Chapter 49 — Systems That Cannot Admit Error

What happens when an institution defines the problem, controls the interpretation, preserves the record and then uses that record to confirm itself, with particular attention to inquiry findings and institutional responses to sexual abuse, secrecy, reputation and accountability.

When an Error Becomes Institutional ↑

There is a difference between being wrong and having an error become institutional.

A person can make a mistake and correct it.

An institution can make a mistake and give it a history.

The difference is not simply size. Institutions record things. They copy them. They refer to earlier decisions. One professional reads what another professional wrote. One department sends information to another. A conclusion becomes part of a history, and the history begins influencing what happens next.

An error acquires a history.

Once that happens, correcting it becomes much more difficult than correcting the original sentence.

The First Assertion ↑

Every institutional history begins somewhere.

Someone observes something. Someone is told something. Someone reaches a conclusion. Someone records an allegation, an impression, a diagnosis, an interpretation or a decision.

Those are not the same things.

An observation is not an allegation. An allegation is not a finding. A diagnosis is not a casual impression. An inference is not a fact merely because it has been entered into a formal record.

The distinction can disappear surprisingly quickly.

A later reader may not know how the information originated. They may see only that it appears in an earlier record. The authority of the institution then begins to attach itself to the assertion.

The question which should remain visible is simple:

Who first said this, and on what basis?

Repetition Is Not Corroboration ↑

Suppose an assertion appears in one record.

A second person reads that record and repeats it.

A third person reads the second record and repeats it again.

Eventually the same assertion may appear in five different documents.

That does not necessarily mean there are five independent sources.

There may still be only one source with five descendants.

This distinction matters because repetition can create the appearance of corroboration. A statement begins to look more certain simply because it has survived long enough to be copied.

The number of times something appears in a record tells us something about its distribution.

It does not, by itself, tell us whether the original assertion was correct.

The Right to Provenance ↑

That is why provenance matters.

When a significant assertion affects a person's life, it should be possible to ask:

Where did it originate? Was the source identified? Was it something directly observed, something reported by somebody else, a professional conclusion, an allegation, an inference or copied history? What evidence supported it? Was contrary information available? Where was the assertion subsequently sent?

Without provenance, institutional information can become detached from the conditions under which it was created.

The record preserves the conclusion while losing the uncertainty.

Years later, somebody reading it may encounter the conclusion without ever seeing the question that should have accompanied it.

How do we know this?

When Age Begins to Look Like Authority ↑

Old information can acquire a peculiar authority.

If something has remained in a record for ten or twenty years, its survival can begin to feel like evidence that it must have been established.

But longevity is not verification.

An error does not become true because nobody corrected it.

Sometimes nobody knew it was there. Sometimes the person concerned never saw the record. Sometimes they challenged it but the challenge did not travel with the information. Sometimes they lacked the documents needed to demonstrate the problem. Sometimes the institution simply continued using what it already had.

The passage of time can therefore create an inversion.

Instead of the institution being required to show why the assertion was sound, the person affected by it can find themselves being asked to disprove something whose original evidentiary basis has become difficult to discover.

The Record Begins to Shape What Comes Next ↑

Once an assertion has entered an institutional record, it does not merely sit there.

It can influence the way later information is interpreted.

A new clinician, administrator, investigator or decision-maker may approach the person through what has already been written.

That is understandable. Records exist partly so that every encounter does not have to begin again from nothing.

But inherited information can also establish the frame through which the next encounter is understood.

Behaviour may be interpreted in light of the existing record. New observations may be given meaning by an earlier conclusion. A later report can then appear to confirm the earlier one, even where both ultimately depend upon the same starting assumption.

The process can become circular:

assertion → copied history → interpretation → decision → new record

The new record then becomes evidence for the next person.

Institutional Inheritance ↑

Most of the people who inherit an institutional error did not create it.

That matters.

A clinician reading an old history may be acting conscientiously. An administrator may reasonably rely upon information supplied by another department. A police officer, lawyer, social worker or government employee may encounter a record whose origins they have no immediate reason to question.

Institutional error does not require a succession of malicious people.

It can be sustained by ordinary people reasonably relying upon information that the system has already taught them to trust.

That is one reason institutional errors can be so persistent.

Responsibility becomes dispersed while the effect remains concentrated upon the person described by the record.

The Person Meets the Record ↑

There comes a point at which the person is no longer dealing only with another human being.

They are dealing with everything the institution already believes it knows about them.

The person may explain that something is wrong.

The record may say otherwise.

The person may remember an event differently.

The record may have been repeated for years.

The person may possess new evidence.

The institution may possess a history.

That creates an imbalance which is easy to underestimate.

The individual has to challenge an assertion. The institution can point to a file.

And the longer the file has existed, the more easily its existence can be mistaken for proof.

When the Error Produces Descendants ↑

The consequences become more serious when the original information has already influenced decisions.

An assertion may affect an assessment.

The assessment may affect treatment.

Treatment may affect behaviour or circumstances.

Those consequences may generate further observations and further records.

The descendants of the original assertion can therefore acquire lives of their own.

By the time the starting point is challenged, the institution may no longer be defending one sentence. It may be defending years of decisions that were built upon it.

This is where correction can become institutionally uncomfortable.

If the original information was wrong, what else has to be reconsidered?

Which later decisions relied upon it?

Who else received it?

What consequences followed?

What would acknowledging the error require the institution to revisit?

Why Institutions Can Find Correction Difficult ↑

There are many reasons an institution may resist reopening an established history.

Some are ordinary. Records may be old. Staff may have left. Systems may have changed. Evidence may be incomplete. Different professional opinions may genuinely exist.

Other pressures can also be present. Correction may require additional work. It may create legal or financial consequences. It may call previous decisions into question. It may expose failures of procedure or supervision.

None of those possibilities proves that resistance in a particular case is deliberate.

But they demonstrate why a correction system cannot depend entirely upon the willingness of the institution whose previous work is being questioned.

A trustworthy institution needs a way to examine its own history without treating every challenge as an attack upon the institution itself.

Correction Is Not an Accusation ↑

Acknowledging an error does not necessarily mean that somebody acted dishonestly, maliciously or incompetently.

People make decisions using the information available to them at the time.

New evidence appears.

Understanding changes.

An interpretation which once seemed reasonable may later prove unsustainable.

A system which treats every correction as an admission of wrongdoing creates an incentive to defend old conclusions.

A system which treats correction as part of learning has a better chance of correcting itself.

Accountability must remain available where negligence, misconduct or deliberate wrongdoing is established.

But ordinary correction should not require an institutional defeat.

Who Carries the Burden? ↑

There is another imbalance.

The institution created the record, stored it, copied it and distributed it.

Yet when the person discovers an error, the practical burden of correction can fall almost entirely upon them.

They may have to locate the original document, determine where the information came from, gather contrary evidence, write applications, pursue complaints and then discover which other records inherited the same information.

The institution has the information infrastructure.

The individual is asked to reconstruct its consequences from outside.

That raises a question which has become increasingly important to me:

What happens to the person when the institution has more capacity to defend its error than the person has to correct it?

The Error Should Carry Its Family Tree ↑

If institutional information can travel, its lineage should be capable of being traced.

A significant assertion should not become an orphaned fact.

There should be some way of determining its source and, where it has materially affected later records or decisions, identifying its descendants.

That does not mean every sentence in every file requires an elaborate audit trail.

It means that information capable of substantially affecting a person's rights, treatment, reputation or access to services should not be allowed to acquire authority while losing its provenance.

The same principle must apply when the information is corrected.

If the original assertion travelled, changing only the original record may leave its descendants untouched.

The error may therefore continue operating after the institution has technically corrected it.

Truth Has to Travel Too ↑

This leads to a principle which seems almost obvious once the problem is visible:

An institution that propagates an error should carry a corresponding responsibility to propagate its correction.

If information was copied into another record, the correction should reach that record where reasonably possible.

If information was supplied to another decision-maker, that person or institution may need to know that its status has changed.

If a decision materially relied upon information later shown to be wrong, there may be a need to reconsider the decision rather than merely amend the historical file.

Otherwise the institution can truthfully say that it corrected the record while the error continues to shape the person's life elsewhere.

The System Does Not Have to Be Infallible ↑

I do not expect institutions to be incapable of error.

That would be an impossible standard.

The more important test is what happens after an error becomes visible.

Can its origin be found?

Can the evidence be examined?

Can uncertainty be restored where certainty was never justified?

Can the information be corrected without requiring the affected person to fight every descendant of the original assertion separately?

Can decisions built upon the error be reconsidered?

And can the correction become as institutionally durable as the mistake once was?

Those questions take me beyond the problem of institutional error itself.

They lead to the question of what correction must actually mean.

Changing a sentence is not enough if the sentence has already travelled.

Correcting a record is not enough if the record has already produced a decision.

And correcting a decision may not be enough if the decision has already produced a consequence.

That is where the next chapter begins.