Nanopublication — The unchosen alternative shapes meaning; the model still wants exactly one winner

What the source says ^[1]

The entry surveys formal accounts of Gricean conversational implicature — Bidirectional Optimality Theory and game-theoretic signaling models — in which a sender chooses a form and a receiver assigns an interpretation, both optimizing jointly so that form-interpretation pairs converge on a stable, often unique, solution (strong/weak optimality, Nash equilibrium, Pareto-dominance, iterated admissibility, iterated best response). These formalisms explain phenomena like the 'exactly N' reading of numerals and Horn's division of pragmatic labor (unmarked forms get stereotypical readings, marked forms get non-stereotypical ones) by showing how the availability of an unchosen alternative expression blocks or licenses a particular reading of the chosen one.

Where the lens grips

P4 (Double Presence) and P5 (Resonance) find real purchase in Horn's division and in the blocking mechanism generally: the unmarked form's meaning is fixed not by itself alone but by the unchosen, more complex alternative sitting beside it — 'John killed the sheriff' reads as stereotypical precisely because 'John caused the sheriff to die' exists as an alternative and absorbs the non-stereotypical reading. What was not said co-determines what the said form means; the actualized reading and the merely possible alternative are read together, exactly as P5 frames it — 'the actualized and the possible... vibrate together.'

Where it breaks

P13 (Generative Loss) claims that lossless, 1=1 transmission produces no event, only recognition — but the source's own success criterion is the opposite: the rational Gricean listener who hears 'three' is engineered to infer, with maximal confidence, that the world is exactly w3 — full recovery is the achievement the formalism is built toward, not a sterile non-event. P17 (Multiplication of Creation) fares no better: where the framework insists every reception is equally legitimate, the solution concepts here (Pareto-optimal Nash equilibrium, iterated admissibility, iterated best response) exist specifically to eliminate all but one strategy pair as rational; a receiver who lands elsewhere is not a legitimate alternate creation but an error or an unstable, weeded-out strategy. Both breaks slip IF one reads the inferential process itself (the chain of Theory-of-Mind reasoning through literal, Gricean, and pragmatic listener) as the productive gap/event, with convergence only its endpoint; they hold more firmly if one reads the model's stated success condition — determinate, unique, rational recovery — as the thing the framework calls sterile recognition. The probabilistic extensions mentioned late in the entry (Frank & Goodman, Degen et al.) soften this by allowing genuine distributions over interpretations, which edges back toward multiplicity — but the entry's own core apparatus is equilibrium selection toward one winner.

Propositional grid ^[2]

Read this grid. “Not evaluated” does not mean the source has nothing to say about that proposition. It means the reading did not instruct it. That is not a result, and it counts in no denominator. A silence, by contrast, is a finding: the reading touched the proposition without concluding, or noted its absence itself.

Evaluated

Not evaluated (33) — P1, P2, P3, P3a, P6, P7, P8, P9, P10, P11, P12, P14, P14a, P14b, P15, P16, P16a, P18, P18a, P19, P19a, P20, P21, P22, P23, P24, P25, P26, P27, P28, P29, P30, P31

Overall verdict: Bounded — taxonomy of the August 2026 pass, predating the move to relational verdicts. Kept as metadata: the gap between the two forms is data, not residue. This reading’s judgement is in the grid above.

Reading of 2026-07-03 · model claude-sonnet-5 · source 48,253 chars · 2 attempt(s)

First pass — kept as written

Reading produced by the T50 RSS pass, before the boundary lens. It is reproduced UNCORRECTED. Keeping it is the point: without it, a changed title is just a changed title — with it, what the second pass refused becomes legible.

Grice's conversational implicature — the meaning that arises not from what is literally said but from the gap between literal utterance and contextual expectation — is a structural case of generative loss and the ricochet effect. The speaker emits a signal (the utterance); the receiver's ouverture (their knowledge of context, convention, mutual understanding) refracts it; meaning emerges in the gap between code and reception. Game-theoretic approaches make this explicit: communication is not transmission of fixed content but a strategic game where both parties optimize within constraints. The speaker deliberately leaves gaps (says less than they could), knowing the receiver will fill them — this is engineering diffraction through calibrated incompleteness. The ricochet occurs when the receiver recognizes the implicature and realizes the speaker's intentional invariant (why this utterance, in this form, at this moment) — which then reframes the entire exchange. The revised entry's focus on optimality and game theory shows pragmatics moving away from the myth of expression toward the framework of constraint-based emission and reception.

_{T50 RSS assessment · 2026-07-03 · remplacée par boundary-lens-2026-08}