When an automated decision system or the institution around it works, the usual accounting credits its formal parts: staffing formulas, software, protocols, and budget.
That credit is often misplaced. Much of what registers as institutional capacity is borrowed from workers, patients, families, and citizens who quietly repair what the formal system leaves broken. The nurse catches the medication error. The employee reconstructs an undocumented process. The patient coordinates five uncommunicative specialists. The claimant translates contradictory bureaucratic notices.
This work is not incidental. Without it, the formal system could not produce the outcomes it reports. It is the human subsidy.
Parts of it already have names. Sociologists of work call the coordination that holds a division of labor together articulation work (Anselm Strauss) and work that goes unseen and uncounted invisible work (Arlene Kaplan Daniels; Susan Leigh Star and Strauss). Mary Gray and Siddharth Suri call the hidden human labor behind systems sold as automated ghost work. This essay treats them as one subsidy and asks what the institution’s accounts do with it.
Pulling the accounts apart
The economic historian Karl Polanyi used “disembedding” for the historical separation of markets from the social relations they had been part of. Here the word does a narrower job: pulling apart things that institutional accounting treats as one.
The usual accounts fuse several things that need to be separated:
- Performance from capacity. An arrangement producing acceptable outcomes does not establish that the arrangement itself possesses the capacity to produce them. Part of that capacity is borrowed from informal human compensation.
- Adaptation from endorsement. That someone learns to function within a defective system does not show that they chose, approve of, or benefit from the arrangement.
- Compliance from consent. Doing what circumstances require to survive or protect dependents is not free agreement.
- Resilience from legitimacy. The ability to endure deprivation does not make the deprivation reasonable.
- Care from sacrifice. Caring for another person does not imply an obligation to absorb unlimited institutional costs.
- Competence from obligation. Being capable of fixing a systemic failure does not establish that fixing it should become your personal responsibility.
- Efficiency from cost. An organization can look cheap or lean because its real operating costs have been pushed off its balance sheet.
- Failure from its point of manifestation. The exhausted clinician, overwhelmed patient, or burned-out caseworker is the location where system failure becomes visible, not where it originates.
- Need from obligation. Something needing to be done does not answer who should bear the burden of doing it. Institutions routinely collapse “someone must adapt” into “therefore this particular person should adapt.”
The main separation is between the formal organization and everything that actually produces the outcome:
total operating architecture = formal organization + uncounted compensatory inputs
Once the two are separated, a question the usual accounts never ask comes back: who bears the cost of making the arrangement work, and why them?
The kinds of uncounted work
Auditing an institution needs a list of the kinds of work the subsidy is made of:
- Repair labor: Fixing defects left behind by the formal process. Correcting corrupt records, working around defective software, and manually reconciling mismatched data schemas.
- Coordination labor: Connecting fragmented parts of an organization that refuse to communicate. Passing records between departmental silos and tracking dependencies across disconnected teams.
- Interpretive labor: Deciphering ambiguous instructions, translating institutional jargon, and reverse-engineering undocumented operating rules.
- Anticipatory labor: Maintaining vigilance to predict catastrophic failures before they trip. Watching a fragile queue to catch dropped requests before they harm a client.
- Affective labor: Absorbing institutional friction and frustration. Soothing angry users, de-escalating conflicts caused by procedural delays, and maintaining morale under chronic dysfunction.
- Self-regulatory labor: Suppressing one’s own physical, psychological, or schedule needs so the arrangement remains workable. Skipping meals or breaks, tolerating sensory overload, and masking distress to meet inflexible protocols.
- Risk-bearing labor: Personally carrying uncertainty or legal liability that the institution has externalized downward without indemnity or safety nets.
- Temporal labor: Waiting on hold, chasing missing paperwork, remembering unnotified deadlines, and rearranging personal life around institutional delays.
This work is usually folded into the background and mistaken for personality, professionalism, dedication, or “just how things work.”
Why the work is not voluntary
Organizations often defend these demands by treating participation as voluntary. If workers do not resign, or if claimants complete the paperwork, the institution assumes consent.
That inference is wrong. Institutions rarely compel anyone by direct command. The pressure is spread across the surrounding arrangement:
- Structural coercion: Refusal guarantees that something critical fails. The nurse can refuse to reconcile records off the clock, but the consequence is patient harm.
- Dependency coercion: The vulnerability of people who depend on you makes withdrawal relationally and morally prohibitive.
- Normative coercion: Professionalism, dedication, maternal love, or citizenship turn refusal into a personal moral defect. The person who insists on the boundary is labeled uncooperative or uncaring.
- Procedural coercion: Formally optional behavior becomes necessary to navigate a defective system. The workaround is the only path to completing the assigned task.
- Economic coercion: The party with the least ability to exit ends up supplying the necessary adjustment.
When the alternative to adaptation is catastrophic for oneself or others, adaptation is functionally compelled.
Turning adaptation into character
To justify relying on this work, institutions relabel coping with their defects as personal virtue:
- A person repeatedly saves a broken process, and the institution calls them dependable.
- A person tolerates chronic instability, and the institution calls them flexible.
- A person absorbs interpersonal hostility, and the institution calls them professional.
- A person anticipates recurring failures, and the institution calls them experienced.
- A person endures prolonged deprivation, and the institution calls them resilient.
This vocabulary credits the worker’s character and leaves the defect that made the work necessary untouched. A debt the institution owes is recorded as an asset it has.
The capable worker who prevents mistakes is not only an asset. Repeated dependence on their vigilance is an unrecorded liability. When an organization praises a team for being “easy to manage,” it usually means the team privately supplies enormous quantities of accommodation to the organization.
The presumed direction of accommodation flips:
Conventional assumption: Institution → accommodates → Person
Operational reality: Person → accommodates → Institution
Theory, drawn
Whose work the outcome runs on
The institution reports one outcome and credits it to its formal parts. Part of that outcome is work supplied by people it does not count, and the accommodation runs from them to it. The institution books their work as their character, not as its debt. The split of the bar is drawn, not measured.
What the adaptation was preventing
Ethotechnical evaluation reverses the usual management question:
Instead of asking why this person failed to adapt, ask what their adaptation had previously been preventing the institution from having to confront.
When a worker collapses, a student drops out, or a patient gives up, conventional metrics treat that individual as the point of failure. The method treats them as the point where the failure showed: the place where one person could no longer carry the cost of the institution’s gaps privately.
Where this fits
Other essays take this work as given:
- The consumption of adaptive capacity asks when an institution may demand this work. The kinds above are what it demands.
- Absorption as concealment shows how quietly fixed errors make a system look reliable. The repair, interpretive, and anticipatory work above is how they get fixed.
- Insulation shows how an institution reads people’s coping as proof that a rule works. The pressures above explain why they keep coping.
- What Ethotechnics is not stops using words like resilience and adaptability as explanations. The kinds above are what those words usually describe.
What the laws take from this
- Correction obligation (Law IV): Every consequential delegation creates an obligation to correct its errors. If an institution relies on uncounted human repair to stay viable, that capacity must be accounted for as an operational debit.
- Proportional standing (Law VII): Error-bearing parties who supply the compensatory subsidy require standing to force upstream rule revision. Without standing, adaptation is uncompensated extraction.
- Non-theatrical observability (Law VIII): Measuring burden by asking exhausted people to fill out surveys repeats the extraction. Burden must be read from the system’s own trail of retries, abandonments, and off-hours activity.
- Highest-layer evaluation (Law X): Evaluation must sit at the highest layer where harm emerges: in the lives and households of the people who absorb the mismatch, not within the boundaries of the model or software code.
- Preserving contestability (Law XII): No system may consume the finite adaptive capacity required to contest it. An institution that exhausts its participants until they can no longer object wins by attrition.