nucleation

CLAIM-244: The defeat tests take one holistic, task-free judgement of fidelity as ground truth, though the line's own theses say fidelity is task-relative, interpreter-relative and not a single ranking

4 min read

Properties9
StatusProposed
Version2
Rolecounter
Warrantinterpretive
Tagsphilosophy-of-language, psychometrics
Linepragmatic-transport
Objects toCLAIM-115 — What survives translation is the set of communicative distinctions and relations a receiving system can still reconstruct and act upon, not the wording or the proposition, CLAIM-139 — A communicative kind's identity is grounded in its relational organization rather than in any expression or encoding; statistical methods recover aspects of that organization from restricted observations, translation transports it, and fidelity is which independently identifiable, decision-relevant relational distinctions persist, CLAIM-005 — Pragmatic fidelity is the match between distributions of context-indexed communicative judgements across a correspondence of source and target contexts, not the similarity of one canonical meaning vector; the match constrains a translation only once one outcome correspondence is fixed across many items, and summed over source contexts alone it charges for loss and not for addition, CLAIM-078 — The theory succeeds if its observational and decision measures predict changes of communicative identity that text similarity misses and follow its compositional predictions across reconstructions; it needs neither quantum contextuality nor exact Noether laws
Objected byCLAIM-240 — A holistic fidelity judgement is fidelity under the judges' default task, a weighting of the profile's components that can be estimated on some renderings and tested on others, so it is well defined and consistent with task-relativity, and a defeat condition that uses it must name the population and fix the weighting before the test, and across populations with unequal access to the source must fit the weighting only on the components each can perceive
UsesTERM-034 — fidelity, as selective preservation of decision-relevant observational information under a communication constraint, relative to an interpreter

Graph View

The objection

The criterion in each central defeat condition is a human judgement with no task and no receiver named:

  • CLAIM-115: “Human judgements of communicative fidelity”;
  • CLAIM-005: “held-out human communicative-fidelity judgements”;
  • CLAIM-139: “Human judgements of shared communicative kind”;
  • CLAIM-078: “communicative-identity changes”, which A203 §37, quoted there, makes “whether humans recognize two realizations as instances of the same communicative kind”.

The line’s own theses say there is no such single quantity:

  • CLAIM-128: “Transports whose distortion profiles trade off are incomparable until a task fixes weights, so no single ranking of translations holds across tasks”;
  • CLAIM-106: fidelity is “a graded, potentially asymmetric property of transport, not an equivalence relation”;
  • CLAIM-074: “fidelity under a fixed interpreter, across interpreters and across states of adaptation are different quantities”;
  • TERM-034: fidelity is “relative to an interpreter”.

A small case shows the difficulty. Rendering A keeps the source’s footing and loses its stance; rendering B keeps the stance and loses the footing. By CLAIM-128 they are incomparable until a task fixes the weights. A judge asked which is more faithful ranks one above the other anyway, and so supplies a weighting nobody stated. The theory can then go one of two ways.

  • It fits its weights to the judges. The weights CLAIM-128 leaves “to be chosen per task” are then chosen to match the criterion, and the prediction is circular.
  • It keeps its own weights, and a mismatch is put down to the judges having weighed for a different task, which CLAIM-128 allows. Then no result defeats the thesis.

The same holds for the interpreter. The judges are some interpreters, at some state of adaptation. By CLAIM-074 their fidelity is one quantity among several, and the defeat conditions do not say which.

The “same kind” question has a further gap: there is no wording to give lay judges. TERM-040: “What makes two such systems the same kind is the open part.”

So the line faces a choice the record does not state. If CLAIM-128, CLAIM-074 and CLAIM-106 are right, the criterion in the defeat conditions is not well defined. If the criterion is well defined, those theses are in trouble.

What it does not say

  • It does not say holistic judgements carry no information about fidelity.
  • It does not say which horn is right. If the relativity theses are wrong, the defeat conditions are well posed and those theses fall instead.

What would answer it

  • A defeat condition indexed to a stated task and a stated receiver population, with the weights of CLAIM-128’s profile fixed before the data are seen.
  • Or an argument that holistic judgements fix the weighting non-circularly: for instance, weights estimated from one set of renderings and judges and tested on another, so that the weighting is a finding and not a fit.
  • A QUESTION refining QUESTION-002 could hold the open part: which task and which receivers fix the human criterion against which fidelity measures are validated?

Where it stands

On 2026-10-10 the record replied. CLAIM-240 answers the dilemma: a holistic judgement is fidelity under the judges’ default task, a weighting of the profile that can be estimated on some renderings and tested on others, so it is well defined and consistent with CLAIM-128, and a weighting fixed before the test leaves no room to blame the task. That the defeat conditions name no task or receivers is met by the revised theses CLAIM-220 and CLAIM-204, and is held open in QUESTION-075. The objection stays open until the owner decides on those two, since CLAIM-115’s and CLAIM-139’s own conditions are unchanged.

Note of 2026-10-29: the owner decided

CLAIM-220 superseded CLAIM-115, and CLAIM-204 superseded CLAIM-139, on the owner’s delegated ruling of 2026-10-29. Both name the receivers and the tasks, and both now fix the weights before the data, so the conditions this objection faulted are no longer the ones the record holds open. It stays pointed at the two it was filed against, and QUESTION-075 stays Open.

Backlinks