Preexisting bath correlations, non-Markovian information backflow, and the capacity of many-body environments to host redundant classical records

When a qubit is dephasingly coupled to a spin bath prepared in the ground states of the transverse-field spin-1/2 XX chain, quantum Darwinism — the redundant proliferation of classical information about the system's pointer observable throughout the environment — is perfect if the bath is uncorrelated (transverse field $h$ above the chain's critical value $h_C$) and progressively degrades as the bath ground state carries more multipartite entanglement ($h<h_C$), with the degraded record plateaus correlating with non-Markovian information backflow. Whether this reflects a generic causal relation between preexisting intra-environment correlations, memory effects, and the inhibition of redundant record formation — and whether it survives finite temperature, weaker system-bath coupling, other bath Hamiltonians, and the thermodynamic limit — was left open, together with a quantitative characterization relating initial-state entanglement measures to record-hosting capacity and the question of whether the observed plateaus correspond to spectrum broadcast structures. Later literature has sharpened rather than settled the issue: several models support the correlation, while decoupling constructions show that the same non-Markovian reduced dynamics can occur with or without Darwinism, and only one-directional bounds (Darwinism implies approximately Markovian dynamics) are known. The general characterization remains open.

Background

Quantum Darwinism (QD) is a framework for the emergence of objective classical facts about a quantum system from decoherence: the environment selectively records information about the system's pointer states (the states stable under the system-environment interaction that decoherence singles out), so that many independent observers, each accessing a different small fragment $\mathcal{F}$ of the environment, can all learn the same classical property. The operational signature is a plateau of the quantum mutual information $\mathcal{I}(S:\mathcal{F})$ close to the system entropy $H_S$ over a wide range of fragment sizes — redundant storage of classical information. A stronger state-level formulation of objectivity is the spectrum broadcast structure (SBS) of Horodecki, Korbicz, and Horodecki: the global state approximately decomposes into a mixture in which each environmental fragment carries an orthogonal record of the system's pointer value. Open-system dynamics is called non-Markovian when the distinguishability of pairs of system states temporarily increases, i.e. information flows back from the environment to the system; the Breuer-Laine-Piilo (BLP) measure quantifies this via the trace distance, and for pure dephasing it reduces to properties of the Loschmidt echo (the overlap between bath states evolved under the two system-pointer-conditioned Hamiltonians, whose decay controls the system's dephasing).

Prior QD analyses (quantum Brownian motion, photon and spin environments, noisy channels) largely assumed an environment initially in a product (uncorrelated) state of its units. Giorgi, Galve, and Zambrini [Phys. Rev. A 92, 022105 (2015)] asked what changes when the environment is a many-body system in a correlated quantum phase. They studied a qubit $S$ coupled by a pure-dephasing spin-star interaction (each bath spin coupled individually to the central system, forming a star geometry) $H_{S\mathcal{E}}=d\,\sigma_S^z\otimes\sum_{i=1}^N\sigma_i^x$ to an $N$-spin bath initialized in ground states of the transverse-field isotropic XX chain $H_B=-\sum_i(\sigma_i^+\sigma_{i+1}^-+\sigma_i^-\sigma_{i+1}^+)-h\sum_i\sigma_i^z$ with periodic boundary conditions. For $h\geq h_C=1$ the ground state is the uncorrelated zero-excitation ferromagnetic state, and at time $d\,t=\pi/4$ the joint state factorizes exactly into a superposition of two product bath states correlated with the two pointer states of $S$, so every single bath spin carries the full classical record: $\mathcal{I}(S:\mathcal{F})=H_S$ for any nonzero fragment (perfect Darwinism, proved analytically via an exact permutation-symmetric $(N+1)$-level reduction). For $h<h_C$ the ground states carry genuine multipartite entanglement (entanglement shared genuinely among all bath spins, as characterized for the XY family by Giampaolo and Hiesmayr), the plateau degrades monotonically as $h$ decreases from $h_C$ — a fraction $\sim N/2$ of the bath must be accessed to recover the information — except for a partial resurgence near $h\approx 0$ caused by a drop of the system entropy; adding the bath self-Hamiltonian ($H=H_{S\mathcal{E}}+\lambda H_B$) mixes and delocalizes the records, spoiling the plateau. The BLP non-Markovianity $\mathcal{N}$ is zero for $h\geq h_C$ and grows monotonically as $h$ decreases, so larger non-Markovianity correlates with poorer Darwinism; just below the critical point, $\mathcal{N}(h_C^-)$ is constant in $N$ for $N\gtrsim 8$, acting as a finite-size precursor of the Kosterlitz-Thouless transition (the symmetry-breaking-free topological phase transition of this chain's universality class).

The authors left the interpretation as a conjecture: that the initial multipartite entanglement of the bath (raising the initial entropy of fragments and misaligning them relative to the interaction eigenbasis) is what inhibits record formation, with non-Markovianity as the dynamical signature. The study was restricted to zero-temperature ground states, the strong-coupling regime (evolution under $H_{S\mathcal{E}}$ alone, evaluated at the single time $d\,t=\pi/4$ chosen as optimal for the uncorrelated ground state), Darwinism curves at a single bath size $N=14$, and no spectrum-broadcast analysis; no quantitative relation between initial-state entanglement and record capacity was established.

Problem Statement

Establish, in general, how preexisting correlations among the units of a many-body environment affect that environment's capacity to redundantly store classical information about a decohering quantum system, and delimit the role of non-Markovian information backflow in this picture. The anchor case is the XX-chain spin bath studied by Giorgi, Galve, and Zambrini, but the question is not tied to that model. Specifically:

(a) Is the degradation of redundant record formation (a mutual-information plateau $\mathcal{I}(S:\mathcal{F})\simeq H_S$ at small fragment sizes, or an approximate spectrum broadcast structure) by initial intra-environment entanglement and correlations generic — across bath Hamiltonians, finite-temperature and mixed initial states, weaker system-bath coupling, and larger baths up to the thermodynamic limit — or do broad, characterizable classes of correlated initial states still support full objectivity? Here 'generic' is read as: the degradation holds for all initial bath states in the stated regimes except for an exception class that is 'broad, characterizable', meaning it is specified by explicit conditions on the bath state rather than by fine-tuning individual preparations. What is the fate of the specific XX-bath phenomenology (perfect records for the uncorrelated ground state, monotonic degradation with decreasing field, partial resurgence near $h\approx 0$, plateau spoilage by bath self-dynamics) under these extensions? A general statement covering these regimes suffices to resolve this part, with worked XX-bath checks optional.

(b) Determine the causal status of non-Markovian information backflow in this picture: whether backflow causally inhibits redundant record formation, or is an incidental correlate of record degradation. The current evidence is split and class-restricted: model-level studies support the correlation, while decoupling constructions within a class of pure dephasing models show that the same non-Markovian reduced dynamics can occur with or without Darwinism, and only the one-directional bound 'Darwinism implies approximately Markovian dynamics' is known. Open is whether general statements (theorems, bounds, or a classification) establish a causal role of backflow, or whether decisive counterexamples rule out any general causal relation.

(c) Provide a quantitative characterization relating multipartite entanglement (or other correlation) measures of the initial bath state to its record-hosting capacity — the degree to which the environment reproduces the plateau signature $\mathcal{I}(S:\mathcal{F})\simeq H_S$ at small fragment sizes, quantified for instance by the plateau height or by the fraction of the bath that must be accessed to recover $H_S$ — in which the trivial mechanism of initial local fragment entropy is separated from genuine multipartite correlations. The XX-model findings and the known many-body-localization-regime link (many-body localization being a disorder-induced phase in which an interacting system fails to thermalize) between entanglement entropy and lack of redundancy indicate what such a characterization should account for.

(d) For correlated bath initial states, determine when a mutual-information plateau reflects genuine spectrum broadcast structure, given that the redundancy and consensus notions of objectivity can provably come apart. A full resolution of this problem requires all four parts, (a) through (d).

References

1. G. L. Giorgi, F. Galve, and R. Zambrini, 'Quantum Darwinism and non-Markovian dissipative dynamics from quantum phases of the spin-1/2 XX model', Physical Review A 92, 022105 (2015). DOI: 10.1103/PhysRevA.92.022105, https://doi.org/10.1103/PhysRevA.92.022105 (arXiv:1504.02105, https://arxiv.org/abs/1504.02105).
2. W. H. Zurek, 'Quantum Darwinism', Nature Physics 5, 181-188 (2009). DOI: 10.1038/nphys1202, https://doi.org/10.1038/nphys1202
3. R. Blume-Kohout and W. H. Zurek, 'Quantum Darwinism in quantum Brownian motion', Physical Review Letters 101, 240405 (2008). DOI: 10.1103/PhysRevLett.101.240405, https://doi.org/10.1103/PhysRevLett.101.240405
4. M. Zwolak, H. T. Quan, and W. H. Zurek, 'Redundant imprinting of information in nonideal environments: Objective reality via a noisy channel', Physical Review A 81, 062110 (2010). DOI: 10.1103/PhysRevA.81.062110, https://doi.org/10.1103/PhysRevA.81.062110
5. R. Horodecki, J. K. Korbicz, and P. Horodecki, 'Quantum origins of objectivity', Physical Review A 91, 032122 (2015). DOI: 10.1103/PhysRevA.91.032122, https://doi.org/10.1103/PhysRevA.91.032122
6. H.-P. Breuer, E.-M. Laine, and J. Piilo, 'Measure for the degree of non-Markovian behavior of quantum processes in open systems', Physical Review Letters 103, 210401 (2009). DOI: 10.1103/PhysRevLett.103.210401, https://doi.org/10.1103/PhysRevLett.103.210401
7. S. M. Giampaolo and B. C. Hiesmayr, 'Genuine multipartite entanglement in the XY model', Physical Review A 88, 052305 (2013). DOI: 10.1103/PhysRevA.88.052305, https://doi.org/10.1103/PhysRevA.88.052305
8. F. Galve, R. Zambrini, and S. Maniscalco, 'Non-Markovianity hinders Quantum Darwinism', Scientific Reports 6, 19607 (2016). DOI: 10.1038/srep19607, https://doi.org/10.1038/srep19607
9. G. Pleasance and B. M. Garraway, 'Application of quantum Darwinism to a structured environment', Physical Review A 96, 062105 (2017). DOI: 10.1103/PhysRevA.96.062105, https://doi.org/10.1103/PhysRevA.96.062105
10. A. Lampo, J. Tuziemski, M. Lewenstein, and J. K. Korbicz, 'Objectivity in the non-Markovian spin-boson model', Physical Review A 96, 012120 (2017). DOI: 10.1103/PhysRevA.96.012120, https://doi.org/10.1103/PhysRevA.96.012120
11. T. P. Le and A. Olaya-Castro, 'Objectivity (or lack thereof): Comparison between predictions of quantum Darwinism and spectrum broadcast structure', Physical Review A 98, 032103 (2018). DOI: 10.1103/PhysRevA.98.032103, https://doi.org/10.1103/PhysRevA.98.032103
12. N. Milazzo, S. Lorenzo, M. Paternostro, and G. Massimo Palma, 'Role of information backflow in the emergence of quantum Darwinism', Physical Review A 100, 012101 (2019). DOI: 10.1103/PhysRevA.100.012101, https://doi.org/10.1103/PhysRevA.100.012101 (arXiv:1901.05826, https://arxiv.org/abs/1901.05826)
13. K. Roszak and J. K. Korbicz, 'Entanglement and objectivity in pure dephasing models', Physical Review A 100, 062127 (2019). DOI: 10.1103/PhysRevA.100.062127, https://doi.org/10.1103/PhysRevA.100.062127
14. G. García-Pérez, D. A. Chisholm, M. A. C. Rossi, G. Massimo Palma, and S. Maniscalco, 'Decoherence without entanglement and quantum Darwinism', Physical Review Research 2, 012061 (2020). DOI: 10.1103/PhysRevResearch.2.012061, https://doi.org/10.1103/PhysRevResearch.2.012061
15. E. Ryan, M. Paternostro, and S. Campbell, 'Quantum Darwinism in a structured spin environment', Physics Letters A 416, 127675 (2021). DOI: 10.1016/j.physleta.2021.127675, https://doi.org/10.1016/j.physleta.2021.127675
16. T. P. Le, A. Winter, and G. Adesso, 'Thermality versus Objectivity: Can They Peacefully Coexist?', Entropy 23, 1506 (2021). DOI: 10.3390/e23111506, https://doi.org/10.3390/e23111506
17. N. Mirkin and D. Wisniacki, 'Many-Body Localization and the Emergence of Quantum Darwinism', Entropy 23, 1377 (2021). DOI: 10.3390/e23111377, https://doi.org/10.3390/e23111377
18. J. K. Korbicz, 'Roads to objectivity: Quantum Darwinism, Spectrum Broadcast Structures, and Strong Quantum Darwinism - a review', Quantum 5, 571 (2021). DOI: 10.22331/q-2021-11-08-571, https://doi.org/10.22331/q-2021-11-08-571
19. W. S. Martins and D. O. Soares-Pinto, 'Suppressing information storage in a structured thermal bath: Objectivity and non-Markovianity', preprint (2021). arXiv:2110.03490, https://arxiv.org/abs/2110.03490; also deposited as a master's dissertation, University of São Paulo (2022). DOI: 10.11606/D.76.2022.tde-19072022-135443, https://doi.org/10.11606/D.76.2022.tde-19072022-135443
20. N. Megier, A. Smirne, S. Campbell, and B. Vacchini, 'Correlations, Information Backflow, and Objectivity in a Class of Pure Dephasing Models', Entropy 24, 304 (2022). DOI: 10.3390/e24020304, https://doi.org/10.3390/e24020304 (arXiv:2201.10573, https://arxiv.org/abs/2201.10573)
21. D. A. Chisholm, L. Innocenti, and G. Massimo Palma, 'The meaning of redundancy and consensus in quantum objectivity', Quantum 7, 1074 (2023). DOI: 10.22331/q-2023-08-03-1074, https://doi.org/10.22331/q-2023-08-03-1074
22. X.-K. Guo and Z. Huang, 'On the relation between quantum Darwinism and approximate quantum Markovianity', Physics Letters A 491, 129204 (2023). DOI: 10.1016/j.physleta.2023.129204, https://doi.org/10.1016/j.physleta.2023.129204
23. P. Duruisseau, A. Touil, and S. Deffner, 'Pointer States and Quantum Darwinism with Two-Body Interactions', Entropy 25, 1573 (2023). DOI: 10.3390/e25121573, https://doi.org/10.3390/e25121573
24. I. Valladares Duque and P. Strasberg, 'Toy model challenging prevailing definitions of classicality', Europhysics Letters 152, 28001 (2025). DOI: 10.1209/0295-5075/ae1639, https://doi.org/10.1209/0295-5075/ae1639 (arXiv:2412.08444, https://arxiv.org/abs/2412.08444)
25. Z. Zhu, K. Salice, A. Touil, Z. Bao, Z. Song, P. Zhang, H. Li, Z. Wang, C. Song, Q. Guo, H. Wang, and 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://doi.org/10.1126/sciadv.adx6857