Voices — Whose voice a CLAIM is in, within the record’s own deliberation. Not whether the record believes it, which is the status: the two vary independently, and the mismatched pairs are the information (ADR-032).
Counter — put forward on an opponent’s behalf, at full strength — a steelman, or an objection anticipated as devil’s advocate. A counter that goes Active means the record was persuaded, and the thesis it opposed needs revisiting.
71 of 547 CLAIM documents. Back to the full index.
| # | Title | Summary | Status |
|---|---|---|---|
| CLAIM-001 | The conditional-diffusion literature already covers directed transport between observational scenarios, including changing covers | The owner at U34, quoting A93’s research opportunity: “I think conditional diffusion literature has us covered here.” Accepted for compositional conditioning and transport between distributions; never shown for changing covers, and limited by A97 itself. | Proposed |
| CLAIM-010 | On this account every control system has a will | The permissiveness objection. Raised by the assistant at A112, and re-derived at least six times afterwards as though new. Answered by agent-indexing (CLAIM-080, CLAIM-116); one code is what makes the recurrences one objection. | Rejected |
| CLAIM-024 v4 | Nothing here is new but a question: planning theory, control theory, executive control and institutional theory already study every mechanism the account names, and where the account has been set beside the expected value of control and Ainslie’s interests it differs only in asking what governs how the driver changes rather than who drives, a question those mechanisms may answer without it | ”The most important objection” (Section 7.1), re-derived three times in the exchange with different criteria each time. The paper concedes the mechanisms (CLAIM-021), and beside the expected value of control and Ainslie the record concedes a difference of question (CLAIM-159). The objection still presses that the mechanisms may answer that question without the account, and planning and institutional theory are not compared. | Proposed |
| CLAIM-040 | Symmetry is a kind of invariance, and Noether’s theorem is the point: invariance under transformation yields conservation | The owner at U37, quoting A102’s “We do not require the structure to remain invariant”: “symmetry is a kind of invariance, no? isn’t that part of the point of Noether Theory?” Conceded at A105. Its Noether half was limited by the assistant in the same reply. | Active |
| CLAIM-089 v2 | Northstar simply prefers speed: the governing relations are revealed preferences of the organization, not a separate organization of efficacy | The redescription objection (Sections 1, 3.5). The paper answers the individual-attitude version; the aggregate revealed-preference version is still open, so this counter stays Proposed. | Rejected |
| CLAIM-110 | Brooks’s architecture, Kelso’s coordination dynamics and Levin’s bioelectric selves are three different ways of understanding what an organized whole is | The assistant’s framing at A8: three accounts of distributed agency as three rival pictures of the whole. Rejected after the owner’s objection at U7, which the assistant conceded at A10. | Rejected |
| CLAIM-134 v2 | A non-eliminative structural realism that keeps relata but denies them any intrinsic identity-fixing feature is the position Chakravartty’s dilemma targets, and granting numerical plurality without relational differentiation, as A194 and A198 do, concedes a non-relational ground of identity | The record’s objection to the owner’s U54 individuation thesis. No one raised it in the exchange. Chakravartty’s fn. 9 names Esfeld and Lam, the position A194 calls “close” (NOTE-149). Leitgeb and Ladyman’s edgeless two-node graph is the hard case the exchange’s examples avoid. The owner’s character move answers the dilemma for kinds and not for identity. | Proposed |
| CLAIM-138 v3 | The Newman objection is the collapse objection at the level of ontology, and the remedy offered for both, independently specified relations, concedes that relational structure alone does not fix communicative identity | The exchange drew the parallel itself, four times (A187 §6, A191 §10, A203 §19, A218 Ch2.10). What it did not draw is the cost. The standard answer to Newman adds non-structural input, and so does admissibility, which sits uneasily with the claim that relations constitute identity. The objection presses on the owner’s individuation thesis only if its relations are read extensionally, and it does not touch the “partly” of relational constitution. | Proposed |
| CLAIM-145 v4 | Trope analysis of the kind A155 proposes is Propp’s and Lévi-Strauss’s transformational analysis, whom the manuscript should cite as precursors | The record’s objection to CLAIM-147; no one raised it in the exchange. A155 presents TV Tropes as “the clearest large-scale empirical setting” for the structuralist side of the theory and names neither Propp nor Lévi-Strauss, whom the record reads as the direct precursors. The circularity half recurs in the exchange’s own controls (A178 Study B and Ch12.9), and A218 drops the control. The prior-art half is close to granted. | Active |
| CLAIM-163 v2 | ”Synchronization” is an analogy without a mechanism: phase synchrony among coupled oscillators has a coupling, an order parameter and transitions, valuations that fail to cross have none of these, and the word adds nothing to “the curves do not cross” | The record’s own anticipated objection to the owner’s framing, and the kind ADR-032 found most often real: same structure, not the same mechanism. Open. TERM-050 defines synchronization without dynamics, so the claims survive with the word replaced, but then the word is only a name. The body says what would turn the analogy into a mechanism. | Superseded — by ARG-016 |
| CLAIM-165 v2 | A combined form constituted by mutual entrainment is not thereby a self in the subject sense: two dancers in phase are one coordination, an autonomous process, and still two subjects | The record’s own objection to CLAIM-172, narrowed in version 2 after the owner’s reply (CLAIM-175). It concedes that the dancing pair is a combined form and an autonomous process. It holds only that nothing in the coupling makes that form a subject, which ADR-024 reserves “self” for. Open: whether that is the sense of “self” worth asking about is itself in question. | Proposed |
| CLAIM-168 | Identifying agents by their will-organization is circular: will-organization is defined for an agent already given, and the agent-versus-device comparison presupposes an identified system, so neither can produce the factorization it is meant to find | The record’s anticipated objection to the owner’s thesis. TERM-023 is “agent-indexed”: a role in an agent already warranted. The will paper supposes its agent (CLAIM-080) so as not to face CLAIM-010. Orseau et al.’s formal agent-device comparison likewise “assume[s] a clearly identified system”. To identify agents by will-organization reverses an order of explanation that was chosen deliberately. Open: CLAIM-174 replies that the test is run on candidate factors, not on agents. | Proposed |
| CLAIM-171 | A stored text means only what its readings make of it: the writer’s projected receiver is one more reading, so an unread text has possible meanings and no fixed one | The record’s anticipated reception-side objection to the owner’s storage claim. Its nearest support in the line is CLAIM-111, a text supporting a distribution over possible communicative situations. No reception theory is held or read. Open: the reply (the joke changed for the reader, not in what it was made to do) works only if the writer’s frame is itself recoverable, and that is QUESTION-031’s problem. | Proposed |
| CLAIM-173 v3 | The threshold criterion measures coupling strength, and coupling is not constitution: a metronome driving a musician, or a nerve driving a muscle, forms a strongly coupled block that clears the threshold without being one system | The record’s anticipated objection to CLAIM-157, and Adams & Aizawa’s coupling-constitution objection (LIT-455) aimed at the new criterion. Near-decomposability is about the size of couplings, not their direction. One-way drive can make a tightly coupled block. The entrainment claim excluded exactly this case. Rejected on the owner’s challenge (CLAIM-167): they should pass, and direction gives a system’s architecture, not its membership. | Rejected |
| CLAIM-181 v4 | The record fixes the grammar an argument is expressed in, though not its shape: its closed vocabulary of schemes, roles and relations formalizes early, and a move that fits no CLAIM, TERM, ARG or declared relation is forced into one or lost until the misfit is noticed and the grammar amended to hold it | Shipman and Marshall’s objection to formal representation (LIT-756), turned on this record, and narrowed on the owner’s correction of 2026-10-11. As the assistant first put it (“a crystal takes the shape of its seed”), it was wrong about crystals. A seed templates the lattice, but the habit is set by growth conditions. The objection therefore reaches the grammar of the argument, and not which claims it comes to hold. Some evidence for it is already in the record: the flourishes it declined to code, and the literature roles that “recurred” but stay in prose (ADR-032). | Proposed |
| CLAIM-184 v3 | Anchors entrench: a structure that persists early can lock in its framing, so that later structure grows on it and alternatives become hard to reach, and retirement by status is the record’s remedy only if superseded anchors stop attracting new structure | The cost side of the owner’s reframing (CLAIM-185), proposed by the assistant on 2026-10-11 and filed at the owner’s request. If persistence is what lets structure grow, it also lets early framings grow structure that is costly to undo. Grounded on Dunitz & Bernstein (unread) and Ostwald (read). The test: do Superseded entries keep drawing new citations? | Proposed |
| CLAIM-188 v4 | Where the model with a candidate structure is singular at the parameters that switch the candidate off, with real log canonical threshold λ below d/2 and multiplicity m, the reification rule’s Gaussian Occam term overcharges keeping the candidate by (d/2 − λ) log n + (m − 1) log log n to leading order, so near the threshold the rule prunes candidates the Bayes factor keeps, and its derivation holds only where that model stays regular | A narrowing of CLAIM-154, filed on 2026-10-11 from a synthesis review that found the record already held the readings that limit it. CLAIM-154’s rule prices a candidate with a Gaussian Occam factor, which is exact in the conjugate case it was re-checked in. The candidates it is meant for (blocks, sectors, couplings with hidden structure) are often singular where they are switched off. There THEORY-087 puts the penalty at λ log n − (m − 1) log log n, not (d/2) log n. On CASE-152’s product model the rule prunes, near the threshold, candidates the exact Bayes factor keeps. CLAIM-154 is narrowed to models that stay regular at the switch-off. | Active |
| CLAIM-189 | Fidelity is the wrong measure of a developing argument: maturing drops and changes things legitimately, so what a record can secure is traceability, change that is announced and located, not preservation | The counter to CLAIM-191, filed with it. The transport line measures fidelity, and CLAIM-034 already says fidelity is one goal among several. A developing argument aims at maturing, and TERM-052 says maturing drops things legitimately. A fidelity profile would score legitimate dissolution as distortion. What the record secures is that change is announced and located: traceability. | Proposed |
| CLAIM-190 v2 | The anchoring measures are computed in the grammar that shapes the argument, so by themselves they cannot certify that drift fell: drift must also be judged under standards the record did not supply | An evaluation counter to the persistence and growth-rate claims, distinct from CLAIM-181. CLAIM-181 says the record shapes the argument. This says it also shapes the measures. Inbound references per code and turns per anchor are counted in the record’s own grammar, so by CLAIM-149 they cannot also certify that what they encourage reduced drift. Proposed from the synthesis review. | Superseded — by ARG-028 |
| CLAIM-194 v4 | A gap or coupling criterion cannot tell a boundary a system produces from one imposed on it: a membrane the cell makes and a beam joining two pendulums pass alike, so the block test cannot by itself supply the individuality minimal agency needs | A counter to the block criterion and to the agency test’s first condition, found by two separate reviews of the unlinked THEORYs in the third synthesis pass. It grants the target-relativity CLAIM-152 states. Its point is different: with the target fixed, a near-decomposability gap is a statistic of coupling, and coupling is the same whether the system made the boundary or something else did. THEORY-072 and THEORY-1119 place the line at production and self-distinction instead. No case has been measured: the point is that a coupling statistic has no term for who made the coupling. | Active |
| CLAIM-203 v4 | The only design for the October thesis builds its outcome in: the act is fixed by whoever writes the items, and the topic-only baseline is built to fail on the critical cells | Found in the record’s audit of the line on 2026-10-10 (empirical lens); no turn of the exchange raises it. CASE-040 “makes the act a manipulated factor rather than an outcome”, so recovering the act recovers the item writer’s intent, close to a manipulation check. Its held-out items are built to “break the familiar correlation”, so a lexical topic model fails on them by construction. And its baseline is topic-only NMF, not the embedding or surface similarity of CLAIM-139’s own defeat condition. Through CLAIM-135 it reaches CLAIM-139’s epistemic clause. CASE-040’s “What it cannot show” covers neither point. It does not say the design is useless as a first pilot. | Superseded — by ARG-011 |
| CLAIM-205 v3 | Functional and relevance-theoretic accounts of translation, as Nida, Nord and Gutt state them, need not fix one context or a static stance profile of the rendering | From the interlocutor’s reading of PR #121 and its reconciliation, relayed by the owner on 2026-10-12. It accepts CLAIM-202 as the line’s principal empirical thesis and objects to its rival set. Narrowed and promoted on 2026-10-30, when Nida, Nord and Gutt were read; the baseline it proposed is CLAIM-504. | Active |
| CLAIM-206 | If a communicative kind can outrun every probe, human same-kind judgements, themselves one probe family, can neither confirm nor defeat the October thesis; if it cannot, the kind does in the tests only what an observational class does, which the record’s term for a kind says it is not | Found in the record’s audit of the line on 2026-10-10 (structure lens, with one point from the empirical lens); no turn of the exchange raises it. The dilemma reaches the epistemic and fidelity clauses CLAIM-139 says the programme does test, not only the ontological clause its “What it does not say” sets aside. Two defeat conditions misfire under it: CLAIM-142’s is met trivially if the judges’ verdicts count as a probe, and CLAIM-139’s third disjunct is either always met or never decidable. It does not say kinds are unreal, and it is not CLAIM-150’s underdetermination of the ontology. | Proposed |
| CLAIM-208 v2 | The compositional predictions are upper bounds whose constants cannot be estimated from serial-reproduction data in a metric fixed beforehand, so the compositional half of the success criterion cannot fail | Found in the record’s audit of the line on 2026-10-10 (empirical lens); no turn of the exchange raises it. The recursion e_n ≤ (∏κ_j)e_0 + Σε_i∏κ_j is a theorem once its constants are the true ones, so a chain cannot depart from it; it can only show a constant was misestimated. A sensitivity κ_i belongs to a reteller’s kernel and needs many retellings of each of several inputs at every generation, where a chain gives one. And CLAIM-059 says amplification “needs a metric other than total variation” (now held as CLAIM-267), so the regime depends on a metric chosen afterwards. CLAIM-078 records that the compositional half was dropped from the manuscript, not that it cannot fail as stated. It does not say error propagation is unimportant. | Active |
| CLAIM-216 v2 | A thesis that conjoins relational constitution, statistical recovery, transport and receivers’ recovery is defeated whole by any one of its empirical findings, though its clauses can hold or fail apart, so the line is better put as separate theses, each with its own defeat condition, resting on one another only where one is a premise of another | The interlocutor’s objection at U62 (A240) to the October thesis (CLAIM-139) and to the record’s restatement of it (CLAIM-204), given as the reason for superseding CLAIM-139: “Separates ontology, statistical recovery, and empirical transport instead of making them rise or fall together”. The record’s replies of 2026-10-10 had already separated the receiver thesis and the incremental thesis; the restatement joined the four clauses again under one defeat condition. Active since 2026-10-30; CLAIM-204 concedes it. It says what would answer it, and does not say an integrative statement is useless. | Active |
| CLAIM-219 v4 | A change in the global compatibility of a fixed scenario is either a change in contextuality proper, which the behavioural data read show none of, or a change in overlap consistency, which is ordinary context dependence and moves by at most 2ε when each context moves by ε | Found in the record’s audit of the line on 2026-10-10 (structure lens); no turn of the exchange raises it. Dzhafarov, Zhang and Kujala (LIT-264) find no contextuality in five kinds of behavioural data, and CLAIM-125 reports that 34 of 90 language systems signal too much to be contextual. If pragmatic data are noncontextual as a rule, CLAIM-105’s defeat condition is met by default; what is left to change is overlap consistency, which is the “trivial sense of contextuality” and is fixed by the local distributions. The only instance of the headline is the stipulated CASE-037. Among the claims LIT-264 is cited only by CLAIM-041, not by CLAIM-105. The objection is open; it does not say a contextual pragmatic data set cannot exist. | Active |
| CLAIM-221 v2 | The relational-constitution thesis’s defeat condition describes a standard rival the record holds no counter for, intention-based accounts of illocution, which plausibly meet it for every act in the four-speakers case except the institutional one | Found in the record’s audit of the line on 2026-10-10 (structure lens); no turn of the exchange raises it. CLAIM-151 is defeated by a held account that individuates teasing, reproach and reprimand by the utterance plus speaker-internal states. That is the intention-based line of Grice, Strawson and Bach and Harnish, none of whom the record holds. On it CASE-018’s friend, parent and comedian differ in communicative intention, and only the judge needs norms, which Searle (LIT-482) already supplies. The record’s own model for CLAIM-111 is of this family. The objection is open; it would narrow CLAIM-151 to institutional acts unless an argument from uptake meets it. | Proposed |
| CLAIM-222 v2 | No design in the programme includes an interlingual translation, and the defeat conditions are disjunctive, so English retellings and model chains could settle a thesis about what survives translation | Found in the record’s audit of the line on 2026-10-10 (empirical lens). The workbench entry for 2026-10-10 notes that the title’s question “is left with no translated text in the programme”; no CLAIM says it, and none draws the consequence. Every stimulus set the designs name is English, or names no language; Villon→Henley and Marot→Hofstadter never enter a design. CLAIM-115’s defeat condition ranges over “translations, retellings or cross-modal renderings”, so a result on English paraphrase alone can meet it. It does not say intralingual results are irrelevant. | Active |
| CLAIM-224 | Under partial relational constitution, a translation always changes the participants, audience and norms that the relational-constitution thesis makes constituents of the act, so whether the act survives turns on a correspondence of relations the structure cannot supply from inside | Found in the record’s audit of the line on 2026-10-10 (structure lens); no turn of the exchange raises it. A translation is addressed to a new audience under new norms, so on CLAIM-151 it changes constituents of the act, and the record’s way out is always a correspondence (“the appropriate relationships”, “corresponding relational organizations”). That makes QUESTION-005 a question about what the act is, not only how to measure it, and the record’s model of translation holds the audience fixed. CLAIM-138 spares CLAIM-151; this applies its point to CLAIM-151 through translation. The objection is open. It does not say acts never survive translation. | Proposed |
| CLAIM-225 v4 | Blackwell’s order measures what an ideal decision-maker could do, not what a given receiver can reconstruct, and the line’s claims that a rendering can improve access describe garblings that real receivers do better with, which is the shape of the Blackwell thesis’s own defeat condition | Found in the record’s audit of the line on 2026-10-10 (formal and structure lenses); no turn of the exchange connects the two claims. A generator that sees only the text makes its image a garbling of the text about the situation, so by Blackwell’s theorem the image is no better for any decision. CLAIM-087 says such an image can improve access. Either that effect is impossible, or judged access is not Blackwell informativeness, and then CLAIM-050’s standard is not the one CLAIM-115’s receiver and its defeat condition use. It meets the shape of CLAIM-050’s defeat condition, not its letter. It is distinct from CLAIM-140, which sets the analyst against the receiver: this sets the ideal receiver against an actual one. | Active |
| CLAIM-226 v2 | The defeat tests pool unnamed judges and treat their disagreement as noise, though the theory makes the act partly constituted by its audience; for historical sources there is no source audience at all, and for interlingual ones population differences are read as distortion | Found in the record’s audit of the line on 2026-10-10 (empirical lens); no turn of the exchange raises it. A theory on which the act depends on its audience predicts structured variation between judges, so averaging a convenience sample measures an unnamed population’s act. The source term of D(τ) for Villon, Henley or Marot can only be a modern reading, and for a live interlingual case the source and target judges are different communities, so the distance confounds population with transport. CLAIM-074 says fidelity differs across interpreters, as a thesis; it is not applied to the tests. It does not say cross-population fidelity is meaningless. | Active |
| CLAIM-230 v3 | Blackwell comparison needs the state each rendering is about, which natural texts do not supply, so a decision measure is fixed by whoever labels the state, and it orders translation procedures, not single renderings | Found in the record’s audit of the line on 2026-10-10 (empirical lens); no turn of the exchange raises it. Testing Blackwell’s order needs the joint law of the situation and the signal. For natural texts the situation is what interpretation recovers, so it comes from the stimulus design, the analyst’s labels or the judges, and decision-relevant information is measured against whoever fixed it. An experiment is a family of distributions, one per state, so a single rendering of a single poem is not one. CLAIM-050’s defeat condition, “one rendering is a garbling of another”, is then not well typed. It does not say a design-based state is illegitimate. | Active |
| CLAIM-232 v3 | Human ratings of stance or footing and human ratings of fidelity of the same renderings share rater construal, scale use and a general impression, which a machine similarity score shares with neither | Found in the record’s audit of the line on 2026-10-10 (empirical lens); no turn of the exchange raises it. Ratings of stance and footing share rater construal, scale use and the reading of one stimulus with ratings of fidelity, and an embedding score shares none of that. The manuscript’s split annotators (§10, as QUESTION-005 reports) remove the shared rater, not the shared construal. It is a second reason, beside CLAIM-233, that the central tests are easy to pass. Open, because a design could answer it. | Active |
| CLAIM-235 v3 | The line’s proposed quantities have no finite-data estimators: plug-in signalling from one or a few occurrences per context is positive in expectation when the true marginals agree, so the contextual fraction is undefined on raw tables, and deficiency over text-valued renderings has no estimator | Found in the record’s audit of the line on 2026-10-10 (formal and empirical lenses); no turn of the exchange raises it. The plug-in signalling of two empirical marginals is positive in expectation when the true marginals agree, so raw tables signal, and the contextual fraction is undefined on them. With one occurrence per context, a non-signalling rank-2 system gives Δ of 0, 2 or 4, with expectation 2: the values NOTE-654 finds in 51 of the 90 systems behind CLAIM-125’s “69 of the 90 systems signal”. The data are author-assigned corpus senses, not judgements. Deficiency over text-valued renderings has no estimator. It does not say signalling is rare in pragmatic judgements, where order effects are themselves signalling. | Active |
| CLAIM-241 v2 | Partial constitution by uptake makes competent participants’ uptake a defeasible standard of a communicative kind, not the criterion: where the act’s other constituents, such as its institutional conditions, diverge from how even informed participants take it, their verdicts can be wrong, and a standard must say whose competence counts, for which class of acts, and how divided verdicts are combined, which a majority does not | The interlocutor’s first refinement at U62 (A240), with its row on CLAIM-217, pressed at full strength. CLAIM-217 moves from “partly constituted by … expected uptake” to “the judges … are the criterion”. Partial constitution gives uptake a vote, not the last word, so where the act’s other constituents diverge from informed uptake the verdicts can be wrong. The claim also names no rule for combining divided verdicts and no class of acts. Open: answering it means saying whether a defeasible standard re-opens the horn CLAIM-217 closed. It does not say uptake is not a standard. | Active |
| CLAIM-243 v2 | Read as description, the manuscript’s thesis is close to true by definition; read as a criterion, it is a schema until the decision family Q is fixed, and its descriptive defeat condition can test neither reading | Found in the record’s audit of the line on 2026-10-10 (structure lens); no turn of the exchange raises it. A distinction no receiver can recover has not survived for anyone, so the descriptive reading is near-analytic. The normative reading leaves which distinctions count to the task family Q, which CLAIM-115’s summary calls “normative and contestable”. The defeat condition fixes observables the thesis declines to fix, and tests them descriptively: CLAIM-085’s split of normative from descriptive, applied to the thesis rather than to model judges. It does not say the thesis is empty. | Active |
| CLAIM-244 v2 | 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 | Found in the record’s audit of the line on 2026-10-10 (empirical lens); no turn of the exchange raises it. A judge asked “how faithful?” or “the same kind?” supplies an unstated task and weighting, and is one interpreter. The theory’s measure must then be fitted to that weighting, which is circular, or a mismatch blamed on the judges’ task, which leaves nothing that defeats. It does not say the judgements are useless, only that the defeat conditions lack the task and receiver index the line’s own theses say fidelity needs. CLAIM-140 is its neighbour: that one moves the criterion to the analyst, and neither says it is task-free. | Proposed |
| CLAIM-246 v2 | Sharing a kind is symmetric and blind to which work is the source, so the October thesis cannot say a translation is of its original, against the line’s own directed, graded fidelity and identity through descent | Found in the record’s audit of the line on 2026-10-10 (structure lens); no turn of the exchange raises it. CLAIM-137 makes realizations one act by a shared invariant, “not resemblance to an original”, and CLAIM-139’s defeat condition tests judgements of shared kind. Sharing a kind is symmetric and says nothing of descent, so an utterance made independently of the source, of the same kind, scores as a translation does. The line’s own theses make fidelity directed and graded (CLAIM-106) and identity partly a matter of transmission (CLAIM-002). The objection is open; it does not say descent must be part of fidelity, only that the kind criterion cannot state it. | Rejected |
| CLAIM-248 v4 | The founding stimuli (Villon, Henley’s Villon, Hofstadter’s Marot) and commentary on them are likely in language models’ training data, and what a context does to one model need not be what it does to another, so a model’s reading of their footing may be recall, and a directional match in one model may not hold in the next | Found in the record’s audit of the line on 2026-10-10 (empirical lens); no turn of the exchange raises it. Agreement in direction is what a model trained on human descriptions of teasing and reproach would show anyway, and Villon, Henley and Hofstadter’s Marot are likely in training data, so a model’s “interpretation” of them may be recall. ARG-008 and CLAIM-085 already bound what model results show about human pragmatics. What is new is that CLAIM-058’s validity bar is too low, and the contamination point. Minor. | Active |
| CLAIM-251 v2 | A change in which questions a receiver population can answer determinately is not by itself a change of cover in the sheaf sense: the cover belongs to the measurement scenario and the responses are an empirical model on it, so such a change is a change of empirical model, which can create or remove a gluing obstruction on a fixed cover but does not change what the cover is, and the operational notion needs its own name until a functorial link is shown | From the interlocutor’s reading of PR #121 and its reconciliation, relayed by the owner on 2026-10-12, on the open tension between CLAIM-197 with CASE-048 and the operational cover of CLAIM-201. It keeps the empirical notion and denies it the sheaf-theoretic name. Open. | Proposed |
| CLAIM-253 v2 | The intertwining defects are defined on unobservable interpretive-state spaces, so they are not identified from judgement data, and the commutator identity bounds observed order effects only with a further, unstated readout-intertwining defect | Found in the record’s audit of the line on 2026-10-10 (formal lens); no turn of the exchange raises it. The commutator identity is correct, but Φ, A and A′ act on interpretive states. CLAIM-046 says hidden-state models that give identical observations can differ, and a worked case shows two such models with defects 0 and the largest possible, so the defect is not a function of the data. What is observed is responses, and the identity reaches the observed order effect only through a readout defect between outcome spaces, which no draft names. It is distinct from CLAIM-011, which is about collapse: this is about identification. | Active |
| CLAIM-254 v2 | Read as the Saussure correction has it, the structuralist premise argues against transport: value is internal to one language state, so what can cross between languages is signification, the aspect the premise demotes | Found in the record’s audit of the line on 2026-10-10 (structure lens); no turn of the exchange raises it. CLAIM-072 corrects the manuscript’s Saussure to value and says the correction “does not undercut the manuscript’s use”. But value is fixed inside one language state, so a change of language is a change of the system that fixes every value, and Saussure’s own example has a term and its translation share a signification and differ in value. Value is also a fact of the language system, while the line’s observables are facts of utterances in use. The objection is open. It does not say relational identity is false, only that Saussure, read correctly, does not support its transport. | Superseded — by ARG-033 |
| CLAIM-258 v2 | Blackwell’s order compares experiments about one parameter, so it does not reach the normal translation case: source and target readers decide about situations with changed coordinates, and where the state is common a translator with context knowledge yields a rendering that is not a garbling of the source, which the order typically leaves incomparable with it and about which the direction the manuscript uses is silent | Found in the record’s audit of the line on 2026-10-10 (formal and structure lenses); no turn of the exchange states it. CLAIM-050’s comparison is over one state Z. Compensation (CLAIM-091) changes the social coordinates, so the target reader decides about a different situation. Where the situation is common, a translator who knows the context makes a rendering that depends on Z other than through the source, so it is not a garbling. Blackwell’s order is still defined then, but typically leaves source and target incomparable, and the direction the manuscript uses says nothing. It does not say decision-relative fidelity is wrong. | Proposed |
| CLAIM-278 v5 | An operator framework flexible enough to redescribe any dynamical process explains none of them until it forbids some outcome, and the shared-organization thesis has not yet said what it forbids | A7 §14’s own first limitation, which the exchange granted and the record files as the standing objection to the explanatory thesis: “A sufficiently flexible operator framework could redescribe almost any dynamical process. The model must earn its value through restrictive assumptions, predictive compression, and intervention-sensitive predictions.” It is CLAIM-024’s objection to the will paper, met by a second account. It is answered by a prediction the framework makes and its rivals do not. | Superseded — by ARG-036 |
| CLAIM-288 v7 | In a population whose operators each maximize an objective the modeller fixed, the norms the organization appears to maintain are set by the modeller, not by the organization: they can become its goals through their role in selecting what persists, but it does not produce or revise them | The record’s objection, raised by no one in the exchange. A2 gives each operator an objective J_i, its evaluation of the collective decision, and A7 and A10 propose sufficiency tests run on such populations. Barandiaran, Di Paolo and Rohde hold that systems optimizing an externally fixed function are not models of agency (NOTE-458, C5), and Di Paolo (2005) that a system’s norms must be norms for the system. The objection undermines the evidence route; it does not refute the thesis. It is answered if the operators’ objectives are themselves produced and revised by the organization. | Active |
| CLAIM-308 v3 | Where reuse cannot be forecast, recurrence weighted by reconstruction cost is the filing criterion, not only evidence for one: filing at the first use at which reconstructing again would take the overpayment past the cost of filing is within 2 − (c_a + u)/c_r of the hindsight optimum on every sequence of uses, at most twice, and no deterministic rule without foresight guarantees materially better, under fixed known costs, with staleness left out and the cost of recognizing what a filed argument answers taken as fixed | Filed at the owner’s request, from the rent-or-buy result of Karlin et al. 1988 (THEORY-348) set against CLAIM-303’s clause that recurrence is evidence of reuse, not the criterion. In CLAIM-303’s own terms the break-even rule needs no forecast and is worst-case optimal; ADR-032’s second-occurrence rule is that rule when filing costs about one reconstruction. It does not reach staleness, which the theorem forbids rather than prices, or drifting costs, and where reuse can be forecast an expected-reuse rule can do better, so CLAIM-303’s criterion survives there. | Proposed |
| CLAIM-311 | A search cut off before an exhaustive pass can establish absence when the arrangement certifies that the unsearched remainder cannot hold the item: a sorted table answers non-membership in a logarithmic number of probes, and the Threshold Algorithm stops early with its answer proved, so absence needs a certificate about the remainder, not exhaustion; similarity search in high dimension is where no such certificate exists | Filed at the owner’s request, from the database readings set against CLAIM-306’s asymmetry: presence by one find, absence “only by exhausting the collection”. Sorted order and the Threshold Algorithm each establish absence from part of the collection, because the arrangement bounds what the rest can hold. That meets CLAIM-306’s own defeat condition, and the claim is conceded: CLAIM-312 restates it. Where the arrangement gives no bound, as in similarity search in high dimension, CLAIM-306’s asymmetry stands. | Active |
| CLAIM-313 v2 | A fixed number of bins realizes one partition only when each item is stored once: with copies, overlapping partitions coexist, as hashing-based similarity search stores each item in many partitions and pays for it in space, so sortings conflict as arrangements only under a one-copy constraint, which binds bricks but not index entries | Filed at the owner’s request, from the readings on locality-sensitive hashing (Indyk and Motwani 1998; Andoni, Laarhoven, Razenshteyn and Waingarten 2017) set against CLAIM-310’s premise that “a set of k bins realizes one partition of the collection at a time”. That holds when each item is in one place. An index can hold many copies, and hashing-based similarity indexes do exactly that, trading the space of the copies against query time on an exact curve. It narrows CLAIM-310 to collections whose items cannot be copied, which includes the LEGO case; it does not touch the claim about descriptions. | Active |
| CLAIM-314 | Reusing a filed argument has a recognition cost that a fixed per-use price leaves out: certifying that stored answers cover a new question is NP-complete even for conjunctive queries, grows with the number and size of the stored definitions rather than being a constant per item, and is undefined for an item filed without a statement of what it answers, so each filing raises the cost of later reuse and per-item break-even decisions are not independent | Filed at the owner’s request, from the readings on answering queries using views (Levy, Mendelzon, Sagiv and Srivastava 1995; Chandra and Merlin 1977; Halevy 2001). CLAIM-308’s guarantee needs a fixed, known per-use cost c_a; recognizing that a stored answer covers a new question is a search whose cost depends on everything stored, so filing one item changes the price of using the others. It undermines the premise CLAIM-308 states, not its arithmetic. The hardness is worst case and over short definitions, partial reuse is cheap, and nothing here is measured on the record. | Proposed |
| CLAIM-337 v2 | Active inference already runs an internal loop: mental action selects among higher-level cognitive states and sets the precision of lower-level observations, structure learning reshapes the generative model without action, and a judgement is behaviour, so a judgement–observation loop may be action and perception one Markov blanket inward | The assistant’s, A1 of the exchange of 2026-10-16, on active inference’s behalf, strengthened at A4. Sandved-Smith et al. cast mental action as policy selection over higher-level cognitive states; Brouillet and Friston make felt fluency an inference about the precision of one’s own processing; Smith et al. let a generative model grow its own structure by inference and learning alone. With Chalmers’s point that judgements are behaviour, the owner’s judgement is active inference’s action relocated. What it does not reach: the diagram. A redescription with no action on the world concedes that action on the world is not part of the loop. | Proposed |
| CLAIM-339 v3 | A judgement–observation loop cut off from action loses one route for evidence it did not select, and the free-energy principle’s reason for minimizing surprise, that a persisting system must keep its states in bounds, needs action on the world, though a loop can confirm itself with action available too | The assistant’s, A1 of the exchange of 2026-10-16, narrowed at A4. Friston et al. (2010) say active sampling is “mandated by ergodic arguments based on the very existence of adaptive agents”, and the enactive account puts a system’s own norms in its capacity to act against tendencies toward its viability boundary. Kübler and Birbaumer’s extinction hypothesis is the empirical form. As first filed it said a loop without action has a self-confirming minimum. NOTE-766 shows the minimum arises with action available, so the counter now claims only that losing action removes one source of evidence the loop did not select. CLAIM-342 answers it. | Active |
| CLAIM-364 v4 | Processes conditioned never to object coordinate without override at the same behaviour and supervisory cost as processes that negotiate, so the two part only under intervention, when a constituent dissents, or in their history | A45’s, U45 of the exchange of 2026-10-17, against the owner’s non-domination criterion (CLAIM-362) read through its economy: “a system whose subordinate processes have been conditioned never to object might exhibit perfect local harmony”, and “A quiet subordinate process isn’t necessarily a satisfied one.” It is the manipulation objection (THEORY-1118), and QUESTION-071, applied to the new criterion. It does not say the criterion fails, only that cost cannot certify it. The authenticity line was set aside on 2026-10-05; whether it reopens is the owner’s call. | Proposed |
| CLAIM-393 | Ordinary and legal attributions of agency to an artifact cannot serve as the independent marks that would defeat the record’s agency test, because they turn on whether its mechanism is understood: a known mechanism invites “This is all just programmed!”, an inscrutable one invites fear, whatever the description lengths | From Boyle’s “inscrutability paradox” in The Line (LIT-1288, read in NOTE-1014): if an apparently general AI comes from well-understood programming, people will doubt it has autonomous consciousness, since it is “merely doing or saying that which we have programmed it to do”; if its workings are inscrutable, they will find it “more mysterious and potentially more threatening” (p. 30; pp. 54, 246, 272). CLAIM-174’s defeater counts a test-passing factor that is “by independent marks, not an agent”. For artifacts, the marks to hand, public and legal judgement, are not independent of what is known about the mechanism. The objection does not say CLAIM-174’s test is wrong, only that this route to defeating it is not open. | Proposed |
| CLAIM-411 v5 | Under the broad, nonphenomenal definition of interpretation the constitutive thesis uses, substantial processing that is not independently shown to be experiential meets the definition, blindsight’s apparently unconscious visual discrimination among it, so interpretation alone does not establish experience, and the thesis must say which further organizational properties, if any, make an interpretation experiential | A3’s “most serious challenge” to the owner’s thesis, which every later reply accepted. Once interpretation is defined broadly and without experience, as the thesis requires, it admits much processing that is not shown to be experiential: “Interpretation ⇏ Conscious experience. At least, not under an unrestricted definition of interpretation.” The objection is that the thesis is incomplete, not that it is false. A7 kept it as the theory’s “open sufficiency condition”. The record already separates access from experience (QUESTION-048, THEORY-577), and the candidate conditions the exchange named (global availability, integration, higher-order or recursive modelling, valuation, perspective) are the question QUESTION-254 asks. | Active |
| CLAIM-419 v2 | The record’s grammar is fixed at any moment but amendable by decision, and in practice a move that fit no scheme or relation has repeatedly been treated as a defect of the grammar and answered by amending it, so a move the grammar cannot hold is not only forced or lost; whether every such move is caught is not shown | The owner’s, at U18 of the exchange of 2026-10-22: “when we’ve encountered representational challenges, we’ve generally interpreted those as a symptom of a represntational weakness of the record and used that opportunity to mature the luria configuration, featureset, or both.” It answers CLAIM-181, which had no reply on file, with the record’s own decisions as evidence. It narrows CLAIM-181 rather than defeating it: the record shows moves that were caught and the grammar amended, and one case (ADR-032’s literature roles) where a recurring move was left in prose. | Active |
| CLAIM-426 v2 | Where a judgement’s circulation helps bring about what it says, its visibility, repetition and attributed plurality are evidence about what it will bring about, so the rule that a judgement’s authority should come from its grounds and not its visibility holds for judgements of what already is, and not without qualification for judgements of what someone or something will become | The owner, on 2026-10-23, after CLAIM-425 was filed: “this is reminding me of ‘hyperstition’”. Land’s definition (LIT-1396): “hyperstitions – by their very existence as ideas – function causally to bring about their own reality.” CLAIM-425 says not to let a judgement’s popularity substitute for its credibility. For a judgement whose popularity helps make it true, popularity is evidence, about the future it is helping to make. The claim narrows CLAIM-425, it does not reverse it: visibility still does not show what was the case. | Proposed |
| CLAIM-431 v3 | A corporation’s functional remorse cannot be told from prudence: any procedure or role duty that makes others’ harms its concern was put there for its own ends, reputation, liability or recruitment, so the line between harm as concern and harm as threat, and with it the line between a remorseful and a psychopathic corporation, does not hold, and the clinical word adds nothing to a description of its incentives | The objection to the record’s reading of the “institutional psychopath”, revised on 2026-10-23 when the reading was revised. The first version said the clinical word adds nothing because a corporation shows only psychopathy’s conduct; once the record assigns functional affect (CLAIM-039), that premise is gone. What remains is sharper: the reading’s distinction between a harm a corporation takes as its own concern (remorse) and a harm it takes only as a threat (fear) may not be drawable for a corporation, whose every concern was installed for its own ends. Velasquez’s “as if” (LIT-1432) in functional form. | Proposed |
| CLAIM-438 v4 | An ordering serving as the negotiation’s potential, as the potential account of valence defines it, tracks ‘wanting’, not ‘liking’: sensitized incentive salience wins a system’s conflicts, persists and resumes while hedonic impact measured by taste reactivity stays flat or falls, so the regime the account calls hedonic valence is motivational valence, and hedonic valence, if it is anywhere in the account, is elsewhere | Raised by the reading of Berridge and Kringelbach (NOTE-1110) filed on 2026-10-23 for the potential account. Sensitized ‘wanting’ meets all three of Frijda’s marks of control precedence, which TERM-076 uses to define potential, while ‘liking’ stays flat or falls. That is CLAIM-440’s own defeater, intensity that does not track control precedence. Hesp et al. (NOTE-1111) concede the same of their account, which puts valence in the system’s model of its own potential rather than in the potential. | Proposed |
| CLAIM-441 v3 | Flow tells against the potential account’s condition that valence is undefined where no pair is contested: on the record’s own account of flow nothing is overridden because an activity’s immediate reward and its longer-run value coincide, yet flow is reported as positive, so either flow’s positivity is intrinsic pleasantness and not goal-conduciveness valence, or flow keeps contested pairs its processes settle among themselves, or the positivity comes after the episode and not during it, or the condition is wrong | The assistant’s, from the comparison of the record’s will and affect lines of 2026-10-23, filed at the owner’s direction. CLAIM-162 explains flow as intrapersonal synchronization in TERM-050’s sense: valuations coincide, so nothing needs overriding. TERM-076 makes goal-conduciveness valence undefined where no pair is contested. Together they make flow’s valence undefined, while flow is reported as positive. The report is second-hand: the flow literature is registered unread (LIT-914), and that it is positive during the episode, rather than in retrospect, is exactly what is unchecked. Four exits are named; the objection holds until one is shown. | Proposed |
| CLAIM-448 v3 | An ordinal potential describes only the winning-move relation it is computed on; where an organization’s legibility regime selects which contests are admitted or recorded, an acyclic record, even a persistent potential, can be the product of selective contestation, so observed coherence is not evidence that the organization’s perspectives are integrated, and an intervention that improves its self-governance may at first increase the disagreement observed | From the exchange of 2026-10-24 (workbench entry of 2026-10-24, U19–U21), after the owner pointed to CASE-124 (“i think we already have a three variable system somewhere, where the third component is described as legibility”) and asked for the synthesis to be tied back: if the legibility regime decides which objections are registered, the record of wins on which CLAIM-440 reads valence can be acyclic for the wrong reason. Objects to reading a prevailing potential as an integrated organization’s valence without a check on what was admitted. | Active |
| CLAIM-449 v2 | The potential account of valence’s first consequence, no frustration, no valence, runs together two relations: perspectives’ contested rankings, which a record of winning moves registers, and demands that are jointly infeasible, which the compatibility complex registers; rankings can contest while every demand is satisfiable, and demands can be infeasible before any contest is recorded, so the consequence needs one of them named, and the absorbed writer, reported as positive with nothing contested, tests whichever is chosen | Raised in the exchange of 2026-10-23 (workbench entry of 2026-10-24, U5), when the owner asked for “cases or theories that aren’t richly connected but which may bear on questions or claims more deeply than we are giving them credit”. The reply proposed “a new counterclaim against consequence 1 of CLAIM-440, explicitly distinguishing evaluative disagreement, contested efficacy, and structural frustration”, grounded in the absorbed writer and the three-demand case. Filed here at the owner’s request to record the exchange. | Active |
| CLAIM-458 v2 | A reception split between populations with unequal access to the source can be partly a difference of access and not of weighting, so a comparison of their default tasks can fit weights only on the components every population compared can perceive | The limit CLAIM-240’s note of 2026-10-10 recorded and did not code, filed as an objection. CLAIM-240 reads a holistic fidelity judgement as a default task’s weighting of the profile’s components. In the split over The Vegetarian, some divergences are invisible to readers without Korean (“Most English readers will simply glide over them unaware”), so their weight is zero by blindness, not by a task. The objection undermines a test of CLAIM-240 across such populations unless it fits weights only where both can perceive. It narrows the claim; it does not deny it. | Active |
| CLAIM-461 | Compensation is not only a re-performance notion: for a reader whose code gives the literal word a meaning the source lacked, the compensating rendering can be the more faithful report, so assigning compensation to re-performance alone needs narrowing | The narrowing CLAIM-212’s note of 2026-10-10 proposed and did not make, filed as an objection so it can carry a status. Translators who rendered “Son of God” with non-biological terms for Muslim readers defend the compensating rendering as the better report: the literal word “add[s] a procreative meaning that was absent from the original”. If that holds, compensation (CLAIM-091) serves report fidelity as well as re-performance. Proposed because the renderings are known only through a critic and the reader evidence is the translators’ own. | Proposed |
| CLAIM-468 v3 | Valence intensity need not be the strength of the prevailing ordering’s potential: in Hesp et al.’s model negative valence is a weakening of the preference ordering’s precision with a fall back to habit, and positive affect is reported to lower a goal’s priority, so for either sign stronger valence can go with a weaker hold of the ordering | The difficulty CLAIM-440’s third clause records against itself (“The clause is open for both signs”), filed as an objection. The clause makes intensity how far the negotiation’s decisions follow the prevailing ordering. Hesp et al. (NOTE-1111, read) model negative valence as a fall in the precision of the preference ordering, with a fall back to habit, and Carver (2003, LIT-1445, unread) is reported to hold that positive affect lowers a goal’s priority. CLAIM-438 objects that the ordering tracks wanting, not liking; this objects to what intensity is. The edge: Hesp’s fear case, confident flight, fits the clause. | Proposed |
| CLAIM-481 | Coupling does not by itself separate time scales: two identical undamped oscillators coupled with stiffness k have normal modes at ω₀ and √(ω₀² + 2k), nearly equal when k ≪ ω₀², and neither decays; separation comes where the coupling dissipates relative differences faster than collective ones, as in ẋ = −ax − k(x − y), ẏ = −ay − k(y − x), whose collective and relative modes relax at rates a and a + 2k, and in a Laplacian flow ẋ = −κLx, where the spectrum of L sets the rates | The answer A20, A21 and A23 gave to the owner’s conjecture at U20 that coupling necessarily induces scale separation (CLAIM-473). Two small computations: conservative coupling splits modes without separating or damping them, and dissipative coupling damps the relative mode at a rate that grows with the coupling. For a Laplacian flow the kernel is exact compatibility and the nonzero spectrum is the hierarchy of relaxation rates, so a spectral gap is where the separation comes from. Conceded, so Active. It gives a mechanism for CLAIM-478’s second half and is not a test of individuality. | Active |
| CLAIM-482 v2 | An organization’s aggregate revealed preference exists exactly when its record of winning moves never cycles, and where it exists it is fitted to one regime of authority, so an intervention on authority that changes the winning moves changes the fitted preference and the preference does not predict the change | The status review of 2026-10-27 (ADR-057) found the record’s later formal work bearing on an objection it never cited. CLAIM-089’s aggregate version says Northstar “simply prefers speed”: the governing relations are the organization’s revealed preference. CLAIM-434, Active and demonstrated, says when such a preference exists: an ordering that every winning move climbs exists exactly when the wins never cycle. So the objection is not refuted in general. It is true of an organization whose record is acyclic, and unavailable for one whose record cycles. Where it holds, the ordering is read off the wins under one regime of authority. CASE-003 changes authority and holds attitudes and evidence fixed, and the wins change. Then the fitted ordering changes too, and nothing in the first ordering predicted the second. The revealed preference redescribes the outcome of the governing relations, which is the reading CLAIM-027’s defeater excludes (“whose content predicts the outcome”). | Active |
| CLAIM-492 v2 | If the self is defined to include whatever an agent cares about, the claim that all motivation serves the self is true by definition and rules nothing out; the thesis has content only with a criterion of the self’s extent that can come apart from what the agent cares about | The classic objection to psychological egoism (Butler; Sober and Wilson, LIT-1683), turned on the owner’s thesis. “Every motive is my motive” is trivially true and does not make every motive self-interested. “My self includes whatever I care about” makes the thesis unfalsifiable. It is answered if the thesis names an independent criterion of the self’s extent, such as overlap, fusion or salient self-categorization, and predicts that motivation follows that criterion and not caring alone. | Active |
| CLAIM-494 v2 | Empathic concern for another produces motivation whose ultimate goal is the other’s welfare, and it does so when perceived self–other merging is controlled, so not all helping is concern for an extended self | The empathy–altruism hypothesis as Batson et al. (1997) defend it against Cialdini et al. (1997): in experiments (two, by the SEP’s account), empathy predicted helping with self–other merging controlled, and merging did not mediate it. Sober and Wilson (1998; LIT-1683) argue on independent grounds that psychological egoism is not established. Filed as the empirical objection to the owner’s thesis; both papers are unread, so the record holds the disagreement, not its resolution. | Proposed |
| CLAIM-505 | A goal’s lowered priority under positive affect need not be a weaker hold of the prevailing ordering: on Carver’s satisficing account the effort released goes to the goal furthest behind, which the prevailing ordering of the system’s states may itself direct | Filed at the twenty-first status review, because the point turned up in three readings of 2026-10-30 (NOTE-1442, NOTE-1438, NOTE-1450). CLAIM-468 reads a goal’s lowered priority under positive affect as a weaker hold of the prevailing ordering. Carver’s own account, as his group restates it, reads the same shift as a reordering of priorities among goals, satisficing across them, which the prevailing ordering may itself direct. | Proposed |
| CLAIM-512 v2 | The diachronic thesis faces two objections: a dilemma, that read narrowly it fails for judgments not about any organization’s continuation, such as that suffering is bad, and read broadly every judgment is met vacuously as a preference between continuations unless the organizational description adds causal specificity, transfer or counterfactual discrimination; and a distinction, that evaluating a history is not prescribing its continuation, so a judgment can be determinate with no future transformation fixed | The objection to CLAIM-518, put by the assistant itself in A26 §6 and A27 §10 of the exchange “Analyze nucleation concepts” (2026-10-31): “If we call every difference between possible worlds a change in some organization, then all normative judgments can be represented in these terms without learning much”, and “A judgment that suffering is bad, for instance, is not obviously about organizational persistence”. Unanswered. A27’s three tests are the form an answer would take. | Proposed |
| CLAIM-513 | Across a family of goals, a world model is not an optional explanation of competent behaviour: in a fully observed, finite, communicating, stationary environment, a goal-conditioned policy within a factor 1 − δ of optimal on every composite goal of depth n fixes every transition probability to within √(2P(1 − P)/((n − 1)(1 − δ))), so observers reading the same policy must ascribe the same world model; what it leaves open is whether a predictive component is realized inside the agent, since a lookup table over histories and goals has the same policy | From the reading of Richens, Abel, Bellot and Everitt (LIT-1891, NOTE-1469), registered as the objection to watch against the owner’s U20 thesis CLAIM-519. It defeats that thesis’s functional reading only. The thesis’s defeated_if is not met: the theorem holds “regardless of … architecture”, makes no matched-resource comparison and no intervention on internals, and its authors say the map runs “from the agent’s policy rather than its activations”. Within its scope, though, calling a world model an explanatory attribution is too weak. The bound was re-derived and checked against an adversarial agent. | Active |