The objection
CLAIM-115 says 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”. There are two ways to read it.
As description. “What survives” is what in fact gets through to a receiver. Then the positive clause is close to true by definition: a distinction that no receiving system can recover has not survived for anyone, and one that a receiver can recover and use has survived for that receiver. What has content is the negative clause, “not the wording or the proposition”, and that is a separate claim, CLAIM-061’s, whose “necessary” half CLAIM-053 already finds unargued in the manuscript.
As a criterion. The thesis says which distinctions a translation ought to keep. Then it depends on which decisions the receiver is to make. CLAIM-050: “the manuscript restricts this to a family Q of tasks — tell teasing from reproach, recognize a position, choose a reply”. CLAIM-115 leaves Q open. Its summary: “the choice of observables and decision tasks is stated to be normative and contestable (§11)”. Its “What it does not say”: “It does not say the observables are fixed in advance”. Until Q is fixed the thesis is a schema, a form that yields a claim for each choice of Q.
The defeat condition fits neither reading. It fixes observables, “(stance, footing, alignment, expected uptake)”, which the thesis declines to fix. And it is descriptive: it asks whether “Human judgements of communicative fidelity” are predicted by one measure better than another. What it tests is a narrower empirical claim, that these four observables predict fidelity judgements better than propositional or embedding similarity. That claim can be true while the thesis, read as description, adds nothing to it, and false while the thesis, read as a criterion, is right about what translations ought to keep.
A descriptive test cannot settle a normative criterion. CLAIM-085 makes the point for language-model judges: “what a translation ought to preserve needs criteria the model cannot supply”. Human judgements of fidelity are also a description of interpreters, so the point carries to them.
CLAIM-115 uses no TERM. Its “receiving system” and “act upon” are bound to
none of the record’s senses of fidelity, such as TERM-034’s, which is
“relative to an interpreter”. So nothing in the record fixes which receiver
the thesis means.
One consequence bears on CLAIM-140, which says the October thesis (CLAIM-139) “replaces what a receiver can reconstruct with what the analyst’s probe system can identify”. On the criterion reading, CLAIM-115’s receiver was already an idealized interpreter facing the manuscript’s Q, not a named audience. The contrast CLAIM-140 draws is then one of degree.
What it does not say
- It does not say the thesis is empty. A near-analytic positive clause can still organize an account, and the negative clause is a substantive claim.
- It does not say CLAIM-140 is wrong. The October thesis does drop the receiver from its wording; this says the receiver in CLAIM-115 was already the analyst’s.
What would answer it
Either of these:
- fix Q, or the receiver population and its tasks, in the thesis, so that it becomes a substantive empirical claim the present defeat condition fits;
- restate the thesis as a normative proposal, with a normative warrant (an argument about what a translation owes its readers) and a defeat condition of the matching kind, such as a case in which keeping the named distinctions is plainly the wrong thing for a translation to do.
Note of 2026-10-12: positions on which receiver, filed on main
This claim says “nothing in the record fixes which receiver” CLAIM-115 means. Since PR #121 branched, main has filed CLAIM-160, which fixes intended meaning by the writer’s projected receiver at the write and measures fidelity for the actual reader at each read. CLAIM-171 objects that the projected receiver is itself one more reading. Neither binds CLAIM-115’s “receiving system” to a TERM, so neither settles which receiver CLAIM-115 means. They are filed positions on a neighbouring question: whose receiver counts.