What institutions owe the people their systems cannot recognize.
Imagine a public-assistance program that requires a recent utility bill to establish residency. An applicant lives with relatives. The bills bear someone else's name. They have other credible evidence of where they live, but the online application accepts only the prescribed document.
The applicant cannot complete the form.
No official has concluded that they live elsewhere. No one has examined their alternative evidence. Yet the practical result is the same as a denial. A requirement introduced to establish a fact has become a condition that prevents the institution from discovering it.
Suppose the software works perfectly. Every field is checked correctly. Every applicant faces the same requirement. Nothing crashes.
Where, then, is the failure?
The window remains. The person who could answer does not. Photo: Elliott Brown, via Wikimedia Commons, CC BY 2.0.
We could say the requirement was poorly designed. But the deeper problem begins when “this is the evidence we have chosen to accept” becomes “this is the evidence that counts,” and eventually “the system won't allow it.” A decision about how to recognize reality starts to exercise authority over what reality can be recognized.
The rule has not literally become a fact. It has acquired a fact's apparent inevitability. The applicant encounters something people decided as something nobody seems able to change.
The decision before the decision
We usually locate institutional power at the moment of judgment: the application is rejected, the claim denied, the transaction blocked. But much of the consequential judgment has already occurred.
Before an institution can decide whether someone qualifies, it must decide what qualification means, what evidence can establish it, and which circumstances can be represented. Here, a protocol means the specified conditions and steps through which a policy becomes an executable process. Software is one way of enforcing it; a paper procedure can enforce it too.
The residency requirement compresses three questions:
Does the applicant live here?
Does the applicant possess a utility bill in their name?
Is lacking that bill sufficient reason to exclude them?
The system checks the second question and enforces a predetermined answer to the third. Neither necessarily settles the first.
This is why formalization does not eliminate discretion. It redistributes it. Judgment moves upstream, into the categories, acceptable inputs, and exceptions someone chose before the applicant arrived.
Consider the reviewer who would willingly examine alternative evidence but never receives it because the form provides nowhere to submit it. The reviewer may be fair, attentive, and experienced. Their virtues cannot correct a circumstance the institution has prevented them from encountering.
The applicant therefore faces more than an unfavorable answer. They cannot make the relevant question legible. The system asks them to prove their circumstances while withholding the means of describing those circumstances accurately.
An institution exercises power through the circumstances it refuses to hear, as well as the decisions it makes.
Every institution must simplify. A program cannot investigate every conceivable detail of every life. But deciding which details to exclude is itself an exercise of authority. It requires justification even when the resulting system has made those details invisible.
What the person is owed
The immediate harm in our example is exclusion from assistance. Yet another harm precedes it: the applicant's account of their own situation cannot enter the decision.
This does not mean institutions must accept whatever people say. Evidence can be incomplete, mistaken, or deceptive. A person can deserve a hearing and still receive a justified refusal.
What matters is whether relevant evidence can change the institution's understanding. If nothing the applicant could explain would make a difference—not because their explanation has been considered and found insufficient, but because the process cannot receive it—then the demand for evidence has become strangely hollow.
The institution asks the person to answer to its categories. It provides no corresponding obligation to answer for those categories.
That asymmetry is the moral center of the problem. The applicant must explain themselves to receive assistance; the institution can withhold assistance by referring to a requirement whose relevance nobody present is authorized to examine.
Institutional authority should carry a reciprocal burden. If a public program demands reasons from a person, it should remain capable of giving reasons to that person: why this condition matters, why their evidence is insufficient, and where a challenge to that judgment can be heard.
The right to be heard cannot amount merely to permission to repeat oneself in an empty room.
The rule protects us, too
It is tempting to answer this problem with more discretion. Let an administrator accept another document. Let a sympathetic person override the system.
But sympathy is an unstable basis for a public entitlement. One applicant may encounter generosity; another may encounter suspicion. A persuasive account may succeed where an equally valid but less fluent account fails. People should not have to become exceptional storytellers to receive ordinary consideration.
Rules protect us from the moods, prejudices, and preferences of those who administer them. Predictability allows people to organize their lives without first discovering which official might be receptive.
The difficulty is that the same rule can protect people from arbitrary judgment while denying them relevant judgment. A uniform requirement may treat everyone identically and still privilege those whose lives already fit its assumptions.
We should therefore distinguish an exception granted as a favor from a correction required by a reason. If alternative evidence establishes residency adequately, accepting it need not be an act of mercy. It can be a more accurate application of the program's stated purpose.
And if the same “exception” recurs, the institution should ask whether it has misnamed the problem. Perhaps these applicants are not exceptional. Perhaps its picture of an ordinary applicant is too narrow.
A rule should protect a person from needing a favor, without preventing them from receiving a reasoned judgment.
A present institution must answer for an inherited rule
The person who selected the documentation requirement may have retired. The software may have been purchased from a vendor. The current administrator may have no authority to alter either.
These facts help explain the persistence of the requirement. They do not settle who must answer for it now.
An institution acts in the present when it enforces an inherited decision. However old the rule, today's rejection is not simply an event left behind by yesterday's policymakers. Someone maintains the process, authorizes its continued use, and relies on its outputs.
Historical authorship and present responsibility are different. An official need not have invented a rule to bear responsibility for maintaining a route through which its defects can be addressed.
This also explains the limits of transparency. A complete audit trail could show exactly why the application could not proceed. It could identify the missing field, the relevant specification, and the policy behind it. The applicant might understand the process perfectly and remain unable to reach anyone empowered to reconsider it.
Explanation answers how the constraint operated. Accountability requires an answer to whether it should have operated that way, and someone with authority to act on that answer.
Responsibility need not belong to one person in full. Engineers, administrators, and policymakers have different obligations and different powers. But a division of responsibility must not become a disappearance of responsibility. The institution should be able to identify who can correct a processing error, who can change the accepted evidence, and who can reconsider the policy itself.
Can the institution be corrected?
An appeal is meaningful only if it can reach the judgment being disputed. Rechecking whether the utility bill was supplied cannot resolve a challenge to making that bill indispensable.
A workable review process needs distinct routes:
- Review the case. Was evidence overlooked, misread, or processed incorrectly?
- Review the specification. Does the process admit the kinds of evidence needed to apply the policy fairly?
- Review the policy. Does the requirement still serve a defensible purpose, given its benefits and the exclusions it produces?
These routes need not have the same decision-maker or timetable. An administrator should not rewrite public policy whenever an applicant objects. But a case that exposes a possible policy defect should be capable of reaching an authority that can examine it, while any available interim remedy is considered at the appropriate level.
Institutions also owe attention to the cost of correction. An appeal requiring months of effort may be formally available and practically useless to someone seeking immediate assistance. The institution cannot assess accessibility solely by asking whether it has published instructions.
Nor should applicants have to diagnose the architecture of their exclusion. Someone may know that the form cannot express their situation without knowing which specification is responsible. Repeated failed submissions, recurring requests for exceptions, and patterns in abandoned applications should prompt institutional investigation.
Such patterns do not prove that a rule is wrong. They create reasons to examine it. Review must remain capable of upholding necessary constraints as well as revising inadequate ones.
The test is whether the institution can learn something from a challenge that its ordinary process did not permit it to see.
The world is not the specification
A public program must decide where its responsibilities begin and end. Some applications will be refused. Some evidence will be insufficient. No system can promise that everyone receives the outcome they seek.
But the limits of an institution's procedure do not define the limits of a person's reality. A missing field is not necessarily a missing fact. An unrecognized circumstance is not necessarily an irrelevant circumstance. A correctly executed rejection is not necessarily a justified one.
What makes a rule dangerous is not simply that it constrains us. All rules do. It is that the institution may stop recognizing the constraint as a choice for which it remains answerable.
The applicant in our example should not have to persuade the institution that they deserve compassion before it can consider evidence of where they live. They need something less sentimental and more demanding: a process through which reality can correct the institution's account of it.
An institution can keep its decisions stable while keeping its reasons open to examination. Its authority need not depend on never being wrong. It depends, in part, on whether being shown wrong can change what it does.
A rule can determine what happens. It cannot, by that fact alone, determine what ought to happen.