{"id":"68fd41e2-3576-484f-b3d5-03adc97dfa21","arxiv_id":"2510.21015","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The impossibility of higher-order quantum interference is generalized into a conjectured principle bounding event-correlation separability by ⌈m/2⌉ mediators.","lead":"This paper argues that the known limits of quantum interference—second-order interference is possible, third-order is not—point to a universal principle: any joint influence of m separated events on a future event can be mediated by about half as many intermediate events. It is a conceptual-physics proposal with a restricted mathematical proof, worth reading for how it connects interferometry to general principles of causality and correlation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Definition 3's no-common-cause condition is formally misstated: as written it forbids future mediators and excludes the motivating double-slit experiment, so the GIP_m class—and the Principle's antecedent—are not well-defined.","rationale":"The reader's weakest assumption correctly identified the extrapolation from unitary, number-preserving semi-general interference experiments to all GIP_m, including QFT, as a major gap. However, the formal flaw in Definition 3's condition 3 is more fundamental: it is not just that the bridge to all GIP_m is unproven, but that the class GIP_m itself is not coherently defined as written. This is a sharp, checkable issue rather than a speculative doubt. It does not require rejecting the paper's core theorems, which are proven for the restricted class of semi-general interference experiments and stand as mathematical results. Rather, it demands a precise revision of Definition 3 before the Principle can be evaluated. The reader's verdict of CONDITIONAL is therefore appropriate, unchanged by my analysis, but the conditions should explicitly include amending condition 3 to restrict z to genuine common causes (e.g., z in the common past of X and y) and stating which refinements count as 'adequate'. My agreement is partial because the reader mentioned Definition 3 as needing rigor but did not identify the specific quantifier error that makes the condition formally incompatible with mediators and with the paper's own cardinal example.","tokens_in":37562,"tokens_out":15491,"duration_ms":152246,"concrete_test":"Take the standard double-slit experiment as a purported GIP_2 and construct the refinement P' that adds a dichotomic event z='which slit the particle passes through' at a location between the slits and the screen. Verify two facts: (i) P'_y(y|X,z) is the single-slit distribution for the indicated slit, so it differs from P'_y(y|X), the two-slit interference distribution; (ii) P'_X(X|z) is not equal to P'_X(X), since z='slit 1' can occur only when slit 1 is open. If both hold, Definition 3, condition 3, is violated, and the double-slit experiment is not a GIP_m under the written definition. If the author responds that this refinement is not 'adequate', then an independent criterion of adequacy for refinements must be supplied; the current definition does not provide one.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing weakness is in Definition 3, condition 3, which quantifies over all z∈R^4 and asserts: if P'_y(ω'_y|ω'_X,ω'_z) ≠ P'_y(ω'_y|ω'_X), then P'_X(ω'_X|ω'_z) = P'_X(ω'_X). Taken literally, this forbids any future mediator of the X→y correlation: a mediator z in C^{(+)}_X is generally correlated with X, so P'_X(X|z) ≠ P'_X(X), and it changes the prediction for y given X, so the antecedent holds. Thus the condition rejects the very type of event the paper later introduces as a 'local completion'. This is not merely an extrapolation gap: the double-slit experiment—the paper's own paradigm GIP_2—violates condition 3 if P' is the refinement that records which slit the particle traverses. Given X alone, P'_y(y|X) is the two-slit interference distribution; given X and z='slit 1', it is the single-slit distribution. Also, P'_X(X|z='slit 1') is concentrated on configurations with slit 1 open. Hence condition 3 fails for a physically meaningful refinement. The intended restriction to common causes in the common past of X and y is not present in the formal statement. Until condition 3 is amended—e.g., by quantifying only over z in the common past of X and y, or at least outside C^{(+)}_X—the class of GIP_m is not well-defined, and the Principle's claim 'every maximal GIP_m...' lacks a coherent domain.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper generalizes the Sorkin hierarchy of interferometric experiments to ‘general interference phenomena’ (GIP_m) defined through probabilistic event models, and proposes a universal Principle: every maximal GIP_m has a physically possible closed bi-local completion, while not every maximal GIP_m has a closed n-local completion when ⌈m/n⌉<2. The supporting analytic results (Lemma 1 and Theorems 1–2) concern non-relativistic semi-general interference experiments with unitary number-preserving transformations; they are rigorous for even m and for odd m under a restricted support condition, and Appendix 5 gives preliminary evidence for a broader odd-m case. The paper explicitly labels the universal principle a conjecture and notes that the QFT case, non-number-preserving CP-maps, and the general odd-m case remain open. The philosophical framing is careful, and the paper’s main theorem is an exact construction, not merely an existence claim.","tokens_in":37956,"tokens_out":8405,"duration_ms":89488,"significance":"If the Principle were established, it would be a genuinely novel causal-statistical principle: an exact quantitative bound on the non-separability of joint influences of m spacelike separated events on a common future event. The restricted theorems are a real contribution: Lemma 1 extracts the structural form forced by maximal interference, and Theorem 1 constructs the mediating measurements explicitly, with a GHJW-based extension to mixed states. The paper is also commendably explicit about what is proven and what is conjectured. However, the formal definition of the GIP_m class has a defect that blocks the entire framework, and the universal step goes far beyond the analytic evidence. With a repaired Definition 3 and a clearer separation between theorem and conjecture, the paper could make a valuable contribution to quantum foundations.","major_comments":[{"comment":"Condition 3 quantifies over all z∈R^4. As written, a future mediator z—e.g. the ‘which slit’ event in the double-slit experiment, or the mediating events y_i in Definition 5—is correlated with X and changes P'_y(y|X,z) relative to P'_y(y|X), so condition 3 is violated. Thus the paper’s own paradigm GIP_2 (the double-slit experiment) is excluded, and every n-local completion T* in Definition 5, whose mediating events lie in C^(+)_X ∩ C^(-)_y, also fails condition 3. The intended no-common-cause condition must be restricted to the common causal past of X and y, e.g. z ∈ C^(-)_X ∩ C^(-)_y (or at least z ∉ C^(+)_X). Until this is fixed, GIP_m is not well-defined and the Principle has no coherent domain.","section":"Definition 3, condition 3"},{"comment":"The analytic evidence covers only unitary number-preserving transformations, not general CP-maps; Theorem 2 assumes input states with support on exactly two spatial configurations (with only Appendix 5 heuristic evidence for three), and the text concedes in §IV.III that the QFT case ‘would require a completely new investigation’. Therefore the universal Principle of §V is a conjecture, not a proved statement. This is acceptable if clearly labeled, but the abstract and §IV.III should state precisely what is proven and what is extrapolated; as written, the phrase ‘analytic evidence’ overstates the support for the claim that any maximal GIP_m has the stated completion property.","section":"§IV.II–IV.III and §V"},{"comment":"The definition of ‘closed’ refers to probabilities P_S(Q) on all spacelike hypersurfaces S, but the PE-model framework of Definition 1 assigns distributions only to finite subsets of R^4. Moreover, the ‘non-dynamical quantities’ Q are never specified; the text itself acknowledges that this ‘would definitely require further elaboration’. Since ‘closed’ appears in the statement of the Principle, a precise definition (or a well-defined approximating family of finite hypersurfaces, and a specification of Q) is needed before the Principle can be evaluated.","section":"Definition 4 (closed GIP_m)"},{"comment":"Condition 5 equates P'_{y'}(ω_{y'}|ω_X) with P'_y(ω_y|ω_X) for y' ∈ C^(-)_y, but these are distributions over different event spaces at different locations. The condition should be reformulated, e.g. in terms of a coarse-graining map or a common refinement, otherwise the non-triviality requirement is not formally meaningful.","section":"Definition 5, condition 5"}],"minor_comments":[{"comment":"The notation P_y(⊕_i ω_{x_i}|ω_X) is overloaded: ω_X is the vector of configurations, and the sum over ω_X is correct, but the conditioning should be written more explicitly to avoid confusion with a single event at X.","section":"Definition 3, condition 2"},{"comment":"The arrow notation used for the implication is nonstandard and should be replaced by ordinary logical implication with parentheses.","section":"Definition 3, condition 3"},{"comment":"The claim that the correlation ‘cannot be mediated by intermediate events arising from number-preserving operations’ is argued only for local projective measurements. The statement should be explicitly qualified to that class, since general number-preserving CP-maps are not analyzed here.","section":"§IV.I, Example 2"},{"comment":"The lemma states that U^(a)|ψ⟩ = G(…), with G an arbitrary unitary, but the proof constructs a specific G of the form (28). The lemma should state this explicit form, or say ‘for some G of the form (28)’.","section":"Lemma 1 statement"},{"comment":"The expression for Π'_b is dense and difficult to parse. A short explanation of the role of each factor (G, E_B, H_B, M^{(B,i)}_{b_i}, Π^{(B)}_b) would greatly improve readability.","section":"Equation (13), Theorem 1"},{"comment":"The paper uses many newly introduced terms (GIP_m, closed, n-local completion, Q). A glossary or a summary table of definitions would help the reader keep track of the dependencies.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern about Definition 3 lands: the formal condition is genuinely inconsistent with the paper’s own examples and with Definition 5. The issue is fixable by restricting the quantifier to the common causal past, so I do not recommend rejection. The broader extrapolation from the restricted theorems to the universal Principle is a scope issue that can be addressed by more careful framing. The paper would be suitable for a quantum-foundations journal after these corrections."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper contains two real existence theorems, and it is unusually honest about the gap between what is proven and what is conjectured. But the definition of the class the principle is about has a formal bug that needs fixing before the main claim can be evaluated, and the bridge from non-relativistic number-preserving unitaries to 'all GIP_m' is very wide.\n\nThe useful content is in Section IV and the appendices. Lemma 1, Theorem 1, and Theorem 2 show that for a maximal semi-general interference experiment with n particles and 2n (or 2n−1 with restrictions) settings, one can construct another experiment with the same input state and unitaries where the correlation is mediated by n number-preserving measurements. The proofs use Helstrom's bound and concrete unitary constructions, and they appear to be rigorous within their stated scope. That is a new result, and it is worth having.\n\nNow the soft spots.\n\nFirst, Definition 3, condition 3. As written it quantifies over all z in R^4 and says that if z changes the distribution of y given X, then z must not be correlated with X. That forbids exactly the intermediate mediating events the paper later introduces: a future mediator z is generally correlated with X and it changes the prediction for y. The stress-test is right about the quantifier. The intended restriction was probably to z in the common past of X and y, but the formal statement does not say that. The double-slit example in the stress-test is less convincing as stated, since a which-slit refinement does not adequately describe the same phenomenon without changing the marginal on y. But the underlying formal problem is real: the class of GIP_m is not well-defined until condition 3 is amended.\n\nSecond, the leap from the proven class to the universal principle is exactly what the paper acknowledges: only unitary, number-preserving, non-relativistic dynamics; odd m restricted to two-support states; QFT explicitly left for later. That is a lot of extrapolation. The paper is commendably explicit about this, but it means the Principle should be read as a conjecture, not a result.\n\nThird, the 'closed' definition via non-dynamical quantities Q is vague; the paper's own footnotes admit it needs elaboration.\n\nWho should read this? People working on the reconstruction of quantum theory and on concepts of non-separability. It is a serious conceptual proposal with a proven core. It deserves a referee, but a referee should demand a fix to Definition 3 and a clear separation of the proven part from the conjectural part.\n\nMy recommendation: send it to review; do not desk reject. With a repaired condition 3 and a more careful statement of what exactly is conjectured, it could be a useful paper.","headline":"Theorems 1 and 2 are genuine, but the universal principle is built on a misstated no-common-cause condition and a very wide extrapolation.","tokens_in":38426,"tokens_out":7384,"would_cite":true,"duration_ms":69534,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Every m-way influence can be mediated by m/2 events, rounded up","keywords":["quantum interference","higher-order interference","non-separability","causal correlations","probabilistic event models","semi-general interference experiments","bi-local completion","spacetime events"],"falsifier":"Exhibit a physically possible maximal GIP_3 (I_3 = 1/2) whose correlation provably cannot be preserved and mediated by two spacelike intermediate events in any closed completion—for example, by showing that any such completion would require changing total energy or the non-dynamical quantities Q.","tokens_in":37433,"feed_emoji":"⚛️","tokens_out":3265,"duration_ms":30011,"temperature":0.7,"pith_summary":"The paper argues that the structure of quantum interference—second-order possible, third-order impossible—implies a universal principle about spatiotemporally localized events. The principle: any joint influence of m mutually spacelike events on a future event can be reproduced through ⌈m/2⌉ mediating events, and in some cases cannot be reproduced through more than that many. If true, non-separable correlations in nature have a precisely quantifiable limit, independent of the underlying objects. The author supports this by proving that every maximal-interference semi-general interference experiment has a \"closed bi-local completion,\" then extends the claim from particles to arbitrary phenomena describable by probabilistic event models.","feed_headline":"Every m-way influence needs only m/2 mediators","feed_subtitle":"Quantum interference's no-third-order rule becomes an exact bound on how non-separable events can be.","key_machinery":"The central machinery is the class of semi-general interference experiments, described by triples (ρ, U(a), Π) with number-preserving local unitary transformations. Lemma 1 shows that maximal interference forces the transformed state into an equal-weight superposition over bipartitions, with the inputs encoded in π-phases. Theorems 1 and 2 then construct, for any such maximal experiment, an alternative experiment sharing the same input state and unitaries but whose final measurement factors through n = ⌈m/2⌉ mutually spacelike, number-preserving intermediate measurements. This concrete construction is abstracted into the definitions of closed GIP_m and n-local completions, yielding the paper","core_discovery":"On the paper's own terms: every maximal GIP_m—a phenomenon with m spacelike binary events whose joint influence I_m reaches the maximum value 1/2—has a physically possible closed bi-local completion. That means there exists another physically possible phenomenon which agrees with the original on the entire causal past, preserves the matter content (non-dynamical quantities Q), and mediates the correlation between the m events and the future event through ⌈m/2⌉ intermediate mutually spacelike events. Conversely, some maximal GIP_m have no closed n-local completion when ⌈m/n⌉ < 2. The possibility of second-order interference and impossibility of third-order interference are thereby transformed","pith_inferences":["A testable corollary: any future theory (e.g., a quantum field theory) that exhibits maximal higher-order interference would be forced to contain hidden intermediate degrees of freedom at the ⌈m/2⌉ level, otherwise it would violate the proposed principle.","The principle suggests a hierarchy of possible worlds: classical worlds mediate through m events, quantum worlds through ⌈m/2⌉, and no worlds allow more non-separability than this—a 'no third-order interference' law of nature.","The paper's modal formulation, comparing pairs of physically possible phenomena, may open a new class of causal-statistical principles beyond the Bell-Tsirelson structure, though the paper only gestures at this direction."],"forward_implications":["If correct, any maximal joint influence of m spacelike events can always be simulated by a chain through ⌈m/2⌉ intermediate spacelike events, without changing the matter content.","No physical phenomenon can require more than ⌈m/2⌉ mediators to reproduce a maximal correlation; the ceiling is exact, not asymptotic.","The principle is formulated for binary events but depends only on causal relations, so it generalizes trivially to arbitrary spacetimes.","Extensions to non-maximal interference and to events with more than two values are left open, with earlier work indicating the same possibility-impossibility structure persists for prime-valued configurations."],"fun_headline_variants":["m-way influence splits at ⌈m/2⌉ mediators","Quantum interference caps influence at half the mediators","No third-order interference means exact mediator bound","Every m-way link needs ⌈m/2⌉ intermediate events","Separability limit: exactly ⌈m/2⌉ mediators for m events"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"What is proven for unitary, number-preserving, non-relativistic particle experiments is assumed to hold for all phenomena describable by probabilistic event models, including quantum field theory and any hypothetical future physics.","fun_headline_variants_meta":{"raw":{"variants":["m-way influence splits at ⌈m/2⌉ mediators","Quantum interference caps influence at half the mediators","No third-order interference means exact mediator bound","Every m-way link needs ⌈m/2⌉ intermediate events","Separability limit: exactly ⌈m/2⌉ mediators for m events"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000224,"raw_usage":{"total_tokens":1286,"prompt_tokens":720,"completion_tokens":566,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":464,"completion_tokens_details":{"reasoning_tokens":481}},"tokens_in":464,"tokens_out":566,"duration_ms":4499,"temperature":1.0,"reasoning_tokens":481,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T08:19:58.257650+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a physically possible maximal GIP_3 (I_3 = 1/2) whose correlation provably cannot be preserved and mediated by two spacelike intermediate events in any closed completion—for example, by showing that any such completion would require changing total energy or the non-dynamical quantities Q.","supporting_citations":[],"review_version":1}