{"id":"d28305d6-1c37-41ce-8b4a-17b41a8ddbea","arxiv_id":"1908.08818","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"Comparing a system-environment evolution with and without an objectivity-enforcing projection yields a lower-bound witness for non-objectivity in strong quantum Darwinism.","lead":"Quantum systems become 'objective' when many observers can learn the same information about them from different parts of the environment. This paper proposes a witness that detects when such objectivity fails, using two runs of a photonic experiment instead of full state tomography.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Core bound holds, but the trace-nonincreasing objectivity operation leaves a normalization ambiguity in PΓ and invalidates the claimed maximum of M.","rationale":"The central mathematical result, W_SQD ≤ M_SQD, is a correct direct application of the trace-norm dual formulation, and a nonzero witness does imply non-objectivity relative to the fixed basis and subspaces. The basis-dependence limitation is explicitly acknowledged in Sec. III and is not a correctness flaw. The real soft spot is the trace-nonincreasing character of Γ: Eq. (10) is a selective projection that can reduce the trace, yet the paper neither states this nor discusses whether PΓ is the joint success-and-outcome probability or the conditional postselected probability. Under the conditional reading, the claimed bound can fail, and the stated maximum of the measure is false. However, because the paper's equations consistently use the unnormalized Γ, the formal bound is valid, and the issue can be resolved by clarifying the probability convention and correcting the normalization claim. Thus the reader's CONDITIONAL verdict remains appropriate, with the condition being a revision that addresses this trace normalization point rather than a change in the scientific conclusion.","tokens_in":18222,"tokens_out":23284,"duration_ms":256169,"concrete_test":"Analytically recompute the witness for the counterexample ρ = |0⟩⟨0|_S ⊗ |+⟩⟨+|_F with Γ defined by the two projectors above, using both definitions of the Γ-branch probability: joint PΓ = tr[M U(Γρ⊗ρ_new)] and conditional PΓ_cond = PΓ / Tr(Γρ). Check whether W ≤ M holds in each case; the conditional case should violate it, and the joint case gives M = √5/2, falsifying the paper's 'maximum value of the measure is 1' claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The inequality W_SQD(M_SE) ≤ M_SQD(ρ_SF(t)) in Eqs. (18)-(23) is correct as written, because it uses the unnormalized (trace-nonincreasing) objectivity operation Γ from Eq. (10). The problem is that the paper never states this trace-nonincreasing character, and the experimental implementation in Sec. IV is a postselected parity check. If an experimenter follows the natural postselection procedure and defines the Γ-branch probability as a conditional probability, PΓ_cond = tr[M_SE U_τ(Γ(ρ_SF(t))⊗ρ_new)] / Tr[Γ(ρ_SF(t))], then the bound W ≤ M can fail. This is not merely academic: the paper also claims after Eq. (11) that max M_SQD = 1, which is false for trace-nonincreasing Γ. For the two-qubit SF with projectors |0⟩⟨0|_S⊗|0⟩⟨0|_F and |1⟩⟨1|_S⊗|1⟩⟨1|_F, take ρ = |0⟩⟨0|_S ⊗ |+⟩⟨+|_F. Then Γ(ρ) = (1/2)|00⟩⟨00| and ||ρ − Γ(ρ)||_1 = √5/2 ≈ 1.118 > 1. The witness inequality itself is unaffected under the joint-probability reading, so the central detection claim survives, but the measure's normalization and the operational meaning of PΓ need to be made explicit.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript introduces a subspace-dependent 'objectivity operation' Gamma_SQD and a trace-distance measure M_SQD of non-objectivity, and constructs a witness W_SQD obtained by comparing the evolution of the system-fragment state with and without Gamma_SQD. The central result is the inequality W_SQD <= M_SQD, proven in Eqs. (18)-(23) via trace-norm duality. The paper also proposes a photonic implementation based on CNOT gates and parity checks, reports numerical simulations (including imperfect gates), and gives a scaling argument showing a potential advantage over full tomography. An analogous witness is developed for invariant spectrum broadcast structure in Appendix B.","tokens_in":18473,"tokens_out":7837,"duration_ms":78938,"significance":"The theoretical bound is derived cleanly and the numerical simulations are consistent with the detection claim; the authors also honestly point out limitations (weakness for full-fragment quantum correlations, overshoot under gate errors, basis-dependence). The proposed witness is a potentially useful step toward scalable experimental tests of quantum Darwinism. However, a normalization/postselection issue in the definition of the measure and in the experimental protocol must be resolved before the experimental claim is sound.","major_comments":[{"comment":"The objectivity operation Gamma in Eq. (10) is trace-nonincreasing for generic choices of the objective subspaces, because the projectors |i><i|_S x Pi_{F|i} need not sum to the identity on SF. Consequently the claim that max_{rho} M_SQD(rho) = 1 is false. For example, with d_S = d_F = 2 and projectors |0><0|_S x |0><0|_F and |1><1|_S x |1><1|_F, the state rho = |0><0|_S x |+><+|_F gives Gamma(rho) = (1/2)|00><00| and ||rho - Gamma(rho)||_1 = sqrt(5)/2 = 1.118 > 1. This should be corrected by either normalizing Gamma so that it is trace-preserving (e.g., by adding a failure outcome) or by explicitly stating that M_SQD is not normalized and can exceed 1.","section":"Sec. III, Eq. (11)"},{"comment":"The witness is defined in terms of unconditional probabilities P1 and PGamma, with PGamma = tr[M_SE U_tau(Gamma(rho_SF(t)) x rho_new)]. The proposed photonic implementation, however, is postselected: only runs in which the parity checks match the system measurement are retained (Fig. 4). If an experimentalist instead estimates the conditional probability PGamma_cond = tr[M_SE U_tau(Gamma(rho) x rho_new)] / Tr[Gamma(rho)], the inequality W <= M can fail, because division by Tr[Gamma(rho)] can amplify the difference. The scaling analysis in Sec. IVD ('in order for there to be a total of C successful runs...') makes clear that the objective branch is treated as conditional. The authors must specify how the unconditional probabilities in Eqs. (15) and (17) are reconstructed from postselected data (e.g., by including failure counts as a null outcome or by multiplying by the success probability), or else provide a modified witness bound valid under postselection.","section":"Sec. IV, Eqs. (15)-(17) and Fig. 4"}],"minor_comments":[{"comment":"The sentence 'Fig. 7 shows the presumes a perfect circuit' should read 'Fig. 7 shows the results for a perfect circuit'.","section":"Sec. IVC"},{"comment":"The expression 'pCNOT & 1/3' should read 'p_CNOT >= 1/3' or 'p_CNOT >~ 1/3'; the ampersand is a typographical artifact.","section":"Sec. IVD"},{"comment":"Consider stating explicitly that Gamma is not trace-preserving in general and that M_SQD is therefore not a normalized measure; this would prevent a natural misreading of Fig. 7, where M_SQD is plotted against 1.","section":"Sec. III, after Eq. (11)"},{"comment":"The phrase 'If the objective operation results in a null state, then all measurement outcomes are zero' is only meaningful if the trace-nonincreasing nature of Gamma is understood; please clarify this sentence.","section":"Sec. IV, Fig. 4 caption"}],"recommendation":"major_revision","confidential_remarks":"The main obstacle to acceptance is the postselection issue in the experimental protocol; the theoretical inequality W <= M is sound as written for unnormalized Gamma. If the authors reconcile the photonic implementation with the unconditional-probability definition (or provide a postselection-valid witness), the paper would be a solid contribution. The self-citation to Ref. [3] is appropriate, and the paper's scope fits the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The one thing to know: this paper proposes a witness for non-objectivity in strong quantum Darwinism that could in principle scale beyond full state tomography in photonic simulators. The main inequality W ≤ M is elementary and correct under the joint-probability reading, but the paper never states that the objectivity operation is trace-nonincreasing, and that omission creates a real operational ambiguity.\n\nWhat is new: the subspace-dependent objectivity operation Γ, the two-run witness, and the point channel that removes the correlation ambiguity in earlier discord/coherence witnesses. The ISBS extension is a nice bonus. Credit where earned: the derivation in Eqs. (18)-(23) is clean trace-norm duality; the simulations are honest, including cases where the witness is weak (full-fragment correlations) and where gate errors make it overshoot the true measure. The authors also explicitly acknowledge the basis-dependence.\n\nSoft spots, in proportion: first, the normalization issue. Γ as defined in Eq. (10) is trace-nonincreasing, not a quantum channel. The claim after Eq. (11) that max M = 1 is wrong as stated; for ρ = |0><0|_S ⊗ |+><+|_F with the two-qubit projectors, M = √5/2 > 1. More importantly, the experimental implementation uses postselection. If PΓ is treated as a conditional probability after successful projection, the bound W ≤ M can fail. The fix is simple: define PΓ as the joint probability including the projection success rate, and say so. But right now the paper leaves it ambiguous.\n\nSecond, the scaling comparison is a back-of-the-envelope estimate. It does not include the cost of the point channel, state preparation, or errors in the parity checks beyond CNOT fidelity. That is fine for a proposal, but the \"outperforms tomography\" claim should be labeled as idealized.\n\nThird, no simulation code or exact input parameters beyond the described CNOT fidelity are shipped, so the figures cannot be independently reproduced from the text.\n\nThe central detection claim survives. This is a proposal paper, not a definitive experiment, and it is reasonably honest about its limitations. It deserves a serious referee. My recommendation: send it to review, and ask the authors to fix the normalization of Γ, clarify the probability convention in the witness, and soften the scaling claim.\n\nWho is it for: people working on experimental tests of quantum Darwinism, and quantum-foundation folks interested in correlation witnesses. I would be happy to cite it in the next year if the normalization patch lands.","headline":"A useful basis-dependent witness for strong quantum Darwinism with a clean core inequality, but a normalization ambiguity in the objectivity operation needs a patch before this is operational.","tokens_in":19009,"tokens_out":4205,"would_cite":true,"duration_ms":39210,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A single comparison of two evolutions lets a photonic experiment witness non-objectivity in strong quantum Darwinism without full state tomography.","keywords":["quantum Darwinism","strong quantum Darwinism","objectivity witness","non-objectivity","spectrum broadcast structure","photonic simulation","quantum state tomography","quantum-to-classical transition"],"falsifier":"A decisive check is to take many randomly generated system-fragment states, estimate $M_{\\mathrm{SQD}}$ by full state tomography, and run the maximal witness on the same states: the paper's bound predicts the witness never exceeds the measure, so any measured violation of $W_{\\mathrm{SQD}} \\le M_{\\mathrm{SQD}}$ would refute the central claim.","tokens_in":17972,"feed_emoji":"⚛️","tokens_out":13576,"duration_ms":123442,"temperature":0.7,"pith_summary":"This paper aims to make the quantum-to-classical transition experimentally testable without full quantum state tomography. It constructs a witness that compares the evolution of a system-environment state with and without an objectivity operation that projects the state into a pre-chosen objective subspace. A non-zero witness value proves the state is non-objective relative to that basis and subspace, and the witness is always a lower bound on a trace-distance measure of non-objectivity. The paper also proposes a photonic implementation and shows numerically that, with sufficiently reliable controlled-NOT gates, the number of measurement runs it needs grows far more slowly with environment size than tomography. A simplified variant for invariant spectrum broadcast structure needs even fewer resources.","feed_headline":"One witness spots non-objectivity by comparing two evolutions","feed_subtitle":"It replaces full state tomography with a single measurement, opening photonic tests to larger environments.","key_machinery":"The central object is the objectivity operation $\\Gamma^{\\mathrm{SQD}}_{SF}(\\rho) = \\sum_i (|i\\rangle\\langle i|_S \\otimes \\Pi_{F|i} \\otimes \\mathbb{1}_{E\\setminus F}) \\rho (|i\\rangle\\langle i|_S \\otimes \\Pi_{F|i} \\otimes \\mathbb{1}_{E\\setminus F})$, a completely positive map that sends any state into the convex set of states satisfying strong quantum Darwinism in the chosen system basis and environment subspaces. The witness is the absolute difference of two probabilities, one from the untouched evolution and one after applying this operation, with a point channel first deleting correlations with the unaccessed environment: $W_{\\mathrm{SQD}}(M_{SE}) = |\\mathrm{tr}[ M_{SE} U_\\tau ( \\rho_{SF}(t) - \\Gamma^{\\mathrm{SQD}}_{SF}(\\rho_{SF}(t)) ) \\otimes \\rho^{\\mathrm{new}}_{E\\setminus F} ]|$. The proof of the bound uses the trace-norm variational formula, expressing the supremum over Hermitian operators with norm at most one as exactly $\\lVert \\rho_{SF}(t) - \\Gamma^{\\mathrm{SQD}}_{SF}(\\rho_{SF}(t)) \\rVert_1$; this is what turns the probability difference into a lower bound on the measure.","core_discovery":"The central claim is that non-objectivity of a system $S$ with respect to an environment fragment $F$, in the framework of strong quantum Darwinism, can be witnessed by a two-branch experiment. In one branch the state $\\rho_{SF}(t)$ is left untouched; in the other it is first acted on by a subspace-dependent objectivity operation that dephases the system in a fixed basis and projects each environment subspace into the corresponding objective subspace. After identical unitary evolution and a fixed final measurement, the absolute difference of the two probabilities is the witness $W_{\\mathrm{SQD}}$. The paper proves $W_{\\mathrm{SQD}}(M_{SE}) \\le M_{\\mathrm{SQD}}(\\rho_{SF}(t))$, where $M_{\\mathrm{SQD}}$ is the trace-norm distance from the objective subspace, so any non-zero witness implies non-objectivity relative to the fixed basis and subspaces. The converse is not claimed: a zero reading does not certify full objectivity, because a state objective in another basis can still give a non-zero witness. The scheme is applied to a photonic qubit simulation using parity checks and CNOT gates, and numerical results show it detects non-objectivity under depolarising noise and imperfect gates.","pith_inferences":["The two-branch comparison is a general template: any non-objectivity or correlation measure defined as a distance to a convex set could be lower-bounded by the same probability-difference construction, so the scheme may transfer to other convex resource theories with minimal changes.","Because the witness reports non-objectivity relative to a fixed subspace, using it on a state that is objective in another basis produces a false positive; a practical protocol may need to scan several candidate subspaces or infer the natural objective basis before trusting a non-zero reading.","The paper's run-count comparison counts only measurement runs and treats the point channel and unitary evolution as deterministic; including state-preparation overheads and CNOT failures will raise the effective resource cost, and the claimed advantage should be re-tested in an end-to-end estimate.","For invariant spectrum broadcast structure, the CNOT-free scheme suggests an immediate experimental test with a five-photon GHZ state: if the witness remains tight for reduced fragments as the paper's simulations show, it would validate the approach at modest scale."],"forward_implications":["A non-zero witness is sufficient to declare the state non-objective relative to the chosen basis and subspaces, without reconstructing the state.","The photonic scheme's run count scales as $C + C(1/p_{\\mathrm{CNOT}})^{2M}$, versus $C\\,3^{1+2M}$ for tomography, so the advantage grows with the number of environment photons once $p_{\\mathrm{CNOT}}$ is above about $1/3$ (about $0.42$ at the simulated gate fidelity).","For invariant spectrum broadcast structure, which applies when all subsystem dimensions are equal, the objectivity operation needs no CNOT gates and the witness requires roughly a constant number of runs independent of fragment size.","The witness gives one combined number for non-objectivity and does not by itself say whether the failure comes from quantum correlations or missing classical correlations; the paper suggests adding a discord witness when the source matters.","A zero witness is not a certificate of full objectivity, since the scheme is system-basis and environment-subspace dependent."],"supporting_citations":[{"why":"Defines strong quantum Darwinism and its equivalence to bipartite spectrum broadcast structure; this is the objectivity notion the witness is built on.","marker":"[3]"},{"why":"Defines spectrum broadcast structure and strong independence; the paper uses it to explain why the objectivity operation preserves environment correlations and why an SBS witness would be much harder.","marker":"[2]"},{"why":"Supplies the two-branch discord-witness method whose structure the objectivity witness follows.","marker":"[65]"},{"why":"Supplies the two-branch coherence-witness construction and the correlation ambiguity that the point channel is introduced to remove.","marker":"[66]"},{"why":"Supplies the near-deterministic CNOT gate based on cross-Kerr nonlinearity whose success probability determines the scaling advantage over tomography.","marker":"[73]"},{"why":"Supplies the non-destructive parity measurement used in the objectivity operation within the photonic proposal.","marker":"[72]"},{"why":"Supplies the four-photon GHZ preparation procedure used to build the initial system-environment state.","marker":"[74]"}],"fun_headline_variants":["Non-objectivity witness needs only two evolutions","Two-branch test detects non-objectivity in quantum systems","Witness non-objectivity without full tomography","Photonic witness tests strong quantum Darwinism"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The scheme assumes that a preferred set of system states and a preferred family of environment subspaces are fixed in advance, and it treats 'objective' as meaning 'lying in those subspaces'; a state that is objective in another basis can therefore be reported as non-objective, and a zero witness does not rule out that possibility.","fun_headline_variants_meta":{"raw":{"variants":["Non-objectivity witness needs only two evolutions","Two-branch test detects non-objectivity in quantum systems","Witness non-objectivity without full tomography","Photonic witness tests strong quantum Darwinism"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000399,"raw_usage":{"total_tokens":2066,"prompt_tokens":903,"completion_tokens":1163,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":519,"completion_tokens_details":{"reasoning_tokens":1104}},"tokens_in":519,"tokens_out":1163,"duration_ms":10058,"temperature":1.0,"reasoning_tokens":1104,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:29:36.151104+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check is to take many randomly generated system-fragment states, estimate $M_{\\mathrm{SQD}}$ by full state tomography, and run the maximal witness on the same states: the paper's bound predicts the witness never exceeds the measure, so any measured violation of $W_{\\mathrm{SQD}} \\le M_{\\mathrm{SQD}}$ would refute the central claim.","supporting_citations":[{"cited_title":"Brunner, R","cited_arxiv_id":null,"evidence_quote":"Supplies the two-branch coherence-witness construction and the correlation ambiguity that the point channel is introduced to remove."},{"cited_title":"Aaronson, inACM SIGACT - STOC 2018(ACM Press, 2018) pp","cited_arxiv_id":null,"evidence_quote":"Supplies the near-deterministic CNOT gate based on cross-Kerr nonlinearity whose success probability determines the scaling advantage over tomography."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the non-destructive parity measurement used in the objectivity operation within the photonic proposal."}],"review_version":1}