Take a water authority set up in 1908 to manage a watershed: allocate draw rights, maintain the works, keep a register. It has done that ever since. The gauge readers are gone and telemetry does their job, the register is a database, and the allocation formula has been amended twice for drought. Hold all of it against the features that justified the original grant and nothing has moved. The authority knows the same kind of thing, about the same class of people, on the same schedule, to the same effect, with the same route to object. All of its changes are the kind the test was built to wave through.1
The adequacy test returns adequate, and that is the correct output. But the nation whose territory the watershed drains was not party to the 1908 statute, was not consulted on it, and has said so continuously for a hundred and eighteen years.
%% title: What the comparator never reads
%% caption: The test runs between the grant and the exercise. Whether the grant was anyone's to issue sits outside the loop, and an arrangement that has not drifted gets a clean pass without anyone reaching that question.
flowchart TB
classDef gate stroke-width:2px;
classDef unread stroke-dasharray: 6 4;
O[Whose was it to grant?]:::unread -.->|never reached| A[Grant at authorization]
A --> B{Exercise moved on<br/>a justifying feature?}:::gate
B -->|no| C[Adequate]
B -->|yes| D[Institution answers<br/>for the distance]:::gate
D --> C
D --> E[Reauthorize, constrain, or stop]
The grant is the ruler
The tests in this sequence work by comparison. They take the original grant as the reference and ask what the institution has done with it since.2
None of these tests evaluates its own reference point. They can show that an institution has wandered from what it was given, but nothing in them can show that what it was given was never anyone's to give.
Worse, the test is most confident exactly where it is least informative. An arrangement that has treated the same people the same way for a century has not moved on any feature, so it clears every gate. Its clean pass reports stability rather than legitimacy. The better an institution is at not changing, the less the test has to say about it, a blind spot I call founding immunity. This sequence of essays built machinery to police what a grant's descendants do, and never built anything that reads the grant itself.
The conscientious version is the problem
The careless institution is not the interesting case. Take the one that reads all of this and agrees with it. The authority accepts that an Edwardian statute cannot speak for a nation that was not at the table. It opens a consultation protocol. It offers seats on the board, a consent requirement for new works, a veto over drilling in specific reaches, a data-sharing agreement with real teeth. Each concession is larger than the last and made in good faith.
Offering to put the relationship on a proper footing assumes the relationship is the authority's to arrange. An earlier essay rejected the phrase effective jurisdiction on exactly this ground, because a critique of power has to apply its own test to its own words. Here the same test catches the remedy instead of the vocabulary.
What the institution hears
An institution built around the adequacy test has one reflex when it is told no, which is to hear a position. Positions can be improved on with more seats, an earlier stage, a broader veto. The process takes the refusal as an input and returns a revision. Because each revision really is more generous than the last, the sequence looks like progress from inside. From the other side it is one claim restated at rising volume: we convene, you attend, and we decide what attending gets you. A refusal goes in and a deferred agreement comes out. Nothing in the machinery can take no as a final answer, because it was never built to produce final answers.
Standing over whether
The standing ladder in these essays has rungs. Being believed is one, and a claim can be fully believed and still change nothing.3 Making an error consequential for the institution that made it is another. Putting the rule itself on trial rather than the clerk who applied it is a third, and the essays have treated that as the top of the ladder.4
One more rung sits above contesting a rule, and it concerns whether this body is the one whose rules run here. Standing to raise that question is constitutive standing. Ordinary standing works inside the relationship, while constitutive standing asks whether the institution had any right to set up the relationship at all.
Below it, every rung can be granted without contradiction. An institution can build an appeal, a binding finding and a docket with the power to revise the rule, and remain entirely coherent. Constitutive standing cannot be granted, because granting it would itself be an exercise of the power in question.
The procedure I do not have
All the other essays in this sequence end in a procedure. It says who must show what and in what order, with the burden on the party that holds the evidence and wants the power. A procedure turns a mood into something that can be decided.
This essay ends without one, and the absence is not a gap I expect to close. Any test for when refusal is correct would be a test, run by somebody, and the somebody would be the party whose standing is being challenged. A court, a tribunal or a treaty body could run it instead, and some claims tribunals do hear founding grievances. Each of them takes its own authority from the order whose founding is in question, so it can weigh the refusal only in the terms that order allows. A framework that rules on everything ends up claiming jurisdiction over its own jurisdiction, and I would rather leave the instrument incomplete than build that.
What survives
Two things survive, and the first is a disclosure. When the adequacy test returns adequate, it should say what it read. This exercise treats the people it governs the way the grant allowed is a narrower sentence than this arrangement is legitimate, and the distance between them is the founding. A verdict that names its blind spot and a clean bill of health are different things. Institutions have every incentive to file the first as the second.
The second is a question about failure behavior, and it can be answered without settling anything about the merits.5 When the answer is no, what does this institution's process do with it? Where no routes back for revision, the process has no refusal state, and every no inside it is a yes with a date attached. That is a fact about the machine, and anyone can check it. Checking it does not require anyone to decide whether the 1908 statute was good.
The authority has not exceeded its grant in a hundred and eighteen years. For all of them the nation has asked who gave it, and nothing the test reads has ever had to answer.
Notes
The adequacy test, and why a portal and a scoring model fall on opposite sides of it: Same Charter, Different Office.
The three principles behind the comparison: Justification Is Not Contagious. The refusal underneath all of them: Five Ways a Fact Becomes a Right.
Credibility and standing come apart: a claim can be believed in full and still reach nothing: Standing as Access Control.
The rungs, and where a signal gains authority over the decision: Filed at the Wrong Size, The Rule Is Never on Trial and Signal Authority.
What an institution is structurally capable of doing after an adverse finding, as distinct from what it says it would do: Stop the Machine, Not the Person.