The claim
A93 §5: “Krifka tells us which changes in conversational standing count as performative updates; Blackwell gives us a way to compare whether different utterances preserve the information necessary to recover or predict those updates.” It “could become one of the manuscript’s distinctive theoretical integrations.”
Where it went
Krifka (2024) was named closest prior art from A33 on (CLAIM-023), cited in C6’s introduction, and ranked ninth in A93’s priorities. The manuscript drops him. It keeps Blackwell over a task family Q (CLAIM-050) whose tasks (“infer whether a statement licenses criticism or shared laughter”) are performative updates in all but name. The pairing would give Q a principled source. It would also do the work of the dropped performative fidelity (CLAIM-054).
Krifka, read
Krifka (LIT-835) gives “change in conversational standing” a precise object: the propositions an update makes true. Brandom’s deontic score (LIT-794) is a perspectival alternative. Neither offers observables, so this claim’s defeat condition is untouched.
Note of 2026-10-30: what holds it back
Its title has Blackwell comparison test whether “a rendering” preserves the information. The Active CLAIM-230 shows that is not well typed: the order compares rendering procedures over a designed distribution of situations. Its own premise, CLAIM-050, was narrowed at version 4 to say so, and to report fidelity only. Krifka (LIT-835, NOTE-602) fixes the changes that count only for commitments, obligations and declared facts, not the speaker–audience relation, and he offers no observables; CLAIM-227 finds no validated measure of footing. A narrowing to procedures would concede CLAIM-230 but frees nothing. The claim waits on its own test: eliciting performative updates with stable response distributions.