PaperlayerView exact revision on arXiv

Source-linked AI summary

Quantum-Bayesian Coherence: The No-Nonsense Version

Christopher A. Fuchs, Ruediger Schack

arXiv:1301.3274v1quant-ph

TL;DR

The paper asks how QBism should understand the Born Rule without objective quantum states. It recasts the rule as an empirically extended Bayesian coherence requirement using a counterfactual SIC measurement, then examines how much quantum-state structure follows.

  • Problem

    QBism needs an account of why probability assignments obey the Born Rule without treating quantum states or probabilities as objective.

  • Method

    The paper represents states with probabilities for a fiducial SIC measurement and formulates the Born Rule through an imaginary conditional lottery, as the urgleichung.

  • Results

    The paper shows that the Born Rule can be written as a normative addition to Dutch-book coherence and derives several features suggestive of quantum-state space.

  • Takeaways & Limitations

    The framework provides a setting for quantifying the quantum-mechanical idea that unperformed measurements have no outcomes.

  • Takeaways & Limitations

    The paper leaves deriving the full formal structure of quantum mechanics, including dynamics and tensor-product structure, for future work.

Abstract

from arXiv · show

In the Quantum-Bayesian interpretation of quantum theory (or QBism), the Born Rule cannot be interpreted as a rule for setting measurement-outcome probabilities from an objective quantum state. But if not, what is the role of the rule? In this paper, we argue that it should be seen as an empirical addition to Bayesian reasoning itself. Particularly, we show how to view the Born Rule as a normative rule in addition to usual Dutch-book coherence. It is a rule that takes into account how one should assign probabilities to the consequences of various intended measurements on a physical system, but explicitly in terms of prior probabilities for and conditional probabilities consequent upon the imagined outcomes of a special counterfactual reference measurement. This interpretation is seen particularly clearly by representing quantum states in terms of probabilities for the outcomes of a fixed, fiducial symmetric informationally complete (SIC) measurement. We further explore the extent to which the general form of the new normative rule implies the full state-space structure of quantum mechanics.

A. Outline of the Paper

The paper develops QBism by supplementing Dutch-book coherence with an empirically motivated Born-Rule relation, then investigates how much quantum-state structure follows from it. It concludes that Hilbert space has been indicated but not derived.

  • II. PERSONALIST BAYESIAN PROBABILITY: Section II reviews personalist Bayesian probability and argues that Bayes’ rule and the Law of Total Probability require an operative conditional lottery.Without such a lottery, Dutch-book coherence alone cannot compel these relations.
  • III. EXPRESSING QUANTUM-STATE SPACE IN TERMS OF SICS: Section III shows that SIC measurements provide a uniquely simple and compact probability representation of quantum states.In SIC terms, unitary evolution also formally resembles classical stochastic evolution.
  • IV. THE BORN RULE AS EMPIRICALLY EXTENDED COHERENCE: Section IV rewrites the Born Rule using an imaginary counterfactual SIC, yielding a relation strikingly similar to the Law of Total Probability.The interpretation treats the counterfactual SIC as supplying conditional probabilities for the intended measurement.
  • V. DERIVING QUANTUM-STATE SPACE: Section V derives features of quantum-state space from the modified Quantum Law of Total Probability, including a generalized Bloch sphere and matching underlying dimensionality.The derivation requires a small number of further assumptions and also produces additional geometric features.
  • VI–VIII. FURTHER ANALYSIS AND OUTLOOK: Sections VI–VIII examine the constants, summarize the research status, and state that Hilbert space, dynamics, and tensor-product structure remain future goals.The paper frames its current ontology around unperformed measurements having no outcomes, tempered by a distinctive form of realism.

II. PERSONALIST BAYESIAN PROBABILITY

Personalist Bayesian probability treats probabilities as agent-relative commitments governed by Dutch-book coherence. Standard conditional-probability rules become normatively binding only when the relevant conditional lottery is actually part of the betting setup.

  • Personalist probability: Probability is a degree of belief expressed through an agent’s willingness to buy or sell lottery tickets at a stated price.The personalist account assigns no correctness to a probability independently of the agent who makes the assignment.
  • Dutch-book coherence: Dutch-book coherence requires probabilities to avoid betting strategies that produce a sure loss, yielding the usual probability calculus.The paper traces this normative framework to personalist Bayesian reasoning associated with Ramsey and de Finetti.
  • Probability sum rule: For mutually exclusive events, coherence requires the probability of their disjunction to equal the sum of their individual probabilities.Otherwise, buying and selling corresponding tickets creates a guaranteed negative balance sheet.
  • Conditional lotteries: Bayes’ rule and the Law of Total Probability are enforced by Dutch-book coherence when a conditional lottery reveals the conditioning event before settling the conditional ticket.The conditional ticket is cancelled and refunded when the conditioning event fails.
  • Counterfactual measurements: If the conditioning measurement is called off, coherence alone does not require retaining the original unconditional probability for the later measurement.The later measurement may be assigned differently because the earlier action could change the situation in which it occurs.
  • Born Rule: The Born Rule adds a precise relation between probabilities for a factual measurement and a counterfactual reference measurement beyond Dutch-book coherence.This added relation is presented as the positive content of the Born Rule within Bayesianism.

III. EXPRESSING QUANTUM-STATE SPACE IN TERMS OF SICS

The paper represents quantum states as a constrained subset of probability distributions over d2 outcomes, using SIC measurements to make that representation especially simple. It then characterizes the geometry and algebraic structure of the valid distributions.

  • SIC measurements: SIC measurements are selected because their symmetry makes the probability representation as simple as possible relative to the simplex geometry.SICs are POVMs with d2 outcomes whose operators satisfy equal-trace and equal-overlap conditions.
  • Probability representation: A minimal informationally complete fiducial measurement maps each density operator injectively to d2 outcome probabilities that completely specify the quantum state.The measurement operators form a basis for L(Hd), so the probabilities determine ρ.
  • Quantum-state space: Only some points in the probability simplex represent quantum states; the valid distributions form a smaller convex set rather than the entire simplex.The mapping from density operators to probability vectors is injective but not surjective.
  • SIC measurements: A SIC is a symmetric informationally complete POVM whose elements are linearly independent and therefore informationally complete.The paper identifies associated projection operators and state vectors as part of the SIC structure.
  • Scope: SIC existence is unresolved in every finite dimension, although analytical constructions are known for several dimensions and high-precision numerical evidence extends further.The paper lists analytical results for dimensions including 2–16, 19, 24, 28, 31, 35, 37, 43, and 48.
  • Geometry of state space: The valid pure-state probability vectors can be characterized through algebraic constraints corresponding to the extreme points of the convex quantum-state set.The full state space is the convex hull of these pure-state vectors.

A. Aside on Unitarity

The paper recasts the Born Rule as a relation among probabilities for a counterfactual SIC measurement, conditional outcomes, and an intended laboratory measurement. This relation resembles classical stochastic probability while imposing additional quantum structure.

  • Unitarity: In SIC representations, unitary evolution is formally close to classical stochastic evolution because its conditional-probability matrix is doubly stochastic.Removing the quantum-specific factor and term would leave classical stochastic evolution.
  • Born Rule in SIC terms: The Born Rule is rewritten in SIC terms for an arbitrary von Neumann measurement, making the paper’s central interpretive move explicit.The reformulation is presented as the heart of the paper.
  • Counterfactual reference measurement: The counterfactual SIC supplies prior probabilities and conditional probabilities for the intended ground measurement, while the SIC itself need not occur.The sky measurement is imagined before the ground measurement and uses Lüders updating for conditional assignments.
  • Empirically extended coherence: Dutch-book coherence alone yields the classical total-probability expression for the hypothetical conditional lottery, but quantum mechanics assigns a different probability when that lottery is nullified.The distinction separates conditional coherence from the unconditional ground-measurement assignment.
  • Empirically extended coherence: The resulting relation formalizes the claim that the unperformed SIC had no outcomes while adding information not supplied by Dutch-book coherence alone.The paper calls this additional constraint empirically extended coherence.
  • State-space consequences: Requiring the ground probabilities to remain proper distributions restricts the allowed sharpness of the reference and conditional distributions, shaping quantum-state space.The paper connects these probability bounds to the structure of extreme states.
  • General POVMs: For a general POVM, the Quantum Law of Total Probability contains a classical total-probability term plus a term depending only on the sum of conditional probabilities.When the ground measurement is another SIC, the relation reduces to a specialized form.

A. Why “Empirically Extended Coherence” Instead of Objective Quantum States?

The paper treats the Born Rule as an empirically based normative addition to Dutch-book coherence, not as a rule imposed by objective quantum states. This preserves subjective state assignments while constraining their probabilistic relationships in measurement contexts.

  • The Born Rule is proposed as an empirically based addition to Dutch-book coherence that guides behavior during physical interactions.It supplies extra normative rules beyond standard probability theory.
  • The proposal restricts allowable probability assignments without diminishing the role of prior beliefs in determining an agent’s assignments.The authors contrast this with objective chance, which would override prior beliefs.
  • Quantum-Bayesian agents may assign different quantum states to the same system because state assignments depend on priors and statistical practice.The paper gives a two-qubit example in which compatible assignments lead to distinct post-measurement states.
  • The added probabilistic relations narrow agents from the full probability simplex to quantum states but do not determine which quantum state an agent should choose.The formalism provides normative constraints, not an objective state selection rule.
  • The authors leave the full formal structure of quantum mechanics and the ontology underlying the Quantum-Bayesian position for future work.They specifically mention dynamics and tensor-product structure as unresolved goals.

V. DERIVING QUANTUM-STATE SPACE FROM “EMPIRICALLY EXTENDED COHERENCE”

The paper studies how far an empirically extended coherence rule can recover quantum-state-space structure without presupposing Hilbert-space quantum mechanics. It formulates priors and conditional measurements abstractly, imposes consistency and maximality, and derives geometric properties.

  • The analysis represents states as probability vectors in the simplex and asks which features of quantum-state space follow from the conceptual framework.The authors treat pure states as an algebraic variety and all states as their convex hull, while adding measurement assumptions.
  • The urgleichung relates sky priors p(i) to ground probabilities q(j) through conditional probabilities r(j|i) for a counterfactual reference measurement.The conditional-probability matrix R is stochastic, though not necessarily doubly stochastic.
  • Consistency requires all admissible state–measurement pairs to satisfy the fundamental inequality ensuring valid ground probabilities.The sets P of priors and R of stochastic matrices are defined relative to this requirement.
  • Maximality makes P and R as non-exclusionary as possible while remaining consistent, although uniqueness is not guaranteed without further assumptions.The paper notes that multiple maximal sets may exist, possibly related by transformations.
  • If P and R are consistent and maximal, both sets are convex and closed, so P has extreme points.Convexity follows because mixtures of admissible priors remain admissible under the inequality.
  • The Principle of Reciprocity identifies posteriors obtained from maximal ignorance with valid priors, and in quantum mechanics these posteriors can represent any quantum state.This connects counterfactual measurement updating to the proposed state space.

A. Basis Distributions

When the sky and ground measurements coincide, the formalism constrains the associated stochastic matrix and, under a spanning assumption, yields distinguished basis distributions. In quantum mechanics, these correspond to SIC states.

  • Setting the ground measurement equal to the sky measurement requires the ground probabilities q(j) to equal the sky probabilities p(j).This condition constrains the special conditional matrix rS(j|i).
  • The authors introduce a matrix M to rewrite the same-measurement constraint in vector form.The resulting relation is used to analyze the allowed priors.
  • Assumption 3 requires the elements of P to span the full probability simplex Δd2.The paper motivates this by saying a smaller simplex could otherwise replace the full one.
  • The Principle of Reciprocity yields at least d2 additional distributions in P besides the uniform distribution.These distributions arise from the relevant inverse-matrix construction.
  • The d2 special distributions are called basis distributions and, in quantum mechanics, are identified with SIC states.They inherit the corresponding SIC-state structure.

B. A Bloch Sphere

The paper uses in-step unpredictability and reciprocity assumptions to derive spherical geometry for extreme points of the state space. The resulting lower-dimensional sphere generalizes the Bloch sphere, while further constraints prevent it from being the full sphere.

  • B. A Bloch Sphere: In-step unpredictability means that a uniform sky distribution implies a uniform ground distribution, expressing the absence of built-in measurement bias.The property links complete ignorance about the reference measurement to complete ignorance about the intended measurement.
  • B. A Bloch Sphere: For ISU measurements with basis distributions as posteriors, the bound requires m ≥ d and is tight for basis states only when m = d.This motivates a stronger assumption about extreme points.
  • B. A Bloch Sphere: Assumption 4 makes every extreme point arise from an ISU measurement with m = d and achieve equality in the relevant bound.The assumption treats basis distributions as prototypes of maximally predictive states.
  • B. A Bloch Sphere: The extreme points of P lie on a sphere, and the probability-simplex constraint places them on a smaller-radius lower-dimensional sphere.The construction yields a sphere analogous to the Bloch sphere.
  • B. A Bloch Sphere: When d = 2, the resulting 2-sphere is isomorphic to the usual Bloch sphere.The paper presents this as the higher-dimensional analogue of the Bloch sphere in the Quantum-Bayesian framework.
  • B. A Bloch Sphere: The resulting constraint involves pairs of distributions and arises from coherence between factual and counterfactual gambles, rather than from a single information function.The authors distinguish this from information-restriction approaches such as the Spekkens toy model.

C. But Only Part of It

The valid extreme states occupy only part of the sphere suggested by the fundamental inequality, and the sphere can extend beyond the probability simplex. The qubit is the only dimension in which the sphere lies completely within the simplex.

  • C. But Only Part of It: The state-space implied by Eq. (92) does not fill the full sphere in Eq. (97), because sufficiently distant pairs cannot both be valid.The extreme points comprise only part of the sphere.
  • C. But Only Part of It: The sphere's radius implies that it can poke outside the probability simplex, so the valid state space requires further trimming.
  • C. But Only Part of It: Only for the qubit, d = 2, does the sphere reside completely within the probability simplex, yielding the Bloch sphere.
  • C. But Only Part of It: Valid distributions have an upper bound on the number of zero components: if n > 1/2 d(d −1), no state can lie on that n-flat.
  • C. But Only Part of It: The zero-component bound matches the best known bound derived using the conventional quantum formalism and is saturated for d = 2 and d = 3.

D. An Underlying ‘Dimensionality’?

The inequality limits how many valid states can be mutually maximally distant: at most d states can attain this relation. This suggests a smaller underlying dimensionality, but does not yet establish the quantum manifold's dimension.

  • D. An Underlying ‘Dimensionality’?: Valid distributions cannot be truly orthogonal, and the maximum number of mutually maximally distant states is n = d.
  • D. An Underlying ‘Dimensionality’?: The Gram-matrix analysis makes n = d maximal because positive semidefiniteness requires n ≤ d; at n = d, only d −1 centered vectors are linearly independent.
  • D. An Underlying ‘Dimensionality’?: The d translated state vectors remain linearly independent, despite the centered vectors having rank d −1.
  • D. An Underlying ‘Dimensionality’?: This produces a significantly smaller apparent dimension than the nearly d^2 points one might initially expect, but the dimension remains only in quotation marks.

E. Summary of the Argument So Far

Under the Urgleichung and four additional assumptions, the authors derive several geometric constraints on valid probability states. These results suggest quantum-state-space structure but do not yet derive full Hilbert-space quantum mechanics.

  • E. Summary of the Argument So Far: The Urgleichung relates sky priors and ground conditional probabilities, while requiring valid ground probabilities 0 ≤ q(j) ≤ 1 yields the fundamental inequality.
  • E. Summary of the Argument So Far: The assumptions include maximal ignorance, reciprocity between posteriors and priors, simplex-spanning states, and extreme-point preparations.
  • E. Summary of the Argument So Far: The derivation places basis distributions among the valid states and bounds every valid component by p(k) ≤ 1.
  • E. Summary of the Argument So Far: The extreme valid distributions lie on a sphere that may extend beyond the simplex, obey a zero-component bound, and contain no more than d mutually maximally distant points.
  • E. Summary of the Argument So Far: These results hint at an isomorphism with quantum-state space, but deriving the required algebraic variety and full quantum mechanics remains unresolved.

VI. RELAXING THE CONSTANTS AND REGAINING THEM

The paper generalizes the Urgleichung by treating its constants as initially arbitrary, then asks which probability-based assumptions recover the quantum values. A special certainty measurement and a universal angle lead back to the original constants without invoking Hilbert-space machinery.

  • VI. RELAXING THE CONSTANTS AND REGAINING THEM: The section asks how additional assumptions can naturally recover α = d + 1 and β = 1 from a generalized Urgleichung.
  • VI. RELAXING THE CONSTANTS AND REGAINING THEM: The generalized setup uses n sky outcomes, arbitrary α and β, and a rule requiring ground probabilities q(j) to remain between zero and one.
  • VI. RELAXING THE CONSTANTS AND REGAINING THEM: Reciprocity and simplex-spanning basis states are retained, while an in-step-unpredictable measurement supplies priors associated with its ground outcomes.
  • VI. RELAXING THE CONSTANTS AND REGAINING THEM: Availability of certainty requires a measurement whose special priors make the agent certain of the corresponding ground outcome and include a basis distribution.
  • VI. RELAXING THE CONSTANTS AND REGAINING THEM: The special priors form equal-angle pairs, and the universal-angle assumption sets cos θ = 1/2 across systems.
  • VI. RELAXING THE CONSTANTS AND REGAINING THEM: The generalized family includes classical probability at q = 0 and real-Hilbert-space relations at q = 1; choosing q = 2 recovers the quantum case.
  • VI. RELAXING THE CONSTANTS AND REGAINING THEM: Eliminating the special-measurement parameters yields the original Urgleichung constants, using only personalist probabilities rather than amplitudes or Hilbert space.

VII. SUMMARY: FROM QUANTUM INTERFERENCE TO QUANTUM BAYESIAN COHERENCE

The paper reframes quantum interference and the Born Rule as additions to Bayesian coherence, expressed through probabilities rather than objective quantum states. It also identifies unresolved questions about deriving the full quantum-state structure and explaining the rule’s origin.

  • Quantum interference is presented as an empirical addition to Dutch-book coherence for probabilities concerning factualizable experiments and an explicitly assumed counterfactual.The authors emphasize that this account does not use probability amplitudes.
  • The Born Rule is interpreted as a relation between probabilities, not as a procedure for assigning probabilities from an objective quantum state.In QBism, quantum states depend on agents’ priors and may diverge even when agents assess the same data.
  • The urgleichung appears to specify a significant fraction of quantum-state structure, but the paper does not recover the full manifold of pure quantum states.The authors therefore suggest it might be treated as a fundamental axiom while leaving the derivation incomplete.
  • The proposed interpretation that the Born Rule represents a cost removed when factualizing a SIC remains speculative.
  • The firm result is a new way to quantify the claim that unperformed measurements have no outcomes.

VIII. OUTLOOK

The outlook situates QBism as a theory for agents interacting with a quantum world rather than a direct theory of the world itself. It argues that quantum theory’s formal rules reflect the character of that world while leaving its deeper ontology and derivation open.

  • The paper presents a deeper understanding of “unperformed measurements have no outcomes” as a first step toward characterizing the universe’s composition.
  • The paper argues that the term “measurement” should be banished from fundamental discussions because it suggests a misleading subject matter for quantum mechanics.The proposed replacement begins from agents’ actions and gambles within the world.
  • QBism treats quantum theory as a theory for agents immersed in and interacting with a quantum world, not as a theory of the world itself.This parallels the paper’s claim that probability theory is not itself a theory of the world.
  • Quantum theory is described as conditioned by the character of the world without directly representing that world.The authors identify confusion over this distinction as a source of longstanding discomfort in quantum foundations.

A. The Paulian Idea and the Jamesian Pluriverse

The Paulian Idea recasts quantum mechanics around agents, external systems, actions, and personal consequences. Within this framework, quantum measurements generate experiences, while the theory organizes Bayesian probabilities for possible actions and their consequences.

  • The Paulian Idea and the Jamesian Pluriverse: Personalist Bayesianism is characterized as a single-user theory in which probability expresses an individual’s relation to the world.
  • The Paulian Idea and the Jamesian Pluriverse: The Paulian Idea extends this single-user character to quantum outcomes, treating measurement outcomes as personal experiences for the agent.
  • The Paulian Idea and the Jamesian Pluriverse: The formal framework takes agents, external systems, actions, and action-consequences as primitive notions.
  • The Paulian Idea and the Jamesian Pluriverse: Quantum mechanics organizes an agent’s Bayesian probabilities for the consequences of all potential actions on surrounding systems.Systems are represented by Hilbert spaces, while actions are represented by POVMs.
  • The Paulian Idea and the Jamesian Pluriverse: For multiple systems, actions can target the joint Hilbert space or one subsystem, and resolving one consequence may update expectations about another action’s consequences.Those latter consequences require an actual action on the second system.
  • The Paulian Idea and the Jamesian Pluriverse: A quantum measurement is an action on a system whose consequence occurs within the agent’s experience, rather than an act of finding a pre-existing outcome.The paper summarizes this distinction by saying that measurement “finds nothing, but very much makes something.”
  • The Paulian Idea and the Jamesian Pluriverse: The urgleichung gives this action-centered theory empirical content by reflecting a world in which consequences are generated rather than waiting to fulfill actions.
Loading 1301.3274v1…