{"id":"0a2e1a08-c0ea-478e-bdf7-e8c367970de2","arxiv_id":"2412.10963","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For a connected measurement space X, the non-signaling polytope of the cone scenario is the join of m copies of the original polytope, and the suspension scenario is characterized by a pullback of two such joins.","lead":"This math paper proves structural decompositions of the polytopes of simplicial distributions on cone and suspension measurement scenarios. It derives Bell inequalities for cone scenarios and constructs new families of contextual vertices, including explanations of known vertices in the (3,2,2) Bell scenario.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 5.14 is stated generally but proved only for Example 5.13; Proposition 5.15 and the advertised suspension-vertex examples rest on the unproven general form.","rationale":"The reader's verdict is CONDITIONAL with moderate confidence; my stress-test supports keeping that verdict. The central cone decomposition (Theorem 1.1/4.11) is well supported: Proposition 4.10 transfers the disjoint-union decomposition of Section 3 to the cone via the connectedness hypothesis, and the noncontextuality and vertex claims follow cleanly from the commuting Diagram (26). The connectedness assumption is explicitly acknowledged, with a counterexample given, so it is a domain restriction rather than a hidden flaw. The main soft spot I find is the mismatch between Lemma 5.14's statement and its proof. Since Proposition 5.15 depends on the general version, the suspension vertex applications are not yet fully established. My algebraic reconstruction suggests the lemma is true, so I do not recommend rejection, only completion of the proof. This partially agrees with the reader: the reader's rationale already mentions Lemma 5.14, but the reader's stated weakest_assumption is connectedness, which I do not consider the most pressing issue. A minor related point is that zero-weight components in the decomposition of Theorem 4.11 need a basepoint convention for the statement to be literally unique; this is easily fixed and does not affect the geometry.","tokens_in":28378,"tokens_out":26620,"duration_ms":236223,"concrete_test":"Enumerate all complete collections for m = 3 and m = 4 (all pairs of permutations α_1, α_2 of Z_m) and solve the linear system ∑λ_j S^{α_1(j)} = ∑μ_j S^{α_1(j)} and ∑λ_j S^{α_2(j)} = ∑μ_j S^{α_2(j)+1}, with ∑λ = ∑μ = 1. If any non-uniform solution occurs, Lemma 5.14 is false and Proposition 5.15 collapses. Independently, supply the missing general proof: the first equation forces λ = μ, after which the second gives λ_{α_2^{-1}(r)} = λ_{α_2^{-1}(r-1)}; because α_2^{-1} ∘ (x ↦ x−1) ∘ α_2 is an m-cycle, λ is uniform. If the enumeration finds no counterexample and the cycle argument checks out, the proof gap is repairable and the conditional verdict stands.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Section 5.3, Lemma 5.14 is the hinge of Proposition 5.15: it asserts that for any complete collection of average distributions, an equality ∑λ_j q_j = ∑μ_j ψ·q_j forces λ_j = μ_j = 1/m. The proof, however, states 'For simplicity, we will give a proof for the case that the complete collection of average distributions is the one given in Example 5.13' and never treats the general case. Proposition 5.15 applies the lemma to restrictions {p_j|L}, where the ordering of {I, S, ..., S^{m-1}} on σ1 and σ2 can be arbitrary, so the general form is genuinely used. Example 5.16 then relies on Proposition 5.15 to identify a contextual vertex. Thus an advertised suspension result is currently unsupported as written. The likely algebra is favorable: writing q_j|_{σ_k} = S^{α_k(j)}, the σ1 equation gives λ_j = μ_j and the σ2 equation gives λ_{α_2^{-1}(r)} = λ_{α_2^{-1}(r-1)}, forcing λ constant; so this is an omitted proof rather than a known counterexample. But the text does not supply that derivation, and Proposition 5.15 cannot be accepted on the strength of the special case alone.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the geometry of simplicial distributions on cone and suspension scenarios. It proves (Theorem 4.11) that for a connected simplicial set X, sDist(CX, ΔZ_m) is isomorphic to the join of m copies of sDist(X, ΔZ_m), with an explicit decomposition that preserves noncontextuality and extremality. It derives Bell inequalities for cone scenarios (Proposition 4.13), gives a pullback description of sDist(ΣX, ΔZ_m) (Proposition 5.2), and constructs contextual vertices in suspension scenarios from deterministic and average collections (Propositions 5.8 and 5.15). The main structural theorems are proved in detail; however, a proof gap in Lemma 5.14 affects Proposition 5.15 and Example 5.16.","tokens_in":28599,"tokens_out":6016,"duration_ms":53244,"significance":"If the gaps are repaired, the paper provides a clean structural tool for studying contextuality: the cone decomposition in Theorem 4.11 and the suspension pullback in Proposition 5.2 are elegant and likely useful for future work. The paper also contributes a new family of Bell inequalities and a topological explanation of some known contextual vertices. The treatment is self-contained and honest about the connectedness assumption, including a counterexample showing that the decomposition fails in general. The main gap is localized and appears fixable, but until it is repaired, the suspension-vertex examples are not fully supported.","major_comments":[{"comment":"The lemma is stated for an arbitrary complete collection of average distributions, but the proof explicitly covers only the collection of Example 5.13. Proposition 5.15 uses the lemma for the restrictions {p_j|_L}, which are not necessarily of that form (Example 5.16 uses PR boxes on the CHSH scenario). Therefore Proposition 5.15 and the identification of the contextual vertex in Example 5.16 are not established as written. The general case is plausible—one can index the restrictions on σ1 and σ2 by powers of S—but the derivation must be written out.","section":"5.3, Lemma 5.14"},{"comment":"The proof shows that each noncontextual distribution satisfies the displayed inequalities, but it does not verify the saturation and violation conditions required by Definition 2.12. These conditions are needed for the inequalities to be 'the Bell inequalities' of the cone scenario. They can presumably be shown by taking p = κ_j(p^(j)) with p^(j) saturating or violating the original inequality, but the argument is omitted.","section":"4.3, Proposition 4.13"}],"minor_comments":[{"comment":"The section title contains a typo: 'Simiplicial distributions' should be 'Simplicial distributions'.","section":"Section 5 heading"},{"comment":"The phrase 'Let L be the line that generated by σ1, σ2, and σ3' should read 'Let L be the line generated by σ1, σ2, and σ3'.","section":"Example 5.9"},{"comment":"The large diagram in the statement is hard to parse; naming the maps and making the pullback square explicit would improve readability.","section":"Proposition 5.2"},{"comment":"The notation A_• = {⟨λ,a⟩ ∈ [0,1] × A ⊔ {•}: λ=0 iff a=•} is ambiguous; the Cartesian product should be with A ⊔ {•} rather than with A.","section":"Definition 3.1"},{"comment":"The statement does not explicitly say that the displayed inequalities are to be written for each 0 ≤ j ≤ m-1 and for each original inequality; this indexing is implicit but should be stated.","section":"Proposition 4.13"}],"recommendation":"major_revision","confidential_remarks":"The proof gap in Lemma 5.14 is the main obstacle; the rest of the paper is largely sound and the missing argument appears straightforward. The paper is within the scope of the journal and should not be rejected outright. I recommend major revision with a request to supply the general proof of Lemma 5.14 and to complete the verification of the Bell-inequality conditions in Proposition 4.13."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth reading for the cone decomposition alone. Theorem 4.11 (Theorem 1.1) — sDist(CX, ΔZ_m) is the join of m copies of sDist(X, ΔZ_m), with noncontextuality and vertices detected componentwise — is proved in detail and is a genuine structural handle on contextuality polytopes. The pullback description for suspensions in Proposition 5.2 is also neat and follows naturally from the cone result. The disjoint-union decomposition (Proposition 3.11) with the connectedness caveat is a solid extension of the classical fact, and the counterexample after Corollary 3.13 shows the author knows the hypothesis is load-bearing. The Bell-inequality derivation in Proposition 4.13 is a clean application.\n\nThe reader's conditional verdict is fair. Lemma 5.14 is the real soft spot. It is stated for any complete collection of average distributions, but the proof explicitly only treats Example 5.13 and never returns to the general case. Proposition 5.15 applies it to restrictions {p_j|_L} where the ordering of {I,S,...,S^{m-1}} on σ1 and σ2 can be arbitrary, so the general form is used. Example 5.16 (the (3,2,2) contextual vertex) depends on it. The stress-test note's suggested algebra — writing q_j|_{σ_k}=S^{α_k(j)} and deriving λ_j=μ_j then λ constant — looks plausible, so I suspect this is an omitted proof, not a false claim. But it needs to be written out. As it stands, the suspension-vertex results in 5.3 are unsupported.\n\nMinor things: the introduction doesn't clearly separate what is new from [11] (Kharoof-Okay on cone spaces), and the section title 'Simiplicial distributions on suspension scenarios' has a typo. Neither affects the math.\n\nBottom line: the core cone material deserves a serious referee and will be cited. Send it out, but the referee should ask for a full proof of Lemma 5.14 (or a restriction of the statement to what is actually proved) before acceptance. If the author supplies the general algebra, the paper is in good shape.","headline":"Core cone decomposition is a genuinely useful structural theorem, but the suspension section contains an unproven general lemma that needs to be fixed before the advertised vertex examples are supported.","tokens_in":29141,"tokens_out":1632,"would_cite":true,"duration_ms":15575,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["55U10","18N50","52B11"],"pacs":[],"model":"deepseek-v4-flash","headline":"Adding a single measurement to a connected scenario decomposes every simplicial distribution into m independent slices, and noncontextuality holds exactly when every slice is noncontextual.","keywords":["simplicial distributions","contextuality","cone construction","suspension","Bell inequalities","non-signaling polytope","décalage","join of convex sets"],"falsifier":"Compute the two sides of Corollary 3.13 for the disconnected space $X = \\Delta^0 \\sqcup \\Delta^0$: the paper itself shows the left side is a product of two joins while the right side is a join of two products, so the isomorphism is false there. A concrete check of dimension or vertex counts for any disconnected $X$ would expose the precise role of connectedness, and finding a connected $X$ where the cone isomorphism fails would refute the main theorem.","tokens_in":28141,"feed_emoji":"📐","tokens_out":10459,"duration_ms":91305,"temperature":0.7,"pith_summary":"This paper asks what happens to the geometry of contextuality when a measurement space is modified by adding one extra measurement point (the cone construction) or by gluing two cones together (the suspension). The central claim is that for a connected measurement space with m-outcome measurements, the non-signaling polytope of the cone scenario is the join of m copies of the original polytope: $sDist(CX, \\Delta\\mathbb{Z}_m) \\cong sDist(X, \\Delta\\mathbb{Z}_m) \\star \\cdots \\star sDist(X, \\Delta\\mathbb{Z}_m)$. Under this decomposition a distribution is noncontextual exactly when each of its m component distributions is noncontextual, and it is an extreme point exactly when one component is extreme and carries all the weight. The same decomposition yields Bell inequalities for cone scenarios from known ones and, after gluing two cones, a characterization of noncontextuality on suspension scenarios plus a construction of contextual vertices. If the picture is right, ordinary topological operations on the measurement space organize the polytope of correlations, making contextuality a topological phenomenon.","feed_headline":"Adding a measurement makes correlations an m-fold join","feed_subtitle":"With a connected measurement space, each distribution splits into m slices, and all slices must be noncontextual.","key_machinery":"The load-bearing machinery is a chain of three identifications. First, the cone–décalage adjunction $sSet(X, Dec_0Y) \\cong sSet(CX,Y)$ turns simplicial distributions on $CX$ into simplicial distributions on $X$ valued in $Dec_0(\\Delta\\mathbb{Z}_m)$. Second, the décalage of the outcome space splits into a disjoint union of $m$ copies: $Dec_0(\\Delta\\mathbb{Z}_m) \\cong \\coprod_{a\\in\\mathbb{Z}_m} \\Delta\\mathbb{Z}_m$. Third, for a connected measurement space, distributions on a disjoint-union outcome space decompose as the join of distributions on the components, with scalars recording the total probability on each component. The join $\\star$ is the coproduct of convex sets, so a vertex of a join is concentrated on a single component, and the hidden-variable map $\\Theta$ commutes with the decomposition. Assembling these identifications gives the cone isomorphism, and the suspension case follows by gluing two such decompositions along $X$.","core_discovery":"On the paper's own terms, the core discovery is a structural identity about simplicial distributions rather than a single new inequality. For any connected simplicial set $X$ and outcome space $\\Delta\\mathbb{Z}_m$, the cone construction induces an isomorphism of convex sets $$sDist(CX, \\$\\Delta$\\mathbb{Z}_m) \\cong sDist(X, \\$\\Delta$\\mathbb{Z}_m) \\star \\cdots \\star sDist(X, \\$\\Delta$\\mathbb{Z}_m)$$ ($m$ copies), where $\\star$ is the join of convex sets, i.e. the coproduct in the category of convex sets. Under this isomorphism a distribution $p$ decomposes uniquely as $(\\langle\\lambda_1,p^{(1)}\\rangle,\\ldots,\\langle\\lambda_m,p^{(m)}\\rangle)$ with $\\sum_j\\lambda_j=1$; $p$ is noncontextual if and only if every $p^{(j)}$ is noncontextual, and $p$ is a vertex if and only if some $\\lambda_j=1$ and $p^{(j)}$ is a vertex. Gluing two copies of the cone along $X$, the suspension scenario inherits a pullback description of $sDist(\\Sigma X, \\Delta\\mathbb{Z}_m)$, and noncontextuality is characterized by matching convex combinations of hidden-variable distributions on the upper and lower cones. From these structures the paper derives Bell inequalities for cone scenarios and gives a topological explanation of two families of contextual vertices appearing in the literature for the $(3,2,2)$ Bell scenario.","pith_inferences":["Beyond the paper: the isomorphism gives a polytope-join rule for resource theories, so monotones such as dimension or vertex count of the non-signaling polytope can be updated exactly when one party is added, without recomputing the whole polytope.","Beyond the paper: the connectedness hypothesis is the real domain boundary; the paper's two-vertex counterexample suggests that disconnected measurement spaces should be handled by decomposing into connected components first and then combining the results with products at the splitting points.","Beyond the paper: because the décalage splitting holds for the nerve of any group, analogous cone and suspension decompositions should exist for other outcome groups, though the arithmetic conditions used to construct deterministic contextual vertices are special to $\\mathbb{Z}_m$."],"forward_implications":["An added party with one $m$-outcome measurement turns the old non-signaling polytope into an $m$-fold join, so dimension, vertex count, and facet structure of the enlarged polytope are computed from the old one.","Bell inequalities for any cone scenario can be written down from Bell inequalities for the base scenario: each original inequality becomes $m$ normalized inequalities indexed by the added party's outcome (Proposition 4.13).","On a suspension scenario, a distribution is noncontextual if and only if each of its upper and lower cone components is noncontextual through hidden-variable distributions whose weighted averages agree exactly (Proposition 5.4).","Complete collections of deterministic or average distributions on a line inside $X$ produce provably contextual vertices of $sDist(\\Sigma X, \\Delta\\mathbb{Z}_m)$, providing a mathematical explanation for the $(3,2,2)$ vertices in the literature (Propositions 5.8 and 5.15).","The cone result quantifies how contextuality survives the addition of a party: no mixture of components can create contextuality unless some component already has it, and no mixture can create a new vertex unless a single component carries one."],"supporting_citations":[{"why":"Introduces simplicial distributions and defines (non)contextuality as the image of the hidden-variable map $\\Theta$, the framework the paper extends.","marker":"[8]"},{"why":"Supplies the convex-monoid structure on $sDist(X,\\Delta\\mathbb{Z}_m)$, the deterministic-vertex facts, and the adjunction isomorphisms used in the cone diagram.","marker":"[22]"},{"why":"Provides the cone–décalage bijection $sSet(X, Dec_0 Y) \\cong sSet(CX,Y)$ that converts cone distributions into distributions on the original space.","marker":"[33]"},{"why":"Brings bundle scenarios and push-forward convex maps, which underpin the disjoint-union decomposition of Section 3.","marker":"[24]"},{"why":"Defines vertex support and closed sets of vertices, and supplies the characterization used to prove the new contextual vertices are extreme.","marker":"[13]"},{"why":"Is the source of the $(3,2,2)$ contextual vertices whose mathematical explanation is given in Examples 5.9 and 5.16.","marker":"[17]"},{"why":"Supplies the PR-box vertex and its $p_+/p_-$ form used in the CHSH examples.","marker":"[16]"}],"fun_headline_variants":["Cone trick turns correlations into an m-fold join","Suspension scenarios reveal hidden join structure","Noncontextuality splits into m independent slices","Topology explains contextual vertices in Bell scenarios"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the measurement space is connected in the combinatorial sense; the paper's own two-isolated-vertices example shows the decomposition fails without this condition.","fun_headline_variants_meta":{"raw":{"variants":["Cone trick turns correlations into an m-fold join","Suspension scenarios reveal hidden join structure","Noncontextuality splits into m independent slices","Topology explains contextual vertices in Bell scenarios"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000244,"raw_usage":{"total_tokens":1610,"prompt_tokens":1103,"completion_tokens":507,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":719,"completion_tokens_details":{"reasoning_tokens":450}},"tokens_in":719,"tokens_out":507,"duration_ms":4432,"temperature":1.0,"reasoning_tokens":450,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T15:27:11.071972+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the two sides of Corollary 3.13 for the disconnected space $X = \\Delta^0 \\sqcup \\Delta^0$: the paper itself shows the left side is a product of two joins while the right side is a join of two products, so the isomorphism is false there. A concrete check of dimension or vertex counts for any disconnected $X$ would expose the precise role of connectedness, and finding a connected $X$ where the cone isomorphism fails would refute the main theorem.","supporting_citations":[{"cited_title":"Simplicial quantum contextuality","cited_arxiv_id":null,"evidence_quote":"Introduces simplicial distributions and defines (non)contextuality as the image of the hidden-variable map $\\Theta$, the framework the paper extends."},{"cited_title":"Simplicial distributions, convex categories and contextuality","cited_arxiv_id":"2211.00571","evidence_quote":"Supplies the convex-monoid structure on $sDist(X,\\Delta\\mathbb{Z}_m)$, the deterministic-vertex facts, and the adjunction isomorphisms used in the cone diagram."},{"cited_title":"Classifying theory for simplicial parametrized groups","cited_arxiv_id":"1203.2461","evidence_quote":"Provides the cone–décalage bijection $sSet(X, Dec_0 Y) \\cong sSet(CX,Y)$ that converts cone distributions into distributions on the original space."}],"review_version":1}