The claim
A33 opened: “A translation is pragmatically faithful to the extent that it preserves the structure of the transformations an utterance induces in its audience’s interpretive state.” A35 §10: “The pragmatic identity of an utterance is not exhausted by its propositional content or by a static specification of its illocutionary force. It is partly constituted by the structure of context-dependent transformations it induces in a communicative system. Accordingly, pragmatic translation fidelity can be modeled as an approximate structure-preserving correspondence between the interpretive and performative dynamics of source and target utterances.” The short form: “A faithful translation preserves not merely what an utterance means, but how its meaning acts upon—and is transformed by—the communicative contexts in which it is encountered.”
The move it made, A35 §5: “a shift from invariance of attributes to invariance of transformations and their composition”. Meaning, at A33, is “what the utterance does to possible interpretive states”, not a fixed vector. A35 §10: “We’ve moved from comparing representations to comparing processes.”
Why it was replaced
It presupposes interpretive states that can be put in correspondence. A35 §8 already weakened that (TERM-033), and A39 made observations primary (CLAIM-046). The revision is CLAIM-083. Its intertwining core survives in the manuscript as CLAIM-117.