Pith. sign in

REVIEW 1 major objections 5 minor 49 references

This paper claims that contextuality is a local-to-global consistency failure for effect-valued measurement data, and proves that internal non-contextuality is equivalent to the existence of coherent classical explanations across all observ

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-01 01:09 UTC pith:ITKZXQUK

load-bearing objection A genuine generalization of sheaf-theoretic contextuality to effect-valued weights, with a sound core theorem and two peripheral blemishes that should be fixed before publication. the 1 major comments →

arxiv 2607.25900 v1 pith:ITKZXQUK submitted 2026-07-28 quant-ph

Effect-valued measurement models and contextuality

Samson Abramsky This is my paper
classification quant-ph MSC 06C1518C2081P1381P15 PACS 03.65.Ta03.65.Ud
keywords contextualityeffect algebraseffect modulesmeasurement modelsPOVMscontextual fractioncone programminguniform dilation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper generalizes the standard sheaf-theoretic analysis of contextuality—whether a family of local probability distributions over measurement contexts can be merged into one global distribution—to distributions whose weights live in a convex effect algebra. That single move puts classical probability, deterministic/possibilistic models, and quantum POVMs and PVMs under one formalism. The key distinction the paper draws is between internal contextuality of an effect-valued measurement model and observable contextuality of the ordinary probabilistic models obtained by evaluating it under states. The central theorem says these coincide, for effect modules whose states separate points, exactly when the statewise classical explanations are coherently representable by a single effect-valued global distribution. Because the framework is built on ordered vector spaces, non-contextuality becomes a cone-feasibility problem, the contextual fraction extends to this setting, and its dual gives generalized Bell-type witnesses.

Core claim

The central claim, stated as Theorem V.4, is that for an effect module A=[0,u]_V with point-separating states, an A-measurement model m is internally A-non-contextual if and only if the family of empirical models {σ*(m)} obtained by evaluating m under all states admits a coherently representable family of classical global explanations. Concretely, for each global assignment t, the statewise probabilities p_σ(t) must themselves be evaluations σ(a_t) of a single effect a_t in A. In vector-space terms this is equivalent to feasibility of the cone program: find d∈V_+^{E(X)} with Σ_t d(t)=u and M_V d=m. The same coherence condition characterizes equality between the effect-valued contextual fract

What carries the argument

The load-bearing construction is the A-valued distribution functor D_A, which assigns to each outcome set X the A-distributions d:X→A with total weight 1; restriction is marginalization, exactly as in the probabilistic case. For effect modules, the representation A=[0,u]_V embeds A as an interval in a real ordered vector space, so an A-distribution is a family of positive vectors summing to the order unit u, and the marginal map M_V packages all compatibility constraints into one linear map. The statewise lifting lemma (Lemma V.2) is the tool that assembles coherent statewise classical explanations into a single A-valued global distribution; its use of point separation is the delicate step t

Load-bearing premise

The reconstruction theorems assume that states separate points in A: if two distinct effects look identical under every state, then statewise classical explanations cannot be uniquely glued into a single effect-valued global explanation, and the equivalence between internal and observable non-contextuality breaks down.

What would settle it

Look for a counterexample to the converse of Theorem V.4: an effect module A whose states do not separate points together with an A-model m such that every statewise image σ*(m) is non-contextual but m has no A-valued global section. A concrete starting point is the quotient of an effect module by the relation 'equal under every state'; checking whether the quotient of a known POVM model, such as three binary POVMs that are pairwise but not jointly triplewise measurable, inherits a global section is a finite-dimensional cone computation. A positive instance would show point separation is indis

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • The probabilistic, deterministic/possibilistic, and quantum POVM/PVM accounts of contextuality all become special cases of one effect-valued formalism.
  • Internal non-contextuality always implies statewise non-contextuality, but the converse holds only under coherence; state-by-state tests alone can miss internal contextuality.
  • Non-contextuality is a cone-program feasibility problem; its dual cone yields operator-valued contextuality witnesses, so Bell-type inequalities appear as state-evaluated special cases.
  • The effect-valued contextual fraction upper-bounds the contextual fraction in every state, and equality with the worst-case statewise value is equivalent to coherent representability of optimal submodels.
  • General POVM-models need not admit uniform dilations to compatible PVM-models, even though individual POVMs always dilate; tensor-product Bell POVM-models do dilate uniformly.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Not spelled out in the paper: because generalized witnesses are operator-valued and act before state evaluation, they could support state-independent or state-robust contextuality certification in quantum experiments.
  • The cone-program formulation suggests a computational route: for finite-dimensional quantum effects the internal contextual fraction is a semidefinite program, so one can numerically compare internal and statewise contextual fractions for concrete POVM scenarios.
  • The uniform-dilation obstruction defines a hierarchy—jointly measurable, extremal, uniformly dilatable—that could be treated as a resource ordering, for example connecting unsharpness to advantage in non-local games or graph homomorphism problems.
  • The graded-monad construction indicates a compositional resource calculus in which Hilbert-space tensor products grade the effect structure; extending it from a monoid of resources to arbitrary symmetric monoidal categories would yield a fully categorical resource theory of contextuality.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

1 major / 5 minor

Summary. The paper generalizes the Abramsky–Brandenburger sheaf-theoretic framework by replacing probability distributions with finitely supported distributions valued in a convex effect algebra A. For a measurement scenario (X,Σ,O) it defines A-measurement models, whose states σ:A→[0,1] push forward to ordinary empirical models. The central distinction is between internal A-contextuality (absence of a global A-valued distribution) and observable/statewise contextuality. Theorem V.4 asserts that, when states separate points, internal non-contextuality is equivalent to the existence, for every state, of a classical global explanation that is coherently representable by A-elements. The paper then formulates non-contextuality as a cone feasibility problem over the ordered-vector-space representation, defines an A-contextual fraction, derives dual generalized Bell witnesses, studies sharp/dilation properties (Specker obstruction, uniform dilation in Bell scenarios), and constructs graded monads from resource-indexed effect monoids.

Significance. If the central results hold, this is a valuable conceptual unification. The internal/observable distinction is natural and new, and the cone-program characterization genuinely generalizes the LP characterization of probabilistic contextuality and the contextual fraction. The reconstruction theorem is proved directly rather than assumed, and the state-separation hypothesis is explicit and satisfied by the main quantum example A=E(H). The discussion of uniform dilation correctly exposes a difference between individual POVM dilation and context-uniform dilation. The paper also honestly flags limitations: factorisability requires effect monoids, convex fraction requires effect modules not plain effect algebras. I am not aware of a competing framework that subsumes both the sheaf-theoretic and GPT aspects at this level of generality.

major comments (1)
  1. [V.A, Lemma V.2 / Theorem V.4] The proof of Theorem V.4 applies Lemma V.2 with Y=E(X), but Lemma V.2 is stated for finite Y. Theorem V.4 has no finiteness hypothesis on the measurement scenario. Consequently the function d constructed in the converse may fail to have finite support and need not be an element of D_A(E(X)); the conclusion "there exists d∈D_A(E(X))" is not established. For the finite scenarios used in Sections VII–IX this is harmless, but as written the central equivalence is gappy. Please add an explicit finite-scenario hypothesis to Theorems V.4 and VIII.2 (which repeats the same use of the lemma), or prove a generalization of Lemma V.2 that derives finite support from the coherence conditions.
minor comments (5)
  1. [XI, R-graded effect monoid] The axiom A_I={0,1} contradicts the quantum-effects example, where A_1=E(C)≅[0,1]. The example as stated is not an instance of the definition. Please weaken the axiom to the unit laws actually needed for the graded monad, or restrict the sharp example to projections.
  2. [IX.A, Example IX.1] The CHSH operator B=A0⊗B0+A0⊗B1+A1⊗B0−A1⊗B1 is not positive, whereas the generalized Bell witnesses obtained by cone duality are elements of the dual positive cone. Thus B is not a special case of the internal witnesses constructed above. Please replace it by a positive success-probability functional (e.g., built from (4I+B)/8) or clearly label the statement as heuristic.
  3. [VIII] The notation CF_obs and NCF_obs can confuse: CF_obs is a supremum of ordinary contextual fractions while NCF_obs is defined as an infimum. A short sentence explaining that these are complementary would improve readability.
  4. [IX] The strong-duality/attainment assumptions in Theorem IX.2 are not trivial for infinite-dimensional ordered vector spaces. Please either state explicit regularity conditions or restrict the dual-witness theorem to the finite-dimensional conic/SDP setting (where they are automatic).
  5. [Miscellaneous] Typos: 'reconstruction prooblem' (Section V.A), 'OV ectu' (Theorem VII.1), and inconsistent apostrophes in 'povm’s' throughout. Also, in Section II.B the phrase 'if a≤a′ ⊥ b, then a⊥b' would benefit from a displayed definition.

Circularity Check

0 steps flagged

No significant circularity: central reconstruction theorem is proved, not assumed; self-citation in graded-monad proof is peripheral.

full rationale

The main derivation chain is self-contained and non-circular. Theorem V.4 defines 'coherently representable' as a separate condition on statewise explanations (existence of a_t in A with p_sigma(t)=sigma(a_t)); the converse direction does not assume the conclusion but uses Lemma V.2 to assemble d(t)=a_t and then derives normalization sum_t d(t)=u and marginal agreement d|_C=m_C from state separation. Lemma V.2 itself is a direct consequence of the explicit 'states separate points' assumption, which is stated at the start of Section V.A and is satisfied by A=E(H). The cone-program formulation (Section VII.B) is a faithful translation of the definition of internal non-contextuality into the ordered-vector-space representation of [23], not a hidden reuse of the target. The dual-witness theorem (IX.2) is a standard strong-duality statement: the dual optimum equals NCF_A(m) by conic duality, and 1 - <beta,m> = CF_A(m) is then just the definition of the contextual fraction; this is a mathematical characterization, not an empirical prediction forced by a fitted parameter. The only same-author citation used in place of a proof is Proposition XI.1, which is deferred to [1]; this is peripheral to the central internal/observable equivalence, and [1] proves the special case without assuming the present results, so it is not load-bearing. The reader's noted correctness issues (signed CHSH operator used as a 'positive' cone witness in Example IX.1; A_1=E(C) not {0,1} in the graded-monad example) are genuine but belong to correctness risk, not circularity, and do not affect Theorem V.4 or the reconstruction argument.

Axiom & Free-Parameter Ledger

0 free parameters · 6 axioms · 0 invented entities

The paper introduces no free parameters fitted to data. Its novel constructs — A-valued measurement models, internal/observable contextuality, uniform dilation, graded effect monoids — are mathematical definitions rather than empirical entities. The axioms listed are the background results and explicit hypotheses that the central theorems rest on.

axioms (6)
  • domain assumption Effect algebra and effect module axioms (Definitions II.1 and II.3), including orthosupplementation and scalar multiplication.
    The entire A-valued framework is built on these structures; they are standard and explicitly defined.
  • standard math Gudder-Pulmannová representation theorem: every effect module is isomorphic to an interval [0,u] in a real ordered vector space with generating order unit.
    Invoked as Theorem VII.1 and used for the cone-program formulation and dual witnesses. Cited from [23] and not proved in the paper.
  • domain assumption The state space S(A) separates the points of A for Theorems V.4 and VIII.2.
    This hypothesis is explicitly assumed in Section V.A and is essential for Lemma V.2's uniqueness and reconstruction step. It holds for [0,1] and E(H), but not for arbitrary effect modules.
  • domain assumption Strong conic duality and dual attainment for the cone program defining NCF_A(m) in Theorem IX.2.
    The dual characterization of the A-contextual fraction relies on this; the paper notes it is automatic in finite-dimensional polyhedral cases and requires Slater-type conditions in general.
  • standard math Naimark dilation and the joint dilation theorem for several POVMs (Theorems X.3 and X.4).
    These are standard operator-algebraic results cited from [11,33] and used for the sharpness/dilation analysis.
  • standard math Gleason's theorem and Busch's theorem identifying states of P(H) and E(H) with density operators.
    Used in Remark II.6 and in the dilation discussion (Section X.A) to transfer POVM statistics to PVM statistics.

pith-pipeline@v1.3.0-alltime-deepseek · 24118 in / 13527 out tokens · 133852 ms · 2026-08-01T01:09:10.542294+00:00 · methodology

0 comments
read the original abstract

We generalize the Abramsky--Brandenburger sheaf-theoretic treatment of contextuality by replacing probability distributions with distributions valued in a convex effect algebra \(A\). This yields a notion of \(A\)-valued measurement model encompassing probabilistic, deterministic, and quantum measurement structures within a single framework. States \(\sigma:A\to[0,1]\) induce ordinary empirical models, allowing observable behaviour to be viewed as arising from effect-valued measurement data. This leads to a distinction between internal contextuality of \(A\)-models and observable contextuality after state evaluation. We analyze the relationship between these notions and identify coherence conditions under which observable classical explanations assemble into internal ones. Using the ordered-vector-space representation of effect modules, we show that non-contextuality is characterized by feasibility of an associated cone program, generalizing the linear-programming formulation of contextuality in the probabilistic case. We also introduce an effect-valued contextual fraction and study its relation to observable contextuality witnesses. We show how cone duality leads to a notion of Bell witnesses for contextuality as a direct generalization of Bell inequalities. Finally, we analyze sharp realizability and uniform dilation for measurement models, clarifying the relationship between general POVMs and projective measurements, and show how resource-indexed effect structures induce graded monads generalizing the quantum monad.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

49 extracted references · 1 canonical work pages

  1. [1]

    Bulletin of Symbolic Logic , volume=

    Team semantics and independence notions in quantum physics , author=. Bulletin of Symbolic Logic , volume=. 2026 , publisher=

    2026

  2. [2]

    Paulsen and Simone Severini and Dan Stahlke , title =

    Vern I. Paulsen and Simone Severini and Dan Stahlke , title =. Journal of Combinatorial Theory, Series A , volume =. 2019 , doi =. 1607.03450 , archivePrefix =

    Pith/arXiv arXiv 2019

  3. [3]

    Information and Computation , volume=

    Effect algebras, presheaves, contextuality and non-locality , author=. Information and Computation , volume=. 2018 , publisher=

    2018

  4. [4]

    International Journal of Theoretical Physics , volume=

    S-dominating effect algebras , author=. International Journal of Theoretical Physics , volume=. 1998 , publisher=

    1998

  5. [5]

    Nature communications , volume=

    An experimental test of noncontextuality without unphysical idealizations , author=. Nature communications , volume=. 2016 , publisher=

    2016

  6. [6]

    Paulsen , title =

    Vern I. Paulsen , title =

  7. [7]

    Journal of Physics A: Mathematical and Theoretical , volume=

    Generalized probability theories: what determines the structure of quantum theory? , author=. Journal of Physics A: Mathematical and Theoretical , volume=. 2014 , publisher=

    2014

  8. [8]

    Physics Reports , volume=

    General probabilistic theories: An introduction , author=. Physics Reports , volume=. 2023 , publisher=

    2023

  9. [9]

    Foundations of Physics , volume=

    Notes on joint measurability of quantum observables , author=. Foundations of Physics , volume=. 2008 , publisher=

    2008

  10. [10]

    Foundations of Physics , volume=

    Effect algebras are not adequate models for quantum mechanics , author=. Foundations of Physics , volume=. 2010 , publisher=

    2010

  11. [11]

    PRX Quantum , volume=

    Certified randomness in tight space , author=. PRX Quantum , volume=. 2024 , publisher=

    2024

  12. [12]

    1993 , publisher=

    Gleason's Theorem and its Applications , author=. 1993 , publisher=

    1993

  13. [13]

    Indiana Univ

    Andrew M Gleason. Indiana Univ. Math. J

  14. [14]

    2011 , publisher=

    The mathematical language of quantum theory: from uncertainty to entanglement , author=. 2011 , publisher=

    2011

  15. [15]

    Physical Review A—Atomic, Molecular, and Optical Physics , volume=

    Contextuality for preparations, transformations, and unsharp measurements , author=. Physical Review A—Atomic, Molecular, and Optical Physics , volume=. 2005 , publisher=

    2005

  16. [16]

    Physical Review A—Atomic, Molecular, and Optical Physics , volume=

    Information processing in generalized probabilistic theories , author=. Physical Review A—Atomic, Molecular, and Optical Physics , volume=. 2007 , publisher=

    2007

  17. [17]

    Theory of Probability & Its Applications , volume=

    Consistent families of measures and their extensions , author=. Theory of Probability & Its Applications , volume=. 1962 , publisher=

    1962

  18. [18]

    Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , volume=

    Combining contextuality and causality: a game semantics approach , author=. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , volume=. 2024 , publisher=

    2024

  19. [19]

    2019 34th Annual ACM/IEEE Symposium on Logic in Computer Science (LICS) , pages=

    A comonadic view of simulation and quantum resources , author=. 2019 34th Annual ACM/IEEE Symposium on Logic in Computer Science (LICS) , pages=. 2019 , organization=

    2019

  20. [20]

    2016 , eprint =

    Roumen, Frank , title =. 2016 , eprint =

    2016

  21. [21]

    and Raussendorf, Robert , title =

    Okay, Cihan and Roberts, Sam and Bartlett, Stephen D. and Raussendorf, Robert , title =. 2017 , eprint =

    2017

  22. [22]

    Contextuality, Cohomology and Paradox , booktitle =

    Samson Abramsky and Rui Soares Barbosa and Kohei Kishida and Raymond Lal and Shane Mansfield , editor =. Contextuality, Cohomology and Paradox , booktitle =. 2015 , url =. doi:10.4230/LIPICS.CSL.2015.211 , timestamp =

    doi:10.4230/lipics.csl.2015.211 2015

  23. [23]

    The Cohomology of Non-Locality and Contextuality , booktitle =

    Samson Abramsky and Shane Mansfield and Rui Soares Barbosa , editor =. The Cohomology of Non-Locality and Contextuality , booktitle =. 2011 , url =. doi:10.4204/EPTCS.95.1 , timestamp =

    doi:10.4204/eptcs.95.1 2011

  24. [24]

    Information and Computation , volume=

    A mathematical theory of resources , author=. Information and Computation , volume=. 2016 , publisher=

    2016

  25. [25]

    Mathematical Structures in Computer Science , volume=

    Resource convertibility and ordered commutative monoids , author=. Mathematical Structures in Computer Science , volume=. 2017 , publisher=

    2017

  26. [26]

    The Quantum Monad on Relational Structures , booktitle =

    Samson Abramsky and Rui Soares Barbosa and Nadish de Silva and Octavio Zapata , editor =. The Quantum Monad on Relational Structures , booktitle =. 2017 , url =. doi:10.4230/LIPICS.MFCS.2017.35 , timestamp =

    doi:10.4230/lipics.mfcs.2017.35 2017

  27. [27]

    2021 , publisher=

    Ji, Zhengfeng and Natarajan, Anand and Vidick, Thomas and Wright, John and Yuen, Henry , journal=. 2021 , publisher=

    2021

  28. [28]

    Tsirelson, Boris , howpublished=

  29. [29]

    Entanglement and non-locality , author=

  30. [30]

    Quantum Measurement , pages=

    Dilation theory , author=. Quantum Measurement , pages=. 2016 , publisher=

    2016

  31. [31]

    Forum of Mathematics, Pi , volume=

    The set of quantum correlations is not closed , author=. Forum of Mathematics, Pi , volume=. 2019 , organization=

    2019

  32. [32]

    Physical Review A , volume=

    Minimal state-dependent proof of measurement contextuality for a qubit , author=. Physical Review A , volume=. 2014 , publisher=

    2014

  33. [33]

    2013 , eprint=

    Quantum contextuality and joint measurement of three observables of a qubit , author=. 2013 , eprint=

    2013

  34. [34]

    2011 , publisher=

    Liang, Yeong-Cherng and Spekkens, Robert W and Wiseman, Howard M , journal=. 2011 , publisher=

    2011

  35. [35]

    International Journal of Theoretical Physics , volume=

    Sharp elements in effect algebras , author=. International Journal of Theoretical Physics , volume=. 2001 , publisher=

    2001

  36. [36]

    Physical Review Letters , volume=

    No quantum realization of extremal no-signaling boxes , author=. Physical Review Letters , volume=. 2016 , publisher=

    2016

  37. [37]

    Theoretical Computer Science , volume=

    Probabilities, distribution monads, and convex categories , author=. Theoretical Computer Science , volume=. 2011 , publisher=

    2011

  38. [38]

    1982 , publisher=

    Fine, Arthur , journal=. 1982 , publisher=

    1982

  39. [39]

    Logical Methods in Computer Science , volume =

    Cho, Kenta and Westerbaan, Bas and van de Wetering, John , title =. Logical Methods in Computer Science , volume =. 2022 , pages=. doi:10.46298/lmcs-18(3:13)2022 , eprint =

    work page doi:10.46298/lmcs-18(3:13)2022 2022

  40. [40]

    Journal of Pure and Applied Algebra , volume =

    Westerbaan, Abraham and Westerbaan, Bas and van de Wetering, John , title =. Journal of Pure and Applied Algebra , volume =. 2021 , doi =. 1912.10040 , archivePrefix =

    Pith/arXiv arXiv 2021

  41. [41]

    Journal of Physics A: Mathematical and General , volume =

    D'Ariano, Giacomo Mauro and Lo Presti, Paoloplacido and Perinotti, Paolo , title =. Journal of Physics A: Mathematical and General , volume =. 2005 , doi =. quant-ph/0408115 , archivePrefix =

    Pith/arXiv arXiv 2005

  42. [42]

    Complete characterization of extreme quantum observables in infinite dimensions , journal =

    Pellonp. Complete characterization of extreme quantum observables in infinite dimensions , journal =. 2011 , doi =. 1010.4972 , archivePrefix =

    Pith/arXiv arXiv 2011

  43. [43]

    Extreme commutative quantum observables are sharp , journal =

    Heinosaari, Teiko and Pellonp. Extreme commutative quantum observables are sharp , journal =. 2011 , doi =. 1106.4295 , archivePrefix =

    Pith/arXiv arXiv 2011

  44. [44]

    New Journal of Physics , volume =

    Abramsky, Samson and Brandenburger, Adam , title =. New Journal of Physics , volume =. 2011 , doi =

    2011

  45. [45]

    Physical Review Letters , volume =

    Abramsky, Samson and Barbosa, Rui Soares and Mansfield, Shane , title =. Physical Review Letters , volume =. 2017 , doi =

    2017

  46. [46]

    Foulis, D. J. and Bennett, M. K. , title =. Foundations of Physics , volume =. 1994 , doi =

    1994

  47. [47]

    and Pulmannov

    Gudder, Stanley P. and Pulmannov. Representation theorem for convex effect algebras , journal =

  48. [48]

    Physical Review Letters , volume =

    Busch, Paul , title =. Physical Review Letters , volume =. 2003 , doi =

    2003

  49. [49]

    Journal of Physics A: Mathematical and Theoretical , volume =

    Pironio, Stefano and Bancal, Jean-Daniel and Scarani, Valerio , title =. Journal of Physics A: Mathematical and Theoretical , volume =. 2011 , doi =

    2011