{"id":"ae3d3fdc-55f0-49f7-baff-4010fbf9792a","arxiv_id":"2602.11300","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Violations of measurement independence imply in-principle signalling when the ensemble reproduces quantum correlations, allowing Bell's theorem to be extended without assuming measurement independence.","lead":"This paper shows that violations of measurement independence—the assumption that hidden variables are independent of measurement choices—can be converted into a testable form of signalling, and uses this to prove a version of Bell's theorem without that assumption. A smart generalist should read it because it reclassifies the so-called conspiracy loophole in Bell tests as a genuine form of nonlocality, with proposed experimental signatures.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 2's proof silently promotes a single-setting sub-distribution to a global sub-ensemble; without operational preparability, the no-signalling Bell theorem overreaches.","rationale":"Read in good faith, the paper's main goal is to show that certain MI violations can be associated with signalling in principle, and the main theorem is explicitly conditional. The mathematical core (Theorems 0/0', 1/1', Lemmas 1–2) is largely coherent; the pointwise equiprobability step in Appendix A is sketched but repairable via a Markov/subsequence argument. The genuinely load-bearing weak point is Theorem 2's transition from a sub-distribution at one setting to a global, operationally preparable sub-ensemble. This is exactly the reader's weakest assumption. The paper itself flags the physical limitation in Section 7 (Bohr, Everett), which supports the conditional verdict. The abstract's phrasing 'by imposing no-signalling one can prove a version of Bell's theorem that does not require MI' is stronger than the theorem's actual assumptions, but the body is more careful and the authors are transparent about the main caveat. Therefore the reader's CONDITIONAL verdict remains appropriate; no adjustment is needed.","tokens_in":25697,"tokens_out":12995,"duration_ms":119227,"concrete_test":"Attempt to fill the gap rigorously: for the context IJ singled out by Lemma 1, define a global sub-ensemble K1 by σ_{I'J'}^{K1}=σ_0^{IJ} if (I',J')=(I,J) and σ_{I'J'}^{K1}=σ_{I'J'}^K otherwise, and check whether there exists a K2 such that Eq. (17) holds for all contexts. If yes, the proof is formally repairable; then test the operational half by taking the toy model (24) (λ=(A,B), σ_IJ^K = quantum probabilities) and asking whether any such K1 is preparable by Charlie's operations alone. If no preparable K1 exists, Theorem 2's conclusion is not guaranteed for this MI-violating model, confirming that the sub-ensemble condition is load-bearing and not a mere technicality.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 5's Lemma 1 guarantees only a sub-distribution of the single distribution σ_K^{IJ} (for some fixed IJ) with weight ≥1/2 and |<A>_0^{IJ}|>ε. Theorem 2's proof then states: 'By Lemma 2, any sub-ensemble with weight 1/2≤α≤1... By Lemma 1, one of these sub-ensembles also satisfies (19)...' This is a quantifier shift: a sub-ensemble is defined (Eq. 17) by distributions for all I,J, whereas Lemma 1's object lives only at one context. The gap can be repaired mathematically by defining K1 to equal σ_0^{IJ} at that context and σ_K^{I'J'} elsewhere, but (a) the paper does not supply this construction, and (b) the real issue is operational: Eq. (17) requires a single mixing procedure that works uniformly over all contexts. The repaired K1 differs from K at exactly one future setting pair, so Charlie must be able to prepare an ensemble that is context-sensitive in an extremely fine-tuned way. The authors' own Section 7 concedes that Bohr-style contexts and Everett divergent worlds violate MI yet do not allow preparation of non-quantum ensembles; for such theories Theorem 2 is vacuous. Consequently the Abstract's claim that 'imposing no-signalling one can prove a version of Bell's theorem that does not require MI' is stronger than what is proven: the theorem needs OI, finite chain-correlation bounds, and the additional, non-trivial, physically contingent assumption of preparable sub-ensembles.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper examines the assumption of Measurement Independence (MI) in Bell's theorem and proposes that certain violations of MI can be understood as 'signalling in principle', making them operationally testable. The authors prove equiprobability theorems (Theorems 0 and 0'), then show that operational violations of equiprobability imply signalling (Theorems 1 and 1'). They introduce Lemmas 1 and 2 connecting CHSH violations to sub-distribution deviations and chain-correlation bounds, leading to Theorem 2/2' as an apparent extension of Bell's theorem without MI, conditional on Outcome Independence, finite chain-correlation bounds, and Charlie's ability to prepare 'appropriate sub-ensembles'. Section 6 applies the framework to the Schulman/Wharton model; Section 7 discusses limitations, including Bohr-style contextualism and Everett 'divergent worlds', where the required ensembles cannot be prepared.","tokens_in":25993,"tokens_out":15595,"duration_ms":144472,"significance":"The paper's novelty is to convert a standard loophole—Measurement Independence—into a potential operational resource with quantitative thresholds. The direct proofs of Theorems 0/0' and 1/1' in Appendices A and B are clear, self-contained, and do not involve parameter fitting. The chain-correlation framework gives robust, experimentally checkable conditions. The authors are also unusually candid about the scope and limitations of their claims, particularly in Section 7. If the main theorem can be made fully rigorous, the paper would be a meaningful contribution to the foundations of quantum mechanics and to 'experimental metaphysics'.","major_comments":[{"comment":"The proof of Theorem 2 contains a quantifier shift. Lemma 1 guarantees only a sub-distribution σ_0^{IJ} of the single context distribution σ_K^{IJ} with weight ≥1/2 and |⟨A⟩_0^{IJ}| > ε. But a sub-ensemble, as defined in Eq. (17), is a collection of distributions for all I,J. The proof states: 'By Lemma 2, any sub-ensemble with weight 1/2≤α≤1... By Lemma 1, one of these sub-ensembles also satisfies (19)...' No construction of such a global K1 is given. A repair would define K1 by setting its IJ component to the sub-distribution from Lemma 1 and all other components equal to K, but this requires Charlie to prepare an ensemble whose distributions differ only at one future setting pair—an extra, non-trivial operational assumption that is not stated in Theorem 2. The authors acknowledge the general difficulty in footnote 28 and Section 7, but the theorem as stated and proved is incomplete.","section":"§5, Theorem 2 proof"},{"comment":"The Abstract claims: 'by imposing no-signalling one can prove a version of Bell's theorem that does not require the assumption of Measurement Independence.' Section 5's opening paragraph similarly says: 'imposing no-signalling allows one to derive the Bell inequalities.' However, Theorem 2 assumes Outcome Independence (Eq. (3)), finite chain-correlation bounds (21)–(22), and the additional premise 'If Charlie can prepare appropriate sub-ensembles of K'. Outcome Independence is not a no-signalling condition, and the theorem does not follow from no-signalling alone. The advertised conclusion is therefore stronger than what is proven. The Abstract and Section 5 should be rephrased to say that, under OI, finite chain-correlation bounds, and preparability of sub-ensembles, violations of MI become signalling in principle.","section":"Abstract and §5 (first paragraph)"},{"comment":"The proof of Lemma 1 contains an unjustified equivalence claim. It reads: 'Now assume that |⟨A⟩_1^{IJ}|, |⟨A⟩_2^{IJ}| ≤ ε. This is equivalent to assuming the bound in (59) for all possible sub-distributions, since for any other sub-distribution there will be partial cancellations.' This equivalence is false: a sub-distribution can select a subset of σ_1 with a larger mean than the mean over all of σ_1. Moreover, the contrapositive of Lemma 1 only bounds sub-distributions with weight ≥1/2; it does not bound the lighter of σ_1 and σ_2. Yet the chain of inequalities (60) uses bounds on both parts regardless of their weights. Since Lemma 1 is used in the proof of Theorem 2, a correct proof or a modified lemma is needed.","section":"Appendix C, Lemma 1 proof"}],"minor_comments":[{"comment":"The text says 'returns the expectation values ... (as defined in (4))', but Eq. (4) defines Parameter Independence, not expectation values. The reference should be to Eq. (2).","section":"§2, near Eq. (2)"},{"comment":"The sentence 'Setting ε=1/2 and contraposing yields the Bell inequalities' could be expanded to show explicitly that this gives the CHSH bound |CHSH| ≤ 2.","section":"§5, after Lemma 1"},{"comment":"The coefficients α, β, γ, δ in (32) reuse letters that earlier denote mixture weights (α) and error parameters (γ). Consider using different symbols to avoid confusion.","section":"§6, Eq. (32)"},{"comment":"The claim that 'this single chain contains all the measurements we need' is terse. A short explanation of how the 12 chain directions include the four Bell-correlation directions would improve readability.","section":"§5, Theorem 2'"}],"recommendation":"major_revision","confidential_remarks":"The core idea is promising and the direct proofs in Appendices A and B are solid. The main obstacles are (i) the incomplete proof of Lemma 1 in Appendix C, (ii) the quantifier shift in the proof of Theorem 2, and (iii) the abstract overstating what is proven. If the authors can repair the Lemma 1 proof, make the global sub-ensemble construction explicit, and carefully qualify the Abstract's claims, the paper could be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know up front. First, the central trick — applying the equiprobability theorem to a trivial hidden-variable model where the hidden variable is Charlie's preparation context — is legitimate and yields a clean conditional result: if a Bell-violating ensemble also has the right chain correlations and Charlie can prepare certain sub-ensembles, then at least one wing can signal. Second, the proof as written has a quantifier slip in Theorem 2, and the abstract promises more than the theorem delivers.\n\nWhat is actually new: the reclassification of MI violations as potentially testable via signalling, the quantitative thresholds, and the extension of Bell's theorem to models that violate MI under an explicitly stated preparability assumption. The appendices are clean; the direct proofs of Theorems 1 and 1' are particularly nice. The Schulman model discussion is a good example, and the Section 7 discussion of Bohr and Everett is honest — it concedes that not every MI-violating theory allows the required sub-ensembles.\n\nWhere it gets soft. The abstract says 'imposing no-signalling one can prove a version of Bell's theorem that does not require MI.' That is too quick. The theorem needs OI, chain-correlation bounds, and the sub-ensemble preparability assumption. No-signalling alone does nothing. The paper knows this — it even notes OI cannot be dropped — but the abstract and the opening of Section 5 overstate. That should be fixed.\n\nThe bigger issue is Theorem 2's proof. Lemma 1 guarantees a sub-distribution of a single σ_IJ, not a sub-ensemble across all contexts. The proof silently promotes one to the other. The gap is repairable: define the sub-ensemble to match the sub-distribution at that one context and equal σ_K elsewhere. That is a mathematically valid sub-ensemble, and the proof then goes through. But the operational cost is real: Charlie must be able to prepare a context-sensitive ensemble that is fine-tuned to the very setting pair that happens to violate. The paper's own Section 5 warns that such mixtures are non-trivial to assume, and Section 7 shows that for Bohr-style and Everett-style theories the theorem is vacuous. So the result is conditional in a way that should be flagged more prominently.\n\nNone of this is fatal. The main claim holds up once you state the full list of assumptions. The paper is a genuine conceptual advance for the foundations community, and it deserves a serious referee. I'd send it to review with a request to fix the abstract and fill the gap in Theorem 2.","headline":"A serious, mostly sound paper that reclassifies measurement-independence violations as operationally testable nonlocality, but the advertised no-signalling Bell theorem overstates what is proven and Theorem 2 needs a small but real repair.","tokens_in":26555,"tokens_out":4108,"would_cite":true,"duration_ms":40537,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P13","81P05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that violations of measurement independence can be made testable by turning them into signalling in principle, and proves a version of the nonlocality theorem without that assumption.","keywords":["measurement independence","hidden variables","nonlocality","signalling in principle","equiprobability theorem","retrocausality","superdeterminism","quantum foundations"],"falsifier":"A concrete check is to construct an explicit hidden-variable model satisfying Outcome Independence, high chain correlations, and a four-term inequality violation larger than 4ε, and then test whether any protocol can prepare the guaranteed sub-ensemble with weight at least 1/2 uniformly across all settings; if no such protocol exists, the signalling conclusion does not follow. The paper's own Section 7 examples—contextual and branching-world descriptions—would also falsify the general claim if they reproduce the Bell violation yet provably cannot prepare non-quantum ensembles.","tokens_in":25509,"feed_emoji":"⚛️","tokens_out":5551,"duration_ms":59227,"temperature":0.7,"pith_summary":"The paper argues that violating the assumption that hidden variables are independent of measurement settings—often dismissed as a 'conspiracy loophole'—need not be a dead end for testing quantum nonlocality. It shows that certain violations of this assumption can be re-expressed as signalling in principle: if Charlie can prepare appropriate sub-ensembles, Alice or Bob can send a message. The main theorem derives a nonlocality inequality from no-signalling plus Outcome Independence, without assuming Measurement Independence. A sympathetic reader should care because this converts a philosophical loophole into an operationally meaningful distinction between forms of nonlocality.","feed_headline":"No-signalling still yields a nonlocality theorem","feed_subtitle":"Measurement independence can be violated, but only if someone can signal in principle.","key_machinery":"The load-bearing object is the equiprobability theorem: for a hidden-variable model reproducing near-perfect (anti)correlations along a chain of alternating spin directions, each hidden variable's local outcomes must be equiprobable unless parameter or measurement independence fails. The paper converts this into a quantitative statement: a chain of 2n directions at angle π/2n bounds the marginal deviation by 2nδ. Combining this with the fact that a Bell-inequality violation larger than 4ε forces a sub-distribution with marginal deviation larger than ε yields the extended theorem. The second essential ingredient is the operational notion of a sub-ensemble: a sub-distribution that Charlie can","core_discovery":"The central claim is Theorem 2: in a hidden-variable model that satisfies Outcome Independence and reproduces high (anti)correlations along four chains of nearby spin directions, any violation of the four-term correlation inequality beyond 4ε guarantees that there exists a sub-distribution deviating from equiprobability. If that sub-distribution can be operationally prepared by Charlie as a sub-ensemble, then at least one of Alice or Bob can signal. The proof pieces together the equiprobability theorem—which bounds hidden-level marginals by chain correlations—with a lemma showing that a Bell-inequality violation forces a large marginal deviation in some sub-distribution. Thus, under the stat","pith_inferences":["Going beyond the paper: if the sub-ensemble preparation condition is dropped, Theorem 2 collapses; the decisive physical question is whether any genuine, non-formal violation of Measurement Independence must allow reliable non-quantum sources.","Editorial extension: the paper's contrast between formal and causal violations suggests a testable hierarchy—an MI-violating theory that can signal in principle is one in which the settings influence the hidden distribution through an actual mechanism, not merely through a re-description of the same quantum state.","The quantitative threshold ε ≥ 4nγ could be turned into a concrete experimental programme: measure chain correlations and the four-term inequality on the same ensemble to determine the maximum chain length n for which no-signalling still forces signalling.","The authors' own limitation examples imply that any attempt to prepare the sub-ensemble must be checked for uniformity across all settings; a model satisfying Theorem 2's premises but lacking such a uniform preparation protocol would be a counterexample to the theorem's applicability."],"forward_implications":["If a hidden-variable model satisfies Outcome Independence and reproduces quantum-like chain correlations, then any Bell-inequality violation above 4ε makes the model signalling-capable whenever Charlie can prepare the relevant sub-ensemble.","No-signalling alone, without Measurement Independence, suffices to derive a nonlocality theorem under the stated chain-correlation and sub-ensemble assumptions.","A concrete retrocausal model discussed in the paper reproduces the quantum correlations but underdetermines the hidden distribution; certain allowed non-quantum distributions violate equiprobability and permit signalling in principle.","In principle, action-at-a-distance (Parameter Independence failure) can be operationally distinguished from measurement-dependence signalling: the latter requires no preferred time ordering between Alice's and Bob's measurements."],"fun_headline_variants":["No-signalling assumption can replace measurement independence in Bell's theorem","Measurement independence violations imply signalling in principle","Nonlocality proof drops independence, only needs no-signalling","Bell's theorem extended by substituting no-signalling for independence"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that Charlie can in principle reliably prepare the non-quantum sub-ensembles that Theorem 2 needs; the proof shows such sub-distributions exist mathematically, but existence in the formalism does not guarantee operational preparability, and the paper itself (Section 7) notes that some MI-violating descriptions—contextual and divergent-worlds pictures—do not allow preparation of any ensembles beyond the quantum ones.","fun_headline_variants_meta":{"raw":{"variants":["No-signalling assumption can replace measurement independence in Bell's theorem","Measurement independence violations imply signalling in principle","Nonlocality proof drops independence, only needs no-signalling","Bell's theorem extended by substituting no-signalling for independence"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000678,"raw_usage":{"total_tokens":2862,"prompt_tokens":627,"completion_tokens":2235,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":371,"completion_tokens_details":{"reasoning_tokens":2182}},"tokens_in":371,"tokens_out":2235,"duration_ms":18622,"temperature":1.0,"reasoning_tokens":2182,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T00:11:10.119253+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check is to construct an explicit hidden-variable model satisfying Outcome Independence, high chain correlations, and a four-term inequality violation larger than 4ε, and then test whether any protocol can prepare the guaranteed sub-ensemble with weight at least 1/2 uniformly across all settings; if no such protocol exists, the signalling conclusion does not follow. The paper's own Section 7 examples—contextual and branching-world descriptions—would also falsify the general claim if they reproduce the Bell violation yet provably cannot prepare non-quantum ensembles.","supporting_citations":[],"review_version":1}