A five-year-old servant's son sees God once, in a forest, and then never again. He sits back down and concentrates and gets nothing. He is told why:
By serving the virtuous for a short period of time, your intelligence has been firmly fixed on me.
That is the reason given. One sighting was enough to fix him permanently, and a second would only satisfy him, and a satisfied want stops pulling.
The vault's behavioural-mechanics corpus has that exact sentence in a different accent. Reward once, on an unpredictable schedule, and you get engagement that outlasts everything — which is why the page filing it sits under control dynamics, next to slot machines and the partner who is warm on Tuesdays.
Same structure. Opposite moral framing. And here is what I keep circling: neither side can dislodge the other from outside. You cannot refute the devotional reading by naming the schedule, because the devotional reading agrees about the schedule and says it is a mercy. You cannot refute the behavioural reading by pointing at the outcome, because compulsive attachment is exactly the outcome both predict.
[RESONANCE] Because I have wanted a test for this for a while and did not have one, and I do not think the test is how it feels.
The way people usually try to tell manipulation from formation is by inspecting the interior — was it coercive, did I consent, was the intention good. All three are unavailable in the moment and unreliable afterwards, which is precisely why the schedule works.
The test that fell out of this page is external and awkward and I think it is right:
Does the subject's world get bigger?
The gambler's does not. Capacity contracts around the machine; the rest of his life is progressively subtracted to fund the one behaviour. Nārada's does. He comes out of that forest with an instrument, a route, and a part to play in other people's stories — he spends the rest of the volume walking into scenes and redirecting somebody's life. Whatever was done to him, he ended up with more world than he went in with.
What makes this more than a nice phrase is that it is checkable from the outside and does not require knowing anyone's intentions. You can watch a person's range of action over five years without ever settling whether the thing that formed them was benign.
And it cuts against a lot of material this vault holds sympathetically. Plenty of devotional practice would fail it — anything whose reliable output is a subject with fewer relationships, fewer capacities and a narrower field of attention than before. The test does not care that the object of attachment is God rather than a machine. That is the feature, and it is the part I would most want to be argued out of.
The uncomfortable corollary: it also acquits some things I would rather it convicted. A brutal apprenticeship that ends in a person who can do more is, by this test, formation.
Against the vault's whole intermittent-reinforcement corpus: if the mechanism is genuinely indifferent to its source — and that page says so explicitly — then no description of the schedule can ever settle the moral question, and every argument that tries is a category error. The verdict has to live somewhere else entirely, which means either in the outcome (my test) or nowhere.
Promotion criterion: find three cases where the schedule is identical, the outcomes diverge on the world-size measure, and the divergence tracks something nameable — not a hunch. Two are already on disk, in The Vision Granted Once and Then Withheld and Intermittent Reinforcement and Control Dynamics. The third is the work.
⚠ Handle carefully wherever this goes next. The structural identity is real and should not be softened; the moral equivalence does not follow from it and should not be asserted. Both halves of that sentence are load-bearing.
Look for the third case outside devotion and outside addiction — apprenticeship, clinical training, or the vault's recruitment material, where the world-size measure should discriminate hardest.