Dying by the Rules: Rational Accumulation, Practical Reasoning, and Premature Death in Street Outreach

 Dying by the Rules: 

Rational Accumulation, Practical Reasoning, and Premature Death in Street Outreach

Wayne Martin Mellinger, Ph.D.

Working Paper — Santa Barbara, California

Abstract

This essay develops an ethnomethodologically-grounded account of premature death among people experiencing unsheltered homelessness in Santa Barbara, California. Drawing on more than two decades of street outreach practice, I argue that early death in what I call the deathworlds of Santa Barbara is not produced by institutional failure but by institutional competence—by the normal, accountable, locally rational operation of multiple organizations whose practical reasoning is incommensurable across institutional boundaries. I introduce the concept of the paradox of rational accumulation to name the structural condition in which locally defensible decisions accumulate into globally lethal outcomes, and I develop a set of ethnomethodologically-grounded mechanisms through which this paradox is practically accomplished: textually-mediated exclusion, the documentary constitution of the partial person, epistemic abandonment, and the misalignment of administrative and biological time. The analysis draws on Harold Garfinkel, Dorothy Smith, Aaron Cicourel, David Sudnow, Michael Lipsky, Alfred Schutz, and Achille Mbembe, among others, and is anchored in the composite case of Rafael—a figure drawn from repeated outreach encounters whose death was produced, step by step, by a system that functioned exactly as designed.

I. The Thesis: Two Arguments, One Mechanism

Every year in Santa Barbara, people die outdoors who did not have to die. They were known to outreach workers. Their names were in databases. Shelter beds existed—not enough, but some. Medical care was available, episodically. Referrals were written. Phone calls were made. Services were documented. And still, people died—on creek beds, under bridges, in doorways, in the parking structures of a city that ranks among the wealthiest in the United States.

The numbers tell part of the story. In 2017, Santa Barbara County recorded 44 deaths of people experiencing homelessness. By 2019–2020, that figure had jumped to 143 over a two-year period—nearly double the rate of just two years prior—even as overall homelessness counts fluctuated only modestly. A rise of this magnitude, absent a comparable rise in the homeless population itself, is not a humanitarian anomaly. It is an institutional outcome. These are not statistical abstractions. They are people whose names outreach workers knew, whose faces they recognized, whose deterioration they witnessed across months and years of regular contact. The question this essay pursues is not why the system failed them. It is something harder: how the system's normal, competent operation produces premature death as one of its regular outputs.

That question requires holding two arguments together simultaneously, because neither is sufficient on its own.

The first argument is structural. Modern institutional systems designed to address complex human needs—homelessness, addiction, chronic illness, poverty—are composed of multiple organizations, each operating with its own logic, its own eligibility criteria, its own temporal rhythms, its own accountability structures. These organizations are not designed to fail. Each is designed, within its mandate, to do specific and bounded things rationally and efficiently. The problem is that their rationalities do not add up. Decisions that are locally defensible—medically appropriate, procedurally correct, fiscally responsible, legally authorized—accumulate into outcomes that are globally irrational and, for the most vulnerable people in the system, fatal. I call this the paradox of rational accumulation.

The second argument is interactional. To understand how this structural paradox is actually produced—encounter by encounter, document by document, decision by decision—we need a fine-grained account of the mechanisms through which institutional actors accomplish their work. That account is ethnomethodological. Following Harold Garfinkel, I argue that institutional harm is generated through practical reasoning: the situated, contingent, context-sensitive cognitive labor through which members of institutions make their actions recognizable, accountable, and defensible within their local settings. Practical reasoning is not irrational. It is not careless. It is the normal operation of competent institutional life. And in the context of homeless services, it is a mechanism of death.

The thesis that unifies these two arguments is this: the paradox of rational accumulation is not an abstract structural condition—it is practically accomplished, in real time, through the ordinary methods of institutional work. Max Weber identified the structural tendency; ethnomethodology reveals the interactional machinery. The deathworlds of Santa Barbara are produced at their intersection.

This essay develops that thesis through theoretical analysis, the composite ethnographic case of Rafael, and a set of analytically distinct mechanisms: practical reasoning across incommensurable vocabularies, textually-mediated exclusion, the documentary constitution of the partial person, epistemic abandonment, the misalignment of administrative and biological time, and the moral condition I call distributed innocence. Each concept names a specific and identifiable way in which locally rational practice becomes globally lethal. Together they constitute a critical interactionist account of how deathworlds are produced and sustained.

II. The Theoretical Architecture

Three bodies of theory converge in this analysis. They are not merely complementary—each does analytical work the others cannot, and together they constitute the theoretical architecture the argument requires. A fourth theoretical tradition—the sociology of death and dying—is brought in at key moments to extend the analysis into territory that the foundational frameworks do not directly address.

Weber and Rationalization.  

Max Weber's analysis of rationalization remains the indispensable structural starting point. Modern institutions, Weber argued, organize themselves around four interrelated values: efficiency, predictability, calculability, and control. In Economy and Society, he traced how these values progressively colonized every domain of social life—law, religion, music, bureaucracy, capitalism—producing organizations of unprecedented power and unprecedented impersonality. His concept of the iron cage (or, in a more precise translation of Stahlhartes Gehäuse, the "shell as hard as steel") named the existential condition that results: a social world organized by instrumental rationality at the cost of the substantive human values that rationality was originally meant to serve.

"The Puritan wanted to work in a calling; we are forced to do so. For when asceticism was carried out of monastic cells into everyday life, and began to dominate worldly morality, it did its part in building the tremendous cosmos of the modern economic order. This order is now bound to the technical and economic conditions of machine production which today determine the lives of all the individuals who are born into this mechanism, not only those directly concerned with economic acquisition, with irresistible force." — Max Weber, The Protestant Ethic and the Spirit of Capitalism (1905/2002:123)

The paradox of rational accumulation is the street-level phenomenology of Weber's iron cage: not a metaphor, but a description of what it feels like to encounter harm produced by rational forces, harm for which no individual is responsible and against which no individual complaint can gain purchase. Rafael is not crushed by tyranny. He is processed to death by a system that is, by every internal measure, working correctly.

What Weber's framework does not provide is a mechanism—an account of how rationalization is practically accomplished in concrete situations. George Ritzer's extension of Weber through the concept of McDonaldization offers some analytical purchase here. Ritzer traces how the logic of efficiency, predictability, calculability, and control—the principles Weber identified as defining the modern order—have been intensified and universalized through the model of the fast-food industry. In social services, McDonaldization produces intake forms that cannot accommodate complexity, eligibility criteria calibrated to average cases rather than extreme ones, case management software that enforces procedural compliance at the expense of clinical judgment, and service delivery timelines that reflect funding cycle logic rather than human need. Each of these is a McDonaldization of care: the application of industrial efficiency principles to situations that require the opposite of efficiency—patience, flexibility, relational continuity, and tolerance for mess.

Garfinkel and Practical Reasoning.  

Harold Garfinkel's foundational work redirected sociological attention from social structures to the methods through which members produce those structures as ongoing practical accomplishments. Social order, on this account, is not a pre-given framework that actors inhabit; it is an achievement—continuously produced and reproduced through the situated, reflexive, accountable practices of members. This means that institutional harm is also an achievement. It is not simply a structural outcome; it is produced, moment by moment, through the practical reasoning of members who are doing exactly what they are supposed to do.

"I use the term "ethnomethodology" to refer to the investigation of the rational properties of indexical expressions and other practical actions as contingent ongoing accomplishments of organized artful practices of everyday life." — Harold Garfinkel, Studies in Ethnomethodology (1967:11)

Garfinkel's concept of accountability is central to this analysis. Members of any social setting are continuously oriented toward the recognizability and justifiability of their actions—toward making what they do accountable as the kind of action it is. In institutional settings, this means that actors are always, simultaneously, doing their work and demonstrating that they are doing their work appropriately. The hospital physician discharges Rafael and documents the medical rationale. The shelter staff member reassigns the bed and notes the procedural basis. The sanitation crew photographs the encampment before removal. Each actor is producing not only an outcome but a record of an accountable action.

John Heritage's synthesis of Garfinkel's program extends this insight by showing how accountability structures are built into the sequential organization of institutional interaction. When a discharge decision is made, the physician does not simply decide—she constructs a decision that is visibly, demonstrably, documentably appropriate. The construction of appropriateness is as much a part of the work as the clinical judgment itself. This means that accountability structures do not merely evaluate action after the fact; they shape what actions are possible in the first place. An institutional actor who cannot make her action accountable within her setting's standards cannot perform the action, regardless of its intrinsic merits.

This has a direct implication for coordination failure. If each actor is accountable for what is accountable within her institutional context, then actions that would be accountable only in the context of the whole system's performance—actions calibrated to the arc of a person's multi-institutional trajectory rather than to the episode within a single institution—are structurally unavailable. The accountability structure of each institution renders the inter-institutional coordination problem invisible to its members, because that problem falls outside the scope of what any single institution is accountable for.

Smith and Institutional Ethnography.  

Dorothy Smith's institutional ethnography provides the theoretical framework for understanding the specific role of texts in coordinating—and, in homeless services, miscommunicating—institutional action. Smith argues that the ruling relations of modern societies operate primarily through textual practices: documents, forms, databases, assessments, eligibility criteria. These texts do not merely record a social reality that exists independently of them; they constitute the realities they purport to describe. An eligibility determination does not record a pre-existing fact about who is eligible; it produces eligibility as an institutional fact that then organizes subsequent action.

"The text-reader conversation is a relation that is peculiar to our time. It is one that is produced at the intersection of the local and the extra-local, of the material and the social, of experience and the relations that organize and determine it. In institutional ethnography, texts are treated as mediators of social relations rather than as mere records or representations." — Dorothy E. Smith, Institutional Ethnography (2005:101)

In homeless services, this constitutive function of texts is absolute. A person's housing status, medical urgency, and eligibility for services exist, for all institutional purposes, in their documentation. A person who is not adequately documented is institutionally a partial person—present in the flesh, on the creek bed, deteriorating visibly—but absent in the systems that allocate resources. Smith's framework makes visible the specific mechanism through which this partial personhood is produced and sustained: not through deliberate exclusion, but through the ordinary operation of textual practices that were never designed to accommodate the kinds of documentary instability that characterize unsheltered homelessness.

Mbembe and the Deathworld.  

Achille Mbembe's concept of necropolitics names the political rationality through which certain populations are exposed to conditions of death—not through direct violence but through structured abandonment and the withdrawal of protection. His concept of the deathworld names spaces where 'vast populations are subjected to conditions of life conferring upon them the status of living dead.' Mbembe developed these concepts primarily in relation to colonial and postcolonial contexts; their application to domestic social policy requires modification, but the core insight is directly relevant: deathworlds are produced not by the absence of governance but by its specific form.

In the Santa Barbara context, the deathworld is produced not through sovereign decision—not through a power that consciously withdraws protection from unsheltered people—but through what I call institutional non-decision: the aggregate of locally rational choices that no single actor intended to produce a space of premature death, but that collectively accomplish precisely that. The mechanism is not violence but procedure. The instrument is not the weapon but the form. Mbembe's concept names the space and its political character; ethnomethodology reveals the practical machinery through which it is produced.

Sudnow and the Social Organization of Dying.  

David Sudnow's Passing On (1967), one of the foundational ethnomethodological studies of death in institutional settings, provides a direct and underutilized theoretical resource for the present analysis. Sudnow traced how hospital staff organized their responses to dying patients through a set of tacit understandings about the social worth of different patients—understandings that were never formally articulated but that shaped, in concrete and consequential ways, which patients received aggressive intervention and which were allowed to decline.

"For the most part, "dying" is not an event that occurs to everyone who enters a hospital, but is an occurrence that may or may not come to be imputed to patients over the course of their stay. The question of whether a person is a dying person or not, while never fully resolvable, is practically resolved by the staff in ways that have important organizational consequences." — David Sudnow, Passing On (1967:68)

Sudnow's point was not that hospital staff were malicious; it was that their practical reasoning was organized by social categories that produced differential outcomes in life and death—categories that included age, diagnosis, social presentation, and perceived social worth, none of which were formally sanctioned criteria for differential treatment but all of which shaped actual practice. Stefan Timmermans's later work on resuscitation decisions extended and complicated this analysis, demonstrating that the social organization of dying decisions is more complex and variable than Sudnow suggested, but confirming the fundamental insight: the decision about who gets aggressive intervention and who gets managed toward death is a social organization problem, not merely a clinical one.

In the context of street outreach, Sudnow's framework points toward a dimension of the paradox that deserves explicit attention: the normalization of early death that occurs when outreach workers and social service staff have observed enough premature deaths that the death of a fifty-year-old unsheltered person no longer triggers extraordinary response. This normalization is not callousness; it is, from within the practical reasoning of the workers, a rational adaptation to chronic resource scarcity and repeated loss. But it is also a mechanism through which the threshold for intervention rises, gradually and imperceptibly, to a level that ensures the intervention arrives too late.

III. Deathworlds as Institutional Accomplishments

Santa Barbara's deathworlds occupy specific geography, and that specificity matters. The creek beds along Cota and De la Vina Streets, running behind storefronts and under the Calle Real overpass, have been sites of unsheltered sleeping, deterioration, and death for decades. The hillside encampments along the 101 corridor—visible from the freeway to tens of thousands of commuters daily—have served as a kind of spatial remainder bin for people displaced from lower State Street by successive waves of hostile urban design: removed benches, locked public restrooms, anti-sleeping spikes, selective enforcement of sit-lie ordinances. The Milpas Street corridor, the Eastside creek paths, the back lots of the Funk Zone: each is a micro-geography of accumulated vulnerability, where the concentration of unsheltered people reflects not random settlement but the systematic operation of displacement pressure from more visible, more commercially valuable urban space.

Loïc Wacquant's concept of territorial stigmatization is useful here. Wacquant traces how advanced marginality in contemporary cities is not merely poverty but a specific spatial condition: the concentration of the most stigmatized populations in the most stigmatized spaces, producing a feedback loop in which spatial stigma intensifies the social stigma of the people who live there, which further concentrates disadvantage, which deepens the spatial stigma. In Santa Barbara, this dynamic operates with a distinctive twist: the city's extraordinary natural beauty and extreme real estate values produce spatial displacement pressures of unusual intensity. There is nowhere in Santa Barbara that is truly beyond the reach of redevelopment and beautification pressures, which means that the spaces available to unsheltered people are continuously shrinking and continuously being pushed toward the city's edges and ecological margins—precisely the spaces that are farthest from services, most exposed to weather, and least visible to the residents and officials who might otherwise advocate for intervention.

The political economy that produces these spatial conditions is not incidental to the paradox of rational accumulation—it is its material foundation. Harvey Molotch's growth machine thesis, developed at the University of California Santa Barbara in direct response to the 1969 Union Oil spill that coated the Santa Barbara coastline, provides the analytical framework. I have traced Molotch's contribution to Critical Interactionism at length elsewhere; what matters here is the specific mechanism his framework identifies: cities are organized by a coalition of land-based elites whose overriding interest is in intensifying the exchange value of land, and this coalition systematically shapes public investment priorities in ways that privilege uses that maximize exchange value over uses that serve residents without property.

In Santa Barbara, the growth machine operates with exceptional intensity. Median home values exceeding two million dollars, a tourism economy generating over two billion dollars annually, and a political culture organized around the preservation of property values and neighborhood aesthetics combine to produce a specific and consequential set of priorities: the suppression of visible poverty, the minimization of shelter and service facilities in commercially valuable areas, and the chronic underfunding of the homeless services system relative to the scale of need it faces. The Santa Barbara County homeless services system operates on a budget that, per capita, ranks among the lowest of comparably-sized counties in California—not because the county lacks resources, but because the political economy of the growth machine systematically redirects public investment away from uses that do not serve exchange value.

Molotch's framework connects the macro-political economy to the micro-level paradox with precision. The paradox of rational accumulation is not simply a coordination failure among well-intentioned institutions. It is a coordination failure that is structurally tolerated—and in some respects structurally produced—by a political economy that benefits from the management of visible poverty rather than its resolution. A managed deathworld is, from the standpoint of the growth machine, preferable to the alternative: the investment in affordable housing, treatment capacity, and outreach infrastructure that would actually dissolve it, but that would require land use decisions and public expenditures that the growth machine resists. The creek beds are not simply where the coordination failure lands. They are where the growth machine parks its human surplus.

These are not hidden spaces. They are intensely, continuously visible—to outreach workers on their weekly routes, to city staff who maintain the creeks, to law enforcement who respond to calls, to the neighbors who submit complaints. What makes them deathworlds is not invisibility but managed visibility: a set of institutionally organized practices that acknowledge these spaces without resolving them, that respond to the people in them without fundamentally altering the conditions that make those spaces lethal.

Managed visibility is itself an institutional accomplishment. Consider the ecology of practices that constitutes a single Santa Barbara creek bed as a tolerated zone of managed deterioration. The outreach organization schedules weekly visits and documents encounters in its Homeless Management Information System database. Santa Barbara County Behavioral Health maintains a High Utilizer list for individuals with five or more emergency room visits in a twelve-month period. Cottage Hospital's social work department conducts discharge planning for homeless patients, documenting housing status and service referrals. The Santa Barbara Police Department's Homeless Liaison Program conducts regular contact with people sleeping outdoors, logging encounters and issuing warnings for ordinance violations. Public Works executes encampment removals on a schedule posted in advance, with cleaning crews and photo documentation for legal compliance. The County's Continuum of Care maintains the Coordinated Entry waitlist, assigning priority scores and tracking placement attempts.

Each of these practices constitutes a form of institutional attention to the creek bed and its inhabitants. Each generates documentation. Each is, within its institutional frame, evidence of appropriate engagement. Together they constitute a system of management that sustains the deathworld by organizing around it rather than dismantling it. The creek bed is cleaned, not closed. The encampment is dispersed, not resolved. The outreach worker returns next week. The High Utilizer returns to the emergency room. The cycle continues.

This is the core ethnomethodological insight about deathworlds that distinguishes the present analysis from purely structural critiques: deathworlds are not the absence of institutional practice but its product. The creek bed is not a space that institutions have failed to reach. It is a space that institutions have reached, organized, documented, managed, and—through the aggregate of their normal operations—reproduced as a zone where people die.

Erving Goffman's analysis of total institutions provides a useful contrast. Goffman described institutions—prisons, mental hospitals, monasteries—that exert comprehensive control over the lives of their inmates, imposing a single institutional logic on every dimension of daily existence. The deathworld is the inverse of the total institution: not a space of comprehensive control but a space of comprehensive non-accountability. Multiple institutions engage with it; none is accountable for it as a whole. The result is, paradoxically, as lethal as the total institution's worst abuses—but far harder to see, to name, and to contest, because it is produced not by a visible sovereign exercising power but by the aggregate of individually defensible institutional decisions.

IV. Rafael: The Methods That Produced His Death

Rafael is a composite figure drawn from outreach encounters in Santa Barbara over more than two decades of weekly street contact. He is not one person. He is the pattern—enacted with variations in name, diagnosis, nationality, and specific sequence of events, but with a structural consistency that amounts to a social type—made analytically visible through a single narrative. The protection of individual privacy and the analytical requirement to illuminate a type rather than a particular person are served simultaneously by the composite method, which Carol Stack and other ethnographers of urban poverty have used to similar purpose.

Rafael is fifty-seven years old. He has lived in Santa Barbara, on and off, for eleven years—long enough to have accumulated institutional history across multiple systems and to be recognizable, by name, to dozens of outreach workers, case managers, and emergency room staff. He sleeps along a creek near downtown. He has Type 2 diabetes, poorly managed: his last A1C, obtained six months ago during a hospitalization, was 11.2. He has hypertension. A chronic infection on his left foot has not healed in four months; the wound began as a small blister, was complicated by the near-impossibility of keeping it clean outdoors, and has progressed to a deep soft-tissue infection that has not responded fully to two courses of oral antibiotics. He is alcohol-dependent—physiologically, not recreationally: he experiences withdrawal symptoms when his consumption drops below a threshold that outreach workers have learned, through observation, to recognize.

He has lost his government-issued identification twice. The first time it was stolen while he slept. The second time it was discarded, along with everything else he owned, during an encampment removal eight months ago. He knows he needs to replace it. He has begun the process twice and been interrupted twice—once by a hospitalization that consumed the weeks during which his paperwork was being processed, and once by a second cleanup that discarded the partially completed forms.

His vulnerability assessment score is high. By every metric the system deploys to identify priority cases, Rafael is a priority case. He is also, within the system's operational constraints, very difficult to house. What follows is an ethnomethodological reading of his encounters with five institutions, attended to the methods through which each institution's members accomplished their work in ways that were locally appropriate and cumulatively lethal.

The Assessment.  An outreach worker administers the VI-SPDAT. Rafael's score places him in the high-acuity tier. This is appropriate: the instrument was designed to identify people like him. His score is entered into HMIS. He is placed on the Coordinated Entry waitlist. The outreach worker has performed her role competently. The documentation is accurate. From within her organization's accountability structure, this encounter is a success: she has engaged a high-acuity client, completed the required assessment, and initiated the intake process.

What the VI-SPDAT cannot capture—and what no instrument currently deployed in Santa Barbara's homeless services system is designed to capture—is the velocity of Rafael's deterioration. The instrument produces a score; scores are static; they represent a moment in time. But Rafael's trajectory is the analytically crucial variable. A person whose vulnerability score is high and whose trajectory is stable is a very different case from a person whose score is equally high and whose trajectory is declining sharply. The assessment system treats them identically. The body does not.

The Documentation Barrier.  Access to housing programs requires government-issued identification. This requirement exists because housing programs are federally and state funded and must demonstrate that they are serving eligible individuals; documentation is the mechanism of eligibility verification. The requirement is administratively rational and legally necessary.

For Rafael, it is an effective barrier. Replacing his lost ID requires a certified copy of his birth certificate from the county in which he was born, several hundred miles away. The vital records office requires a written request, a fee, and proof of identity—which Rafael cannot easily provide because he lacks identification, creating a circular requirement that is familiar to anyone who has worked with unsheltered people on document recovery. A support letter from a social service organization can sometimes substitute; obtaining one requires an intake appointment; the appointment requires advance scheduling; the schedule requires a phone, which Rafael has intermittently. Processing time: four to six weeks, assuming no complications.

This is what Dorothy Smith would call a textually-mediated exclusion: Rafael's disqualification from housing is not a judgment about him as a person. It is a consequence of his relationship to a cascade of documents. The ruling relations of the housing system operate through texts, and Rafael's relationship to those texts is structurally precarious—not because he is careless or incompetent, but because his living conditions make document retention nearly impossible. Each cleanup resets his documentary existence. Each reset extends his time on the creek. Each extension advances his deterioration. The documentary requirement that exists to ensure fair allocation of public resources functions, in practice, as a mechanism of lethal delay for the people most in need of what those resources provide.

The Medical Discharge.  His foot worsens. An outreach worker arranges transport to the emergency room. Rafael is admitted, the wound debrided, intravenous antibiotics administered. Six days. The attending physician discharges him when the acute infection is clinically stable. This is medically appropriate, legally defensible, and fiscally rational under current hospital reimbursement structures, which incentivize discharge once acute criteria are met. The discharge summary documents his condition, recommends follow-up with a wound care specialist in two weeks, and notes that the patient is experiencing homelessness.

The discharge summary is a document that performs accountability within the hospital's institutional context with precision. It demonstrates that the standard of care was met, that the clinical decision was appropriate, that the physician acted responsibly. What it cannot do is follow Rafael to the creek. The wound care recommendation exists in the medical record system. There is no mechanism by which it reaches the outreach worker who will find Rafael the following week—the two organizations use different, non-interoperable electronic health record systems, and there is no data-sharing agreement between them. The text circulates within the institution that produced it. The recommendation dies with the discharge.

Sudnow's analysis of dying in institutional settings is illuminating here by extension. The hospital's practical organization of the discharge process is calibrated to what the hospital is accountable for: acute medical status, standard of care, appropriate documentation. It is not calibrated to what will happen to the patient after discharge, because post-discharge outcomes are not within the hospital's accountability structure. The result is a form of institutionally organized non-seeing: the physician who signs the discharge order is not indifferent to what will happen to Rafael on the creek. She is simply not organized, accountable, or resourced to see it as her problem.

The Shelter Miss.  A bed opens at a low-barrier shelter. Rafael's name is on the prioritized list. Shelter intake requires arrival between 4:00 and 5:00 p.m. His infected foot limits his mobility. He has been sitting on concrete in a shaded area for most of the day. He cannot walk the required distance within the window. He does not arrive in time. The bed is reassigned to the next person on the list. The shelter staff member documents the reassignment and its procedural basis. From within the shelter's accountability structure, this is a successful bed placement.

The intake window exists because it is operationally rational: it allows staff to complete intake processing before the evening shift ends, ensuring that the bed is assigned and the client is processed before the night staff arrive. No one designed it to exclude people with mobility impairments. But the intake window, combined with the assumption that clients have reliable transportation and unimpaired physical capacity, functions as a systematic exclusion of exactly the population the shelter is most urgently needed by. This is the paradox of rational accumulation at the level of a single procedural rule: locally sensible, globally lethal.

The Cleanup.  A sanitation crew executes a scheduled encampment removal on Rafael's creek. The area was posted 72 hours in advance, as required by the legal framework that emerged from the Ninth Circuit's Martin v. City of Boise decision. Rafael's sleeping bag, his remaining antibiotics, and the partially completed birth certificate paperwork—representing eight months of accumulated effort across two separate attempts—are discarded. The crew followed procedure. The posting requirement was satisfied. The removal is legally authorized and photographically documented.

The 72-hour notice was posted on a sign at the encampment's periphery in English only. Rafael reads Spanish more fluently than English. The posting was not explained to him. There is no requirement that it be explained, because the posting requirement was designed by legal counsel to satisfy due process standards in federal court, not to ensure that the affected person understood what was coming. Three months of documentary work was discarded in an afternoon. The process reset.

An outreach worker finds him at a new location the following week. The VI-SPDAT is re-administered—protocol requires reassessment when a client has been inactive in the system for more than 90 days, and the cleanup disrupted his case continuity. He is placed back on the list.

Six months after the first encounter, Rafael is hospitalized with a severe systemic infection. Eleven days. Discharged when medically stable.

He returns to the creek.

One year after the first encounter, Rafael is dead. Cause of death on the coroner's report: sepsis secondary to soft tissue infection, with contributing factors of diabetes mellitus type 2 and chronic alcohol use disorder.

No single decision caused his death. Each member of each institution performed their accountable role competently. Each decision was locally rational, appropriately documented, and professionally defensible. The system worked.

Rafael died.

Rafael is one type of person who dies in Santa Barbara's deathworlds. The mechanisms are more general than his specific case. Consider Maria.

Maria is forty-three years old. She has lived in Santa Barbara her entire life—was born here, went to school here, raised a child here before her housing became unstable. She has a diagnosis of schizoaffective disorder, bipolar type, first made during a psychiatric hospitalization eleven years ago. She has been hospitalized involuntarily under 5150 holds seven times in the past four years. She does not have diabetes or a foot infection. Her body, at forty-three, is not visibly deteriorating in the way Rafael's is. But she is dying too, on a slower and less linear schedule.

Maria's encounters with the institutional system differ from Rafael's in their specific texture, but the structural mechanics are identical. Her documentation is intact—she carries her ID in a small waterproof bag she has learned, through hard experience, to keep on her person at all times. Her VI-SPDAT score is high. She is on the Coordinated Entry waitlist for permanent supportive housing, which is the appropriate placement given her diagnosis and history.

The waitlist for permanent supportive housing in Santa Barbara County currently runs to approximately four years. This is not because the housing authority is indifferent or incompetent. It is because the supply of permanent supportive housing units in the county is radically insufficient relative to the number of people who need them, and because the political economy of the growth machine has, for decades, successfully resisted the construction of new affordable and supportive housing in the locations where it would be most effective—near services, transportation, and the social networks that support recovery.

While Maria waits, she cycles. A psychiatric crisis produces a 5150 hold. She is transported to the psychiatric emergency unit, stabilized on medication, and discharged—typically within seventy-two hours—when she is no longer an immediate danger to herself or others. This is the legal standard for involuntary hold; it is also, for Maria, a revolving door. The medication that stabilizes her in the hospital is prescribed for outpatient use, but filling the prescription requires a pharmacy visit, which requires transportation, which requires a functioning bus pass, which requires having collected the cash for it before the hospitalization, which she has not done because the hospitalization was unplanned.

She is discharged without medication. Within ten days her symptoms return. She is visible on State Street, talking to people who are not there, frightening the tourists who are. A call is made to the police non-emergency line. Officers respond—professionally, with the training the department has invested in mental health crisis response—and connect her with a mobile crisis team. The team engages her and recommends voluntary hospitalization. She declines. She has the legal right to decline. The team documents the encounter and refers her back to her case manager.

Her case manager carries a caseload of thirty-eight clients. She calls Maria twice in the following week. Maria does not answer—her phone has been stolen again. The case manager documents the contact attempts. Maria is marked as 'difficult to engage.'

Six weeks later, Maria is found unresponsive in a park. She has been assaulted. She is hospitalized with a traumatic brain injury. She survives, but the injury produces cognitive changes that make her housing case significantly more complex. Her VI-SPDAT score is re-administered. It is higher now. She moves up the waitlist.

She does not reach the top before her next crisis.

Maria and Rafael died by different mechanisms—his primarily biological, hers primarily social and psychiatric—but the structural logic is identical. In both cases, the paradox of rational accumulation operates through the same set of mechanisms: incommensurable institutional vocabularies, textually-mediated exclusion, epistemic abandonment, the misalignment of administrative and biological time, and distributed innocence. The specific institutions are different; the specific documents are different; the specific bodies are different. The machine is the same.

The fact that the pattern repeats across radically different individual circumstances is the strongest evidence that the argument presented here names a structural condition rather than an individual story. Rafael and Maria are not exceptional cases. They are the norm. They are what the system produces, reliably, when it works as designed.

V. Incommensurable Vocabularies: The Mechanism of Coordination Failure

The Rafael narrative reveals, at the level of institutional interaction, the proximate mechanism of the paradox: the incommensurability of institutional vocabularies. Each institution Rafael encountered maintained its own practical vocabulary—the categories, distinctions, and criteria through which its members made sense of their work and produced accountable actions. These vocabularies were not merely different; they were constitutively incompatible in ways that made genuine coordination across institutional boundaries structurally unavailable.

The hospital's vocabulary was clinical: acute, stable, responsive to treatment, appropriate for discharge. These categories were calibrated to the institution's core function—the management of acute medical crises—and to its accountability structure, which required documentation of medical decision-making legible to peer reviewers, insurance auditors, and malpractice assessors. Within this vocabulary, Rafael's discharge was appropriate: his acute condition had been addressed to clinical standard. The vocabulary had no category for "will certainly re-injure under current living conditions" because that is a social determination, not a clinical one, and social determinations fall outside the institution's accountable scope.

The shelter's vocabulary was operational: capacity, intake, waitlist position, behavioral compliance, bed management. Within this vocabulary, the bed reassignment was appropriate: the slot had been offered; the client had not arrived within the established window; the next person on the list was served. The vocabulary had no category for "could not arrive due to mobility impairment caused by a condition that is itself a product of unsheltered living" because such distinctions were not operationally necessary. The intake window was a rule; the rule applied; the rule was followed.

The housing program's vocabulary was administrative and documentary: eligible, documented, verified, prioritized, compliant, active in system. Within this vocabulary, Rafael was, at any given moment, either progressing through the documentation requirements or not. The vocabulary had no category for "documentation is repeatedly destroyed by sanitation procedures mandated by the same municipal government that administers the housing program" because the housing program and the sanitation department operated through entirely separate administrative chains with no coordinating accountability structure.

Aaron Cicourel's cognitive sociology is directly relevant here. Cicourel demonstrated that institutional actors necessarily simplify the complexity of their clients' situations through typification—the assignment of cases to pre-existing categorical schemes that render them administratively manageable. Typification is not a failure of intelligence or empathy; it is a practical necessity. No institution could function if it attempted to respond to the full particularity of each case. But typification means that what gets seen—and what gets acted upon—is determined by the categorical scheme of the institution doing the seeing. Rafael is visible to the hospital as an acute medical case; to the shelter as a waitlist position; to the housing program as a documentation file; to law enforcement as a potential ordinance violation; to Public Works as an encampment requiring removal. No institution's typification scheme includes the Rafael who is moving toward death across the full arc of his encounters with all of them simultaneously.

This is the mechanism of coordination failure: not bad faith between institutions, but a structural condition in which each institution's way of constituting its clients is shaped by its own accountability requirements and is therefore constitutively blind to what other institutions perceive. The vocabularies do not merely fail to communicate; they actively prevent communication by defining the same person in categorically incompatible ways. A hospital social worker who calls a shelter to discuss a patient's discharge is operating with a clinical typification of the patient's situation; the shelter intake coordinator is operating with an operational one. The conversation produces referral documentation. It does not produce coordination. The patient goes back to the creek.

VI. Longitudinal Witnessing and the Knowledge That Cannot Travel

Street outreach workers occupy a structurally distinctive position in the multi-institutional field that addresses homelessness. They are, in a precise sense, boundary workers: affiliated with one institution but operating in the interstices between institutions, accompanying clients across thresholds that other institutional actors rarely cross, accumulating knowledge through sustained relational presence that no other actor in the system possesses. The specific form of knowledge that sustained presence generates is longitudinal—knowledge of trajectory, of change over time, of the relationship between setback and decline. I have elsewhere called this capacity professional vision as it develops in street outreach: a distinctive way of seeing that is trained by repeated exposure to deterioration and sharpened by the outreach worker's structural position at the edge of the institutional field.

The outreach worker sees Rafael not at a single clinical moment but across months and years of regular contact. She notices the weight loss—a few pounds here, a few there—long before it registers on a medical chart. She observes the change in his gait months before the infection reaches crisis. She has learned, through repeated observation, that his alcohol consumption escalates predictably when his housing documentation is destroyed, because she has watched the connection between documentary setback and the collapse of hope play out across multiple cycles. She has accumulated, through the practice of sustained relational presence, something that approximates a clinical trajectory—a dynamic picture of Rafael's deterioration that is more complete and more predictive than anything contained in any single institutional database.

This longitudinal knowledge is, from a public health standpoint, extraordinarily valuable. It is the kind of knowledge that could interrupt the paradox of rational accumulation—if it could be translated into institutional action. The problem is structural: the outreach worker's knowledge has no institutional form. It does not appear in the discharge calculus of the hospital, which requires medical status determinations produced by licensed clinicians. It does not appear in the eligibility assessments of the housing program, which requires VI-SPDAT scores and documentary verification. It does not appear in the encampment removal procedures of Public Works, which requires legal compliance with posting requirements. The knowledge exists; the channels through which it could travel to consequential decision points do not.

Michael Lipsky's analysis of street-level bureaucracy captures part of this condition. Street-level bureaucrats face a chronic gap between the resources available and the needs they encounter, and they develop coping strategies—routinization, rationing, the management of their own expectations—that allow them to continue functioning under conditions of chronic inadequacy. The outreach worker who has watched enough Rafaels move through the system and die develops a version of these coping strategies: a practical orientation toward what can be accomplished within the system's constraints, combined with a management of grief that makes continued work possible. This is not callousness. It is what Lipsky would recognize as a rational adaptation to chronic structural inadequacy.

But the outreach worker's position includes a dimension that Lipsky's framework does not fully capture: she is not merely managing inadequate resources. She is in possession of knowledge that the institutions with resources do not have, and she lacks the institutional standing to compel those institutions to act on it. She can document. She can refer. She can advocate in case conferences where her voice is one among many and carries no formal weight. She cannot interrupt a discharge. She cannot override a cleanup schedule. She cannot force a housing program to waive a documentation requirement for a client she has watched deteriorate for eighteen months.

I call this condition epistemic abandonment: a structural arrangement in which the person with the most complete and most accurate knowledge of a client's situation has the least institutional authority to act on that knowledge. The outreach worker knows. The institutions that can act do not know. The knowledge cannot travel, because the institutions have not built the channels through which it could, and the accountability structures that govern each institution provide no incentive to build them.

The moral injury produced by this condition is significant and underexamined. To work in street outreach over years is to accumulate a particular kind of knowledge: knowledge of who will die, often, before they die. Outreach workers with substantial experience develop something that functions like clinical foresight—a pattern-recognition capacity built from repeated exposure to the sequence of deterioration. They know what comes after the second hospitalization without housing placement. They know what the change in gait means. They know, in a way that is practically certain but institutionally weightless, that Rafael is dying. And they cannot stop it.

VII. Documentary Practices and the Production of the Partial Person

Dorothy Smith's institutional ethnography provides the theoretical framework for one of the most consequential and least visible mechanisms in homeless services: the systematic production of partial persons—individuals whose institutional existence is incomplete because their documentary existence is incomplete, and who are therefore only partially accessible to the systems that are theoretically designed to serve them. I have developed an extended analysis of what I call the documentary constitution of organizational reality in the context of homeless services elsewhere; the present section draws on and extends that analysis.

In Smith's framework, the ruling relations of modern societies operate through textual practices: the production, circulation, and interpretation of documents that coordinate action across organizational settings. These texts do not simply record a pre-existing social reality; they constitute the realities they purport to describe. This constitutive function is not unique to homelessness services—it characterizes all modern bureaucratic organization—but it has a specific and particularly consequential form in systems that serve populations whose relationship to documentation is structurally precarious.

To be housed in Santa Barbara's Coordinated Entry system, one must be documented. To be documented, one must have government-issued identification. To have identification, one must have a primary identity document—a birth certificate, a passport, a court order. To obtain a birth certificate, one must know where one was born, have access to the appropriate vital records office, pay the required fee, provide some secondary identification or a letter from a qualified social service organization, and maintain the paperwork through the processing period without losing it to theft, weather, or encampment removal. Each of these requirements is individually modest. Together they constitute an effective barrier for people whose lives are characterized by the precise kind of instability that makes documentation difficult to obtain and nearly impossible to retain.

Don Zimmerman's foundational study of record-keeping in a public welfare agency, conducted in the Los Angeles area and published in Stanton Wheeler's On Record (1969), demonstrated that caseworkers maintained client files through ongoing practical reasoning that diverged significantly from the formal rules governing record-keeping. Workers organized their records to accomplish the practical purposes of their work—demonstrating eligibility, managing caseloads, satisfying supervisory review—rather than to provide comprehensive accounts of their clients' actual situations. The records were organizationally adequate rather than phenomenologically accurate: they captured what the organization needed to capture, not what was true about the client's life.

In contemporary homeless services, this practical organization of record-keeping has been partially formalized and automated through electronic HMIS systems. The Homeless Management Information System is a federally mandated data infrastructure that requires participating organizations to enter client data in standardized formats, enabling aggregate reporting for HUD compliance purposes. The system produces longitudinal data—in theory. In practice, the data is episodic: it records contacts and transactions within each participating organization but does not produce the kind of trajectory data that would make the arc of a person's multi-institutional deterioration visible. A person can have dozens of HMIS entries across multiple organizations over several years, each recording an appropriate encounter, none recording the overall pattern that those encounters, taken together, constitute.

The result, in ethnomethodological terms, is a documentary system that produces institutional persons who are constituted by the categorical requirements of the funding and accountability systems the documentation serves, rather than by the actual characteristics of the people being documented. The Rafael who exists in the HMIS database is a Rafael composed of VI-SPDAT scores, service contacts, referral outcomes, and housing placement attempts. He is organized by the reporting categories that HUD requires and that local funders audit. He is not organized by the trajectory of his deterioration, because HMIS is not designed to capture trajectories. The partial person is not produced by individual caseworker negligence; he is produced by the architectural requirements of the data systems themselves.

There is a further dimension to this problem that deserves explicit attention: the temporality of documents. Documents are produced at moments in time and, unless actively updated, remain fixed at those moments. A vulnerability assessment conducted eight months ago reflects Rafael's situation eight months ago. The waitlist position assigned on the basis of that assessment reflects a priority calibration derived from a static score. The person who appears in the system and the person who is deteriorating on the creek are increasingly, over time, different people—and the gap between them widens exactly as the need for intervention becomes most urgent.

VIII. The Phenomenology of Administrative Time

The paradox of rational accumulation is lived, from inside the experience of unsheltered homelessness, primarily as a temporal phenomenon. Understanding this requires distinguishing between two different temporalities that coexist, without coordination, in the experience of people navigating the homeless services system: administrative time and biological time.

Administrative time is the temporality of institutional process. It is measured in weeks between case reviews, months on waitlists, processing periods for documents, program cycles calibrated to funding streams. Administrative time is not designed to be slow; it reflects genuine organizational demands. But it is calibrated to organizational rhythms—to what institutions need to function—not to the rhythms of human biology and deteriorating health. A four-to-six-week processing period for a birth certificate is not unreasonable by administrative standards. It is potentially lethal by biological ones.

Biological time moves differently. An untreated infection does not pause while a birth certificate is being processed. A person sleeping on wet concrete in January does not stop losing core temperature while a shelter bed is being held pending documentation. Cognitive decline associated with alcohol dependence, poor nutrition, and chronic sleep deprivation does not suspend itself while a case manager works through a caseload backlog. The body has its own temporal logic, and it is not synchronized with the administrative logic of the institutions that are, in principle, organized to address its needs.

The gap between administrative time and biological time is the temporal mechanism of premature death in deathworlds. It is not that the system is too slow in some absolute sense. It is that the system moves at a speed calibrated to its own operational needs, and the bodies of the most vulnerable people in it move faster toward death than the system moves toward intervention. This gap is not accidental; it is structural. It would persist even if every individual worker in every institution performed their work with perfect competence and maximum efficiency.

Alfred Schutz's phenomenology of time-consciousness provides the conceptual tools for understanding how this temporal gap is lived from the inside. Schutz distinguished between multiple temporal registers: the vivid present of direct embodied experience; the world of predecessors constituted by memory and history; and the world of contemporaries known only through typification and inference. In the context of homeless services, this last distinction is crucial: the Rafael who is present to the outreach worker in the vivid present of their face-to-face encounter, with his specific weight, his specific gait, his specific affect on a specific cold morning, is a different Rafael from the one who exists for the hospital's discharge planner, the shelter's intake coordinator, and the housing program's case manager as a contemporary known only through records, scores, and categorical typifications. The institutions act on the documentary contemporary. The deteriorating embodied person waits.

What repeated encounters with this temporal mismatch produce in the person experiencing it is a rational reorganization of temporal orientation. When the system has repeatedly demonstrated that its long-term promises are not reliably fulfilled before the body fails, orienting toward the long term becomes irrational. The next meal, the next safe place to sleep, the next day: these become the appropriate planning horizons, because they are the horizons within which reliable planning is actually possible. The shortening of time horizons that institutional actors often read as evidence of the client's pathology—their presumed inability to engage in "future-oriented thinking," their supposed impulsivity or present-orientation—is, from within the client's experiential logic, a reasonable adaptation to a system whose promises have consistently expired before they could be kept.

The generation and repeated deferral of hope is one of the most consequential and least examined mechanisms in this system. "You're almost at the top of the list." "The document should arrive this week." "There's a new program opening next month." These statements are made in good faith by workers who believe them. They are also temporal commitments made on behalf of systems whose timelines the speaker cannot control. When those commitments are not fulfilled—and in a system operating under conditions of chronic scarcity, they are frequently not—the cost is not merely disappointment. Repeated unfulfilled temporal commitments reorganize the person's practical engagement with the system in ways that make future engagement more difficult: missed appointments begin to seem rational; paperwork maintenance ceases to seem worth the effort; withdrawal from outreach contact becomes, from inside the experience, a reasonable response to repeated investment with negligible return. The system reads these behavioral patterns as evidence of client pathology—as symptoms of the mental illness or substance use disorder that the system is ostensibly addressing. They are, more accurately, rational adaptations to the system's demonstrated unreliability.

IX. Distributed Innocence and the Limits of Accountability

The most morally disorienting feature of the paradox of rational accumulation is what I call distributed innocence: the condition in which genuine harm is produced in the complete absence of culpable actors. Every person who interacted with Rafael acted within their accountable role. Every institution that processed him followed its established procedures. No individual made a decision that was unreasonable within its institutional context. Rafael died. I have described this configuration elsewhere as the homeless system functioning as a death regime—not a system designed to kill, but one whose normal operation produces death as a statistically predictable output, and whose accountability structures are organized in ways that prevent that output from registering as anyone's failure.

Distributed innocence is not a moral alibi. It is a precise description of a structural condition that requires precise analytical treatment. The standard responses to institutional harm—identify the responsible actor, investigate the deviation from procedure, implement corrective training, reform the policy that was misapplied—all presuppose that harm results from a failure to follow procedures correctly. They are responses to implementation failures. Distributed innocence names a different kind of problem: a design failure, in which procedures followed correctly produce harm as a regular output.

The distinction matters enormously for how we respond. Implementation failure analysis produces retraining, supervision, and protocol revision within individual institutions. It leaves the coordination structure—the relations between institutions—untouched, because it diagnoses problems at the institutional component level rather than at the level of the whole system. Design failure analysis requires a different and harder response: not the repair of components but the redesign of the architecture that coordinates their interaction.

Lipsky's street-level bureaucracy framework illuminates the position of front-line workers within this architecture. Street-level bureaucrats face a chronic gap between available resources and encountered needs, and they develop coping strategies that allow continued functioning under chronic inadequacy. In homeless services, these coping strategies include: the routinization of client engagement (the same assessment protocol administered in the same sequence regardless of the client's changing circumstances); the normalization of inadequate outcomes (the gradual upward revision of the threshold for "acceptable" results as repeated inadequacy becomes the baseline expectation); and what Lipsky calls creaming—the informal prioritization of clients who are most likely to produce successful outcomes by institutional measures, at the cost of clients whose needs are greatest but whose probability of measurable success is lowest.

Creaming is particularly consequential in the context of the paradox of rational accumulation. The people most likely to be creamed—most likely to be de-prioritized in practice even when formally prioritized on paper—are precisely the people like Rafael: high-acuity, multiply-diagnosed, documentarily unstable, with records of service engagement that produced no housing placement. From within the practical reasoning of individual workers operating under resource constraints, the decision to invest limited effort in clients with higher probabilities of successful placement is rational. From the standpoint of the whole system's performance, it is a mechanism through which the paradox of rational accumulation intensifies: the people most in need of intervention receive it least effectively.

The deepest ethical problem posed by distributed innocence is systemic: it produces a condition of non-accountability for outcomes. Each institution is accountable for the quality of its component performance. No institution is accountable for what the whole system produces. The whole system is where people die. Until an actor with genuine authority is designated as accountable for the arc—for what happens to Rafael across all of his institutional encounters, in their cumulative effect on his survival—the paradox will continue to operate, and the deaths will continue to be produced by institutions that are, individually, functioning correctly.

The Behavioral Wellness Commission, the Santa Maria/Santa Barbara County Continuum of Care, the Board of Supervisors: these are the bodies with the jurisdictional authority to create the cross-institutional accountability structures that could interrupt the paradox. They have not done so. This is not, primarily, because they are indifferent to the deaths. It is because the diagnostic framework available to them—the framework of implementation failure—does not reveal the design failure that the paradox names. Until the diagnostic framework changes, the institutional response will remain calibrated to the wrong level of the problem.

X. Conclusion: Naming the Machine

The paradox of rational accumulation is not a critique of individual actors. It is an analysis of a machine—a system that operates, component by component, with competence and good faith, and produces premature death as one of its regular outputs. The machine is not designed to kill. It is designed to manage: to allocate scarce housing resources fairly, to treat acute medical crises appropriately, to enforce public space ordinances legally, to conduct outreach ethically. Each design decision is defensible. The machine, in aggregate, is lethal.

The ethnomethodological contribution to this analysis is to show that the machine operates through people—through the practical reasoning of members who are doing their jobs, producing accountable actions, meeting the standards of their professional settings. The machine is not an abstraction floating above the social world; it is assembled, encounter by encounter, in the situated interactions through which institutional actors accomplish their work. This means it can, in principle, be disassembled—not through moral exhortation directed at individual workers, but through the redesign of the accountability structures that shape practical reasoning. Change what actors are accountable for, and you change what they can see, and what they can do.

Weber saw the iron cage closing. Garfinkel showed us the hands that built it, one reasonable decision at a time. Smith showed us how the texts through which those decisions were recorded became the architecture of exclusion. Sudnow showed us how the organization of dying is a social accomplishment, not merely a biological fact. Mbembe named the spaces where people are left to die. Lipsky showed us the workers trapped inside the machine, doing what the machine requires. What remains is to name the machine clearly enough that it cannot be mistaken for a service system—to say, in the chambers where decisions are made, that the system is not failing. It is working. And its normal operation produces death.

In Santa Barbara's deathworlds, naming the machine means saying clearly to the Board of Supervisors, to the Behavioral Wellness Commission, to the Continuum of Care board, to the leadership of every hospital, shelter, housing program, and law enforcement agency in the county: the people dying on the creek beds are not dying because the system failed to reach them. They are dying because the system reached them—assessed them, waitlisted them, referred them, discharged them, removed their encampments, reset their documentation, and processed them again—and the cumulative effect of that reaching was to leave them, month by month, more damaged and more exposed than they were before. The system did not abandon them. It managed them toward death.

Structural redesign requires, at minimum: a unified longitudinal tracking system that follows individuals across all institutional encounters and makes the arc of deterioration visible to every actor in the system simultaneously; cross-institutional accountability structures that designate a responsible actor for each high-acuity individual's overall trajectory; discharge protocols that make post-discharge outcomes, not merely discharge appropriateness, the measure of hospital performance for homeless patients; housing eligibility criteria redesigned around the actual documentary circumstances of people experiencing total instability; and encampment removal protocols that treat document preservation as a legal and ethical obligation rather than an operational afterthought.

These are not radical proposals. They do not require new resources at the scale that would make them politically impossible. They require something harder: the willingness to be accountable for what the system produces, not only for what each component intended. They require saying, out loud and on the record, that competent professional practice within each institution is necessary but not sufficient, and that an accountability structure that ends at the institutional boundary is an accountability structure designed to produce the outcomes we currently observe.

Proof that this is achievable exists. Houston, Texas—not a city typically associated with progressive social policy—has come closer than any other major American city to ending chronic unsheltered homelessness, not by solving the growth machine problem or eliminating poverty, but by building precisely the cross-institutional accountability infrastructure this essay has argued for: a unified data system that follows individuals across all institutional encounters; a single point of accountability for each high-acuity person's trajectory; and a governance structure in which hospital systems, law enforcement, housing agencies, and outreach organizations are jointly accountable for outcomes at the level of the whole system rather than only for the performance of their individual components. Houston did not do this by spending more money than other cities. It did it by reorganizing accountability. The deaths on Santa Barbara's creek beds are not inevitable. They are a choice—a choice embedded in the structure of the system, reproduced through the ordinary practical reasoning of its members, and sustained by a political economy that has not yet been compelled to bear the cost of what it produces.

Everyone was being reasonable.

And people died anyway.

Rafael did not die because anyone failed him. He died because everyone succeeded—at their piece of the machine. That is the paradox. And until we are willing to say it plainly, in the rooms where decisions are made, we will keep building the same machine and watching it produce the same deaths, year after year, on the same creek beds, among people whose names we know.

Notes

1.  Mortality figures are drawn from Mellinger (2026), which cites Santa Barbara County data: 44 deaths in 2017, rising to 143 over the 2019–2020 two-year period. As Mellinger notes, this rise occurred even as overall homelessness counts fluctuated only modestly—meaning the increase in deaths reflects a deterioration in the conditions under which people remain unsheltered, not simply an increase in the population at risk. Official counts use a narrow definition requiring the deceased to have been verified homeless at time of death, producing significant undercounting; deaths occurring after hospital discharge or in locations without witnesses are frequently not captured.

2.  Garfinkel's concept of practical reasoning must be distinguished from the colloquial sense of 'being practical.' In his usage, practical reasoning refers to the specific, situated, context-sensitive methods through which members of any social setting make sense of and act within that setting—methods that are neither formal-logical nor irrational but occupy a third territory that Garfinkel called 'practical action and practical reasoning.' See Garfinkel (1967:1–34) for the foundational discussion, and Heritage (1984:111–152) for the clearest secondary exposition.

3.  The translation of 'Stahlhartes Gehäuse' as 'iron cage' is Talcott Parsons's, from the 1930 translation. Peter Baehr and Gordon Wells translate it as 'shell as hard as steel,' arguing that the Parsons translation imports a metaphor of imprisonment that is not in Weber's original. For the present analysis, the distinction matters: the Parsons translation emphasizes entrapment; the Baehr-Wells translation emphasizes rigidity and impermeability. Both are relevant to the paradox of rational accumulation, which entails both.

4.  Ritzer (1993, 1998). The McDonaldization thesis has been criticized for overstating the uniformity and completeness of rationalization processes; see Smart (1999) for a range of critical responses. For present purposes, the thesis is used not as a comprehensive account of contemporary social organization but as an analytic lens for the specific mechanisms through which efficiency, predictability, and calculability operate in homeless service delivery.

5.  Garfinkel (1967). The best single introduction to ethnomethodology's program and its implications for institutional analysis remains Heritage (1984). For applications to organizational settings specifically, see Boden (1994) and Zimmerman (1969, 1970).

6.  Heritage (1984:8–44, 236–280). Heritage's analysis of how accountability is built into the sequential organization of talk and interaction is the most accessible point of entry into Garfinkel's accountability concept for readers coming from outside conversation analysis.

7.  Smith (1987, 2005). The institutional ethnography program Smith developed at OISE/University of Toronto from the 1970s onward has generated a substantial body of applied research; see Campbell and Gregor (2002) for a methodological introduction, and DeVault and McCoy (2006) for a review of the field. For applications in the area of social services and public health specifically, see Mykhalovskiy and McCoy (2002).

8.  Mbembe (2003). The concept of necropolitics has been extended and applied by a range of scholars; see Puar (2007) on biopolitics and disability, and Gilmore (2007) on carceral geography. For applications to urban homelessness in the U.S. context, see Stuart (2016), whose ethnographic study of Skid Row in Los Angeles provides an important empirical complement to the present analysis.

9.  Mbembe (2003:40). The full passage reads: 'new and unique forms of social existence in which vast populations are subjected to conditions of life conferring upon them the status of living dead.' The application to unsheltered homelessness requires specification: people on the creek beds of Santa Barbara are not fully living dead in Mbembe's sense—they retain legal personhood, access to emergency services, and (in principle) the full rights of citizenship. Their deathworld status is a matter of statistical probability rather than sovereign decree. The concept nonetheless names something real about the organization of their situation that more anodyne sociological language does not.

10.  Sudnow (1967). The study was conducted at two California hospitals in the early 1960s, giving it particular regional relevance to the present analysis. Timmermans (1999) provides the most important subsequent engagement with Sudnow's framework, arguing that his account overstates the role of social worth in resuscitation decisions and understates the role of situational contingency and interactional negotiation. The present analysis draws on Sudnow's broader framework—the social organization of dying in institutional settings—rather than specifically on his social worth thesis.

11.  Timmermans (1999:3–18). Timmermans's ethnographic study of CPR practice across multiple hospital settings demonstrates that resuscitation decisions involve complex interactional negotiations among staff that cannot be reduced to social worth calculations, but that social factors—including age, diagnosis, social presentation, and the presence of family members—consistently shape the character and intensity of resuscitation efforts. The implication for the present analysis is that even the most 'technical' institutional decisions about life and death are organized, in part, by social categories that operate below the threshold of explicit policy.

12.  The concept of hostile urbanism, developed in the literature on 'defensive architecture' and 'hostile design,' refers to the deliberate modification of public space to render it unusable for unsheltered people. In Santa Barbara, documented examples include: the removal of benches from lower State Street following the 2019 public camping ordinance; the installation of metal dividers on park benches throughout the downtown core; the removal of public restroom facilities in Alameda Park; and the periodic deployment of sprinkler systems timed to activate during overnight hours at locations known to be sleeping sites. For the broader literature, see Mitchell (2003) and Davis (1990).

13.  Molotch, Harvey. 1976. "The City as a Growth Machine: Toward a Political Economy of Place." American Journal of Sociology 82(2):309–332. Molotch's foundational article was directly shaped by his experience as a faculty member at UC Santa Barbara during and after the 1969 Union Oil spill—an event that made visible, in dramatic form, the tension between growth machine interests and the quality of life of ordinary residents. His subsequent book with John Logan, Urban Fortunes: The Political Economy of Place (1987), extended the framework into a comprehensive theory of urban development. For the application to homelessness and land use specifically, see Culhane (2008) and the critical housing geography literature in Dear and Wolch (1987).

14.  Mellinger (2025c) provides an extended analysis of Molotch's contribution to Critical Interactionism, tracing how the growth machine thesis connects to the interactionist tradition through Molotch's collaborative work with Deirdre Boden on the compulsion of proximity. The present analysis extends Mellinger's argument into the specific context of homeless service underfunding. Santa Barbara County homeless services budget figures are drawn from the County's annual Adopted Budget documents and from HUD Continuum of Care funding award data.

15.  Wacquant (2008:1–30, 237–251). Wacquant's comparative analysis of advanced marginality in Chicago's Woodlawn neighborhood and the French banlieues identifies territorial stigmatization as one of the four 'novel logics' characterizing the new urban poverty. His framework is adapted here to the Santa Barbara context, which differs from his cases in the extreme compression of spatial inequality—the distance between the most expensive residential neighborhoods and the creek beds where people sleep is, in Santa Barbara, a matter of blocks—and in the specific role of the tourism economy in intensifying displacement pressure on visible poverty.

16.  The organizational ecology described here reflects the author's direct observation across more than two decades of weekly street outreach in Santa Barbara. Institutional procedures, program names, and operational details are drawn from firsthand familiarity with the local system; specific organizational names have been generalized to protect working relationships.

17.  Goffman (1961:1–124). Goffman's analysis of total institutions remains valuable not primarily as an empirical description of contemporary institutions—most of which have been substantially de-institutionalized since the 1960s—but as a conceptual resource for analyzing the relationship between institutional organization and individual identity. The contrast between the total institution and the deathworld is an analytical device that illuminates the specific character of harm produced by non-coordination rather than by comprehensive control.

18.  On the composite method in ethnographic research, see Stack (1974) and, for a more recent methodological discussion, Lofland et al. (2006:63–78). The composite method is not simply a privacy protection device; it is an analytical strategy for making visible the structural patterns that underlie individual variation, allowing the researcher to construct a 'typical case' that reveals the generative structure more clearly than any individual case could.

19.  The VI-SPDAT (Vulnerability Index – Service Prioritization Decision Assistance Tool) was developed by OrgCode Consulting and Community Solutions as a rapid assessment instrument for Coordinated Entry systems. It has been widely adopted across U.S. Continuum of Care programs but has also been subject to significant criticism; see Brown et al. (2018) for evidence of racial bias in VI-SPDAT scoring, and Kube et al. (2019) for reliability and validity concerns. The instrument's fundamental architectural limitation for purposes of the present analysis is its static character: it captures a moment, not a trajectory.

20.  The circular documentation problem—cannot obtain ID without documentation, cannot obtain documentation without ID—is a well-documented structural barrier in homeless services. See National Homelessness Law Center (2019) for a legal analysis; see Henwood et al. (2015) for an empirical examination of documentation barriers in Coordinated Entry systems.

21.  On professional vision in street outreach specifically, see Mellinger (2025a), which develops the concept of outreach workers' distinctive ways of seeing through Goodwin's framework of professional vision and applies it to the politics of recognizing and responding to homelessness. The term 'professional vision' originates in Goodwin (1994); Mellinger's application extends it to the relational and longitudinal dimensions specific to street outreach practice.

22.  Cicourel (1968, 1974). Cicourel's concept of typification draws on Alfred Schutz's phenomenological sociology while extending it to the specific cognitive operations of institutional actors. His early work on juvenile justice (1968) demonstrated that the categories through which juvenile justice workers classified young people had powerful self-fulfilling properties—the categories shaped intervention, which shaped outcomes, which confirmed the original categorization. The parallel in homeless services is direct: the categories through which case managers classify clients shape the interventions offered, which shape outcomes, which confirm or complicate the categorization.

23.  Lipsky (1980:xi–xix, 81–105). The coping strategies Lipsky documents are rational responses to structural conditions, not individual moral failures. This point is worth emphasis because the discourse around homeless service worker 'burnout' and 'compassion fatigue'—while not wrong—tends to individualize and medicalize what is, in Lipsky's framework, a structural phenomenon produced by the chronic gap between mandated obligations and available resources.

24.  The concept of epistemic abandonment is related to but distinct from Miranda Fricker's concept of 'epistemic injustice' (2007), which names conditions in which a person is wronged specifically in their capacity as a knower—either through testimonial injustice (having their credibility unjustly diminished) or hermeneutical injustice (lacking the conceptual resources to understand their own experience). Epistemic abandonment as used here refers not to the wrong done to the outreach worker whose knowledge is discounted, but to the structural arrangement through which the knowledge most relevant to preventing harm is institutionally inaccessible to the actors empowered to act on it.

25.  The moral injury literature, developed primarily in the context of military combat and, more recently, healthcare (see Litz et al. 2009; Talbot and Dean 2018), identifies moral injury as distinct from post-traumatic stress: where PTSD involves the psychological aftermath of overwhelming fear or horror, moral injury involves the psychological aftermath of 'perpetrating, failing to prevent, bearing witness to, or learning about acts that transgress deeply held moral beliefs.' Long-term outreach workers occupy a distinctive moral injury position: they witness preventable deterioration and death that they have not perpetrated, cannot prevent, and have attempted to avert through appropriate professional action. The injury comes from competent helplessness—from doing everything right and watching it not be enough.

26.  Zimmerman (1969). The study was conducted at a public welfare agency in the Los Angeles area. It remains one of the foundational demonstrations of how organizational record-keeping practices constitute, rather than merely record, organizational reality—showing that caseworkers' files were organized around what the institution needed to accomplish rather than around accurate representation of clients' circumstances. Zimmerman's later work on practical reasoning and rule use (1970) extends this analysis to the general problem of how members accomplish compliance with organizational rules through situated judgment rather than mechanical application.

27.  Schutz (1967:39–96). The relevance of Schutz's phenomenology to institutional analysis has been developed most fully by Heritage (1984:35–75) and, in the context of time and social organization specifically, by Zerubavel (1981). For the application to healthcare temporality and patient experience specifically, see Frankenberg (1992).

28.  The waitlist figure of approximately four years for permanent supportive housing in Santa Barbara County is drawn from the author's familiarity with the Coordinated Entry system and is consistent with publicly available CoC performance data. The political dynamics of supportive housing siting resistance in Santa Barbara—documented most visibly in the repeated defeat of proposed Navigation Center locations in 2019–2022—reflect the growth machine's resistance to land uses that do not serve exchange value, which Molotch's framework predicts.

29.  Martin v. City of Boise, 920 F.3d 584 (9th Cir. 2019). The decision held that the Eighth Amendment's prohibition on cruel and unusual punishment bars cities from criminalizing the act of sleeping in public when no adequate shelter alternative is available. The decision established a legal framework for encampment management that requires advance notice and prohibits certain punitive measures, but does not prohibit encampment removal when shelter is available, and does not address the destruction of belongings during removal.

30.  The Houston model is documented in Khadduri et al. (2010) and, most recently, in the Community Solutions Built for Zero initiative, which Houston anchors. The key structural features—unified HMIS data across all participating organizations, a 'by-name list' that makes every individual's trajectory visible in real time to all actors, and a governance structure with a designated 'system owner' accountable for overall outcomes—are precisely the cross-institutional accountability infrastructure that the paradox of rational accumulation analysis identifies as necessary.

31.  Lipsky (1980:83–105). The creaming phenomenon has been documented extensively in employment services, welfare programs, and education; see Dias and Maynard-Moody (2007) for a review. In homeless services specifically, evidence of informal prioritization of more 'manageable' clients in resource allocation decisions has been documented by Willse (2015) and Padgett, Henwood, and Tsemberis (2016).

32.  Mellinger (2026b). The death regime concept draws on Mbembe's necropolitics but specifies the mechanism as institutional non-decision rather than sovereign decision—the accumulation of locally rational choices that produce structured abandonment without any actor intending that outcome. See also Mellinger (2026c) on conditional dignity and the organization of care, which develops the related concept of 'ordinary cruelty' as the moral phenomenology of this arrangement.

33.  Mellinger (2026d). The documentary constitution analysis developed there traces how HMIS data architecture constitutes homeless individuals as partial persons in ways that systematically underserve the highest-acuity cases. The present section synthesizes that analysis within the broader paradox of rational accumulation framework.

References

Boden, Deirdre. 1994. The Business of Talk: Organizations in Action. Cambridge: Polity Press.

Brown, Molly, et al. 2018. "Reliability and Validity of the VI-SPDAT in Discriminating Housing Program Placements." Journal of Social Distress and Homelessness 27(2):109–117.

Community Solutions. 2021. Built for Zero: Progress Report. New York: Community Solutions. Available at: community.solutions/built-for-zero.

Cicourel, Aaron V. 1968. The Social Organization of Juvenile Justice. New York: Wiley.

Cicourel, Aaron V. 1974. Cognitive Sociology: Language and Meaning in Social Interaction. New York: Free Press.

Dear, Michael J., and Jennifer Wolch. 1987. Landscapes of Despair: From Deinstitutionalization to Homelessness. Princeton: Princeton University Press.

Culhane, Dennis P. 2008. "The Costs of Homelessness: A Perspective from the United States." European Journal of Homelessness 2(1):97–114.

DeVault, Marjorie L., and Liza McCoy. 2006. "Institutional Ethnography: Using Interviews to Investigate Ruling Relations." Pp. 15–44 in Institutional Ethnography as Practice, edited by Dorothy E. Smith. Lanham, MD: Rowman & Littlefield.

Dias, Janice Johnson, and Steven Maynard-Moody. 2007. "For-Profit Welfare: Contracts, Conflicts, and the Performance Paradox." Journal of Public Administration Research and Theory 17(2):189–211.

Frankenberg, Ronald. 1992. "Your Time or Mine? An Anthropological View of the Tragic Temporal Contradictions of Biomedical Practice." Pp. 1–29 in Time, Health and Medicine, edited by Ronald Frankenberg. London: Sage.

Fricker, Miranda. 2007. Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press.

Garfinkel, Harold. 1967. Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall.

Gilmore, Ruth Wilson. 2007. Golden Gulag: Prisons, Surplus, Crisis, and Opposition in Globalizing California. Berkeley: University of California Press.

Goffman, Erving. 1961. Asylums: Essays on the Social Situation of Mental Patients and Other Inmates. Garden City, NY: Anchor Books.

Henwood, Benjamin F., et al. 2015. "Examining Provider Perspectives on Barriers to Coordinated Entry for Homeless Individuals with Substance Use Disorders." Journal of Substance Abuse Treatment 57:91–96.

Goodwin, Charles. 1994. "Professional Vision." American Anthropologist 96(3):606–633.

Kube, Andrew, Sanmay Das, and Patrick W. Fowler. 2019. "Allocating Interventions Based on Predicted Outcomes: A Case Study on Homelessness Services." Proceedings of the AAAI Conference on Artificial Intelligence 33:622–629.

Lipsky, Michael. 1980. Street-Level Bureaucracy: Dilemmas of the Individual in Public Services. New York: Russell Sage Foundation.

Litz, Brett T., et al. 2009. "Moral Injury and Moral Repair in War Veterans: A Preliminary Model and Intervention Strategy." Clinical Psychology Review 29(8):695–706.

Lofland, John, et al. 2006. Analyzing Social Settings: A Guide to Qualitative Observation and Analysis. 4th ed. Belmont, CA: Wadsworth.

Logan, John R., and Harvey L. Molotch. 1987. Urban Fortunes: The Political Economy of Place. Berkeley: University of California Press.

Molotch, Harvey. 1976. "The City as a Growth Machine: Toward a Political Economy of Place." American Journal of Sociology 82(2):309–332.

Mellinger, Wayne Martin. 2026. "Rising Tide of Homeless Deaths a Tragic Reality We Must Face." Noozhawk, February 23. Available at: noozhawk.com.

Mellinger, Wayne Martin. 2024. "The Two Sides of Tourism in Santa Barbara." Noozhawk, May 18. Available at: noozhawk.com.

Mellinger, Wayne Martin. 2026b. "The Homeless System as Death Regime: An Ethnomethodological Re-Specification." Doing Modernity, January 28. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 2026c. "Conditional Dignity and the Organization of Care: The Machinery of Ordinary Cruelty." Doing Modernity, January 28. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 2026d. "The Documentary Constitution of Organizational Reality: A Research Program." Doing Modernity, March 14. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 2026e. "The Organization of Abandonment: Homelessness and the Institutional Production of Premature Death." Doing Modernity, March 9. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 2026f. "Abandonment in Real Time: Recognition, Rupture, and the Moral Economy of Interaction." Doing Modernity, January 13. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 2026g. "The Career of Disposability: Street Outreach, Fentanyl, and the Sociology of Abandonment." Doing Modernity, January 13. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 2025a. "Seeing Like an Outreach Worker: Professional Vision and the Politics of Homelessness." Doing Modernity, December 28. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 2025b. "The Layered Moral Economy of Oppression: Completing the Santa Barbara School's Unfinished Revolution." Doing Modernity, November 20. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 2025c. "Harvey Molotch and the Foundations of Critical Interactionism." Doing Modernity, July 3. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 2025d. "Deirdre Boden and Harvey Molotch: The Compulsion of Proximity." Doing Modernity, November 3. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 2025e. "When the Doctor is a 'Lady': Candace West, the Santa Barbara School, and the Micro-Politics of Gender." Doing Modernity, December 31. Available at: doingmodernity.blogspot.com.

Mellinger, Wayne Martin. 1992a. "Talk-as-Work: The Case of Paramedic Calls for Emergency Field Orders." Current Research in Occupations and Professions 7.

Mellinger, Wayne Martin. 1992b. "Accomplishing Fact in Police 'Dispatch Packages': An Analysis of the Situated Construction of an Organizational Record." Perspectives on Social Problems 4.

Mellinger, Wayne Martin. 1994. "Negotiated Orders: The Negotiation of Directives in Paramedic-Nurse Interactions." Symbolic Interaction.

Mykhalovskiy, Eric, and Liza McCoy. 2002. "Troubling Ruling Discourses of Health: Using Institutional Ethnography in Community-Based Research." Critical Public Health 12(1):17–37.

National Homelessness Law Center. 2019. Housing Not Handcuffs: Ending the Criminalization of Homelessness in U.S. Cities. Washington, DC: National Homelessness Law Center.

Padgett, Deborah K., Benjamin F. Henwood, and Sam J. Tsemberis. 2016. Housing First: Ending Homelessness, Transforming Systems, and Changing Lives. New York: Oxford University Press.

Puar, Jasbir K. 2007. Terrorist Assemblages: Homonationalism in Queer Times. Durham, NC: Duke University Press.

Ritzer, George. 1993. The McDonaldization of Society. Thousand Oaks, CA: Pine Forge Press.

Ritzer, George, ed. 1998. The McDonaldization Thesis: Explorations and Extensions. London: Sage.

Schutz, Alfred. 1967. The Phenomenology of the Social World. Evanston, IL: Northwestern University Press.

Smart, Barry, ed. 1999. Resisting McDonaldization. London: Sage.

Smith, Dorothy E. 1987. The Everyday World as Problematic: A Feminist Sociology. Boston: Northeastern University Press.

Smith, Dorothy E. 2005. Institutional Ethnography: A Sociology for People. Lanham, MD: AltaMira Press.

Stack, Carol B. 1974. All Our Kin: Strategies for Survival in a Black Community. New York: Harper & Row.

Stuart, Forrest. 2016. Down, Out, and Under Arrest: Policing and Everyday Life in Skid Row. Chicago: University of Chicago Press.

Sudnow, David. 1967. Passing On: The Social Organization of Dying. Englewood Cliffs, NJ: Prentice-Hall.

Talbot, S.G., and Wendy Dean. 2018. "Physicians Aren't 'Burning Out.' They're Suffering from Moral Injury." STAT News, July 26.

Timmermans, Stefan. 1999. Sudden Death and the Myth of CPR. Philadelphia: Temple University Press.

Wacquant, Loïc. 2008. Urban Outcasts: A Comparative Sociology of Advanced Marginality. Cambridge: Polity Press.

Weber, Max. 1978. Economy and Society: An Outline of Interpretive Sociology. 2 vols. Edited by Guenther Roth and Claus Wittich. Berkeley: University of California Press.

Weber, Max. 2002. The Protestant Ethic and the Spirit of Capitalism. Translated by Peter Baehr and Gordon C. Wells. New York: Penguin.

Willse, Craig. 2015. The Value of Homelessness: Managing Surplus Life in the United States. Minneapolis: University of Minnesota Press.

Zerubavel, Eviatar. 1981. Hidden Rhythms: Schedules and Calendars in Social Life. Chicago: University of Chicago Press.

Zimmerman, Don H. 1969. "Record-Keeping and the Intake Process in a Public Welfare Agency." Pp. 319–354 in On Record: Files and Dossiers in American Life, edited by Stanton Wheeler. New York: Russell Sage Foundation.

Zimmerman, Don H. 1970. "The Practicalities of Rule Use." Pp. 221–238 in Understanding Everyday Life, edited by Jack D. Douglas. Chicago: Aldine.