviewing and conversational procedures, but we have said little about how we positioned ourselves in relation to the homeless. The position or role the field researcher claims or is assigned, it has been argued, "is perhaps the single most important determinant of what he (sic ) will be able to learn," for it "largely determines where he can go, what he can do, whom he can interact with, what he can inquire about, what he can see, and what he can be told."83

With that in mind, we positioned ourselves in relation to the homeless in the role of a "buddy-researcher." Although not discussed explicitly by other researchers, it is a role that has been used in a number of studies of street-corner men, hoboes, and tramps.84 In this role, one of us hung out with the homeless on a regular basis, making the daily institutional rounds with them as individuals and in small groups. As a friend, the buddy-researcher provided his companions with minor necessities on occasion, such as small loans that were not expected to be repaid, clothes, rides in an old Toyota, and a sympathetic ear for their hopes, troubles, and fears. The buddy role entailed receiving as well as giving. The homeless shared some of their resources with the researcher, who as a friend was expected to accept such offers.85

In keeping with the buddy role, the researcher tried to avoid distinctive dress by wearing old clothes similar to those worn by most homeless, although his were generally cleaner. He also avoided the use of

academic English.86 It should be emphasized, however, that the researcher did not attempt to pass as a homeless individual. He frequently brought up his researcher status by mentioning his research and university affiliation and by asking questions homeless people were less likely to ask (e.g., personal questions about an individual's past). The research role gave a credible reason for inquiring into such personal matters, while the buddy role generated enough trust and goodwill that the homeless responded.87

The researcher, by virtue of his status as a friend of homeless individuals, could participate in most activities with them. However, although many of his experiences on the streets could approximate those of the homeless, they could never be quite the same, since the option to leave the streets was always available. On the day following a murder at the Salvation Army shelter, for example, the buddy-researcher could participate in the discussion of the event with a group of homeless men while waiting in the dinner line, but he had not had to experience the trauma of the event. Later that evening, when one small group of these men huddled nervously together for protection in the Sally shelter while others, who were afraid to stay at the shelter after the murder, went out on the streets in search of less dangerous sleeping places, the researcher returned to his home for the night. And when the buddy-researcher was arrested with two homeless men one evening, he experienced the arrest and jailing but was bailed out later the same night, whereas the two homeless men did not have that option.88 In sum, although the researcher's role granted him access to a variety of street situations and experiences, the fact that he did not fully disengage himself from his other roles allowed for the ever-present prospect of escape from the streets, thus diluting the direct experience of homelessness.

Rather than using his personal experiences on the streets as the primary data base, then, the buddy-researcher collected data from the homeless themselves. Their behavior and conversations were recorded in a stepwise fashion, beginning with mental and jotted notes in the field and culminating in a detailed field narrative based on elaboration of these notes. These narrative elaborations constitute the ethnographic data log on which much of the book is based.

The data derived from our field observations and encounters were supplemented by taped, in-depth, life-history interviews with six homeless individuals who had been on the streets for lengths of time ranging from two months to fourteen years. Each of these individuals was a key informant, with whom we had numerous contacts.

Although it might be argued that much of the behavior and talk secured during field encounters represented a reaction to the researcher's presence rather than a naturally occurring phenomenon among the homeless, our field experiences suggest that this was not the case. Most of the homeless were apprised of our researcher status, but they typically lost sight of it as we continued to spend time on the streets with them. This forgetfulness was forcefully illustrated one night when one of us gave an ill, homeless woman a ride to a health clinic. On the way back from the clinic to the abandoned warehouse where she was going to spend the night, she asked, "Are you sleeping in your car these days or down at the Sally?" The researcher had explained his situation to this woman many times during the previous two and a half months, but she had forgotten or had not fully believed what he had told her. This should not be surprising, however, in light of the dramaturgical thesis that individuals tend to respond to and identify others more in terms of their proximate roles or actions than in terms of their claims to the contrary.

Tracking Strategy In addition to studying the homeless ethnographically, we tracked a random sample of 767 homeless through a network of core institutions with which they had varying degrees of contact. This data source provided a detailed portrait of the demographics and disabilities of the homeless in Austin, and it also enabled us to learn more about the institutional contacts and experiences of the homeless and to cross-validate information on each case by comparing independent institutional records.89 Implementation of this tracking strategy was contingent on the resolution of four problems: constructing a reasonable sampling frame; securing identifying information on each case; selecting the core institutions; and negotiating access to their institutional records.

The first problem was resolved by using as our sampling frame all homeless adults who had one or more contacts with the local Salvation Army between January 1, 1984, and March 1, 1985. This yielded an unduplicated count of 13,881 homeless persons. The decision to use this population as our sampling frame was based on the assumption that the total number of homeless who had contact with the Salvation Army at any given time during this fourteen-month period comprised a reasonable approximation of the number of homeless who had been in Austin for one or more days during that same time period. Several con-

siderations justified this assumption. Foremost was the fact that the Salvation Army operated Austin's only public shelter for the homeless and was the only facility in Austin that provided free showers, breakfast, and dinner.90 Consequently, we assumed that nearly all homeless men and women in Austin would have had occasion to utilize the Salvation Army at least once.

Our ethnographic research revealed that this assumption was essentially correct. Only two subgroups of homeless were underrepresented: women, and a scant number of street men whose daily routines did not encompass the Salvation Army. Women were underrepresented for two reasons. First, the majority who used the Salvation Army had children and utilized its family services. As a consequence, their records were kept in a separate family file. We chose not to sample from this file because the difficulty of disentangling individual from family data seemed excessive for what comprised less than 9 percent of the individuals who had contact with the Salvation Army in 1984. The other reason for the low proportion of women in the sample is that the majority of childless women had developed means of survival independent of the Salvation Army. The younger ones in particular attached themselves to men with some income or resources that allowed them to stay away from the Salvation Army, a point we will return to later. The other subgroup underrepresented in the sampling frame was the relatively small number of homeless men whose daily routines seldom brought them into contact with the Salvation Army. Since these subgroups, taken together, comprised a comparatively small proportion of the homeless in Austin, and since they were represented in our field sample, we felt justified in using as the sampling frame the 13,881 homeless who had registered at the Salvation Army during the period of our sample.

The problem of securing identifying information to track cases through the other selected institutions was resolved by the Salvation Army's practice of requiring all first-time users to fill out a registration card that asks for name and Social Security number, if any, as well as some demographic and background information. Upon receiving permission from Salvation Army officials to use this identifying information for tracking purposes, we randomly drew 800 cards and then negotiated access to the records of six other local and state agencies. These included Austin's major city hospital, Caritas, the city police department, the Texas Employment Commission, the state department of mental health and mental retardation (MHMR), and the local community mental

TABLE 1.2 NUMBER OF HOMELESS IN TRACKING SAMPLE, BY AGENCY Agency Number of Cases Percentage of Sample Salvation Army 767a 100.0 Texas Employment Commission 348 45.4 Caritas 294 38.3 City police department 248 32.3 City hospital 181 23.6 Texas MHMR b 84 11.0 Austin/Travis Country MHMR b 78 10.2 a Number of usable cases from a sample of 800 drawn randomly from a population of 13,881 individuals who registered at least once at the Salvation Army between January 1,1984, and March 1, 1985. b MHMR=Mental Health and Mental Retardation.

health center. These agencies were selected because of their centrality in the lives of the homeless living in or passing through Austin and because each had amassed data relevant to much that has been hypothesized about today's homeless. We searched each agency's files for some record of contact with our sample cases. When a match was made, all data on that case were coded onto a surveylike instrument that had been constructed on the basis of prior inspection of the record forms of the agencies. Once these data were computerized and cleaned, we were left with a usable tracking sample of 767 cases. Table 1.2 shows the percentage of the total sample matched at each agency.

In sum, our data were derived from three sources: (1) ethnographic encounters with 168 homeless individuals on the streets of Austin over a two-year period; (2) records of seven local and state agencies through which we tracked 767 homeless; and (3) the community itselfthat is, the agencies, governmental units, commercial establishments, and ecological niches associated with the lives of the homeless. These data sources were tapped by a mixture of procedures: participant observation and informal interviews in the case of the homeless; participant and nonparticipant observation, coupled with formal and informal interviewing, in street agencies and settings; and a systematic survey of institutional records. In order to tap the various data sources by these means, the traditional "long ranger" approach characteristic of most urban ethnography, was inappropriate.91 Instead, a team of researchers was

developed, consisting of four individuals with different but complementary and overlapping tasks. Two members of the team were responsible for the tracking data, another assumed the role of primary ethnographer among the homeless, and the fourth functioned in a number of capacitiescoordinating the research, negotiating access to agency records and interviewing agency personnel, and functioning occasionally as a street ethnographer as well as a detached observer or sideline coach in relation to the primary street ethnographer.

Regarding the relationship between the roles of the buddy-researcher and the detached observer, it is important to note that rarely was a day, an evening, or a series of days in the field not followed by a debriefing session that involved discussion of field experiences with their methodological and theoretical implications, and development of plans for subsequent outings. Conscious and reflective enactment of these two roles enabled us simultaneously to maintain involvement and detachment, thereby facilitating management of the insider/outsider dialectic characteristic of much ethnographic research.92

Capturing Social Processes and Change

A fourth major research concern was to capture processes of psychological adjustment, interpersonal relations, and material adaptation as they unfolded over time. To capture such changes requires that the researcher have extended access to the same settings and that the principal actors, both individuals and organizations, be observed at different points in time.

We were able to pursue this objective. Since ethnographic fieldwork involves prolonged and persistent observation within a bounded sociohistorical context, it is ideally suited for sustained contact with the same individuals across time and for grasping changes in behavior and orientation among those individuals. We had repeated contacts with many homeless individuals in a range of settings, in some cases extending for nearly two years. All totaled, we had 492 ethnographic encounters with our field sample of 168 homeless adults, averaging three encounters per person, with a high of twenty-five.

The tracking strategy, which was longitudinal in design, also enabled us to trace the institutional contacts and careers of the homeless across time, in many cases over several years. All totaled, the 767 individuals in the tracking sample had 30,400 contacts with the seven core institutions, over a span of time ranging from one day to twelve years.

The longitudinal character of both the ethnographic and tracking strategies enabled us to observe, not only different patterns of adaptation with variation in length of time on the streets, but also the transition from one pattern to another and what events and experiences were affecting these changes.

Open-Ended Research

The final concern underlying our research was that it not foreclose discovery of unanticipated findings and data sources.93 Indeed, several of our eventual focal concerns were not fully anticipated initially. Upon entering the field, we coded our observations broadly, into twenty-five focal settings and thirty cultural domains.94 In time, however, variation in the number of data entries contained in each respective category showed that some of the settings and domains were more central than others to the daily lives of the homeless.95 Those files that bulged with data became the foci of our analysis. Thus, our discussions of work, social relationships, and meaning and identity do not rest so much on a priori concerns as on what we actually observed and heard as the research progressed.

In addition to the emergence of a number of focal concerns, the possibility of tracking the homeless institutionally surfaced only after we had been in the field for several months. Initially, we thought we would conduct only an ethnographic study. In time, however, the study broadened methodologically, as we became aware of the possibility of tracking the homeless and then took the steps necessary to compile the tracking sample.

The Issue of Generalizability

In the preceding pages we have provided a rationale for the case study we conducted, we have placed that study in historical context, and we have described our underlying research concerns and how we dealt with them methodologically. Whatever the strengths of the resultant research, it is confronted by a vexing issue that haunts every case study. We refer to the issue of generalizability or typicality. To what extent is our portrayal of street life in Austin and the adaptive strategies and routines of the city's homeless representative of homelessness elsewhere, and to what extent is it peculiar to homelessness in Austin? If

we cannot answer this question, then of what utility is the case study we conducted?

Satisfactory answers to questions of the first kind, which are prompted by the positivistic contention that the production of generalizable findings is social science's most basic activity, depend on the degree of comparability among the cases in question. The more similar the relevant community contexts and the more alike the populations being studied, the greater the generalizability from one case to another. Since nearly all of the research on the homeless across the country is based on crosssectional surveys, with relatively little attention paid to the community contexts, it is difficult to assess the extent to which these contexts differ from one case to another. Austin certainly differed from other cities in having only one shelter for the street homeless and in enjoying a booming economy during a portion of the period during which the research was conducted. But, as we see in subsequent chapters, striking similarities also appear.

The issue of demographic comparability is easier to assess, since there is a wealth of demographic data on the homeless across cities. Table 1.3 compares the homeless in the tracking sample with the homeless in other samples compiled during the mid-1980s in a number of southern and western cities. In each case, the majority of the homeless are male, under forty years of age, and single or unattached. Considerable similarity also emerges in terms of education and military experience. However, ethnic composition seems to vary with city size and minority base: the larger the city and its minority base, the greater the proportion of minorities in its homeless population. All in all, though, the table shows that the homeless in the tracking sample are quite similar to the homeless elsewhere, or at least to the homeless in the South and the far West.

This apparent demographic comparability between the homeless in Austin and a number of other cities suggests that some of what we report may hold for the homeless more generally. But we do not want to overstate the case for generalization, especially statistical generalization.96 After all, the virtue of ethnographically based case studies of the kind we have attempted to conduct resides in their dense contextualization, their concern with process and the fluidity of social life, and their attention to the voices and experiences of individuals in their daily lives. Adherence to these values can result in in-depth examination and possible debunking or refinement of existing folk and theoretic presumptions about the life-style, subculture, or people in question. While the findings of such case studies may not be fully generalizable in the

TABLE 1.3 COMPARISON OF HOMELESS SAMPLES IN THE UNITED STATES IN THE MID-1980S Selected Demographic Variables Tracking Samplea (N:767) San Antoniob (N:139) El Pasoc (N:197) Phoneixd (N:195) Los Angelese (N:269) Portlandf (N:131) Nashvilleg (N:117) Birminghamh (N:150) Gender: Male 90% 90% 97% 86% 78% 85% 85% 78% Age: Mean 35 34 33i 37 37 38 35i 40 Under 40 73% 71% 61% 63% 60% 54% Over 60 1% 3% 3% 6% 7% 7% 6% Ethnicity/race: Anglo 75% 60% 64% 49% 77% 79% 69% Black 12% 15% 9% 32% 6% 21% 31% Hispanic 12% 25% 17% 10% 4% Native American 1% 9% 6% 10% Other 1% 3% 3% Marital status Married 8% 14% 3% 13% 6% 12% 14% 7% Unattached (single, divorced) 92% 86% 97% 87% 94% 88% 86% 93%

(table continued on next page)

(table continued from previous page)

Selected Demographic Variables Tracking Samplea (N: 767) San Antoniob (N: 139) El Pasoc (N: 197) Phoenixd (N: 195) Los Angelse (N: 269) Portlandf (N: 131) Nashvilleg (N: 117) Birminghamh (N: 150) Education Completed high school 58% 49% 60% 64% 47% 30% 57% Some postsecondary 18% 18% 34% 38% 24% 21% Military veteran: Yes 49% 46% 36% 46% 42% Current residential status: Local resident 9% 39% 7% 14% 23% 36% 51% From within state 39% 26% 7% 5% 15% From out-of-state 52% 74% 61% 86% 81% 77% 64% 35% a Percentage male is based on the total population of men, women, and children registered at least once at the Salvation Army in 1984. The remaining figures are based on the tracking sample. These figures correspond closely with those of Baumann et al. (1985), based on their August, 1984, street survey of homeless in Austin. They also found a small proportion of women on the streets (only 7 percent), a mean age near 35, 74 percent Anglo, and 91 percent unattached. b Based on a survey conducted in an emergency shelter over a three-week period in February and March, 1984. Residential data based on place of birth (San Antonio Urban Council, 1984). c Based on a survey conducted in an emergency shelter during March and April, 1983 (E1 Paso Task Force on the Homeless, 1984). d Based on a survey conducted in food lines, shelters, and urban camps in March, 1983. Residential data based on response to question concerning place of origin (Brown et al., 1983). e Based on a survey conducted in shelters, in soup lines, on skid row, and in other areas from December, 1983, to May, 1984 (Robertson et al., 1985; Ropers, 1988). f Based on a survey conducted in 1983 (Caulk, 1983). g Based on survey enumeration data gathered in 1986 and 1987 (Lee, 1989). h Based on survey and enumeration data collected in February, 1987 (LaGory et al., 1989). The 31 percent under ethnicity/race was in the original called simply "nonwhite." i Median.

conventional sense, they can identify presumptions and generalizations that do not fit the case and can thereby alter existing understandings.97

This leads to a second, and perhaps more important, function of the ethnographic case study: its up-close, naturalistic focus can put its readers in touch with the lives of others and thereby reduce the distance between "us" and "them."98 The generation of such an understanding is much too valuable for social scientists to ignore, especially when considering marginal populations and subcultures, such as the homeless, that are frequent objects of stigmatization and dehumanization. We thus leave the issue of the generalizability of our findings to the reader to ponder, but we do hope that the subsequent chapters reduce the distance between the homeless, as we came to know them, and those who take hearth and home for granted. If an ethnographic case study can enlarge the span of sympathy and universe of discourse of its readers while simultaneously prompting reconsideration or extension of existing theoretic presumptions, then it must be judged a worthwhile endeavor. It is our hope that this one does a bit of both.

A Note on Tense, Voice, and Language

Although it is customary for academic texts to be written in the past tense, this convention has not been followed strictly by ethnographers. Instead, much ethnographic writing is couched in the present tense, and for good reason. The present tense creates a greater sense of immediacy and realism, it captures more accurately the character of events as they were observed and experienced by the researcher, and it more accurately preserves the voices of the informants. Because of these considerations, it seemed to us that the present tense would often serve our interests better than the academic past tense. Accordingly, we have written much of the text in the present tense. However, we have not ignored recent criticisms of the tendency of the ethnographic present to freeze its subjects and their activities in time and space.99 Our attention to this issue should become clear in a number of subsequent chapters, particularly the final one, and in the epilogue.

A related critique of traditional ethnography, as well as of more traditional research procedures, argues that researchers should try to examine their objects of research from as many perspectives as possible.100 Others have made a similar call recently, suggesting that ethnographic research and writing should be "dialogic" and "poly-

phonic" and that the voices of the array of actors pertinent to a particular social world should be privileged over that of the researcher.101

What is being called for at least in part, it seems, is the kind of multiperpectivl strategy that we pursed by triangulating our research in terms of informants, situations, and researchers. We have thus endeavored to be polyphonic in the sense of including the voices of a number of different sets of actors relevant to street lifein Austin. But we have also explicitly and consistently sought to put into the foreground the voices of the homeless themselves, since it is their adpative behaviors and experiences that are the central concerns of this book. Of course, not all of their voices are equally privileged. Some are featured more than others for reaons that will be elaborated in the next chapter, but those privileged voices are, we believe, representative of the cross-section of homeless we came to know.

We do not presume, however, that we have privileged their voices in an uncontaminated fashion. It is our book, not our homeless informants'. We are the choreographers or narrators, so to speak. We recognize that our discussion provides second-order interpretations of the homeless people's own interpretations of their experiences.102 But we do believe that our descriptions andinterpretations are reasonable approximations of what we were privy to, in that they are restrained both empirically and methodologically. We did not create the settings, the characters, the behaviors, or the life histories and experiences that constitute the empirical basis for the book; we tried to let the voices of the homeless speak loudly and clearly throughout; we refrained from editing their talk or excising their often colorful and profane language; we characterized them as individuals we and others knew, that is, some as sweet and gentle, others as cussed and rough; and we attempted to keep in the forefront of our consciousness the often dehumanizing consequences of the social scientist's antiseptic gaze. These empirical and procedural restriants notwithstanding, most of what we observed and heard was filtered through the sociological eyeglasses we brought to the field. This is neither new nor particularly damning, for it could hardly be any other way. Ethnography is and always has been an interpretive enterprise involving the mediation of frames of meaning between the world of those studied and that of some imagined audience.103 We can only hope, then, that the subsequent chapters prove to be effective mediations in the sense that they are not overladen with distortions and that they form a meaningful bridge between the world of the homeless and that of our readers.

Continues...

Excerpted from Down on Their Luck: A Study of Homeless Street People by David Snow and Leon Anderson Copyright 1993 by David Snow and Leon Anderson. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.