Care

Outlying

A wet network runs on oaths. Not rules, not targets, not policies. Oaths — the kind sworn by a person, in a role, binding them to something, with identity as the collateral.

A rule can be worked around. A target can be hit without changing anything. An oath binds. A person who breaks an oath stops being the kind of thing that person is. That is why oaths are wet. They live in the interior, not in the paperwork.

And a wet network — a hospital, a school, a legal system, a service — does not have one oath. It has three. They are not written down together. They are not sworn by the same people. And they are not always compatible.

The three oaths

The practitioner's oath binds the person at the point of contact — the doctor, the nurse, the teacher, the solicitor — to the one in front of them. It refuses aggregation. It says: this person, this moment, act in their interest. It is the oldest oath in every profession, and it is the one that cannot be measured, because the thing it binds you to is a person and a moment, and neither one can be counted.

The manager's oath binds the person running the service — the ward manager, the head of department, the operations lead — to the operation. It says: keep the thing running. It aggregates by necessity, because a service is not a person. It hits targets, manages flow, keeps the budget, protects the staff, holds the line on the waiting list. It is the only oath in the stack that produces an auditable artefact, because running a service is what produces dashboards.

The leader's oath binds the person holding the institution — the chief executive, the medical director, the head of the trust — to the population and the long arc. It says: this service must still exist in ten years, and it must be fair to everyone it serves, and it must be sustainable. It operates at a timescale nobody in the network can measure directly. It is also the oath that sounds best in speeches, which is not the same as being the oath that is acted on.

All three are wet. All three are sworn. All three are right. And all three bind the swearer to a different object over a different horizon:

None of them is a lesser oath. None of them is a mistake. And no node in the network can hold all three at once, because the three objects they bind to are not always compatible, and the three horizons they run on are not measurable against each other.

The collision at the point of care

In a well-structured network, the three oaths live in three layers. Practitioners keep the practitioner's oath. Managers keep the manager's oath. Leaders keep the leader's oath. Each layer translates the oath above into terms the layer below can act on, and absorbs the contradictions on the way down.

But wet networks do not always translate cleanly. When the manager's oath becomes operationally dominant — because it is the only one that produces metrics, and the network acts on what it can measure — the manager's rules get pushed down to the practitioner's layer. When the leader's frame becomes rhetorically dominant — because it is the one that sounds best — it also gets pushed down, as culture, as expectation, as tone.

Which means the practitioner ends up carrying all three oaths at the point of care, in a single wet node, in the moment of contact with a person who cannot wait for the network to sort itself out.

This is what we saw earlier, in the piece about the doctor who froze. A training rule from the manager's layer — do not perform a procedure you have not been signed off for — collided with the practitioner's oath — act in this patient's interest. The doctor was not carrying two rules. He was carrying two oaths, from two layers, neither of which had a precedence rule against the other, and no way to hand either one back up.

That is not a failure of the doctor. It is a failure of the network to keep its oaths where they belong. The manager's oath belongs in the manager's layer. The leader's oath belongs in the leader's layer. When they are pushed down to the practitioner's layer, they are being carried by someone who never swore them and has no structural authority to refuse them.

One oath becomes the default

If a wet network is run on three oaths, and only one of them produces artefacts, the network will gradually come to act as if that one is the real oath and the others are aspirational.

Not because anyone decided it. Because the network can only act on what it can see, and the manager's oath is the only one that shows up in a report. The practitioner's oath produces no artefact — you cannot audit a moment of attention. The leader's oath produces no artefact until the long arc arrives, which is usually after the current leadership has moved on. The manager's oath produces dashboards, spreadsheets, flow charts, plans.

So the network runs on the manager's oath by default. It becomes the operational reality of the service, and the other two become the rhetoric.

Which is where conformance culture comes from. It is not a departure from the manager's oath. It is the manager's oath, taken seriously and operationalised. The problem is not that managers are bad. The problem is that the manager's oath is the only one the network can act on, and so it slowly becomes the only one the network keeps.

How the wrong people end up in the wrong layer

If the manager's oath is the dominant one, then the network will tend to promote people who are good at it. That sounds correct. But the selection mechanisms do not work the way they appear to, and the same family of mechanisms operates at every layer.

They are all the same mechanism, wearing different faces. The network selects for what it can see. And the thing it can see is never quite the thing that matters.

The Peter mechanism. People are promoted for competence in the role they are leaving, not the role they are entering. A good practitioner is promoted to manager — and a good practitioner's competence is the practitioner's oath, not the manager's. A good manager is promoted to leader — and a good manager's competence is the manager's oath, not the leader's. So each layer ends up filled with people whose real strength is the layer below. The network is not staffed by fools. It is staffed by people wearing the wrong oath for their skill, because the promotion was based on the previous one.

The artefact mechanism. The person whose work produces a visible object — a report, a slide, a dashboard, a metric — is promoted over the person whose work produces an effect that cannot be shown. This is the mechanism that produces conformance specifically. It is not that targets are chosen for their value. It is that targets can be shown. And so the visible surface of the work rises, and the invisible work that produced it does not.

The silence mechanism. The person quietly holding the network together is invisible, because the network is holding. The practitioner who prevents a bad outcome is invisible, because nothing happened. The manager who absorbs pressure before it reaches the ward is invisible, because the pressure did not arrive. The network can only see the failures it failed to prevent. So it selects against the people who prevent them.

The burnout mechanism. The people who hold the practitioner's oath most strongly are the most likely to burn out, or to leave. Which means the network is filtering them out by attrition. The ones who stay are the ones who learned to keep the operational oath without the practitioner one, because that is the only way to survive the pressure. The network does not need to fire the wrong people. It only needs to wait.

There are others. Every wet network has its own. The point is not the list. The point is the family: the network selects for what it can see, and what it can see is never the oath that matters. The selection mechanisms are how the manager's oath becomes dominant without anyone choosing it, and how the middle layer — the layer that should be holding all three oaths and absorbing the contradictions — gradually empties of the people best equipped to do it.

This is nobody's fault and everybody's problem

The manager who is good at communication is not wrong. She is doing what the network rewarded her for. The leader whose oath sounds best in speeches is not a fraud. He is articulating the only oath that can be articulated. The practitioner who holds the practitioner's oath and is invisible to the network is not a martyr. He is doing his job, and being penalised for it by a structure that cannot see the job he is doing.

None of the individuals in this picture are bad. The picture itself is a structural consequence of running a wet network on visible artefacts, with wet oaths that produce no artefacts of their own.

Which means the fix cannot be to find better people. There is no pipeline of better people. There is only the question of whether the network can be made to see what it currently cannot, and to hold the three oaths in the layers where they belong instead of pushing them down into the node that has to face the person.

The outlier

An outlier is a data point that falls outside the expected range. In a metric-driven system, the outlier is a problem. The patient who costs too much. The case that takes too long. The employee whose work does not show up in the report. The person whose needs do not fit the protocol.

But to be outlying also means to lie outside — to be at the periphery, on the boundary, in the place where the map stops and something else begins.

The outlier is the person the three oaths fail to cover. Not because any one of the oaths is wrong. But because the network's selection mechanisms have left no one who can hold all three at once, and the outlier is standing at the point where all three meet and none of them wins.

The patient whose case doesn't fit the operational frame. The practitioner who keeps the practitioner's oath and is therefore invisible to the manager's layer. The manager who tries to hold both the operational oath and the practitioner one, and who is not visible from above. The leader who attempts to keep the leader's oath and is overruled by the quarterly reporting cycle.

All of them are outlying. All of them are on the boundary of the network's visibility. All of them are people the network has failed to see, at the point where seeing them would have mattered most.

This is why Care has the name it has, and why the site is called outlying.one. The domain was chosen before the argument was written, and only later did the two become the same thing.

An outlier in a data set is a point that the model cannot accommodate. An outlying one is a person for whom the network's three oaths have not been reconciled, because the network has been shaped — by the selection mechanisms, not by anyone's intention — into a form that cannot reconcile them.

The outlying one is not a problem to be managed. They are the place where the network's own invisible layer becomes visible. Which is the same claim the poem made at the beginning of Care, and the same claim the dry systems made in the middle of it, and the same claim the timeline in the previous piece has been making for thirty-two years.

The outlying one is not the exception. The outlying one is the point where the network is finally seen.

What comes next

One wet network has developed a mechanism for holding its own three oaths together. It is not medicine, and it is not management. It is law.

Precedent is the piece about that mechanism — how case law pins ambiguity, and how the same structure could hold the three oaths in a form the network can actually see.