The Subordination Test — What May and May Not Be Added v1.0

 

The Subordination Test — What May and May Not Be Added v1.0

Theoretical foundations: Grant C. Sterling (Eastern Illinois University). Analysis and synthesis: Dave Kelly. Prose rendering: Claude (Anthropic). 2026.

Companion to The Corpus as a Thinking System, Section VII.

---

I. The Premise That Decides It

Complete happiness must lie wholly within our control, and only assent does. So the system’s field is not narrow by modesty — it is narrow because it is exactly coextensive with what is up to us. Anything added either falls inside that field, in which case it was always the system’s business, or falls outside it, in which case adding it means admitting something not up to us into a system whose whole claim is that nothing of that kind belongs there.

---

II. The Test

An added function is harmless when it is subordinate: used as material, governed by the one criterion, never itself a measure of success.

It is fatal when it becomes criterial: a second thing the practitioner can succeed or fail at, so that a correct assent accompanied by a bad forecast counts as a partial failure.

The seven TSA runs are the evidence. Every one of them failed at the point where a capacity that ought to have been subordinate became the measure — the intensity rating, the Brier score, the condition of the world.

---

III. The Five Omissions Taken in Turn

Deliberation about what to do. Already inside the field, not outside it. Clause (b) governs the impulse to act, and acting on a correct judgement is an assent. What the system declines is not deliberation but a technique for it — and that decline is a real omission, because the decision point’s assumption that the practitioner already knows does not always hold. Adding provision here does not hamper the system; it completes clause (b). Whether the corpus already carries that provision under the discipline of action is an open question and is not asserted here.

Correcting factual error. Outside the field and safely added, because it cannot compete. No one thinks getting a fact right is the good; the temptation to make it criterial is weak. A man may use Descartes’s precepts all morning and the system is untouched, provided he does not begin scoring himself on them. This is the clearest case of harmless addition.

Institutional and collective action. Outside the field and addable at the level of material — what is appropriate to do in an office, a city, a church. The corpus already does this in its executive and organizational material. The danger is not in the subject matter but in the drift the Meadows run displays: the good of the whole is exactly the kind of end that, once admitted, begins to measure. Add the subject; refuse the criterion.

Relief as a criterion. Cannot be added at all. It is not a function but a rival measure, and admitting it converts the system into the cell the two therapies occupy. This is the one case where addition is destruction rather than hindrance, and the reason is the first premise: relief is not up to us, so a system measured by it has forfeited the completeness that made the whole argument go.

External verification. Cannot be added either, and for a related reason. The act is unobservable and produces no score; any proxy invented for it — a count, a rating, a log — would become the thing improved. The forecasting run shows the mechanism exactly: once the output is a number, the number is what the practice optimises. This absence is not a defect to be repaired but a consequence of the act being the kind of thing it is.

---

IV. The Rule

Three of the five may be added and two may not. The three are added the same way: as subordinate material, with the single criterion left untouched, and with an explicit statement that competence in them is not a measure of anything.

---

V. What This Settles About the Corpus’s Shape

The system is not comprehensive in the sense of covering everything a man needs. It is comprehensive in the sense of covering, completely, the one thing that is his.

A practitioner needs facts, deliberation, and ways of working with others, and he is free to take them wherever they are best had — including from the audited systems, most of which are competent in precisely the fields the corpus abstains from. What he may not do is let any of them tell him how he is doing.

---

Theoretical foundations: Grant C. Sterling (Eastern Illinois University). Analysis and synthesis: Dave Kelly. Prose rendering: Claude (Anthropic). 2026.