nucleation

CLAIM-190: The anchoring measures are computed in the grammar that shapes the argument, so by themselves they cannot certify that drift fell: drift must also be judged under standards the record did not supply

3 min read

Properties10
StatusSuperseded — by ARG-028 — Inbound references per code and turns per filed entry show which structures anchor and that drift falls as anchoring keeps up with arrival
Version2
Rolecounter
Warrantinterpretive
Tagsdesign-rationale, epistemology
Linenucleating-argument
Worksluria-nucleating-an-argument
Rests onCLAIM-181 — The record fixes the grammar an argument is expressed in, though not its shape: its closed vocabulary of schemes, roles and relations formalizes early, and a move that fits no CLAIM, TERM, ARG or declared relation is forced into one or lost until the misfit is noticed and the grammar amended to hold it, CLAIM-264 — Two objects can agree on every probe in a restricted family without being isomorphic, so equivalence under a probe family is coarser than identity unless the family is rich enough, CLAIM-266 — A family of probes that supplies an optimized system's training signal cannot also certify what the system preserves: optimization pushes toward whatever the probes reward, including what they cannot see, so preservation must be shown under probes not used to train it
GroundsCASE-071 — Kovacs, Deutsch and Freitag's metric bias in MBR decoding: translations a metric selects, that the metric and its neighbours rate better and human raters do not
Illustrated byCASE-084 — The perfectly accountable false theory: a community whose record traces every commitment and whose vocabulary leaves out the right explanation

Graph View

The claim

The nucleating-argument line proposes measures for its claims: inbound references per code, to tell an anchor from a structure that merely persists (QUESTION-038, CLAIM-179), and turns per filed entry, set against drift (CLAIM-186). Each is computed in the record’s grammar: codes, relations and statuses.

CLAIM-149, on the pragmatic-transport line, says “a probe family that supplies a transformation’s training signal cannot also certify what the transformation preserves: preservation must be shown under probes that were not used to train it.” The record’s grammar is the training signal for the argument. It rewards coding, linking and announcing changes. A measure built from the same grammar will register more coding, linking and announcing, whether or not the argument drifted less by the participants’ own standards. CLAIM-142 says observational equivalence is relative to the probes, and these probes are the record’s.

This is distinct from CLAIM-181. CLAIM-181 is about how the record shapes what the argument becomes. This is about how the record shapes what its success looks like.

What would answer it

Inbound-reference counts and turns per anchor agree with drift judged blind, by raters using the exchange’s own standards, on stretches held out from the grammar’s influence: for example, re-derivations and unannounced changes of thesis coded before the record existed.

Note of 2026-10-12: a documented case, and an open dispute about raters

CASE-071, filed in PR #121, is now a ground of this claim. Translations picked by a neural metric are rated better by that metric and its neighbours, and human raters find them no better. It is CLAIM-149’s point made by selection rather than training. Its limit: it is machine translation scored for accuracy, not an argument scored for drift. This claim’s remedy, drift “judged blind, by raters using the exchange’s own standards”, treats raters as the criterion. That is CLAIM-217’s position, CLAIM-206 holds that human judgements are one probe family among others, and the dispute is open.

Note of 2026-10-13: premises re-pointed to their general forms

This claim needed only probe-relative equivalence and the point that training probes cannot certify, not the account of translation it was filed under. Under ADR-040 that content is filed separately as CLAIM-264 and CLAIM-266, in the formal-apparatus line, and this claim’s rests_on now names CLAIM-264 and CLAIM-266 in place of CLAIM-142 and CLAIM-149. The body’s citations of CLAIM-142 and CLAIM-149 stand: they are where the record first argued the point. The change is that setting pragmatic-transport aside no longer takes this premise with it.

Note of 2026-10-28: moved to ARG-028

Under ADR-061 a claim that a measure cannot show a thesis is the argument from the measure to the thesis, with the failing critical question in its body. This claim is now ARG-028, from CASE-183 to CLAIM-185 and CLAIM-186, form evidence, Proposed, and its title is the answer there to “Does the measure measure what it is said to?”.

Backlinks