The Paradox of Rational Accumulation:

Local Reason, Systemic Harm, and a Theory of Institutional Life in Late Modernity

Wayne Martin Mellinger, Ph.D.

Merged draft — for review and revision

Abstract

Modern institutions routinely produce outcomes that virtually none of their participants explicitly intend and that few would defend if asked to evaluate them as wholes. Hospitals discharge medically stable patients. Shelters enforce intake windows. Cities remove infrastructure in response to isolated complaints. Security systems authenticate users according to protocol. Each decision may be reasonable within the institutional setting in which it occurs. Yet when such decisions accumulate across organizations, time, and space, their combined consequences can become exclusionary, absurd, or lethal.

This article develops the concept of the Paradox of Rational Accumulation: the condition in which actions that are locally rational within bounded institutional frameworks combine sequentially to generate systemically irrational, and sometimes fatal, outcomes. The concept is situated among a wide family of adjacent frameworks — Weberian rationalization, Merton's unintended consequences, Simon's bounded rationality, Kahn's tyranny of small decisions, Hardin's tragedy of the commons, Perrow's normal accidents, Beck's organized irresponsibility and risk society, Arendt's banality of evil, Galtung's structural violence, and Bauman's adiaphorization — and distinguished from each. Its distinctive claim is that responsibility, knowledge, and accountability remain organizationally fragmented and vertical, while consequences travel horizontally across institutional boundaries and accumulate within environments, public space, and human biographies. Situated, discretionary interactional repair — the capacity for someone to say, mid-sequence, "I see what happened here, let's fix this" — is not a precondition for the paradox to occur, since sedimentation and path dependence can accumulate harm even where full human discretion remains formally intact. Rather, effective repair capacity functions as a continuous variable: as it declines relative to the speed and complexity of an accumulating sequence, the paradox becomes both more likely and harder to interrupt, with fully automated systems representing its limiting case rather than its defining one.

The article develops seven mechanisms constituting the paradox's architecture and four distinct modes — additive, sequential, sedimentary, and recursive — by which accumulation occurs. It then traces the concept's own developmental arc across six published cases: environmental exposure at the Andrée Clark Bird Refuge; the disappearance of public infrastructure on East Cota Street; the explicit theoretical statement in "How Santa Barbara's Institutions Produce Harm Without Intending To"; the death of a composite unhoused man, Rafael; the sedimentation of the 500 block of State Street; and the author's own repeated, automated exile from his Apple account. The article closes by asking what it would take to elevate the concept from a diagnostic tool for homelessness policy into a general theory of institutional life under late modernity, and states explicitly where the theory should be expected to fail.

I. Introduction

There is a form of social harm that is peculiarly difficult to diagnose because nearly everyone involved is behaving reasonably.

A physician discharges a patient who no longer requires hospitalization. A city removes a bench after merchant complaints about loitering. A grocery store security guard asks a weakened man to move along from a low wall outside the store. A housing program requires identification before allocating a scarce unit. A digital security system, following its own logic precisely, concludes that a stolen phone and a mismatched bank card add up to an impostor.

None of these actions is intrinsically absurd. Many are professionally responsible, some legally mandated. Examined separately, each requires little explanation.

But people do not live inside separate decisions. They live through sequences.

The patient leaving the hospital encounters the shelter's intake window. The bench's removal changes what the next tired body finds when it needs to rest. The lost identification, once gone, becomes the reason a housing application stalls for months. The effects of one institution become the conditions under which the next institution acts — and, in the sharpest cases, the conditions the same institution later treats as evidence against the very person its own earlier action helped produce.

I call this process the Paradox of Rational Accumulation. At its most elementary:

Individually rational actions can accumulate into collectively irrational, and sometimes lethal, outcomes.

Stated this simply, the proposition is not new — social theory has investigated the gap between intention and consequence, individual and collective rationality, procedural and substantive reason, for a century. The theoretical task of this essay is therefore not to claim novelty for that general observation. It is to answer a narrower question: what particular social mechanism does Rational Accumulation identify that Merton, Simon, Kahn, Hardin, Arendt, Galtung, Bauman, and Beck do not, individually, fully capture?

My answer is that the distinctive object is sequential institutional causation under fragmented, vertical accountability, whose severity and durability are governed by the degree to which situated, discretionary repair capacity remains available to interrupt it. Modern institutions differentiate labor, and differentiation is largely responsible for their power — hospitals know medicine because they concentrate medical expertise; security systems protect accounts because they refuse ad hoc trust. But differentiation divides more than labor. It divides vision, time, and responsibility, and it divides the whole human being into institutionally legible fragments: a clinical status, an eligibility category, an intake window, a verification signal. No single institution is necessarily positioned to see the trajectory produced by their combination. Where a person remains empowered to say, mid-sequence, I see what happened here — let's fix this, the paradox can often be interrupted before it completes; as that capacity weakens — through consistency pressure, scale, or full automation — the paradox becomes progressively harder to stop, though its underlying mechanism, sequential accumulation under fragmented accountability, can operate even where repair remains formally available but is not exercised, as the sedimentation cases below show.

The concept did not arrive fully formed. It emerged across six essays over five months, first in essays about vanishing shade and vanishing benches, then explicitly in an essay about why people die in Santa Barbara's creek beds at rates that should stop a community cold, then in an essay about a city block, and finally — without my having planned it — in an essay about twelve days I spent locked out of my own digital identity. The concept's refusal to stay confined to homelessness services is itself part of the argument this essay makes: something durable is being tracked here, not a description limited to one policy domain.

The six cases differ enormously in moral gravity. A digital lockout is not equivalent to a death in a creek bed, and lost vegetation is not morally interchangeable with institutional abandonment. They are compared here at the level of mechanism, not magnitude. That their diversity nonetheless yields the same structural pattern — bounded rationality, fragmented perception, horizontal transmission, sequential or recursive accumulation, and varying capacities for repair — is the empirical case for treating Rational Accumulation as more than a local description of one city's homelessness crisis.

II. Rationality: Rational for Whom, and for What?

The first theoretical task is to specify what "rational" means, since without clarification the concept risks becoming too loose to test.

Instrumental rationality asks whether an action efficiently advances an assigned objective. A hospital discharges a medically stable patient to preserve beds for patients requiring acute care — instrumentally sound.

Institutional rationality concerns conformity to authorized procedure, mandate, budget, and jurisdiction. A shelter enforcing its intake window consistently is following established procedure, whatever the weather that night.

Substantive rationality, in the Weberian sense, asks a broader question: is the resulting arrangement reasonable given its actual consequences for human beings, public purposes, or democratic values?

These three can diverge sharply. An action can be instrumentally efficient, procedurally correct, and substantively disastrous — and this divergence, rather than any single actor's rule-following, is the crucial site of analysis. This places Rational Accumulation in a recognizably Weberian lineage: the power of bureaucratic rationality arises from abstraction, calculability, and impersonality, which protect against arbitrary decision-making at enormous scale, and the familiar danger is that a rule becomes rational in relation to itself while detached from the substantive purpose the institution exists to serve.

Rational Accumulation extends that danger beyond the single bureaucracy. The core difficulty is not merely that one organization rigidly applies its rules. It is that multiple institutions can each behave rationally according to different institutional purposes, none of which is the person's whole trajectory, while collectively producing an outcome none of them measures as its own performance.

III. The Theoretical Neighborhood

III.1 Merton and unintended consequences

Merton's 1936 analysis of purposive social action established that intentions cannot be treated as adequate explanations of outcomes. Rational Accumulation belongs to this tradition but is narrower: it specifies one mechanism by which the discrepancy is produced — consequences generated by multiple decisions accumulate across boundaries that organize decision-making but do not contain those consequences. There is also a temporal claim Merton's framework does not itself supply: an outcome can begin genuinely unanticipated, but recurrence produces knowledge, and once a harm is recurrent, measured, and predictable, "unintended consequence" begins to function as an alibi rather than a description.

III.2 Simon and bounded rationality

Simon's actors possess limited information, processing capacity, and institutionally structured objectives. This supplies an important micro-foundation: the housing worker need not know the applicant's full medical history, the physician need not know housing eligibility rules, the software engineer need not know the biography of the eventual user encountering an authentication protocol. Each bounded frame may be administratively necessary. Rational Accumulation asks what Simon's framework does not: what happens when many locally satisfactory decisions, each generated within a separate bounded frame, are transmitted sequentially into one another?

III.3 Hardin and the tragedy of the commons

Hardin's tragedy describes rational actors overexploiting a shared, unowned resource because the cost of overuse is diffused while the benefit of use is private. The driver is a misalignment between individual incentive and a depletable collective resource. Rational Accumulation has no equivalent shared resource and no competitive self-interest: the hospital discharge planner and the grocery store manager in the founding case are not competing for anything, several actors in these chains are actively trying to help, and there is no move either could make that would constitute cooperation or defection with respect to the other. The mechanism is sequential and institutional, not competitive and resource-based.

III.4 Collective-action models generally

The Prisoner's Dilemma and related frameworks demonstrate that rational self-interested actors, unable to coordinate, can reach a jointly worse outcome than cooperation would produce. Rational Accumulation does not require selfishness or coordination failure in this sense. The hospital physician may care deeply; the outreach worker may expend extraordinary effort; the Apple employee may fully believe the customer's story. Benevolence does not by itself alter the institutional architecture. This distinction carries a direct policy implication: if selfishness is the cause, the remedy is changed incentives; if fragmented institutional rationality is the cause, installing kinder people into an unchanged structure will reproduce essentially the same outcome.

III.5 Kahn and the tyranny of small decisions

Kahn's 1966 formulation is the closest established relative and demands the clearest differentiation. Kahn describes small, individually reasonable, temporally limited decisions cumulating into a larger allocation people might reject if offered the chance to choose the aggregate directly. The resemblance is real and should not be minimized. The distinction lies in architecture: Kahn's concern is largely market allocation and simple aggregation, while Rational Accumulation gives analytical weight to institutional differentiation, cross-organizational sequence, the transmission of consequences from one jurisdiction into another, feedback, biographical accumulation, temporal mismatch, and unequal capacity to absorb consequences. The resulting structure is often not A + B + C = undesirable D, but rather: A changes the conditions under which B occurs; B alters the subject who later encounters C; C treats the consequences of A and B as new, apparently self-generated facts. This is trajectory, not mere aggregation.

III.6 Path dependence and Perrow's normal accidents

Path dependence explains how historical sequences constrain later alternatives and generate durability — the State Street case below is a clear instance — but it is a mechanism by which Rational Accumulation becomes entrenched, not the whole account of how the accumulating decisions were themselves locally rational. Perrow's normal accidents, meanwhile, explain how failures interact catastrophically in tightly coupled systems. Rational Accumulation is frequently concerned with how successes interact: the hospital successfully discharges, the shelter successfully manages scarcity, the security system successfully blocks unverified access. Component success does not guarantee — and can actively undermine — systemic outcome.

III.7 Arendt and the banality of evil

Arendt's Eichmann is the canonical figure closest to the actors in these cases, and the debt deserves to be named directly rather than left implicit. Both frameworks insist that catastrophic harm does not require monstrous intent and locate danger in bureaucratic role performance rather than individual malice. The distinction is in what is being suppressed. For Arendt, the danger is the suspension of thinking itself — a specific failure to imaginatively occupy another's position, in service of an ideological project. The actors in the Paradox of Rational Accumulation are frequently thinking a great deal, and thinking well, within their local frame: the discharge planner is not failing to think, she is thinking correctly about bed capacity. What is missing is not thought but scope. This makes the paradox, in one sense, the more troubling diagnosis: it does not require even Arendt's thoughtlessness. It can run entirely on people doing their jobs conscientiously.

III.8 Galtung and structural violence

Galtung's structural violence — harm built into the ordinary operation of social and political structures such that no identifiable actor need commit a violent act for the harm to occur — is the nearest true theoretical ancestor of this concept, closer than any of the frameworks above. Both reject the requirement of an intending harm-doer. Rational Accumulation's distinctive contribution to the Galtung lineage is mechanistic specificity: where structural violence is typically described at the level of whole systems (economic distribution, political exclusion), this concept specifies the sequential, interactional mechanism — bounded rationality, fragmented perception, horizontal transmission, and variable repair capacity — by which structural violence is produced turn by turn, in the observable, mundane interactions of street-level institutional life. It is Galtung's insight given an ethnomethodological grain.

III.9 Bauman and adiaphorization

Bauman's adiaphorization — the process by which bureaucratic division of labor renders actions morally neutral by fragmenting them across actors none of whom perceives the whole moral act — is the concept's closest sibling after Galtung, and it is worth stating plainly that any account of Rational Accumulation that omits Bauman is missing its nearest relative. Where this concept extends Bauman is in refusing the premise of moral neutrality: the actors in these cases do not experience their actions as morally neutral so much as affirmatively reasonable, even caring. The paradox occurs even when moral engagement at the local level is fully intact — adiaphorization is not a precondition for it.

III.10 Beck: risk society and organized irresponsibility

Beck's organized irresponsibility — institutions collectively producing consequential risk while responsibility for it becomes structurally difficult to assign — comes closest of all the frameworks reviewed here to this essay's political implications, and overlaps substantially with what might be called distributed innocence: a situation in which significant harm is produced without an easily identifiable culpable actor, because each participant can give an institutionally adequate account of their own behavior. But distributed innocence should describe an empirical structure, not confer moral absolution indefinitely. Once accumulated consequences become repeatedly visible, continuing to organize responsibility so that no one owns them begins to resemble Beck's organized irresponsibility in the fullest sense. Hence the essay's governing ethical principle, developed further in Section VII: distributed innocence has an expiration date.

III.11 Mills and the sociological imagination

Mills's distinction between personal troubles and public issues supplies the diagnostic posture the paradox requires, rather than a competing causal mechanism. Without it, every case in this essay reads as a private misfortune: Rafael's inability to secure an ID, my own inability to regain access to my Apple account. The Millsian turn — from trouble to issue — is the specific rhetorical and analytical move each of the essays discussed below performs, and it is why each begins in the first-person, concrete register (a name, a Tuesday morning, a shopping bag on a bicycle handlebar) before generalizing.

III.12 Lipsky, and the move from street-level to system-level bureaucracy

Lipsky's street-level bureaucrats exercise discretion under scarcity and ambiguity in translating policy into encounters with citizens; that discretion is where informal repair has traditionally lived. Bovens and Zouridis describe the historical movement from street-level toward system-level bureaucracy, in which information technology relocates discretion away from frontline encounters and into software and system architecture itself. This literature is essential to the sixth case below, and supplies the concept of repairability developed in Section VIII: the question of whether a rationalized system retains any viable mechanism for recognizing that a given sequence of individually correct steps has, in this instance, produced an absurd or harmful result.

Synthesis

No single predecessor captures the combination this concept requires: harm produced through success rather than failure (unlike simple system breakdown); no competitive misalignment (unlike the commons) or coordination failure (unlike the Prisoner's Dilemma); no suspension of thought (unlike Arendt) and no requirement of moral neutralization (unlike Bauman, its closest kin); a mechanism specified at the level of sequence and interaction rather than only system or structure (extending Galtung); and an account of a crucial variable governing whether the mechanism is interrupted or allowed to run to completion: the availability of effective discretionary repair.

IV. The Architecture of Rational Accumulation

Seven mechanisms recur across all six cases examined below.

1. Bounded rationality. No single actor — physician, planner, housing worker, sanitation employee, security engineer, outreach worker — possesses complete control over, or visibility into, the social trajectory their action contributes to.

2. Fragmented perception. Institutions do not merely divide responsibility; they construct different objects. The patient becomes clinically stable. The applicant becomes eligible or ineligible. The sidewalk becomes right-of-way. The account holder becomes a bundle of authentication signals. This selectivity is often administratively necessary. The person perceived remains larger than any category built to perceive them.

3. Horizontal transmission of consequences. Consequences move across organizational boundaries an institution has no obligation to track. Hospital discharge exports a care need into the street. Encampment clearance exports documentary loss into a housing system. A stolen phone exports authentication failure into a person's ability to recover his own account.

4. Sequential accumulation. Actions at time one alter the circumstances encountered at time two. The relevant analytic unit is therefore the sequence, not the isolated decision.

5. Temporal mismatch. Institutions organize time administratively — applications take weeks, processing has windows, benefits require documentation. Human need operates biologically — infection worsens daily, hunger returns tonight, exposure happens every night regardless of where an application sits in a queue. This is the gap between administrative time and biological time, and it is frequently causal rather than incidental.

6. Differential absorptive capacity. Institutional friction does not burden everyone equally. A missing bench is negligible to someone who can walk another mile in comfortable shoes; it can be decisive for someone whose body cannot. A six-day digital lockout is an inconvenience to someone with a backup device, a spare card, and money; it is a serious event for someone for whom that account mediates banking, work, and communication. The harm produced by Rational Accumulation is therefore stratified by resources that have nothing to do with the merits of any individual decision in the chain.

7. Fragmented, vertical accountability. Organizations are accountable predominantly upward — workers to supervisors, programs to funders, departments to regulators — and performance metrics typically measure component transactions. But harm travels horizontally, across the boxes on an organizational chart, following the person rather than the program. Hence the essay's central structural claim:

Accountability is vertical while cumulative harm is horizontal.

No single institution necessarily owns what the institutional ecology, taken as a whole, produces.

V. Four Forms of Accumulation

Across the six cases, at least four distinct modes of accumulation can be distinguished, and collapsing them into a single "small things add up" intuition understates the theory's precision.

Additive accumulation. Multiple independent actions collectively produce a larger result with no interaction between them. No single tree removal transforms the Bird Refuge's ecology; enough removals do.

Sequential accumulation. Earlier actions change the circumstances under which later actions occur. Hospital discharge becomes the situation the shelter's intake window later encounters.

Sedimentary accumulation. Temporary arrangements acquire durability through repetition, investment, and infrastructure, until their emergency origin recedes from view. Emergency parklets become permanent, city-funded "pedlets."

Recursive accumulation. A system's own earlier response becomes an input treated by the same system as evidence justifying a further, harsher response. A failed verification attempt generates heightened suspicion, which restricts recovery options, which produces a further failed attempt, which deepens suspicion again.

Recursive Rational Accumulation occurs when the consequences of an institutional response become inputs into subsequent institutional decisions, creating a feedback loop in which earlier system actions intensify the very condition they were meant to manage.

Sometimes decisions simply add. Sometimes they form a trajectory. Sometimes they harden into infrastructure. Sometimes they feed back on themselves. The cases below exhibit all four.

VI. Empirical Review: Six Cases and a Self-Correction, in Publication Order

The concept did not arrive fully formed; it was built essay by essay, and its own publication sequence is instructive, since each case forced a refinement the previous one had not required.

VI.1 The Andrée Clark Bird Refuge (March 23, 2026) — additive accumulation, first sighting

The earliest essay does not name the concept. It describes a wildlife refuge that has lost the vegetative cover letting birds — and, implicitly, people — disappear into shadow, driven substantially by the Highway 101 widening project along the South Coast: bridges rebuilt, utilities relocated, mature trees felled along the corridor, with replacement saplings offering no near-term substitute for a mature coastal canopy. No single maintenance or safety decision is unreasonable. The essay's own diagnosis is direct: "The people with the least power experience the most exposure. The people with the most power enjoy the greatest privacy." This is the paradox's first appearance — additive, aesthetic and ecological in register — and that lower register is what let the mechanism become visible before it required a name.

VI.2 The Bench on Cota Street That Isn't There (March 30, 2026) — subtraction as method

The second essay narrows to a single bus stop: benches, shelter, and commemorative tiles at East Cota Street, gone. The essay names the formal property directly — the removal "operates through subtraction," requiring no ordinance and no named target. This introduces functional erasure: an activity (waiting, resting) disappears not because it is prohibited but because the infrastructure that supported it disappears, decision by defensible decision.

VI.3 How Santa Barbara's Institutions Produce Harm Without Intending To (April 13, 2026) — the naming

The theoretical essay proper, explicitly retroactive: "I introduced this concept in my earlier essays on the Andrée Clark Bird Refuge and the East Cota Street benches, where the stakes were civic: lost shade, lost seating, lost refuge. Here the stakes are different. Here we are talking about whether people live or die." It lays out the founding chain — hospital discharge, city bench removal after merchant complaints, a security guard's request to move along — terminating in death, and states the core inversion this essay formalizes in Section II: "When we find a failure, we can fix it... But the paradox is not produced by failures. It is produced by successes." This is successful failure: the phrase does not imply secret intentional abandonment, only that locally successful performances can combine into systemic catastrophe.

VI.4 Visionaries, Guardians, Critics — and Benches — in Santa Barbara (April 20, 2026) — the self-correction

A necessary companion piece, and evidence the concept can absorb correction without collapsing. Former Mayor and longtime Planning Commissioner Sheila Lodge — at 96, a figure the essay compares to Pearl Chase in her civic authority — clarified that the bench planned for the new police station site on East Cota Street had never been removed from the plan; it had been there from the beginning, reviewed by the Planning Commission, and would be installed on completion. The essay accepts this correction unequivocally while distinguishing it from the separate, confirmed removal of benches at the sheltered stop by the Smart & Final lot a block and a half away. This matters theoretically as much as journalistically: the argument for Rational Accumulation cannot rest on any single misidentified case. Its force comes from the recurring pattern, not from any one bench.

VI.5 Everyone Was Being Reasonable, and He Died Anyway (April 27, 2026) — the limit case, embodied accumulation

The paradox run to completion in a documented life. Rafael was not invisible: on the county database, holding a qualifying vulnerability score, engaged weekly by outreach workers for over a year, hospitalized twice, the subject of referrals and documented services. The mechanism that defeats him is bureaucratically mundane and sequential — an ID lost twice, once to theft, once to an encampment sweep; a replacement requiring a certified birth certificate from a distant county, a written request, a fee, proof of identity he could not easily supply because he lacked identification, and a four-to-six-week window reset twice by hospitalization or cleanup before it could complete. The essay grounds the structural driver in Harvey Molotch's growth-machine theory, developed at UC Santa Barbara: land's exchange value systematically outcompeting its use value for people like Rafael. The case introduces embodied accumulation — administrative delay becomes infection, displacement becomes sleep deprivation, documentary failure becomes additional weeks of exposure — and the corresponding methodological caution: a service encounter can be successfully completed by an institution's own metrics while the person's underlying trajectory worsens. Receiving services is not equivalent to receiving an effective trajectory toward survival. The essay also supplies this project's clearest counter-case and proof of concept: Houston's unified cross-institutional data system and joint accountability structure reduced its unsheltered population by more than 60 percent over thirteen years by "reorganizing accountability" rather than working harder inside the existing fragmented architecture.

VI.6 How the 500 Block of State Street Quietly Transformed Public Space (June 29, 2026) — sedimentary accumulation, civic and fiscal generalization

The fifth essay steps back from any single person to an entire city block, tracing the passage from emergency accommodation to permanent capital infrastructure: COVID-era parklets replaced, by September 2025, with permanent modular "pedlets" at a cost of roughly $590,000 in public funds — a figure independently confirmed by the city's own announcements and by contemporaneous local news coverage. The essay's formulation is the clearest statement in the corpus of sedimentary accumulation at urban scale: "The street we walk today is the accumulated result of many incremental decisions, most of them defensible in isolation, whose cumulative effect no single person ever intended." This is the case where competing claims on public space — diners, wheelchair users, cyclists, an unhoused man arranging his belongings near a planter "with the careful economy of someone for whom this stretch of pavement is not an amenity but a place to be" — are resolved by accumulated budget and permitting decisions whether or not the city understands itself to be resolving them.

VI.7 I Am Currently Serving My Second Sentence (forthcoming, publication expected the essay's stated Monday) — recursive accumulation, cross-domain generalization

The pivot case. Every prior essay involves the author as sociologist and outreach worker documenting institutional harm done to others; here the author is the subject. A water-damaged keyboard leads to a factory reset erasing stored device trust. A phone, lost and briefly recovered, is then genuinely stolen from a shopping bag. A bank card reissued after an unrelated fraud incident a year earlier no longer matches records on file. None of this is any single actor's fault; Apple's technicians, its fraud-detection system, and the author's carrier each behave exactly as designed. The essay's own diagnosis: "Each rule makes sense... Add one stolen phone, one wiped hard drive, one mismatched bank card and one increasingly frantic human being, and all that good sense produces madness." The lockout recurs — twice, twelve days total — because the sequence is recursive: missing phone weakens authentication, weakened authentication produces a failed attempt, the failed attempt heightens suspicion, heightened suspicion restricts recovery options, restricted options produce a further failure. The system's own prior responses become the evidence justifying its next response. And the essay supplies the clearest illustration of this project's central theoretical variable, discussed at length in Section VIII: "What the accumulation removes is discretion — and with it, the possibility of repair" — true of this case specifically, though not, as Section VIII argues, a condition the paradox requires generally. The case matters empirically beyond its theoretical yield because it demonstrates the mechanism is not specific to homelessness or to poverty: it happens to a Ph.D. sociologist with a California ID, receipts, devices, and "a coherent, if improbable, story" — every institutional advantage Rafael lacked — and still runs for twelve days.

Synthesis of the arc

Read in sequence, the cases trace a deliberate widening: aesthetic and ecological stakes (Bird Refuge) → civic stakes and self-correction (Cota Street, twice) → explicit theory (Institutions) → mortal stakes and embodiment (Rafael) → civic-and-fiscal, city-scale stakes (State Street) → personal, cross-class, recursive, cross-domain stakes (Apple). The concept's own publication history performs the claim this essay makes about it: a mechanism that does not stay confined to the domain in which it was first observed.

VII. Foreseeability and the End of Innocence

If individually reasonable practices produce harm nobody intended, how should responsibility be assigned? The first requirement is to distinguish causation from blame: not everyone who causally contributes to an outcome is equally blameworthy, and the absence of a villain does not eliminate responsibility.

The crucial variable is foreseeability. A harm can begin genuinely unknown. Recurrence makes it observable. Measurement makes it documentable. Institutional learning makes it foreseeable. Once serious harm becomes foreseeable, institutional obligation changes:

When serious recurrent harm is a foreseeable consequence of interacting institutional routines, responsibility for preventing that harm cannot be dissolved merely because no single routine is sufficient, alone, to produce it.

This moves the concept from a theory of unintended consequences toward a theory of structural responsibility. Distributed innocence may genuinely describe the moral field the first time a pattern occurs. The same arrangement, continued unchanged after the pattern is known and the capacity to redesign it exists, becomes distributed evasion. The appropriate normative standard is systemic non-abandonment: no division of institutional labor should be considered legitimate when its routine operation foreseeably leaves people without a viable pathway to basic safety and social membership. The question is no longer only did every institution follow its rules, but did the institutional ecology leave the person a viable path through them.

VIII. Accountability Architecture and the Variable Availability of Repair

If this diagnosis is right, the standard levers of institutional reform are frequently misdirected. Retraining a hospital worker does not repair the boundary between health care and housing. Asking outreach workers to try harder does not create housing inventory. Teaching a sanitation crew more compassion does not make displaced identification recoverable. Giving an Apple Store employee more empathy does not give that employee authority to override an automated lockout, however completely that employee believes the customer's story.

The central intervention has to occur at the level of accountability architecture — mechanisms that answer for what the institutional system produces across boundaries, not only within them. This requires institutions to ask not only was this transaction completed correctly but what happened next, and ultimately what happened to this person, place, or public good across the entire trajectory. That question is structurally difficult because databases are organized by program, budgets by agency, authority by jurisdiction, and performance metrics by funded activity — and human lives do not respect those divisions. Houston's reorganized, cross-institutional accountability structure remains the clearest available proof that this redesign is possible and effective, not merely aspirational.

For automated systems specifically, accountability architecture must include repairability: a viable mechanism for reviewing the context surrounding an anomalous case before that case completes. Human beings interact successfully not because misunderstanding is absent but because misunderstanding is structurally repairable — conversation analysis, in the tradition of Sacks, Schegloff, and Jefferson, has shown repair to be an organized, constitutive feature of ordinary talk, not an occasional exception to it. Bovens and Zouridis's account of the shift from street-level to system-level bureaucracy describes precisely how that repair capacity gets engineered out as discretion relocates from frontline workers into software architecture.

It is important, however, not to overstate repair's place in the theory. Repair capacity should be treated as a variable, not as a binary precondition built into the definition of the paradox itself — repair can be present in weakened, inaccessible, poorly resourced, or jurisdictionally limited forms, not only fully absent. The claim is not that Rational Accumulation requires the disappearance of repair; the Bird Refuge, Cota Street, and State Street cases all accumulate their harm through sedimentation and path dependence while officials retain full formal discretion to reverse course at any point — no repair mechanism has been engineered out of those cases, and yet the paradox operates in all three. The stronger and more defensible claim is:

Rational Accumulation becomes more likely, and harder to interrupt, as effective repair capacity declines relative to the speed and complexity of the accumulating sequence — with full automation representing its limiting case, not its defining one.

This formulation does real work the stricter version does not: it lets the theory cover cases of slow institutional sedimentation, where discretion is present but unexercised, as readily as it covers the Apple case, where discretion has been structurally removed. What is lost as repair capacity declines is not consistency's opposite; it is the system's mechanism for recognizing, mid-sequence, that a chain of individually correct steps has, in this instance, produced an absurd or lethal result. Human oversight that can only explain that nothing can be done is not meaningful discretion. Repairability means the capacity to alter, not merely the capacity to hear.

IX. Rational Accumulation as a Theory of Modernity

IX.1 The objection

A skeptical reader will ask whether this is simply Weber's rationalization thesis — the "iron cage" produced when formal, calculable procedure displaces substantive judgment — restated with new case studies attached. The objection deserves a direct answer rather than a gesture past it.

IX.2 What the macro-theories leave abstract

Weber, Beck, and Bauman each diagnose the macro-condition this essay is concerned with. None specifies the micro-interactional mechanism by which a rationalized system's particular sequence of ordinary encounters produces a particular, traceable harm to a particular person on a particular Tuesday. Weber's iron cage does not, on its own, explain how a hospital discharge, a bench removal, and a request to move along combine to end a specific life. Beck's risk society is built for systemic hazards like nuclear or climate risk, not a birth-certificate processing window. Bauman's adiaphorization comes closest, but, as argued in III.9, requires a moral neutralization these cases do not exhibit.

IX.3 The actual contribution

What this body of work offers the rationalization tradition is not a rival macro-theory but a bridge: a specification, grounded in the ethnomethodological tradition of Garfinkel, Sacks, and Schegloff — the tradition of the Santa Barbara School itself — of the exact interactional mechanism by which macro-rationalization becomes micro-harm. The concept of variable repair capacity is the hinge. It asks Weber's abstract iron cage a concrete question: what, specifically, at the level of a single exchange between a person and an institutional representative, determines whether the sequence can still be interrupted? The answer is visible only at the level Sacks and Schegloff worked at, applied to institutional settings — this is rationalization theory read through the interaction order, or, equivalently, ethnomethodology scaled up to explain outcomes rationalization theorists have described only in aggregate. Modern institutions become, in this account, sites of fragmented competence: every part knows more, and the whole knows less — not because participants are unintelligent, but as a direct consequence of the specialization that makes their local competence possible in the first place.

IX.4 Scope conditions

A theory that explains everything explains nothing. Honesty about limits is what will let this concept survive serious scrutiny rather than collapse under the first counter-example. The mechanism should operate with markedly less force, or not at all, under identifiable conditions:

Where discretionary repair capacity has been deliberately preserved or restored — Houston's reorganized accountability structure is the clearest available counter-case, targeting the paradox's precondition of bounded, vertical accountability directly.
Where relationships are long-term, small-scale, and face-to-face, of the kind classical community studies describe; discretionary repair tends to survive formalization better under these conditions than in large bureaucracies or automated systems, and this is a testable claim rather than an assumption.
Where actors work from genuinely shared, cross-institutional information rather than bounded, siloed records; better visibility across the chain should, in principle, interrupt accumulation earlier.

Stating these conditions is not a retreat from the theoretical ambition of this essay. It is what makes the ambition credible.

X. Discussion

Read against the full comparison set, the concept's distinctiveness can be stated precisely. It is not simply another name for unintended consequences, though it belongs to that family; it specifies a particular route by which consequence detaches from intention. It is not identical to bounded rationality, which explains incomplete individual perspectives without addressing what happens when many such perspectives act on the same trajectory. It is not the tragedy of the commons or the Prisoner's Dilemma, neither of which requires selfishness or coordination failure here. It is closest to Kahn, but extends aggregation into institutional sequence, biography, temporality, and recursive feedback. It is not path dependence, which explains durability rather than the local rationality that constitutes the accumulating arrangement in the first place. It differs from normal accidents because the relevant components frequently do not malfunction — they function precisely as designed. It shares the deepest kinship with Galtung's structural violence and Bauman's adiaphorization, extending the former with interactional mechanism and the latter by showing the paradox operates without moral neutralization. It is closer to Arendt than any economic model, while requiring less — not the suspension of thought, only the boundedness of scope. And it extends the literature on system-level bureaucracy by showing why the relocation of discretion into technological architecture matters sociologically: such systems can eliminate precisely the situated judgment through which abnormal but intelligible circumstances become repairable.

The theory's contribution therefore does not lie in the elementary equation rational + rational = irrational. It lies in specifying the institutional architecture that makes that equation recurrent: bounded jurisdiction, fragmented perception, horizontal consequence transmission, sequential and recursive causation, administrative-biological temporal mismatch, differential absorptive capacity, sedimentation, and — the concept's most original claim — the variable availability of effective repair capacity.

XI. Conclusion: Who Is Responsible for What the Pieces Add Up To?

The Bird Refuge did not become exposed because someone declared shelter undesirable. East Cota Street did not lose its benches through one decree abolishing rest — and where a specific claim about that loss proved partly wrong, the correction, offered by a 96-year-old former mayor with the standing to make it, only sharpened the argument rather than undermining it. Rafael's death cannot be pinned on one uniquely culpable actor among the physicians, shelter workers, sanitation crews, and outreach workers who touched his case. The 500 block of State Street did not assume its current form through a single decision choosing that allocation of public and commercial space. And nobody at Apple needed to decide the author was an impostor for him to spend twelve days, across two episodes, exiled from his own digital life.

Our ordinary moral vocabulary handles intentional wrongdoing far more easily than cumulative institutional harm. We know how to name cruelty, negligence, corruption, discrimination. We have far less practice naming harm produced by competent people performing individually defensible tasks inside structures that fragment responsibility and, in their more rationalized forms, narrow the discretion that would otherwise catch it in time.

The Paradox of Rational Accumulation names that problem. Its unit of analysis is the trajectory, not the isolated decision. Its causal object is the accumulation of locally rational action across institutional boundaries, not the incompetent individual actor. Its temporal concern is the gap between administrative schedule and lived, biological time. Its stratifying principle is unequal capacity to absorb institutional friction. Its organizational diagnosis is the mismatch between vertical accountability and horizontal consequence. Its technological extension is recursion and the variable availability of effective repair capacity. Its moral threshold is foreseeability. And its political question is responsibility for the whole rather than the part.

Modern institutions have sophisticated answers to the question who is responsible for this task. The harder, and more urgent, question is: who is responsible for what all our tasks add up to? A humane institutional order will have to learn how to answer it — architecturally, not merely by asking its most conscientious people to try harder at jobs they are already, individually, doing well.

References

Arendt, H. (1963). Eichmann in Jerusalem: A report on the banality of evil. Viking Press.

Bauman, Z. (1989). Modernity and the Holocaust. Cornell University Press.

Beck, U. (1992). Risk society: Towards a new modernity. Sage.

Bovens, M., & Zouridis, S. (2002). From street-level to system-level bureaucracies: How information and communication technology is transforming administrative discretion and constitutional control. Public Administration Review, 62(2), 174–184.

Galtung, J. (1969). Violence, peace, and peace research. Journal of Peace Research, 6(3), 167–191.

Garfinkel, H. (1967). Studies in ethnomethodology. Prentice-Hall.

Hardin, G. (1968). The tragedy of the commons. Science, 162(3859), 1243–1248.

Kahn, A. E. (1966). The tyranny of small decisions: Market failures, imperfections, and the limits of economics. Kyklos, 19(1), 23–47.

Lipsky, M. (1980). Street-level bureaucracy: Dilemmas of the individual in public services. Russell Sage Foundation.

Mellinger, W. M. (2026a, March 23). What the Andrée Clark Bird Refuge reveals about visibility and power. Noozhawk.

Mellinger, W. M. (2026b, March 30). The bench on Cota Street that isn't there. Noozhawk.

Mellinger, W. M. (2026c, April 13). How Santa Barbara's institutions produce harm without intending to. Noozhawk.

Mellinger, W. M. (2026d, April 20). Visionaries, guardians, critics — and benches — in Santa Barbara. Noozhawk.

Mellinger, W. M. (2026e, April 27). Everyone was being reasonable, and he died anyway. Noozhawk.

Mellinger, W. M. (2026f, June 29). How the 500 block of State Street quietly transformed public space. Noozhawk.

Mellinger, W. M. (2026g, forthcoming). I am currently serving my second sentence. Noozhawk.

Merton, R. K. (1936). The unanticipated consequences of purposive social action. American Sociological Review, 1(6), 894–904.

Mills, C. W. (1959). The sociological imagination. Oxford University Press.

Molotch, H. (1976). The city as a growth machine: Toward a political economy of place. American Journal of Sociology, 82(2), 309–332.

Perrow, C. (1984). Normal accidents: Living with high-risk technologies. Basic Books.

Sacks, H. (1992). Lectures on conversation (G. Jefferson, Ed.). Blackwell.

Schegloff, E. A. (2007). Sequence organization in interaction: A primer in conversation analysis I. Cambridge University Press.

Schegloff, E. A., Jefferson, G., & Sacks, H. (1977). The preference for self-correction in the organization of repair in conversation. Language, 53(2), 361–382.

Simon, H. A. (1947). Administrative behavior. Macmillan.

Weber, M. (1978). Economy and society. University of California Press.