The central problem of political life is neither power nor freedom in the abstract. It arises from a stubborn, material fact: human beings can exercise agency only through relationships and institutions on which they depend.
Much of political philosophy begins by treating dependence as an embarrassing defect. The subject is imagined as an unencumbered chooser, self-directing and whole, and authority enters the scene only when someone interferes with that autonomy. Actual social life bears no resemblance to this fable. People depend on employers for income, states for legal recognition and public goods, families and networks for physical care, professionals for diagnostic judgment, infrastructures for movement, and technical systems for basic access to civil society.
To understand this dependence, political theory must split it into two distinct realities. There is, first, ontological interdependence: the bodily fact that human beings are mortal, vulnerable creatures requiring generational care, shared language, physical infrastructures, and mutual solidarity. This interdependence is permanent. But overlaid upon this is a second, historically manufactured dependence: the condition of proletarianization, where people are systematically severed from direct access to the means of subsistence, left with nothing to sell but their labor-time. Capitalism takes our universal interdependence and converts it into a coercive trap, holding survival hostage to market discipline. When political thought confuses these two realities, it naturalizes capitalist compulsion as if it were simply human biology.
The goal cannot be to abolish interdependence so that an isolated individual can stand alone; interdependence is irreducible. Nor can the answer be that because dependence exists, whatever institution coordinates us acquires a blank check to dictate the terms of life.
The question underneath everything this publication examines is simpler and more severe:
What makes dependence compatible with agency rather than domination?
That question demands more than an inventory of choices. Choice is necessary, but choices always occur inside structured fields. If losing a job forfeits healthcare, residential stability, immigration status, and the ability to feed dependents, an employer does not need to issue an explicit threat to exercise coercive force. If a credit engine or an algorithmic classifier determines who appears before employers, landlords, or audiences, its classifications reorganize human possibilities without anyone barking an order. If a welfare bureau decides whether a life counts as disabled, unemployed, or fraudulent, it settles what a person can do by decreeing what kind of case they are.
The choice is real, yet it operates entirely within possibilities manufactured by institutions, material distributions, and administrative classifications. People are formed by structures without being exhausted by them. They can judge, refuse, organize, build counter-institutions, and act in ways that shatter bureaucratic expectations.
The human person is socially constituted, but normatively irreducible.
To say a person is socially constituted means there is no untouched self outside history, language, embodiment, and material relations. To say they are normatively irreducible means no institutional description ever exhausts their moral status. The worker is never merely labor power. The patient is never merely a diagnosis. The claimant is never merely an entry on an intake log. The citizen is never merely a demographic bucket. The person subjected to an administrative system retains an ethical standing that the system's representations cannot absorb.
The Representational Turn in Politics
Most theories of power concentrate on who commands whom. A forensic account of institutions requires a prior question:
Through what representation does the institution know the person over whom it exercises authority?
That sounds epistemic, but it is the fulcrum of power.
Every institution coordinates at scale by simplifying. It cannot encounter every human being in their unrepeatable particulars. A tax agency requires income brackets. An emergency department requires triage protocols. A university requires admissions thresholds. A disability office requires functional criteria. A technology platform requires measurable telemetry. A state requires census categories and legal definitions.
These abstractions are unavoidable. No serious politics proposes abolishing categories to encounter billions of people in pure, unmediated intuition. James C. Scott documented how centralization requires legibility; Herbert Simon showed that human and organizational rationality is inevitably bounded. Simplification is the only way collective work happens.
The political crisis begins the moment an administrative abstraction becomes the basis for allocating rights, burdens, protections, punishments, or survival. At that intersection, the institution issues a binding decree: For the purposes that determine your material life, this model is what you are.
For most people, the model functions with rough adequacy. The political test arrives at the boundary: the condition the diagnostic code excludes, the caregiving schedule the shift-clock penalizes, the non-standard worker the benefit algorithm flags as fraudulent, the resident physician whose training demands violate the duty-hour audit, the applicant whose history the credit scoring pipeline misreads.
At that boundary, an institution chooses between two postures:
It can treat the human being as an anomaly, an edge case, a deviant, or an error relative to an otherwise complete system.
Or it can treat the mismatch as evidence that the system’s model of reality is deficient.
That choice determines where the burden of correction falls. If the model is presumptively sovereign, the human being must contort their body, schedule, and household until they fit its classifications, absorbing the friction in silence. If lived experience retains epistemic standing, the institution's model becomes an object of mandatory revision.
This is the political meaning of the remainder. The remainder is whatever human reality the governing representation cannot contain.
When reality exceeds the category, does the institution possess the right to discard the remainder, or does the remainder have standing to indict the representation?
Domination Without a Tyrant
This distinction separates structural domination from ordinary mistakes. All administrative theories acknowledge error. They provide customer-service hotlines, quality-assurance audits, and standard appeals. But treating error as an accidental flaw assumes the machine was built correctly and merely misapplied its rules.
Representational incompleteness is structural. Because no model can anticipate the full heterogeneity of human life, failure is not an anomaly; it is an invariant property of governance through abstraction. If representational incompleteness is structural, correction cannot be bolted onto an institution as an afterthought. The capacity to be revised by those it misrepresents must be written into the institution's basic theory of authority.
This clarifies the relationship to republican non-domination. In Philip Pettit's classic formulation, freedom requires the absence of arbitrary power—typified by the master whose slaves remain unfree even if the master happens to be kind, because their security depends entirely on his discretionary mood.
In contemporary life, the dominating entity is rarely a personal master. It operates across two intertwined vectors: agentless administrative domination and impersonal market compulsion.
Bureaucracy provides the first. A claims engine rejects a life-saving treatment because a metric threshold was missed. The claims clerk explains that the system permits no override. The hospital administrator points to software compliance. The software vendor cites contractual specifications. The regulatory agency notes that the standard was adopted under proper administrative procedures. Everyone involved can be conscientious, polite, and entirely devoid of personal malice. No single human actor exercises arbitrary will.
Yet behind this polite bureaucracy sits the second vector: compulsion without jurisdiction. The hospital administrator, the clinic director, and the logistics manager are themselves disciplined by market competition and financialized solvency. Nobody holds formal jurisdiction over the market, yet its impersonal imperatives—capital accumulation, cost containment, margin protection—compel institutions to externalize their costs onto human bodies. The market forces institutions to become dominating in order to survive.
When administrative abstraction fuses with market compulsion, agentless domination becomes total. The system reorganizes life; its model is false; the human carries the catastrophe; and the market ensures that no actor inside the institution can afford to treat that failure as evidence against the business model.
This is agentless domination: power exercised through an institutional representation insulated from correction. The Horizon post-office scandal in the United Kingdom and the automated MiDAS fraud system in Michigan were not failures of rogue tyrants; they were automated representations granted sovereign authority over human lives while the humans carrying the remainder were stripped of the standing to challenge the machine's model of truth.
[Institutional Abstraction] ──acts upon──► [Human Subject]
│ │
(produces) (generates)
▼ ▼
[Sanitized Metric] [Material Remainder]
│ │
(denies return path) (absorbs cost in body)
▼ ▼
[Insulated Representation] ◄──────────── [Agentless Domination]The Ladder of Administrative Response
Designing institutions that do not dominate requires distinguishing five fundamentally different postures an institution can take toward the governed:
- Transparency: The institution permits you to see how it operates. You can examine the rulebook, read the algorithmic feature weights, or view your classification. Transparency shows you the architecture of the cage; it provides no key.
- Voice: The institution permits you to speak. It conducts town halls, hosts feedback panels, and invites public comments. As Albert Hirschman recognized, institutions frequently grant performative voice to defuse the threat of exit, installing microphones that connect to no administrative levers.
- Appeal: The institution allows you to ask whether an existing rule was misapplied. An appeal tests the clerk against the manual. If the clerk deviated, you receive an exception. But if the manual itself is defective, the appeal confirms the injury.
- Participation: The institution invites you to help design or select its processes. You vote for representatives, join advisory committees, or help shape initial guidelines. But participation settles legitimacy at the origin of a policy; it cannot anticipate how the rule behaves when applied to unforeseen contexts.
- Corrective Standing: The institution provides an enforceable return path through which your encounter with its failure acquires binding authority to force revision of the governing categories.
Corrective standing is simultaneously moral, epistemic, and political. It is moral because the person possesses a dignity that no administrative tag can extinguish. It is epistemic because the experience of systemic failure constitutes vital empirical information about the boundaries of the model. It is political because that information carries the legal and institutional power to compel institutional change.
Yet here lies the central hazard of reform: corrective standing must not degenerate into adaptive domination. Institutions excel at absorbing critique to stabilize extraction. A platform or an insurer can install user panels, ombuds, and algorithmic audits precisely to vent revolutionary friction, making the underlying machinery of accumulation run smoother and longer. Corrective standing achieves emancipation only when it holds the power to contest the systemic imperative itself—to eliminate the classification, halt deployment, and decommodify the resource. If standing merely allows subjects to negotiate the calibration of their own exploitation, it remains a shock absorber for the machine.
An institution's legitimacy does not depend solely on who founded it, how its leaders were elected, or how clean its code was on the day of deployment.
The legitimacy of an institution depends on whether the knowledge through which it governs remains corrigible by those whom that knowledge misrepresents.
Toward a Theory of Government: Six Constructive Principles
Asserting that institutions must be corrigible is easy; specifying how they govern without dissolving into chaotic paralysis is the real task. An institution with infinite elasticity cannot coordinate complex societies. If every individual grievance rewrote the general code, no standard could stabilize social expectations.
A mature political philosophy must specify the mechanics of bounded, defeasible authority. It must answer six fundamental design questions:
[ Encounter with Mismatch ]
│
▼
1. STANDING: Who holds the right to trigger contest?
│
▼
2. THRESHOLDS: When does variance indict the category?
│
▼
3. AUTHORITY: Who arbitrates between model and subject?
│
▼
4. SCALE: How does revision propagate without collapse?
│
▼
5. POWER: Who protects the return path from capture?
│
▼
6. FINALITY: Under what conditions does a rule hold?1. Standing: Who is entitled to trigger correction?
Standing cannot be restricted to accredited experts, property holders, or designated institutional representatives. Standing belongs to those upon whom the institution’s representation acts and who bear the material cost of its failure.
In administrative systems, standing must be grounded in the incidence of remainder. When a patient, a worker, an insured claimant, or a tenant demonstrates a persistent mismatch between their material reality and the administrative category governing them, their status as the bearer of that unabsorbed cost confers prima facie standing. Standing is not a request for charity; it is an assertion of epistemic sovereignty over one's own conditions of survival.
2. Thresholds: When does an exception indict the rule?
Every system encounters one-off frictions that do not signal a broken rule. How does an institution distinguish between an idiosyncratic variance and a structural failure?
The architecture requires a two-tiered trigger:
- The Variance Track (Immediate Relief): An individual mismatch triggers an immediate, low-friction administrative remedy that prevents the person from absorbing material catastrophe while the dispute is examined.
- The Diagnostic Tripwire (Mandatory Review): When individual variances cluster—whether across geographic zones, demographic lines, job classifications, or operational units—the system crosses an automatic statutory threshold. Once crossed, the institution loses the right to treat the cases as isolated exceptions. The accumulated remainder automatically opens a mandatory revision docket, compelling the agency to place the category itself on trial.
3. Authority: Who decides between conflicting claims?
No institution can act as the sole judge of its own representations. When an agency, an employer, or an algorithmic system evaluates whether its own categories are adequate, institutional self-preservation guarantees that it will classify systemic failures as user error or unpreventable bad luck.
Authority over contested categories must be polycentric, drawing on Elinor Ostrom's insights into federated commons governance. Adjudication requires independent tribunals: citizen juries, mixed worker-and-public councils, and structural ombuds equipped with independent subpoena and investigative powers. When an institutional model is contested, the issuing entity sits as an interested party, not the judge.
4. Scale: How does correction propagate without systemic collapse?
Institutions must act. A healthcare system cannot freeze admissions because a diagnostic category is contested; an electricity grid cannot stop dispatching power while a tariff formula is renegotiated.
The solution is bounded, defeasible authority operating through staged dampening:
- The institution acts provisionally on the best available abstraction. Action is permitted because coordination is urgent.
- But its decisions carry provisional status whenever challenged through accredited return paths.
- While structural revision proceeds through asynchronous batch updates, the institution must provide statutory safe harbors: interim resources, provisional benefits, and liability shields that insulate affected humans from harm while the model is rebuilt. Governance maintains continuity at scale, while the human beings at the edge are protected from bearing the cost of system latency.
5. Power: Preventing capture of the return path
Without deliberate institutional design, correction mechanisms are swiftly colonized by actors with the greatest stamina, money, and administrative literacy. Pamela Herd and Donald Moynihan demonstrate that administrative burden is an unacknowledged tool of rationing and inequality.
To prevent the capture of corrigibility:
- Invert the Burden of Proof: For vulnerable populations, the evidentiary burden must sit on the institution. If an automated system terminates a benefit or reclassifies a worker, the institution must prove its model's empirical adequacy before the penalty takes effect, rather than demanding that a precarious human produce mountains of documentation to prove their existence.
- Structural Advocacy: Well-funded, permanent public advocates must be assigned to defend the remainder, auditing case drop-offs and representing claimants without charging them for the defense of their basic standing.
6. Finality: When may an institution legitimately say, "The rule stands"?
Corrigibility does not mean every challenge must prevail. Collective life routinely requires enforcing rules that impose burdens and constrain choices. When is a refusal to revise legitimate?
An institution may sustain a challenged rule only under three strict conditions:
- It must publicly demonstrate that retaining the contested classification preserves a vital collective capability (in Amartya Sen’s sense) that narrower tailoring would destroy.
- It must demonstrate that it has exhausted feasible modifications to accommodate the remainder without compromising the core coordination function.
- Most critically, it must provide material dampers and compensatory protections to those who fall outside the rule's boundary. The institution can say, "The general rule must stand, but because our abstraction cannot fit your reality, the collective will absorb the cost of your difference rather than abandoning you to it."
Finality is legitimate only when accompanied by material non-abandonment. The moment an institution says "the rule stands" and leaves the remainder to be crushed by the consequences of the mismatch, it ceases to govern; it dominates.
The Socio-Technical Horizon
This framework illuminates why the governance of artificial intelligence belongs to the same intellectual problem-space as democratic theory and labor rights.
Machine learning intensifies the ancient bureaucratic problem by orders of magnitude. A traditional bureaucracy simplifies human lives into dozens of coarse paper categories. A modern predictive system compresses thousands of behavioral signals into high-dimensional vectors, generating probabilistic risk scores at millisecond speeds.
Because these models are mathematically sophisticated, they masquerade as objective truth. Yet their representations remain radically incomplete. A model trained on past distributions cannot anticipate emergent social realities; an algorithmic classifier optimizing for an administrative proxy systematically excludes whatever falls outside its measurement apparatus.
Delegated machine agency creates the acute contemporary frontier of domination:
- The system acts at planetary scale with minimal latency.
- Its representations are opaque even to the engineers who configured them.
- Responsibility is fragmented across developers, operators, data contractors, and deployment managers.
- The human bearing the failure cannot locate a conscious actor with the authority to alter the program.
AI safety cannot be reduced to technical alignment or demographic debiasing. A model that distributes exclusions with pristine mathematical parity remains an instrument of domination if the people it misrepresents possess no power to contest the metric itself.
Model-level corrigibility—the technical capacity to halt, redirect, or modify an agent—must be integrated with institutional corrigibility: the socio-technical architecture that ensures affected communities have the standing to inspect, contest, and shut down automated systems that treat human lives as expendable noise.
This returns us to the foundational moral asymmetry that must anchor all social design:
Systems are instruments. Human beings are not instruments of the systems that coordinate them.
An institution can be scientifically sophisticated, collectively owned, historically necessary, and impeccably administered. But its authority is always derivative. Its knowledge is always partial. Its abstractions are always defeasible.
The Tripartite Horizon of Emancipation
This analysis exposes the central architectural boundary separating our institutional theory from classical political economy.
Marx achieved an indispensable breakthrough: diagnosing how human collective activity becomes alienated and class-divided, confronting producers as an external, hostile power governed by capital accumulation. His positive horizon was associated producers consciously governing their common conditions—what we formulate as Collective Non-Alienation.
Yet that framework carries a foundational vulnerability: the assumption that overcoming capitalist class relations resolves the general problem of domination. It does not.
Abolishing private ownership, wage labor, and production for profit leaves the core political questions unanswered:
- Who administers?
- Who classifies?
- Who decides?
- Who can refuse?
- Who holds standing to correct the institution when its models fail?
- And how much of a person's life may collective purposes legitimately claim?
A post-capitalist society remains fully inside the problem of organized power. Every coordination system—no matter how democratic, communist, or egalitarian in origin—relies on simplified categories, operational models, administrative divisions, and delegated authorities. There is no institutional form whose emancipatory origin makes it self-certifying.
To prevent coordination from converting into dominion, institutional legitimacy requires an uncompromising Non-Conversion Principle:
- Expertise does not become sovereignty: Knowing more about an aquifer, an electrical grid, an algorithm, or a disease does not confer the right to govern the person. Epistemic competence never equals political jurisdiction.
- Provision does not become command: Distributing medicine, food, housing, or compute does not grant entitlement to dictate how the recipient lives.
- Dependence does not become jurisdiction: The existence of a material or physical vulnerability licenses support, never subordination.
From these boundaries emerges a cardinal rule:
Collective authorship does not entail collective ownership of individual agency.
A democratically governed workplace, commune, planning board, or federation does not acquire unlimited jurisdiction merely because the collective created it. Freedom cannot be collapsed into collective control alone.
An emancipated society must satisfy three independent, irreducible tests:
- Collective Non-Alienation: People must possess meaningful, conscious collective control over the social structures and conditions of reproduction they sustain together. The world we build must not confront us as an autonomous, market-driven necessity.
- Vertical Non-Domination: No person, administrative hierarchy, or algorithmic classifier may possess uncontrolled, unanswerable power over another merely because an organized dependency exists. Dependence must never convert into sovereign command.
- Personal Non-Absorption (Discretionary Sovereignty): Collective purposes—even genuine, democratically ratified collective purposes—must leave persons substantial capacities, time, relationships, uncommitted space, and futures that are not permanently appropriated by the collective. Freedom requires the protected power of exit, refusal, time sovereignty, and private projects. What "we" cannot claim from "me."
These three requirements are bound together by an institutional mechanism that cuts across all of them: Epistemic Corrigibility.
Whatever institutions implement these ideals must themselves remain structured around their own possible error. The emancipatory ideal is not a society that finally knows how to govern itself correctly, but an institutional architecture capable of discovering that its own governing representations are wrong.
That capacity cannot depend on managerial benevolence or vanguard enlightenment. It requires that the human beings subjected to administrative and algorithmic classifications possess the procedural and material standing to force those representations to confront their remainder.
Marx asks how our own powers cease confronting us as alien powers.
The Crumple Zone adds the question no post-capitalist design can evade:
And once those powers become genuinely ours, how do we stop “ours” from becoming a new justification for ruling each of us?
That is where this theory departs from classical political economy. The task of politics is not merely to construct common power, but to build coordination systems strong enough to absorb collective vulnerability and sustain shared life—while ensuring that the collective never acquires sovereign title over the individuals who compose it.