Quantum Darwinism explains objective classical reality through redundant environmental records: decoherence imprints pointer-state information on many disjoint environment fragments, so observers sampling different fragments agree about the system. Touil, Yan, and Zurek (2025) made this agreement quantitative, proving that for branching states decohered by the environment remainder, fragments with sufficient information about the system inevitably agree on its pointer state, and proposing the refined mutual information I(F:F′∣S)=I(F:F′)−I(F:F′∣S)\mathfrak{I}(\mathcal{F}:\mathcal{F}'|\mathcal{S})=I(\mathcal{F}:\mathcal{F}')-I(\mathcal{F}:\mathcal{F}'|\mathcal{S}) and its normalization C=I/HS\mathfrak{C}=\mathfrak{I}/H_{\mathcal{S}} as consensus measures, with nonnegativity proved only for branching states. They explicitly left open the detailed analysis when environment subsystems are pre-correlated or interact so that the global state is no longer branching, or when decoherence by the environment remainder is ineffective — regimes where I\mathfrak{I} can become negative. The open problem is a complete characterization of these consensus measures and their successors for such generic environments, an operational interpretation of their negative values, and branching-free analogues or replacements of the consensus theorems.

Background

Decoherence theory explains why quantum superpositions of certain preferred system states are unstable: interaction with an environment continuously monitors the system, suppressing superpositions between the stable "pointer states" selected by this monitoring (a process called einselection). Quantum Darwinism adds that the information about the pointer state is deposited redundantly across many disjoint subsystems of the environment, called fragments, so that many observers can each learn the state of the system by intercepting their own fragment without disturbing it. The standard diagnostic is the mutual information I(S:F)=HS+HF−HSFI(\mathcal{S}:\mathcal{F})=H_{\mathcal{S}}+H_{\mathcal{F}}-H_{\mathcal{S}\mathcal{F}} (with HH the von Neumann entropy) plotted against fragment size: a "classical plateau" at I(S:F)≃HSI(\mathcal{S}:\mathcal{F})\simeq H_{\mathcal{S}} for small fragments signals redundancy, quantified by the number RδR_\delta of fragments that each carry a (1−δ)(1-\delta) share of the information. A closely related structural formulation, the spectrum broadcast structure, requires the environment fragments conditioned on a pointer value to be uncorrelated product states.

For globally pure states produced by decoherence, the joint system-environment state takes the branching form ∣ΨSE⟩=∑nqn∣sn⟩∣Fn⟩∣Fn′⟩∣E/FFn′⟩|\Psi_{\mathcal{S}\mathcal{E}}\rangle=\sum_n\sqrt{q_n}|s_n\rangle|\mathcal{F}_n\rangle|\mathcal{F}'_n\rangle|\mathcal{E}_{/\mathcal{F}\mathcal{F}'_n}\rangle, where ∣sn⟩|s_n\rangle are orthogonal pointer states and different branches carry distinguishable environment records. Touil, Yan, and Zurek asked the consensus question directly: do two observers accessing disjoint fragments F\mathcal{F} and F′\mathcal{F}' agree about the state of S\mathcal{S}? Their Theorem 1 shows that for completely decohered branching states (the environment remainder removes inter-branch coherences, so HSF=HSF′=HSFF′=HSH_{\mathcal{S}\mathcal{F}}=H_{\mathcal{S}\mathcal{F}'}=H_{\mathcal{S}\mathcal{F}\mathcal{F}'}=H_{\mathcal{S}}), if each fragment's information about the system has deficit δ\delta and δ′\delta', then I(F:F′)=(1−δ−δ′+δ~)HSI(\mathcal{F}:\mathcal{F}')=(1-\delta-\delta'+\tilde\delta)H_{\mathcal{S}} with 0≤δ~≤min⁡(δ,δ′)0\leq\tilde\delta\leq\min(\delta,\delta'): fragments that know enough inevitably agree ("those who know enough about the system agree about what they know").

Raw mutual information I(F:F′)I(\mathcal{F}:\mathcal{F}'), however, can include spurious correlations the fragments share through direct interaction rather than through the system. To filter these out, the authors proposed the refined mutual information, I(F:F′∣S)=I(F:F′)−I(F:F′∣S)\mathfrak{I}(\mathcal{F}:\mathcal{F}'|\mathcal{S})=I(\mathcal{F}:\mathcal{F}')-I(\mathcal{F}:\mathcal{F}'|\mathcal{S}) (the quantity known, up to sign conventions, as interaction information in classical information theory), and the consensus measure C(F:F′∣S)=I(F:F′∣S)/HS\mathfrak{C}(\mathcal{F}:\mathcal{F}'|\mathcal{S})=\mathfrak{I}(\mathcal{F}:\mathcal{F}'|\mathcal{S})/H_{\mathcal{S}}, together with observable-dependent versions I(F:F′∣SΛ)\mathfrak{I}(\mathcal{F}:\mathcal{F}'|\mathcal{S}_\Lambda) in which conditioning is on a measured observable Λ\Lambda of the system. For branching states, I(F:F′∣S)=0I(\mathcal{F}:\mathcal{F}'|\mathcal{S})=0 and I\mathfrak{I} coincides with I(F:F′)I(\mathcal{F}:\mathcal{F}'). Their Theorem 2 proves that I(F:F′∣SΛ)\mathfrak{I}(\mathcal{F}:\mathcal{F}'|\mathcal{S}_\Lambda) is nonnegative for branching states whenever one fragment and the environment remainder have orthogonal states across branches, independent of the measurement applied on S\mathcal{S}. Outside such conditions, however, I\mathfrak{I} can be negative — when conditioning on S\mathcal{S} increases the fragments' mutual information, either through classical correlations or, in the quantum case, because measuring S\mathcal{S} reveals that F\mathcal{F} and F′\mathcal{F}' are entangled.

These tools were demonstrated in an analytically solvable decoherence model (the c-maybe model of imperfect conditional gates) and in numerical simulations of an all-to-all spin environment, H=σzS⊗∑idiσzi+∑j≠kgjkσzj⊗σzkH=\sigma_z^{\mathcal{S}}\otimes\sum_i d_i\sigma_z^i+\sum_{j\neq k}g_{jk}\sigma_z^j\otimes\sigma_z^k with random couplings drawn with standard deviations satisfying Δd≫Δg\Delta_d\gg\Delta_g (system-environment couplings dominate intra-environment couplings), N=20N=20 environment qubits, and an initially uncorrelated pure product state — the interacting-environment setting previously used by Riedel, Zurek, and Zwolak to exhibit the rise and fall of redundancy. There, a consensus plateau coincides with the quantum Darwinism window; after the onset of relaxation, intra-environment interactions scramble the records, and the refined mutual information decays and can dip negative for fragments of size ∣F∣≥4|\mathcal{F}|\geq 4, partly because the environment remainder E/FF′\mathcal{E}_{/\mathcal{F}\mathcal{F}'} is then too small to decohere SFF′\mathcal{S}\mathcal{F}\mathcal{F}'.

The analysis nevertheless rests on the branching structure (or on the quantum Darwinism window of the hierarchy Δd≫Δg\Delta_d\gg\Delta_g). Known results already delineate the tension: intra-environment interactions destroy "perfect" quantum Darwinism (Duruisseau, Touil, and Deffner); branching is the only zero-discord state structure — quantum discord being the standard measure of the genuinely quantum part of the correlations between two subsystems — and small discord only forces states close to branching (Touil, Anza, Deffner, and Crutchfield); and redundancy and consensus are distinct notions that can diverge in states with nonlocally encoded information (Chisholm, Innocenti, and Palma). But no characterization exists for genuinely non-branching situations — environments whose subsystems possess pre-existing correlations, environments whose subsystems interact strongly, or regimes where decoherence by the environment remainder is ineffective — and the authors state that a detailed analysis of such cases is beyond the scope of their work.

Problem Statement

Consider a system S\mathcal{S} monitored by a multipartite environment E\mathcal{E}, with two disjoint fragments F,F′⊂E\mathcal{F},\mathcal{F}'\subset\mathcal{E}, and define the refined mutual information I(F:F′∣S)=I(F:F′)−I(F:F′∣S)\mathfrak{I}(\mathcal{F}:\mathcal{F}'|\mathcal{S})=I(\mathcal{F}:\mathcal{F}')-I(\mathcal{F}:\mathcal{F}'|\mathcal{S}) and the consensus measure C(F:F′∣S)=I(F:F′∣S)/HS\mathfrak{C}(\mathcal{F}:\mathcal{F}'|\mathcal{S})=\mathfrak{I}(\mathcal{F}:\mathcal{F}'|\mathcal{S})/H_{\mathcal{S}}, including the observable-dependent variants conditioned on a measurement Λ\Lambda on S\mathcal{S}, as in Touil-Yan-Zurek (arXiv:2503.14791). Existing guarantees for these measures — Theorem 1 (fragments with sufficient information agree on the pointer state) and Theorem 2 (nonnegativity of I\mathfrak{I}) — hold only for branching states, with branches decohered by the environment remainder. The problem is to extend the operational theory of inter-observer consensus to joint system-environment states without this structure, jointly addressing:

(a) Characterization. Determine the behavior of I\mathfrak{I} and C\mathfrak{C} (or of a justified successor measure with an explicitly stated definition) for the non-branching regimes the source leaves open: (i) environment subsystems with pre-existing correlations (prior classical correlations or entanglement between fragments, including correlations with systems other than S\mathcal{S}); (ii) dynamically interacting environment subsystems, beyond the coupling hierarchy Δd≫Δg\Delta_d\gg\Delta_g and beyond the scrambling window already simulated; and (iii) regimes where the environment remainder E/FF′\mathcal{E}_{/\mathcal{F}\mathcal{F}'} fails to decohere SFF′\mathcal{S}\mathcal{F}\mathcal{F}', including large-fragment and late-time limits. For each regime, establish which quantitative statements about consensus survive, and which fail, with proofs or explicit counterexamples.

(b) Operational interpretation. Give a well-defined operational meaning to the values of the consensus measures in these regimes — in particular to negative refined mutual information, which arises when conditioning on S\mathcal{S} increases the fragments' mutual information, both through classical correlations and through entanglement revealed by measuring S\mathcal{S}. State verifiable conditions under which I\mathfrak{I} and C\mathfrak{C} remain faithful indicators of agreement about pointer states outside branching states, and what claim about observer agreement, if any, a negative value licenses.

(c) Theorem analogues or replacements. Prove or refute analogues of the source's theorems for the non-branching classes above — e.g., whether sufficient fragment information about the system still forces agreement on some observable in the presence of pre-correlations, and under what structural conditions refined-mutual-information nonnegativity holds — or establish that no such guarantee is possible and provide the replacement guarantee a consistent consensus framework should use there.

The common objective uniting these deliverables is a single operational theory of inter-observer consensus in quantum Darwinism that is valid beyond the branching ansatz: which state classes support well-defined consensus, what the measures mean there, and where the branching-based picture breaks. A complete resolution therefore addresses all of (a), (b), and (c) jointly: the characterization (a) for each of the three regimes (i)-(iii), each supported by a proof or an explicit counterexample; the operational interpretation (b), including the meaning of negative values; and the theorem analogues or replacements (c), which may be discharged either by a positive analogue or by an impossibility result together with the replacement guarantee. Answers restricted to branching states, or to any single regime with the others explicitly left unaddressed, do not complete the problem.

References

  1. A. Touil, B. Yan, W. H. Zurek, "Consensus About Classical Reality in a Quantum Universe", arXiv:2503.14791 (2025), DOI 10.48550/arXiv.2503.14791, https://arxiv.org/abs/2503.14791
  2. A. Touil, B. Yan, D. Girolami, S. Deffner, W. H. Zurek, "Eavesdropping on the Decohering Environment: Quantum Darwinism, Amplification, and the Origin of Objective Classical Reality", Phys. Rev. Lett. 128, 010401 (2022), DOI 10.1103/PhysRevLett.128.010401, https://arxiv.org/abs/2107.00035
  3. C. J. Riedel, W. H. Zurek, M. Zwolak, "The rise and fall of redundancy in decoherence and quantum Darwinism", New J. Phys. 14, 083010 (2012), DOI 10.1088/1367-2630/14/8/083010, https://arxiv.org/abs/1205.3197
  4. A. Touil, F. Anza, S. Deffner, J. P. Crutchfield, "Branching States as The Emergent Structure of a Quantum Universe", Quantum 8, 1494 (2024), DOI 10.22331/q-2024-10-10-1494, https://arxiv.org/abs/2208.05497
  5. P. Duruisseau, A. Touil, S. Deffner, "Pointer States and Quantum Darwinism with Two-Body Interactions", Entropy 25, 1573 (2023), DOI 10.3390/e25121573, https://arxiv.org/abs/2309.03299
  6. D. A. Chisholm, L. Innocenti, G. M. Palma, "The meaning of redundancy and consensus in quantum objectivity", Quantum 7, 1074 (2023), DOI 10.22331/q-2023-08-03-1074, https://arxiv.org/abs/2211.09150
  7. D. A. Chisholm, L. Innocenti, G. M. Palma, "Importance of using the averaged mutual information when quantifying quantum objectivity", Phys. Rev. A 110, 012218 (2024), DOI 10.1103/PhysRevA.110.012218, https://doi.org/10.1103/PhysRevA.110.012218
  8. A. Kiely, D. A. Chisholm, A. Touil, S. Deffner, G. Landi, S. Campbell, "Metrological approach to the emergence of classical objectivity", Phys. Rev. A 113, 022403 (2026), DOI 10.1103/hn78-7xx3, https://arxiv.org/abs/2510.12313
  9. B. Ferté, D. Farci, X. Cao, "Decoherent Histories with(out) Objectivity in a (Broken) Apparatus", Phys. Rev. Lett. 136, 090404 (2026), DOI 10.1103/8qzx-xpfz, https://arxiv.org/abs/2508.16482
  10. J. Gu, "Hamiltonian thresholds for objective records", Phys. Rev. A 114, L020201 (2026), DOI 10.1103/fm9j-m9lh, https://arxiv.org/abs/2608.00365
  11. E. Doucet, S. Deffner, "From compatibility of measurements to exploring quantum Darwinism on NISQ", Quantum Sci. Technol. 11, 035008 (2026), DOI 10.1088/2058-9565/ae753f, https://arxiv.org/abs/2601.05350
  12. S.-I. Tang, E. Doucet, A. Touil, S. Deffner, A. Sone, "Information processing in quantum thermodynamic systems: An autonomous Hamiltonian approach", Phys. Rev. E 114, 024144 (2026), DOI 10.1103/rwkz-dc5z, https://arxiv.org/abs/2511.08858
  13. H. Arora, B. K. Das, B. Suri, V. Madhok, "Quantum-to-classical transition and the emergence of trajectory level Darwinism with measurements distributed in time: a path integral approach", arXiv:2505.16889 (2025), https://arxiv.org/abs/2505.16889
  14. Z. Zhu, K. Salice, A. Touil, Z. Bao, Z. Song, P. Zhang, H. Li, Z. Wang, C. Song, Q. Guo, H. Wang, R. Mondaini, "Observation of quantum Darwinism and the origin of classicality with superconducting circuits", Science Advances 11, eadx6857 (2025), DOI 10.1126/sciadv.adx6857, https://arxiv.org/abs/2504.00781
  15. J. Torvinen, E. Keski-Vakkuri, N. Pranzini, "Quantum Darwinism and the quality of Petz recovery", arXiv:2605.06848 (2026), https://arxiv.org/abs/2605.06848
  16. A. Maity, K. Onggadinata, T. S. Koh, "Exact Tradeoff Between Quantum Error Correction and Quantum Darwinism: An Information-Theoretic No-Go Theorem", arXiv:2608.03944 (2026), https://arxiv.org/abs/2608.03944
  17. A. Batin Tank, "Functional Information in Quantum Darwinism: An Operational Measure of Objectivity", arXiv:2509.17775 (2025), https://arxiv.org/abs/2509.17775