The claim
If persistent structures are what more complex structure grows on (CLAIM-185), the first structures to persist get built on first. An early framing, given a code, gathers claims that rest on it, objections to it and terms defined against it. Replacing it later means re-anchoring all of that.
The physics read for this line has the extreme case. Dunitz & Bernstein (LIT-917, unread) are cited for crystal forms that could no longer be made once a more stable form had appeared, because seeds of the new form were everywhere. Ostwald (LIT-920, NOTE-709) shows the mechanism at the bench: once a room has been used for alum, a supersaturated drop cannot be kept free of seeds, and a single stray seed revives a sterilized batch (pp. 291, 324). Luria names the same danger on the record side. It is named for a mnemonist whose total recall crowded out meaning (LU-ADR-011).
The record’s remedy is retirement by status. A Superseded claim stays, with its successor named, so that what pointed at it still resolves, and a reader is sent on. That works only if new structure stops growing on the retired anchor.
What would answer it
Two findings would. The first: Superseded entries draw almost no new citations after retirement, other than citations of them as history, so supersession does turn growth off. The second: a framing replaced late in an argument costs no more to replace than one replaced early.
The check can be run on this record now. Count the citations of each
Superseded CLAIM and TERM made after its retirement date, and separate
those acknowledged as history (inactive-ok) from the rest.
Note of 2026-10-11 (later): constancy against resilience
THEORY-132 (Holling) is added as a ground. Managing a system for constancy can erode its resilience, its capacity to absorb change and persist within its domain. That is a stronger form of this counter. A record that enforces identity strictly maximizes constancy, and may reduce the argument’s capacity to reframe. CLAIM-195 draws the same point from the constraint-set side.
Note of 2026-10-12: the bar for lock-in
The will-organization line’s TERM-020 (protective lock-in) sets the bar this counter has to clear: a regime that “generates the conditions that justify its own continuation and reduces the capacity to leave it”. That needs a loop, and TERM-020 excludes inertia and habit. Path dependence alone, Ostwald’s ubiquitous seeds, does not qualify. THEORY-178 adds that, in the one regulatory network studied, a lasting state needs a positive circuit. THEORY-136 says what would show one: hysteresis or a lasting shift after a temporary push. This counter has not shown a loop. A candidate would be that forcing moves into the grammar miscodes them, the miscodes produce correction entries, and those entries feed more filing. That is unshown.
Note of 2026-10-29: what holds it back
The count under “What would answer it” has not been run: citations of each Superseded CLAIM and TERM made after its retirement, with those acknowledged as history set apart. Nor has the loop TERM-020 requires been shown (note of 2026-10-12), so the claim has not yet cleared its own bar for lock-in.
Note of 2026-10-31: the dense field, in organization studies
Bencherki, Brummans and Vézy (LIT-1902, pp. 1356–1357) state this counter’s worry as a property of fields. Each being-in-becoming is a “field of relation, more or less densely assembling the preindividual action it has integrated thus far”. Dense fields, “crystals or well-oiled bureaucracies”, “cannot be reorganized easily when they meet other fields”. Loose ones, “water or a project team”, take up what they meet. The paper offers no measure of density and no account of how a field becomes dense, so it adds a name for lock-in and not a mechanism. It does place lock-in where this claim does: in what has accumulated, not in any single anchor. The crystal that is the paper’s dense field is also the record’s seed (CLAIM-185, note of this date). No status, title or warrant changes.
Note of 2026-11-01: lock-in as generative entrenchment
The mechanism this claim names has a standard source the record now holds as a registration: Wimsatt’s generative entrenchment (LIT-1926, unread), as Griesemer and Shavit quote it (LIT-188, §2; NOTE-173). An entrenched element is one “that later-evolved interactions will depend on for their existence and character, so that perturbations of it will tend to be catastrophic to the whole system while changes in the later interactions tend to lead to tolerable system changes”. That is this claim’s “replacing it later means re-anchoring all of that”, with the asymmetry stated. The paper also calls the outcome “highly contingent, even if contingently irreversible” (§2): nothing makes a given anchor entrench, but once it has, the change is unlikely to be undone. The record’s remedy, retirement by status, works against that asymmetry only if superseded anchors stop gaining dependents, which is this claim’s own test. Same structure, not shown same mechanism.
No status, title or warrant changes.